Targeting large phosp(iii)azane macrocyles [{P(μ-NR)} 2(LL)]n (n ≥ 2)

Article abstract of DOI:10.1039/b607332h

The condensation reactions of the dimer [ClP(μ-NR)]₂ (7) with organic diacids [LL(H)₂], possessing linear orientations of their organic groups, result in the formation of phospha(iii)zane macrocyles of the type [{P(μ-NR)}₂

Full text of DOI:10.1039/b607332h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Targeting large phosp(III)azane macrocyles [{P(l-NR)}2(LL)]n (n ≥ 2)
Fay Dodds, Felipe Garc´ıa, Richard A. Kowenicki, Simon P. Parsons, Mary McPartlin and Dominic S. Wright*
Received 23rd May 2006, Accepted 10th July 2006
First published as an Advance Article on the web 21st July 2006
DOI: 10.1039/b607332h
The condensation reactions of the dimer [ClP(l-NR)]2 (7) with organic diacids [LL(H)2], possessing
linear orientations of their organic groups, result in the formation of phospha(III)zane macrocyles of the
type [{P(l-NR)}2(LL)]n of various sizes. The series of macrocycles [{P(l-NtBu)}2{1,5-(NH)2C10H6}]3
(1), [{P(l-NCy)}2(1,5-O2C10H6)]n [n = 3 (2a); n = 4 (2b)], [{P(l-NtBu)}2{1,4-(NH)2C6H4}]4 (3),
[{P(l-NtBu)}2(1,4-O2C6H4)]3 (4), [{P(l-NCy)}2(1,4-O2C6H4)]3 (5) and
[{P(l-NtBu)}2{(NH)C6H4OC6H4(NH)}]2 (6) can be related to classical organic frameworks, like
calixarenes.
the synthesis of macrocyles larger than dimers—whose small,
Introduction
elliptical cavities appear to be unsuitable for the coordination
of hosts. We report here the syntheses of [{P(l-NtBu)}2{1,5-
Phosph(III)azane dimers of the type [ClP(l-NR)]2 are excellent
(NH)2C10H7}]3 (1),7 [{P(l-NCy)}2(1,5-O2C10H7)]n [n = 3 (2a);
building blocks for the formation of a range of inorganic
macrocycles. This is at least in part due to the fact that the
n
=
4
(2b)], [{P(l-NtBu)}2{1,4-(NH)2C6H4}]4 (3),8 [{P(l-
NtBu)}2(1,4-O2C6H4)]3 (4), [{P(l-NCy)}2(1,4-O2C6H4)]3 (5) and
[{P(l-NtBu)}2{(NH)C6H4OC6H4(NH)}]2 (6), containing a variety
of spacers, and on the observation of host–ndash behaviour which
mirrors that found commonly in related calixarenes.
thermodynamic preference for the cis-isomers results in pre-
organisation of the reactions to cyclise rather than form polymers.1
For some time it has been known that the condensation reactions
of these dimers with bifunctional organic acids LLꢀ(H2) give
cyclic oligomers of the type [{P(l-NR)}2{LLꢀ}]n. However, until
recently there was only unequivocal evidence2 for the existence
of monomers of this type (n = 1).3 Monomers appear to be
particularly prevalent where more flexible aliphatic spacers are
involved (Fig. 1). More recently, it was found that macrocyclic
dimers (n = 2) can be obtained from these reactions,4 depending
on the reaction conditions employed and the orientation of the
organic substituents present in the organic spacer. In particular,
rigid aromatic spacers possessing ortho- or para-substituents
normally give rise to dimeric macrocycles of this type (Fig. 2).4b,c
Our interest in these species stems not only from their ease of
synthesis but also from the multifaceted behaviour of macrocyles
of this type, which can function potentially as P-donors (at
their peripheries),5 or as metal or anion hosts (within their
cavities).6 In order to realise the full potential of these ligands,
our attention has been focused most recently on exploring
Results and discussion
The syntheses and structures of the trimer [{P(l-NtBu)}2{1,5-
(NH)2C10H7}]3 (1)7 and tetramer [{P(l-NtBu)}2{1,4-(NH)2-
C6H4}]4 (3) (Fig. 3),8 containing N-functionalised bridging organic
groups, have been communicated by us previously. Both complexes
are generated by the condensation reactions of [ClP(l-NtBu)]2
with 1,5-diamino-naphthalene or phenylenediamine in the pres-
ence of Et3N. Compounds 1 and 3 are of interest in that they are
the only examples of macrocycles of this type bigger than dimers.
The presence of endo-directed N–H groups in these macrocycles
potentially allows these species to behave as H-bonding, anion
receptors. In the current study we wanted, in particular, to explore
the synthesis of larger macrocycles of this type containing donor
O-functionality which (like crown ethers) might be capable of the
coordination of cations. The new macrocycles [{P(l-NtBu)}2(1,4-
O2C6H4)]3 (4) and [{P(l-NtBu)}2{(NH)C6H4OC6H4(NH)}]2 (6)
Chemistry Department, Cambridge University, Lensfield Road, Cambridge,
UK CB2 1EW
Fig. 1 Monomers (n = 1) formed by various organic spacers.
The Royal Society of Chemistry 2006 Dalton Trans., 2006, 4235–4243 | 4235
This journal is
©

View Article Online
Fig. 2 Dimers (n = 2) formed by various organic spacers.
Fig. 3 Structures of (a) the bowl-shaped trimeric macrocycle 1 and (b) the folded macrocyle 3.
were obtained in a similar manner to 1 and 3 from the reactions
of the dimer [ClP(l-NtBu)]2 with the appropriate organic diacids
in the presence of Et3N as the Brønsted base. [{P(l-NCy)}2(1,5-
O2C10H7)]3 (2) and [{P(l-NCy)}2(1,4-O2C6H4)]3 (5) were obtained
from reactions involving the novel dimer [ClP(l-NCy)]2 (7)
(Scheme 1).
the Cy groups within the P2N2 units of 2 will allow the formation
of larger oligomers, by reducing the overall steric congestion at
the periphery of the macrocycles. A further difference between 1
and 2 is the far lower activation energy for the interconversion
of the exo- and endo-P atoms in 2 (obtained from 31P NMR
spectroscopic studies).9 Whereas the activation Gibb’s free energy
for this process is similar for the two oligomers of 2 at DG‡ ≈
43–45 kJ mol−1, two distinct resonances for these environments
are found for the analogous N-functionalised trimer 1 at room
temperature, with no interconversion between exo- and endo-P
environments in toluene solvent up to a temperature of 333 K. The
DG‡ found for the oligomers of 2 compares to 31 kJ mol−1 found
for the dimer [{P(l-NtBu)}2(OCH2C(Me)2CH2O)]2, containing a
flexible OCH2C(Me)2CH2O linker.4c
The major P-containing products formed in the reactions
of hydroquinone with [ClP(l-NtBu)]2 or [ClP(l-NCy)]2 are the
trimers [{P(l-NtBu)}2(1,4-O2C6H4)]3 (4) and [{P(l-NCy)}2(1,4-
O2C6H4)]3 (5), as shown by an in situ 31P NMR spectrum of
the reaction mixtures (Fig. 6). The room-temperature 31P NMR
spectra of both species consist of singlets (d = 136.6 for 4 and
141.2 for 5). These resonances split into two singlets for both
compounds at low temperature (ca. 240 K for 4; ca. 250 K for 5)
with associated activation Gibb’s free energies (DG‡ ≈ 43 kJ mol−1
for 4; 41 kJ mol−1 for 5) which are very similar to those found for the
O-functionalised trimer (2a) andtetramer (2b). Again, comparison
can be made with the behaviour of the previously reported trimer
(3a) and tetramer (3b) formed in the reaction of phenylenediamine
with [ClP(l-NtBu)]2. The exo- and endo-P atoms in both of these
The macrocycles 2, 4, 5 and 6 were characterised using a com-
bination of NMR and mass spectroscopic techniques. Although
elemental analysis was also employed in the case of 4 and 6,
satisfactory analysis was difficult to obtain on these species despite
repeated attempts. The reaction of 1,5-dihydroxy-naphthalene
with [ClP(l-NCy)]2 produces a ca. 1 : 1 mixture of two distinct
macrocylic oligomers (Fig. 4). Their presence can be discerned
from the room-temperature 31P NMR spectrum in CDCl3 which
consists of two singlets at d = 138.6 and 134.9. These resonances
both split into 1 : 1 singlets at 250 and 265 K, respectively,
resulting from the resolution of the exo- and endo-P atoms of
the macrocylic framework (Fig. 5). Mass spectroscopy shows that
these species are a trimer (2a) (m/e = 1244.0, [trimerH]+) and a
tetramer (2b) (m/e = 1656.6, [tetramerH]+) (Fig. 4). Fractional
crystallisation of 2 led to the isolation of crystals of the trimer
2a, which was characterised by X-ray crystallography (see later).
Interestingly, the closely related reaction of [ClP(l-tBu)]2 with 1,5-
diamino-naphthalene gives only the trimer [{P(l-NtBu)}2{1,5-
(NH)2C10H6}]3 (1) (as shown by in situ 31P NMR studies). This
difference is probably due to a combination of the different
geometric demands of the N atoms and different steric demands
of the tBu groups in 1. In particular, the lower steric demands of
4236 | Dalton Trans., 2006, 4235–4243
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Scheme 1
Fig. 4 Structures of the trimer and tetramer of 2.
Fig. 5 exo- and endo-P atoms in the trimer and tetramer of 2.
oligomers do not interconvert at room temperature. In the case of
the tetramer 3b the activation Gibb’s free energy for this process
(DG‡ = 83 kJ mol−1) is considerably higher than that found for
the various oligomers of 2, 4 or 5. Despite repeated attempts, we
were unable to obtain crystals of 5 for X-ray analysis, although
mass spectroscopy suggests the presence of a trimer (observed
as a low-abundance [trimerH]+ peak). However, suitable crystals of
Fig. 6 Trimeric arrangement of 4 and 5.
Dalton Trans., 2006, 4235–4243 | 4237
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
the closely related macrocyle 4 were grown from n-pentane and the
compound confirmed to be trimeric in the solid state (see later).
rings of the NH–C6H4–O–C6H4–NH linker. Thus, the aromatic
ring next to the exo-P centres H-atoms are inequivalent to the
aromatic ring next to the endo-P atom, with the result that two
sets of doublets are observed in the aromatic region of the room-
temperature 1H NMR spectrum. In addition, the exo- and endo-
directed N–H groups also appear as two separate doublets (as a
result of coupling to their respective P centres).
Low-temperature X-ray structures of the new macrocycles
2a·thf·n-pentane, 4 and 6·2toluene provide unequivocal evidence
of the structures of the compounds, supporting the prelimi-
nary analytical and spectroscopic results described above. For
completeness, the structure of the dimeric precursor 7 was also
obtained. Details of the data collections and refinements of
these compounds are provided in Table 1 (Experimental section).
The three compounds 2a, 4 and 7 all presented difficulties in
refinement, due to problems either with the quality of the crystals
or withinherentcrystallographic features. As aresult, bondlengths
and angles are of little value for detailed comparative purposes.
Where appropriate, selected bond lengths and angles are given in
the figure captions to each of the structures.
Fig. 7 Structure of the macrocyclic dimer 6.
The reaction of the O/N-functionalised spacer 4,4ꢀ-oxydianiline
with [ClP(l-NtBu)]2 was undertaken as a potential alternative
strategy towards larger phosphazane macrocycles. It transpires,
however, that the only product of this reaction is the dimer
[{P(l-NtBu)}2{(NH)C6H4OC6H4(NH)}]2 (6) (Fig. 7). Although
similar in connectivity to previously reported dimers of this type,
nonetheless the size of the organic spacer results in a large
macrocyclic cavity. NMR spectroscopy of 6 provides extremely
clear evidence of the macrocyclic nature of its structure. The room-
temperature 31P NMR spectrum of 6 consists of two singlets (d =
101.8, 99.6) for the exo- and endo-P atoms of the macrocyle (the
resonance at d 99.6 splitting into a 1 : 1 doublet as a result of
coupling to an N–H proton). The rigidity of the macrocyle at
room temperature results in the inequivalence of the two aromatic
The structure of 2a is that of a trimer, in which the P2N2
ring units are linked into a cyclic structure by 1,5-O2C10H6 groups
(Fig. 8). This is very similar to that of the nitrogen analogue [{P(l-
NtBu)}2{1,5-(NH)2C10H6}]3 (1) reported previously (Fig. 3a).7
Like 1, macrocyclic molecules of 2a have approximate C3 symme-
try in which the naphthyl groups are arranged in a bowl (or cone)
shape which is reminiscent of the conformation of a calixarene.10
The cavity of 2a is considerably smaller than that of 1 [with a mean
˚
radius of ca. 3.5 A measured from the three endo-O atoms O(2),
˚
O(4) and O(6) to their centroid; cf . ca. 4.4 A in 1 measured with
respect to the three endo-N(H) atoms7]. The more compact nature
of the structure of 2a is also evident in the far greater inclination
Table 1 Crystal data and structure refinements for 2a·thf·n-pentane, 4, 6·2toluene and 7a
2a·thf.·n-pentane
4
6·2toluene
7
Empirical formula
Fw
C79H112N6O8P6
1459.57
Monoclinic
P21/c
13.205(3)
33.184(7)
C42H66N6O6P6
936.83
Monoclinic
P21/c
17.410(4)
15.925(3)
C54H72N8O2P4
989.08
Monoclinic
P21/n
17.832(4)
13.419(3)
C12H22Cl2N2P2
327.16
Monoclinic
C2/m
13.707(3)
5.4813(11)
Crystal system
Space group
˚
a/A
˚
b/A
˚
c/A
18.385(4)
102.98(3)
37.093(7)
100.04(3)
23.472(5)
92.05(3)
21.039(4)
91.74(3)
b/◦
Z
V/A
4
8
4
2
3
˚
7850(3)
1.235
0.194
3128
10126(4)
1.229
0.260
3984
5615(2)
1.170
0.180
2114
1580(5)
1.375
0.599
6888
q(calc)/Mg m−3
l(Mo-Ka)/mm−1
F(000)
Refl. collected
Crystal size/mm
h range/◦
Indep. refns. (Rint
Absorp. corr.
49553
25649
22915
5658
0.16 × 0.16 × 0.14
3.51–21.05
8441 (0.048)
Semi-empirical from
equivalents
0.971, 0.906
8441/61/879
1.071
R1 = 0.079, wR2 = 0.197
R1 = 0.096, wR2 = 0.207
0.699, −0.558
0.25 × 0.05 × 0.05
3.53–19.00
7246 (0.157)
Semi-empirical from
equivalents
0.991, 0.722
7235/1032/901
1.132
R1 = 0.169, wR2 = 0.373
R1 = 0.198, wR2 = 0.373
0.981, −0.573
0.42 × 0.16 × 0.16
3.57–24.00
8650 (0.089)
Semi-empirical from
equivalents
0.997, 0.867
8650/0/580
1.036
R1 = 0.063, wR2 = 0.146
R1 = 0.117, wR2 = 0.165
0.698, −0.426
0.18 × 0.08 × 0.05
3.60–25.00
1545 (0.054)
Semi-empirical from
equivalents
0.971, 0.898
1545/15/116
1.088
R1 = 0.051, wR2 = 0.123
R1 = 0.079, wR2 = 0.134
0.142, −0.146
)
Max., min. transmission
Data/restraints/parameters
Goodness-of-fit on F2
R indices [I > 2r(I)]
R indices (all data)
Absolute structure parameter
−3
˚
largest peak and hole/e A
a
˚
Data in common; T = 180(2) K, k = 0.71073 A.
4238 | Dalton Trans., 2006, 4235–4243
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
C–H · · · arene contacts with the naphthylene rings (Fig. 9b).6,11
The inclusion of toluene and thf molecules in 1 and 2a respectively
is similar to that found previously in the structures of clathrates of
calix[4]arenes, e.g., in the [p-tBu-MeO-calix[4]arene·Na·toluene]+
cation12 and [p-tBu-OH-calix[4]arene·toluene].13 The remaining
thf and n-pentane molecules in 2a are located between the
molecules in the lattice.
The structure of the phosph(III)azane precursor to 2a, the dimer
[ClP(l-NCy)]2 (7), was also problematic due to its disorder across a
crystallographic mirror plane so that, at first sight, the presence of
either the cis or the trans isomer appears equally possible. However
only three atoms of the P2N2 unit lie in the mirror plane and
the fourth, P(1), lies slightly out of the plane (see Experimental
section), and due to the proximity of the mirror plane to P(1)
the two P–Cl bond lengths were constrained to be equal in the
refinement. This evidence of a puckered P2N2 ring unit strongly
suggests a cis geometry of the Cl atoms with respect to the P2N2
ring unit (as depicted in Fig. 10).1,14 The Cy ring is a less sterically
demanding R-group than tBu (and has low electronegativity1) so
Fig. 8 Structure of trimeric molecules of 2a (thf and n-hexane molecules
in the lattice and H-atoms have been removed for clarity).
of the planes of the naphthyl rings with respect to the mean plane
of the molecule (i.e., through the six exo- and endo-O atoms which
˚
are within 0.67 A of a plane in 2a). In 2a, these naphthyl rings are
inclined at 49.4–74.2◦ to the major◦plane, whereas in 1 a shallower
and more regular angle of ca. 35.5 is observed.6 Thus, the bowl-
shaped cavity of 2a is deeper and narrower than that of 1.
An interesting feature of the structure of 2a is the inclusion of
a thf molecule within the macrocyclic cavity (Fig. 9a). The thf O-
atom, which was assigned tentatively on the basis of initial peak
height and atomic displacements parameters, is directed outwards
from the core so that the predominant guest–host interaction
appears to be hydrophobic, a situation similar to that found in
1,7 in which the Me group of one toluene molecule is orientated
towards the centroid of the macrocycle, forming long-range
Fig. 10 Structure of the dimer 7, the precursor to 2a, showing one
component of the 50 : 50% disorder present in the crystal. H-atoms
◦
˚
have been omitted for clarity. Selected bond lengths (A) and angles ( );
mean P–Cl 2.167(1), P(1)–N(1) 1.711(2), P(1)–N(2) 1.686(2), P(2)–N(2)
1.702(2), P(2)–N(1) 1.680(2); N(1)–P(1)–N(2) 81.1(1), N(2)–P(2)–N(1)
81.5(1), P(1)–N(1)–P(2) 97.9(1), P(2)–N(2)–P(1) 98.0(1), puckering of P2N2
ring 167.6.
Fig. 9 Space-filling diagrams of a) the thf inclusion in 2a, and b) the toluene inclusion in 1 (the included thf and toluene molecules are highlighted in
darker shades).
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4235–4243 | 4239
©

View Article Online
that the cis geometry (which is also found for [ClP(l-NtBu)]2, the
precursor to 1, 3, 4 and 6) is not unexpected, but nevertheless it
suggests that pre-organisation of the phosph(III)azane precursor is
one of the factors directing cyclisation in the condensation reaction
producing 2a.
from molecule B (effectively ‘self-inclusion’, shown in Fig. 11b); a
similar interlocking occurs between molecule B and a Bu group
of molecule A in the asymmetric unit of the next unit cell at x, 1
+ y, z and so generates infinite chains of interlocked macrocycles
running parallel to the b-axis of the unit cell. This association
t
Despite repeated attempts we were unable to obtain good
quality crystals of 4, but nonetheless the X-ray study established
that like 2a, 4 is a trimeric macrocycle, confirming the results ob-
tained from mass spectroscopy, and the connectivity is established
unequivocally. Two independent, chemically equivalent molecules
(A and B) are found in the asymmetric unit and Fig. 11a shows one
of these molecules. All the trimeric macrocycles have approximate
C3 symmetry but in contrast to 17 and 2a which have bowl-shaped
arrangements, the planes of the aromatic spacers in 4 adopt a
more propeller-like conformation in both independent molecules.
Although again the individual bond lengths in 4 are not suitable for
detailed comparisons it is apparent that several structural features
are very similar to those found for 2a. In particular, the six O-
atoms of both molecules of 4 are almost coplanar [maximum
involves a combination of relatively short C–H · · · p-arene (3.12–
10
˚
˚
3.19 A; cf . 2.85–3.15 A estimated for a van der Waals interaction )
˚
˚
and C–H · · · O (2.87–2.89 A; cf . 2.70–2.95 A estimated for a van
der Waals interaction10) contacts.
The solid-state structure of 6 is that of a dimer which is closely
related to those which we and others have reported previously
(Fig. 12a).4 The more congested nature of the cavities of 6 means
that no inclusion of solvent into the molecule is observed. Instead,
molecules associate into layers via weak N–H · · · O H-bonding
◦
˚
˚
[N · · · O 3.16 A (H · · · O 2.53–2.66 A), N–H · · · O 115.9–128.0 ]
(Fig. 12b),10 between which tBu groups interlock to form voids in
which two toluene molecules per macrocyle are accommodated.
The association of phosph(III)azane macrocycles via H-bonding
has been seen before in [{P(l-NtBu)}2{2,6-(NH)2C5H3N}]2, in
which N–H · · · P interactions also result in the formation of
layers.4b
˚
˚
deviation 0.65 A (mean); cf . 0.67 A in 2a] and the size of the
˚
cavities [radius 3.4 A (mean) measured to from the three endo-O
atoms to their centroid] is significantly smaller than that found in
˚
˚
the closest N-analogue 1 of 4.4 A and close to the radius of 3.5 A
in 2a).
Conclusions
One interesting difference between the two molecules 17 and
2a, and 4 is the absence of solvent inclusion within the cavity
of the macrocycle in the latter. We assume that this is due to
a combination of the conformational differences between the
cavities of these macrocycles and to the more sterically crowded
nature of the periphery of 4 (containing Bu rather than Cy
groups). Instead of coordinating a molecule of solvent, the two
independent molecules of 4 interlock using one of the tBu groups
The current study has shown that larger phosp(III)azane macro-
cycles of the type [{P(l-NR)}2(LLꢀ)]n with n > 2 can be readily
prepared. Unlike the dimeric counterparts, trimeric macrocycles
of this type have the ability to act as hosts to neutral guests, as wit-
nessed in the structures of 2a and 4. The macrocycle 2a is the first
in which a substituent other than tBu has been incorporated into
these species. Further coordination studies in this area are planned.
t
Fig. 11 a) Structure of one of the molecules of 4 (molecule A) (H-atoms omitted for clarity), and b) interlocking of molecule A and B in the crystal lattice
(the included tBu group of molecule B is highlighted in blue).
4240 | Dalton Trans., 2006, 4235–4243
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
standard for 1H NMR and 85% H3PO4/D2O as the external
standard for 31P NMR). In situ 31P NMR spectroscopic studies
on reaction mixtures in non-deuterated solvents were recorded
using an internal D6-acetone capillary to obtain a lock.
Synthesis of 2
To a solution of 1,5-dihydroxynaphthalene (0.158 g, 1.0 mmol)
◦
in thf (30 ml) and Et3N (1.0 ml) at −78 C was added dropwise
[ClP(l-NCy)]2 (0.158 g, 1.0 mmol) in thf (30 ml) and the mixture
stirred (1.5 h). The resulting suspension was allowed to warm
to room temperature and then brought to reflux (48 h). The thf
was removed under vacuum and the white residue extracted with
n-pentane and filtered through Celite. The filtrate was reduced
in volume under vacuum until precipitation of a white solid
occurred. This was dissolved by the addition of a few drops of
thf and heating. Storage of the solution at room temperature gave
colourless needles of 2. Yield 0.07 g (16%). H NMR (+25 ◦C,
1
CDCl3, 500.2 MHz), d = 8.3–7.1 (mult., 6H, C–H naphthyl group),
3.14 (br.s., 2H, a-C–H of Cy), 2.3–0.8 (mult., 22H, Cy group). 31
P
NMR (+25 ◦C, CDCl3, 202.48 MHz), d = 138.6 (s.) (trimer) [splits
into two resonances at ca. 250 K at d = 139.4 (d., 3J31P–H = 21.0 hz)
3
and 124.8 (d., J31P–H = 21.0 hz) (DG‡ ≈ 42.8 kJ mol−1)], 134.9
(tetramer) [splits into two resonances at ca. 265 K at d = 145.0
(s.) and 125.2 (s.) (DG‡ ≈ 44.7 kJ mol−1)]. Mass spectroscopy (+ve
ion) m/e = 1656.6 (tetramerH+), 1244.0 (trimerH+).
Synthesis of 4
Fig. 12 a) Structure of one of the independent molecules of 6 (toluene
molecules and H-atoms are omitted for clarity), and b) association of
molecules into a layer structure via N-H · · · O H-bonding. Selected bond
To a solution of hydroquinone (1.0◦mmol, 0.11 g) in thf (20 ml)
and Et3N (4.0 mmol, 0.6 ml) at −78 C was added [ClP(l-NtBu)]2
(1.0 mmol, 0.28 g) in thf (20 ml) dropwise. The mixture was allowed
to warm to room temperature and stirred (1 h) before being heated
to reflux (16 h), resulting in a colourless suspension. The solvent
was removed under vacuum, n-pentane added (40 ml) and the
suspension filtered through Celite. The filtrate was reduced in
volume until precipitation of a colourless solid occurred. This
was warmed back into solution and storage at room temperat◦ure
gave colourless crystals of 4. Yield 0.14 g (45%). 1H NMR (+25 C,
CDCl3, 500.2 MHz), d = 7.30 [d. (2JHH = 7.8 hz), 2H, C–H aryl],
◦
t
˚
lengths (A) and angles ( ); P–l-N( Bu) range 1.713(3)–1.726(3), P–N(H)
range 1.686(3)–1.702(3), P · · · P range 2.587(1)–2.593(1); P–N(H)–C range
123.5(3)–127.2(3), C–O(1,2)–C range 115.8(2)–118.0(2); N · · · O 3.16
(H · · · O 2.53–2.66), N–H · · · O 115.9–128.0, puckering of P2N2 rings
17.9–18.9.
Experimental
General experimental
t
7.20 [d. (2JHH = 7.8 hz), 2H, C–H aryl], 1.28 [s., 36H, Bu] 31P
Compounds 2–7 were prepared under dry, O2-free N2 on a vacuum
line. [ClP(l-NtBu)]2 was obtained from the 3 : 1 reaction of tBuNH2
with PCl3, respectively, in thf in the manner described in the
literature.6a Thf, toluene and n-pentane were dried by distillation
over sodium or sodium/benzophenone prior to the reactions. Et3N
was dried over CaH2. 2–7 were isolated and characterised with
the aid of an N2-filled glove box fitted with a Belle Technology
O2 and H2O internal recirculation system. Elemental analyses
were performed by first sealing the samples under argon in air-
tight aluminium boats (1–2 mg) and C, H and N content was
analysed using an Exeter Analytical CE-440. P analysis was
obtained using spectrophotometric means. Elemental analysis of
all of the compounds proved to be difficult and frequently the
P and N contents were significantly lower than the theoretical
values. This may be due to the formation of PN ceramics during
combustion.15 Where possible, mass spectroscopy was also used to
augment elemental analysis. 1H, 31P NMR spectra were recorded
on a Bruker ATM DRX500 spectrometer, in dry deuterated
CDCl3 (using the solvent resonances as the internal reference
NMR (+25 ◦C, CDCl3, 202.48 MHz), d = 136.6 (s.) (trimer) (other
unidentified minor species at d = 139.9–142.4) [resonance for the
trimer splits into two resonances at d = 145.2 (d.) and 123.7 (d.)
at ca. 240 K (DG‡ ≈ 43 kJ mol−1)]. Elemental analysis, found
C 54.2, H 7.4, N 7.8, P 16.2; cald. for 4 C 53.8, H 7.1, N 9.0,
P 19.9%.
Synthesis of 5
To a solution of hydroquinone (0.11 g, 1.0 mmol) in thf (40 ml) and
Et3N (0.5 ml) at −78 ◦C was added [ClP(l-NCy)]2 in thf (40 ml).
The mixture was stirred (1.5 h) then allowed to warm to room
temperature and brought to reflux (24 h). The solvent was removed
under vacuum and the white residue extracted with n-pentane
(40 ml) and filtered through Celite. The solvent was removed under
vacuum to yield a white powder of 5. Yield 0.098 g (27%). 1H NMR
(+25 ◦C, CDCl3, 500.2 MHz), d = 5.94 (s., aromatic C–H), 1.7–0.9
(mult., Cy group). 31P NMR (+25 ◦C, CDCl3, 202.48 MHz), d =
141.2 (s.) [splits into two singlets at ca. 260 K at d = 149.6 and
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4235–4243 | 4241
©

View Article Online
127.3 (DG‡ ≈ 45 kJ mol−1)]. Mass spectroscopy (+ve ion) m/e =
1093.7 (trimerH+) (low abundance with higher m/e peaks being
observed up to 1712.3).
close to lying half way between the c-glide planes (i.e., y for all
ten C-atoms is 0.53). As a consequence of these factors some
of the interatomic distances in the chemically equivalent bonds
show a rather large discrepancy. Rather high ADP indicated some
conformational disorder of the two thf and the n-pentane solvate
molecules. The carbon atoms of one thf molecule and four in the
n-pentane were resolved into two components. The crystal of 4
has two trimeric macrocycle molecules per equivalent position,
and the very fine needle crystals diffracted poorly at high angle
and refinement was very difficult. This was apparently mainly
due to special relationships between equivalent atoms in the two
molecules, i.e., each atom in one molecule has nearly the same
z-coordinates as the corresponding atom in the second molecule
and each pair is separated by almost exactly 1/2 in y. To help in the
refinement equivalent atoms in the two molecules were constrained
to have the same bonding parameters, and constraints were also
applied to the anisotropic ADP to prevent non-positive definite
results. Despite the unsatisfactory refinement the main features of
the macrocyclic molecule are clearly established. The problem in 7
is that that the molecule is disordered across the crystallographic
mirror plane, with the P2N2 unit lying in the symmetry plane.
This arrangement gave an unreasonably short P(1)–Cl(1) bond
with a P(1) displacement ellipsoid very elongated in a direction
perpendicular to the plane. It was concluded that the P2N2 unit
was in fact slightly folded with two disordered components of P(1)
sited too close to the mirror to be resolved in the initial Fourier
maps. To obtain a better model in refinement the half-occupancy
P(1) was moved slightly off the plane away from the Cl(1) atom and
the two P–Cl lengths were constrained to be equal within an e.s.d.
Synthesis of 6
To a solution of 4,4ꢀ-oxydianiline (2.0 mmol, 0.40 g) in thf (40 ml)
and Et3N (8.0 mmol, 2.0 ml) at −78 ◦C was added a solution
of [ClP(l-NtBu)]2 (2.0 mol, 0.55 g) in thf (30 ml) dropwise. The
mixture was stirred at −78 ◦C (2 h) before being allowed to warm
to room temperature and stirred (16 h). The solvent was removed
under vacuum, the dry solid produced dissolved in toluene (40 ml)
and filtered through Celite. Removal of the solvent under vacuum
gave a colourless powder of 5. Yield 0.45 g (56%). 1H NMR
(+25 ◦C, CD2Cl2, 500.20 MHz), d = 7.25 (d. 1H, J = 8.6 Hz,
C–H), 7.03 (d. 1H, J = 8.7 hz), 6.72 (d. 1H, J = 8.7 hz), 6.62 (d.
2
1H, J = 8.6 hz), 4.87 (d., 1H, N–H, JP–H = 36.6 hz), 4.67 (d.,
1H, N–H, 2JP–H = 5.7 hz), 1.32 (s., 18H, tBu). 31P NMR (+25 ◦C,
CD2Cl2, 202.48 MHz), d = 101.8 (s.), 99.6 (s.) [splits into a 1 : 1
doublet (2JP–H = 36.6 hz) at low temperature] [an in situ 31P NMR
study of the reaction solution (using d6-acetone capillary to obtain
a lock) showed that 5 is the only soluble P-containing compound
formed]. Elemental analysis, found C 56.5, H 7.1, N 12.5, 14.1;
cald. for 6 C 59.7, H 7.0, N 13.9, P 15.4%.
Synthesis of 7
To cyclohexylamine (11.5 ml, 100 mmol) in thf (200 ml) at −78 ◦C
n
was added BuLi (62.5 ml, 1.6 mol dm−3 in hexanes, 100 mmol).
The solution was allowed to warm to room temperature and stirred
(1 h). This was then added dropwise to a solution of PCl3 (8.75 ml,
100 mmol) in thf (200 ml) at −78 ◦C. The mixture was stirred
(2 h) before being allowed to warm to room temperature and
stirred (48 h). The mixture was filtered through Celite, the solvent
removed under vacuum and toluene was added (150 ml). The
mixture was filtered and the toluene removed under vacuum to
give an orange oil. Diethyl ether was added (40 ml) and then
removed under vacuum. This was repeated. A viscous semi-solid
was finally produced which was sublimed onto a water-cooled cold
finger in a Schlenk tube (150 ◦C, 0.1 mmHg). Yield 1.64 g (20%).
31P NMR (+25 ◦C, CDCl3, 202.48 MHz), d = 221.0 (s.). Mass
spectroscopy (+ve ion) m/e = 327.2 [MH+].
˚
of 0.005 A. From the disordered model with mirror symmetry the
molecule might be in a cis- or trans-configuration, but the evidence
that the P2N2 unit is folded, is consistent with the presence of the
cis-isomer. The crystals of 6 consist of hydrogen bonded sheets of
molecules parallel to the ab plane with two toluene solvates per
molecule lying between the sheets. One of the toluene molecules
is disordered with two overlapping components of ca. 70 : 30%
occupancy. The hydrogen atoms for 2a, 4 and 7 were placed in
calculated positions with displacement parameters set equal 1.2
Ueq (or1.5 Ueq for methyl groups) of the parent carbon atoms.
The nitrogen-bonded hydrogen atoms in 6 were directly located
and included in the refinement without restraint, all other H-
atoms being included in idealised sites. For all four structures
anisotropic displacement parameters were assigned to all full
occupancy atoms in the final cycles of full-matrix refinement based
of F2.16
Crystal structures of 2a·thf·n-C5H12, 4, 6·2toluene and 7
Crystals of 2a·thf·n-C5H12, 4, 6·2toluene and 7 were mounted
directly from solution under argon using an inert oil which protects
them from atmospheric oxygen and moisture. X-Ray intensity data
were collected using a Nonius Kappa CCD diffractometer. Details
of the data collections16 and structural refinements are given in
Table 1. For all four crystals the positions of the non-hydrogen
atoms were located by direct methods. The overall features of the
four new molecules are clearly established but only the data for
compound 6 refined well. The three compounds, 2a, 4 and 7 pre-
sented problems in refinement. The crystal of 2a shows a trimeric
macrocyclic molecule with an n-pentane and two thf solvate
molecules in the lattice. Refinement of 2a was rather unsatisfactory
due to unfortunate alignment of two of the naphthyl units. Ring
C(20) · · · C(29) is almost exactly parallel to the ab plane and ring
C(40) · · · C(49) to the ac plane; additionally, the latter ring is very
CCDC reference numbers 608772–608775.
For crystallographic data in CIF or other electronic format see
DOI: 10.1039/b607332h
Acknowledgements
We gratefully acknowledge the EPSRC (F.D., F.G., R.A.K.,
M.McP., D.S.W.) and Wolfson College (Fellowship for F.G.) for
financial support.
References
1 I. Silaghi-Dumitrescu and I. Haiduc, Phoshorus, Sulfur Silcon Relat.
Elem., 1994, 91, 21.
4242 | Dalton Trans., 2006, 4235–4243
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
2 The first evidence for dimeric macrocycles came from the structural
characterisation of oxidation products, P. Kommana and K. C. K.
Swamy, Inorg. Chem., 2000, 39, 4384.
3 S. S. Kumaravel, S. S. Krishnamurthy, T. S. Cameron and A. Linden,
Inorg. Chem., 1988, 27, 4546; M. Vijjulatha, S. Kumaraswamy, K. C. K.
Swamy and U. Engelhardt, Polyhedron, 1999, 18, 2557; P. Kommana
and K. C. K. Swamy, Inorg. Chem., 2000, 39, 4384.
4 (a) P. Kommana, K. V. P. P. Kumar and K. C. K. Swamy, In-
dian J. Chem., Sect. A, 2003, 42, 2371; (b) F. Garcia, R. A. Kowenicki,
I. Kuzu, L. Riera, M. McPartlin and D. S. Wright, Dalton Trans., 2004,
2904; (c) F. Garc´ıa, J. M. Goodman, R. A. Kowenicki, M. McPartlin,
L. Riera, M. A. Silva, A. Wirsing and D. S. Wright, Dalton Trans.,
2005, 1764.
5 A. D. Bond, E. L. Doyle, F. Garc´ıa, R. A. Kowenicki, M. McPartlin,
L. Riera and D. S. Wright, Chem. Commun., 2003, 2990.
6 (a) A. Bashall, A. D. Bond, E. L. Doyle, F. Garc´ıa, S. J. Kidd, G. T.
Lawson, M. C. Parry, M. McPartlin, A. D. Wood and D. S. Wright,
Chem.–Eur. J., 2002, 8, 3377; (b) F. Garc´ıa, J. M. Goodman, R. A.
Kowenicki, I. Kuzu, M. McPartlin, M. A. Silva, L. Riera and D. S.
Wright, Chem.–Eur. J., 2004, 10, 6066.
9 D. H. Williams and I. Flemming, Spectoscopic Methods in Organic
Chemistry, McGraw-Hill, London, 1989, p. 103.
10 M. A. McKewey, M.-J. Schwing-Weill and F. Ardaud-Neu, Compre-
hensive Coordination Chemistry; Cation Binding by Calixarenes, ed.
G. W. Gokel, Pergamon, Oxford, 1996, vol. 1, ch. 15, p. 537.
11 J. E. Huheey, E. A. Keiter and R. L. Keiter, Inorganic Chemistry:
Principles of Structure and Reactivity, Harper and Collins, New York,
4th edn, 1993, p. 290.
12 S. G. Bott, A. W. Coleman and J. L. Atwood, J. Am. Chem. Soc., 1986,
108, 1709.
13 G. D. Andreetti, R. Ungaro and A. Pochini, J. Chem. Soc., Chem.
Commun., 1979, 1005.
14 The only cyclodiphos(III)azane having a planar P2N2 ring unit and
a cis conformation is [ClP(l-NPh)]2, H.-J. Chen, R. C. Haltiwanger,
T. G. Hill, M. L. Thomson, D. E. Coons and A. D. Norman, Inorg.
Chem., 1985, 24, 4725; A. R. Davies, A. T. Dronsfield, R. N. Haseldine
and D. R. Taylor, J. Chem. Soc., Perkin Trans. I, 1973, 379.; Of
the sixty disubstituted cylophosph(III)azanes found on the Cambridge
Crystallography Data Base, forty were cis and puckered, nineteen were
trans and planar and only the former was cis and planar, see: E. M.
Doyle, Ph.D. Thesis, University of Cambridge, 2003.
7 F. Dodds, F. Garcia, R. A. Kowenicki, M. McPartlin and D. S. Wright,
Chem. Commun., 2005, 3733.
8 F. Dodds, F. Garcia, R. A. Kowenicki, M. McPartlin, A. Steiner and
D. S. Wright, Chem. Commun., 2005, 5041.
15 J. E. Mark, H. R. Allcock and R. West, Inorganic Polymers, Oxford
University Press, Oxford, 2005, p. 133.
16 G. M. Sheldrick, SHELX-97, University of Go¨ttingen, Germany, 1997.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 4235–4243 | 4243
©