Selection of the cis and trans phosph(III)azane macrocycles [{P(μ-N tBu)}2(1-Y-2-NH-C6H4)]2 (Y = O, S)

Article abstract of DOI:10.1039/b502200b

The 1:1 reactions of [ClP(μ-N^tBu)]₂ with the difunctional aromatic amines 1,2-1-YH-2-NH₂-C₆H ₄ in the presence of Et₃N give the dimeric phosph(III)azane macrocycles [{P(μ-N_tBu)

Full text of DOI:10.1039/b502200b

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Selection of the cis and trans phosph(III)azane macrocycles
[{P(l-NtBu)}2(1-Y-2-NH-C6H4)]2 (Y = O, S)
Felipe Garc´ıa,a Jonathan M. Goodman,a Richard A. Kowenicki,a Mary McPartlin,b
Luc´ıa Riera,*a Mar´ıa A. Silva,a Andreas Wirsinga and Dominic S. Wright*a
a Chemistry Department, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1EW.
E-mail: dsw1000@cus.cam.ac.uk; Tel: +44 1223 763122
b Department of Health and Biological Sciences, London Metropolitan University, London,
UK N7 8DB
Received 14th February 2005, Accepted 22nd March 2005
First published as an Advance Article on the web 13th April 2005
The 1 : 1 reactions of [ClP(l-NtBu)]2 with the difunctional aromatic amines 1,2-1-YH-2-NH2-C6H4 in the presence of
Et3N give the dimeric phosph(III)azane macrocycles [{P(l-NtBu)2(1-Y-2-HN-C6H4)]2, predominantly as the cis
isomer in the case of Y = O (1·cis) and as the trans isomer for Y = S (2·trans). Model M.O. calculations suggest that
the selection of the cis and trans isomers is not thermodynamically controlled. The alternative isomers 1·trans and
2·cis are generated exclusively by the deprotonation of the model intermediates [(1-Y-2-NH2-C6H4)P(l-NtBu)]2 [Y =
O (3), S (4)] with nBuLi followed by cyclisation with [ClP(l-NtBu)]2. The solid-state structures of 1·cis/trans (50 : 50),
2·cis, 3 and 4 are reported.
Introduction
Results and discussion
There has been considerable recent interest in phosphorus(III)–
nitrogen macrocycles containing sufficiently large cavities, in
particular, for the coordination of catalytically-active metal
centres.1 Recent studies have shown that these species can be
obtained in high yields by thermodynamic or kinetic selection,1,2
without the need for conventional dilution techniques which
are common in the synthesis of organic counterparts. Our
interests in this area have focused on the preparation of
sterically hindered, toroidal arrangements containing dimeric
P2N2 ring units.2,3 For example, we showed that the condensation
reactions of the cyclophosph(III)azane dimers [ClP(l-NtBu)]2
and [H2NP(l-NtBu)]2 give the cyclic tetramer [{P(l-NtBu)}2(l-
An unusual bias towards the formation of the cis isomer of the
macrocyle [{P(l-NtBu)}2(1-O-2-NH-C6H4)]2 (1) was observed
in 31P and 1H NMR spectroscopic studies of the 1 : 1 stoichio-
metric reaction of [ClP(l-NtBu)]2 with 2-amino-phenol (1-OH-
2-NH2-C6H4) in THF in the presence of excess Et3N. In situ
1
31P{ H} NMR studies of the reaction mixture show that the
ratio of the cis isomer to the trans isomer is on average 60 :
40 [by integration of the 31P NMR spectrum]. The product
is isolated in 58% yield after extraction of the crude reaction
mixture with n-pentane, and consists of up to ca. 70 : 30 mixture
1
of the cis : trans isomers (according to integration of the H
NMR spectrum) [see Synthesis of 1; Method A, 1·cis/trans
(70 : 30), Experimental section]. Bearing in mind the inherent
inaccuracy of the integration of the in situ 31P NMR spectrum
compared to that for the 1H NMR spectrum, we do not believe
that significant enrichment of the cis isomer occurs on extraction
into n-pentane. However, extraction using toluene does give the
NH)]4 or pentamer [{P(l-NtBu)}2(l-NH)]5 (Fig. 1(a) and
(b), respectively), depending on the presence or absence of halide
ions as a template. Recent studies have found that reactions of
difunctional organic diacids (LLH2) with [ClP(l-NtBu)]2 give
a range of related, hybrid organic/inorganic macrocycles of the
type [{P(l-NtBu)}2(LL)]2 (Fig. 1(c)).3–5 This synthetic procedure
provides a highly versatile class of new ligands with an extremely
broad and potentially tunable range of donor functionality
for specific applications. We report here investigations of the
reactions of the dimer [ClP(l-NtBu)]2 with non-symmetric
organic diacids (LLꢀH2; L = Lꢀ) and a surprisingly easy set of
synthetic strategies for the selective formation of cis and trans
isomers of the resulting macrocycles [{P(l-NtBu)}2(LLꢀ)]2.
2a
2b,c
1
cis isomer exclusively in the absence of the trans. The H and
31P NMR spectra of the two isomers of 1 are very diagnostic
(Fig. 2). In particular, two singlet resonances are observed in
1
the 31P{ H} spectrum of 1·cis at d 171.1 and 133.7, whereas the
1
31P{ H} spectrum of 1·trans exhibits two doublets at d 173.8
and 148.4. Interestingly, the latter resonance at d 148.4 splits
1
into an apparent triplet in the H-coupled 31P NMR spectrum
2
2
(i.e., with JPNP ∼ JPH), allowing assignment of this resonance
Fig. 1 Frameworks of (a) the tetramer, (b) the pentamer, (c) hybrid organic/inorganic macrocycles (LL = difunctional organic spacer).
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3 T h i s j o u r n a l i s T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
1 7 6 4
©

View Article Online
Fig. 2 1H and 31P NMR assignments for the cis and trans isomers of 1.
to the P atoms bonded to the NH group within the macrocycle.
2·cis and trans are shown in Fig. 4, with their relative energies
being listed in Table 1. Particularly at the higher DFT level,6
the optimised structures of 1·trans and 2·cis are very similar
to the solid-state structures determined by later X-ray analysis,
the calculated structure of 1·trans having a centrosymmetric
core arrangement in which the planes of the 1-O-2-NH-C6H4
spacer groups are eclipsed and 2·cis having a highly distorted
structure in which P2N2 ring units are non-coplanar. Although
the differences in absolute energies between the cis and trans
forms of 2 are slightly basis set dependent, i.e., the optimised
gas-phase structure of 2·cis being 2.69 kcal mol−1 more stable
than the 2·trans at the HF/3-21G level but 0.51 kcal mol−1 less
stable at the B3LYP/6-31G* level, the key point is that the
differences in energy between the cis and trans isomers of 1 and
2 are small and unlikely to be large enough to provide significant
1
As expected, the H NMR spectrum of 1·cis also shows two
t
1 : 1 singlets for the two Bu environments (whereas only one
environment is present in 1·trans).
The analogous reaction of 2-aminothiophenol (1-SH-2-
NH2-C6H4) with [ClP(l-NtBu)]2 gives the macrocycle [{P(l-
NtBu)}2(1-S-2-NH-C6H4)]2 (2) mainly as the trans isomer before
extraction of the crude reaction residue with n-pentane (>90%
by integration of the in situ 31P NMR spectrum). After extraction
into n-pentane and crystallization, the product isolated in 56%
yield is at least 90% 2·trans (according to integration of the
1H NMR spectrum) (see Synthesis of 2; Method A, 2·trans,
1
Experimental section). The 31P{ H} and 1H NMR spectroscopic
data for 2·trans is similar to that of 1·trans (Fig. 3), being
characterised by two doublets at d 185.1 and 149.3 (2JPNP
=
39.6 Hz) (with the latter splitting into an apparent triplet in the
1H-coupled NMR spectrum), while the minor cis isomer (2·cis)
1
is characterised by two singlet resonances in the 31P{ H} NMR
Table 1 Relative energies of the HF/3-21G and DFT/6-31G* opti-
mised structures (all values are in kcal mol−1
)
spectrum (at d = 231.1 and 128.0). Like the 1·cis, 2·cis exhibits
two tBu resonances in the 1H NMR spectrum.
Level
1·cis
1·trans
2·cis
2·trans
The unexpected formation predominantly of 1·cis with the
1-O-2-NH-C6H4 spacer and the almost exclusive formation of
2·trans with the closely related 1-S-2-NH-C6H4 spacer led us
to wonder what thermodynamic and/or kinetic factors were
at work in the selection of the cis and trans isomers in these
reactions. Ab initio calculations in the gas phase [HF/3-21G)
and DFT B3LYP/6-31G*] and in H2O and THF solvents [DFT
B3LYP/6-31G*] were undertaken in an attempt to address this
issue. The gas-phase optimised structures of 1·cis and trans and
HF/3-21G
3.27
0.00
2.69
0.00
B3LYP/6-31G*
Gas phase
H2O
Thf
1.96
3.82
3.32
0.00
0.00
0.00
0.00
0.00
0.00
0.51
1.97
1.12
Fig. 3 1H and 31P NMR assignments for the cis and trans isomer of 2 (the N–H resonance for 2·cis was not observed).
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3
1 7 6 5

View Article Online
Fig. 4 B3LYP/6-31G* optimised structures of (a) 1·cis, (b) 1·trans, (c) 2·cis, (d) 2·trans.
3 and 4 are characterised by singlets in their 31P{ H} NMR
spectra (at d = 144.8 and 221.2, respectively). Significantly, no
further coupling to any N–H protons is observed in the fully-
coupled 31P NMR spectra of both compounds and no splitting
of the N–H resonances is seen in the 1H NMR spectra at room
temperature. The fact that the O and S atoms of the 1-O-2-NH2-
C6H4 and 1-S-2-NH2-C6H4 ligands in 3 and 4 are bonded to the
P centres (not the NH) is confirmed further by the observation
of two (symmetric and asymmetric) N–H stretching bands in
the IR spectra of solid 3 and 4 (at 3487 and 3390 and 3466
and 3368 cm−1, respectively), showing that the NH2 groups are
retained in both complexes.
1
thermodynamic control of the reaction product. Instead, kinetic
factors are likely to be decisive in the selection of the isomers.
Further support for this conclusion is suggested by the fact that
1·cis is calculated to be less stable than 1·trans at both levels of
theory in the gas phase and in solution, yet the cis isomer is the
predominant species formed in the reaction of [ClP(l-NtBu)]2
with 1-OH-2-NH2-C6H4.
A possible explanation for the formation of cis and trans iso-
mers of 1 and 2 is shown in Scheme 1. The 1 : 1 reaction of [ClP(l-
NtBu)]2 with 1-YH-2-NH2-C6H4 (Y = O, S) initially gives the
monosubstituted intermediate [ClP(l-NtBu)(1-Y-NH2-C6H4)]
which can cyclise directly into the trans isomer. However, the
predominant formation of 1·cis in the reaction of 1-OH-2-
NH2-C6H4 with [ClP(l-NtBu)]2 suggests that cyclisation occurs
via the symmetric disubstitution product [{P(l-NtBu)}2(1-O-2-
NH2-C6H4)2] (3), which preorganises the cyclisation step for the
formation of the cis isomer on reaction with another molecule of
[ClP(l-NtBu)]2. The predominant formation of the trans isomer
of 2 and the cis isomer of 1 in these reactions could be accounted
for by the different relative rates of the reactions involved with
Y = O or S.
The reaction of 3 with [ClP(l-NtBu)]2 in the presence of excess
Et3N was monitored in an in situ 31P NMR study. After ca.
3 h at reflux, only the cis isomer is observed together with
unreacted 3 and [ClP(l-NtBu)]2. The final reaction solution
obtained after 16 h reflux consists of a ca. 85 : 15 mixture
of the cis isomer to the trans (determined by integration of
the 31P NMR spectrum). This distribution can be compared
to the 60 : 40 mixture of the cis to trans isomers obtained
in the original 1 : 1 stoichiometric reaction of [ClP(l-NtBu)]2
with 1-O-2-NH-C6H4 (also determined by integration of the 31
NMR spectrum). Thus, it appears that the bias towards the cis
isomer might indeed be due to the formation of intermediate 3
in this reaction. Although less dramatic, the analogous reaction
P
In order to provide a partial test of this hypothesis, the model
intermediates 3 and 4 were prepared by the 1 : 2 reactions
of [ClP(l-NtBu)]2 with 1-YH-2-NH2-C6H4 (Y = O, S) in the
presence of Et3N (in 57 and 43% yield, respectively) (Scheme 2).
1 7 6 6
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3

View Article Online
Scheme 1
Scheme 2
of the symmetric dimer 4 with [ClP(l-NtBu)]2 provides further
confirmation of this, giving a ca. 50 : 50 mixture of the cis and
trans isomers of 2 after 1 h reflux in toluene (together with
unreacted starting materials and other minor products). After
12 h reflux, mainly the trans isomer of 2 can be detected together
with a considerable amount of decomposition. This, however,
does not suggest that interconversion of the cis and trans isomers
occurs in this case, merely that the formation of the trans
isomer predominates with increased reaction time. Accordingly,
Fig. 5 Structure of 5.
1
31P{ H} NMR studies show that no interconversion of the cis
of the trans isomer of 1 in the reaction of 3 with [ClP(l-NtBu)]2.
However, the presence of two 1 : 1 doublets in the P(III) region
of the spectrum of 4 after reflux in THF for 16 h (d = 125.1 and
87.8) shows that a completely different rearrangement pattern
to that found for 3 is occurring. The coupling constant for these
and trans isomers of 1 and 2 occurs on prolonged reflux of
these compounds in toluene in the presence of Et3N or in
the presence of Et3NHCl. The formation of the trans isomers
of 1 (minor) and 2 (major) in the reactions involving 3 and
4 must therefore arise from rearrangement of the symmetric
frameworks of both intermediates during cyclisation (not by
interconversion of the cis and trans isomers of 1 and 2). We
find that samples of 3 and 4 do indeed exhibit extensive thermal
rearrangement, although NMR studies fail to provide definitive
explanations for the observed distribution of the isomers of 1
and 2 formed. In the case of 3, after 3 h reflux in THF the
previously reported P(V) compound [P(H){1-O-2-NH2-C6H4}2]
(5) is the only soluble P-containing species produced (Fig. 5).7
Compound 5 is characterised by a singlet at d = −47.7 in the
2
doublets is characteristic of a JPNP coupling (30.6 Hz). These
features suggest the formation of the asymmetrical species (6)
or cyclic monomer (7) (Scheme 3). The formation of 6 and/or
7 would provide a means of scrambling the stereochemistry of
4 from cis to trans. Scrambling of the ligands of 4 provides a
further explanation for the predominant formation of 2·trans
in the reaction of [ClP(l-NtBu)]2 with 1-SH-2-NH2-C6H4 (in
addition to that suggested earlier in Scheme 1). Attempts to
elucidate the reaction mechanism by in situ 31P and H NMR
proved inconclusive, owing to the presence of several species in
the reaction mixture.
1
1
31P{ H} NMR spectrum at room temperature, which splits into
a doublet of triplets in the fully-coupled spectrum [1JP–H
=
2
1
827.4, JP–H = 18.1 0.3 Hz]. The H NMR spectrum of this
solution shows doublets at d 8.70 (1JP–H = 829.2 Hz, P–H) and
d 4.41 (2JP–H 18.8 Hz, N–H).8 Further reflux of this solution
(up to 16 h) results in more of 5 being generated, together
with minor amounts of 1·cis. Although of interest, the thermal
decomposition characteristics of 3 do not explain the formation
Further evidence of the ability for rearrangement of the
intermediates 3 and 4 is seen in their reactions with nBuLi
followed by reaction with [ClP(l-NtBu)]2. A similar synthetic
approach has been used to obtain the dimeric macrocyle
[{P(l-NtBu)}2(l-NtBu)]2 via deprotonation of [tBuN(H)P(l-
NtBu)]2 with BuLi followed by reaction with [ClP(l-NtBu)]2
n
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3
1 7 6 7

View Article Online
Scheme 3
Scheme 4
(Scheme 4).9 Surprisingly, in situ 31P NMR studies reveal that
the reactions involving 3 and 4 in this way generate exclusively
the trans isomer 1·trans and the cis isomer 2·cis, respectively
(see Synthesis of 1 and 2, Methods B, Experimental section).
1·trans and 2·cis are isolated in 67 and 40% yields (respectively)
after extraction of the crude residues with n-pentane. This is
the opposite to the product distribution found for either the
reactions of [ClP(l-NtBu)]2 with the diacids 1-OH-2-NH2-C6H4
and 1-SH-2-NH2-C6H4 or with 3 and 4, in which the cis isomer
of 1 and the trans isomer of 2 are the major products. We assume
that the formation of these isomers occur via rearrangement of
the intermediate dianions according to Scheme 5. Unfortunately,
we have been unable to obtain unambiguous evidence for the
nature of the intermediates in these reactions from in situ 31P or
7Li NMR studies or to isolate the dilithiates of 3 and 4 in pure
and/or crystalline form. An explanation for the formation of the
other isomers of 1 and 2 by this method is therefore not obvious,
but it is reasonable to assume that the Li+ cations play a role in
preorganisation and/or templating of the respective isomers.
Despite repeated attempts to obtain the crystal structures of
both the cis and trans isomers of 1 and 2 from various solvents
using various conditions, good quality crystals of 1 could only
be obtained as a 1 : 1 co-crystal of the cis and trans isomers
[1·cis/trans (50 : 50)] [obtained from Synthesis of 1; Method A,
see Experimental section] and 2 could only be crystallised as the
cis isomer (2·cis) [obtained from Synthesis of 2; Method A, see
Experimental section]. The low-temperature X-ray structures of
compounds 1·cis/trans (50 : 50), 2·cis, 3 and 4 were obtained.
Details of the data collections and structure refinements are
given in Table 2. Selected bond lengths and angles for 1·cis/trans
(50 : 50), 2·cis, 3 and 4 are listed in the footnotes to Figs. 5, 6, 7
and 8, respectively.
The X-ray structural study of 1·cis/trans (50 : 50) shows a
centrosymmetric macrocyclic structure (Fig. 6). This arises from
a random distribution of the cis and trans isomers throughout
the lattice, so that in the observed structure the carbon atoms of
both isomers occupy the same sites but the O and N atoms of the
bridging 1-O-2-N-C6H4 ligand are disordered. The disorder was
modeled as a 50 : 50 mixture of the two isomers, with the trans-
isomer in one orientation superimposed with two orientations
of the cis-isomer [Fig. 6(a) shows the trans isomer and 6(b)
shows one orientation of the cis-isomers]. It can be noted that
the bulk composition of the sample from which the 50 : 50 co-
crystal [1·cis/trans (50 : 50)] was obtained was confirmed to be
1
70 : 30 cis:trans in a separate H NMR study (consistent with
previous studies of the product of Method A, see Experimental
section). The fact that the crystal composition is different from
the bulk composition is not contradictory; it is simply that the
only material that is sufficiently crystalline for X-ray analysis is
the 50 : 50 co-crystal (with the rest of the material being rich
in the cis isomer). This is further supported by recollection of
the X-ray data on a separate crystal and by obtaining unit cell
parameters on other crystals from this sample [all of which are
1·cis/trans (50 : 50)].
The overall structure of 1·cis/trans is very similar to that of the
closely related macrocycle [{P(l-NtBu)}2{1,2-(NH)2-C6H4}]2.3
Like the latter, the two parallel phenyl rings of the 1-O-2-N-C6H4
˚
ligands in 1 are almost co-planar (to within 0.1 A) and the O
and N atoms are orientated endo with respect to the macrocyclic
˚
cavity. The macrocylic core of 1 measures ca. 5.5 A between the
centroids of the two P2N2 ring units and is slightly greater than
t
˚
that found in [{P(l-N Bu)}2{1,2-(NH)2-C6H4}]2 (ca. 5.3 A). The
displacement of one of the tBu groups out of the mean plane of
the [P(l-NtBu)]2 ring units appears to be steric in origin.
1 7 6 8
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3

View Article Online
Scheme 5
The structure of 2·cis shows a far more distorted arrangement
than is found for 1·cis/trans (Fig. 7). In particular, the mean
planes of the two P2N2 ring units are inclined with respect to
each other by an angle of 43.4◦ rather than being parallel as in
the structure of 1. The mean planes of the phenyl rings of 2 are
also not parallel, being inclined by ca. 25◦ with respect to each
other (cf. the parallel almost co-planar arrangement of the rings
in 1).
The molecular structures of the dimer precursors 3 (Fig. 8(a))
and 4 (Fig. 9) are similar to each other, having chemically
Table 2 X-Ray crystallographic and data processing parameters for complexes 1·cis/trans (50 : 50), 2·trans, 3, and 4a
1
2
3
4
Formula
M
C28H46N6O2P4
622.59
Triclinic
¯
P1
C28H46N6P4S2
654.71
Triclinic
¯
P1
C20H30N4O2P2
420.42
Triclinic
¯
P1
C20H30N4P2S2
452.54
Monoclinic
Crystal system
Space group
Unit cell dimensions
P21/n
˚
a/A
8.8285(18)
10.307(2)
10.362(2)
68.10(3)
73.40(3)
75.00(3)
825.9(3)
1
1.252
0.263
9.0091(18)
9.3028(19)
20.968(4)
87.66(3)
89.25(3)
79.64(3)
1727.2(6)
2
1.259
11.308(2)
13.578(3)
15.342(3)
103.49(3)
96.66(3)
91.07(3)
2272.5(8)
4
14.756(3)
9.0309(18)
17.947(4)
—
97.74(3)
—
2369.9(8)
4
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
3
˚
U/A
Z
Dc/Mg m−3
1.229
0.213
1.268
0.373
l/mm−1
0.367
F(000)
332
696
896
960
Crystal size/mm
0.23 × 0.12 × 0.10
3.52 to 21.00
−8/8, −10/10, −10/10
6042
1767 (0.057)
1767/ 1 /196
1.202
0.42 × 0.18 × 0.05
3.55 to 27.487
−11/11, −12/12, −26/27
22098
7837(0.0472)
7837/ 0/381
1.038
0.28 × 025. × 0.16
3.55 to 23.00
−12/12, −14/14, −16/16
19081
6291 (0.0768)
6291/2/550
1.064
0.44 × 0.32 × 0.05
3.53 to 27.49
−19/18, −10/11, −17/23
17832
5395 (0.0559)
5395/10/272
1.032
h-range/◦
Limiting hkl indices
Reflections collected
Independent reflections (Rint
Data/restraints/parameters
Sb on F2
)
Final R indicesb
[I > 2r(I)]
R1 = 0.063, wR2 = 0.173 R1 = 0.0443, wR2 = 0.1021 R1 = 0.0523, wR2 = 0.1427 R1 = 0.0576, wR2 = 0.1450
R1 = 0.071, wR2 = 0.175 R1 = 0.0745, wR2 = 0.1142 R1 = 0.0886, wR2 = 0.1642 R1 = 0.0988, wR2 = 0.1696
(All data)
Weights a, bb
0.0409, 3.5315
0.450 and −0.315
0.0458, 0.7090
0.303, −0.358
0.1058, 0.0000
0.607, −0.518
0.0759, 2.0077
1.083, −0.326
−3
˚
Max and min Dq/e A
Data in common k = 0.7103 A, T = 180(2)K. b S = [Rw(Fo − Fc2)2/(n − p)]2 where n = number of reflections and p = total number of parameters,
a
2
˚
R 1 = R ꢂFo − |Fcꢂ/R|Fo|, wR2 = R [w(Fo − Fc2)2]/R[w(Fo2)2]2, w−1 = [r2(Fo)2 + (aP)2 + bP], P = [max(Fo2,0) + 2(Fc2)]/3.
3
2
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3
1 7 6 9

View Article Online
Fig. 6 Structure of the macrocycle 1·cis/trans (50 : 50), showing (a) the trans-isomer and (b) one o◦f the two orientations of the cis-isomers.
˚
H-atoms (except those attached to N) have been omitted for clarity. Selected bond lengths (A) and angles ( ); P(1)–N(1) 1.707(5), P(1)–N(2) 1.709(5),
P(1)–N(3) 1.694(5), P(2)–N(1) 1.691(5), P(2)–N(2) 1.714(5), P(1) · · · P(2) 2.580(2), C(31)–N(3) 1.399(6), C(36)–O(1A) 1.390(7), P(2A)–O(1A) 1.687(5),
N(1)–P(1,2)–N(2) mean 81.6, P(1)–N(1,2)–P(2) mean 98.4, N(1,2)–P(1)–N(3) mean 102.4, N(1,2)–P(2)–O(1) mean 101.8, P(1)–N(3)–C(31) 123.9(3),
C(36)–O(1A)–P(2A) 122.7(4). (Symmetry transformations used to generate equivalent atoms A: 2 − x, 1 − y, 2 − z).
˚
Fig. 7 Structure of the macrocycle 2·cis. H-atoms (except those attached to N) have been omitted for clarity. Selected bond lengths (A) and angles
(◦); P(1)–N(1) 1.713(2), P(1)–N(2) 1.717(2), P(1)–S(1) 2.176(1), P(2)–N(1) 1.710(2), P(2)–N(2) 1.706(2), P(2)–S(2) 2.188(1), P–N(1,2)–P mean 98.07,
N–P(1,2)–N mean 81,9, N(1,2)–P(1,2)–S(1,2) mean 102.6, C(51)–S(1)–P(1) 105.09(7), C(61)–S(2)–P(2) 101.22(8), P(3)–N(3) 1.728(2), P(3)–N(4)
1.709(2), P(3)–N(5) 1.700(2), P(4)–N(3) 1.721(3), P(4)–N(4) 1.708(2), P(4)–N(6) 1.703(2), N(3,4)–P(3,4)–N(5,6) mean 102.3, C(66)–N(5)–P(3)
130.7(2), C(56)–N(6)–P(4) 131.8(2).
symmetrical molecules in which the O-centers of the two 1-
NH2-2-O-C6H4 groups (in 3) and the S-centers of 1-NH2-2-S-
Conclusions and closing remarks
The current study concerns the first report of the devel-
opment of regioselective synthetic routes for the synthe-
sis of asymmetrically-bridge macrocycles of the type [{P(l-
NtBu)}2(LLꢀ)]2 (where LLꢀ is a bifunctional organic bridge). The
formation of such species and the selectivity of the reactions
forming them is highly dependent on the specific reaction route
employed and apparently on the combination of heteroatoms
present in the bridging group (LLꢀ). Overall, the synthetic
(as well as the theoretical) results of this study illustrate that
the products are kinetically rather than thermodynamically
controlled. To summarise these findings, the cis isomer of 1
(1·cis) is obtained as the major product in the direct reaction of
[ClP(l-NtBu)]2 with 1-OH-2-NH2-C6H4 or 3; the latter giving
the greatest selectivity of ca. 85 : 15 cis to trans isomer. The trans
isomer (1·trans) is generated exclusively in the deprotonation of
C6H4 (in 4) are bonded to the two P-centers of the respective
P2N2 core. The pendant NH2 groups are not involved in any
intramolecular donor or H-bonding interactions in either com-
pound. However, in the case of 3 the two crystallographically-
independent molecules in the asymmetric unit, which differ
in the orientation of the phenylene rings, are linked to each
other by a H-bonding interaction involving two NH2 groups. A
further interaction involving the remaining two NH2 groups, of
independent molecules in adjacent unit cells, leads to a helical
H-bonded polymer propagated along the diagonal of the unit
cell. Consequently the four NH2 groups of the two independent
molecules act either as H-bond donors [in the cases of N(5)
and N(8)] or H-bond acceptors [in the cases of N(6) and N(7)].
The two N–H · · · N interactions involved are within the range of
values expected for hydrogen-bonding; N(5)–H · · · N(7) 2.35(4)
[155◦] [N(5) · · · N(7) 3.174(5)]; N(8)–H · · · N(6A) 2.24(2) [176◦]
[N(8) · · · N(6A) 3.172(4)].
n
3 with BuLi followed by cyclisation of the presumed dianion
with [ClP(l-NtBu)]2. In contrast, the trans isomer of 2 (2·trans) is
1 7 7 0
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3

View Article Online
Fig. 8 (a) Structure of 3, the dimeric precursor to 1·cis, (b) association of molecules in the la◦ttice parallel to the unit cell diagonal. H-atoms (except
˚
those attached to N) have been omitted for clarity. Selected bond lengths (A) and angles ( ); P(1)–O(1) 1.652(2), P(1)–N(1) 1.705(3), P(1)–N(2)
1.714(2), P(2)–O(2) 1.660(2), P(2)–N(1) 1.703(2), P(2)–N(2) 1.700(3), P(3)–O(5) 1.661(2), P(3)–N(3) 1.702(2), P(3)–N(4) 1.710(3), P(4)–O(6) 1.674(2),
P(4)–N(3) 1.697(3), P(4)–N(4) 1.713(2), N–P–N mean 81.4, P–N–P mean 97.7, N–P–O mean 103.6; (b) N(5)–H · · · N(7) 2.35(4) [155] [N(5) · · · N(7)
3.174(5)], N(8)–H · · · N(6A) [176] [N(8) · · · N(6A) 3.172(4)] (Symmetry transformations used to generate equivalent atoms x + 1, y − 1, z).
˚
Fig. 9 Structure of 4, the dimeric precursor to 2·cis. H-atoms (except those attached to N) have been omitted for clarity. Selected bond lengths (A) and
angles (◦); P(1)–N(1) 1.710(3), P(1)–N(2) 1.704(3), P(1)–S(2) 2.1783(12), P(2)–N(1) 1.711(2), P(2)–N(2) 1.705(3), P(2)–S(2) 2.1814(12), P(1) · · · P(2)
2.594(1), C(31,41)–S(1,2) mean 1.77, C–N(3,4) mean 1.32, N–P(1,2)–N mean 81.06, P(1)–N(1,2)–P(2) mean 98.9, N(1,2)–P(1,2)–S(1,2) mean 104.3,
P(1,2)–S(1,2)–C(31,41) mean 99.5.
generated as the major product of the reaction of [ClP(l-NtBu)]2
with 1-SH-2-NH2-C6H4 or with 4. However, the cis isomer (2·cis)
is the only product of the deprotonation of 4 with nBuLi followed
by cyclisation of the presumed dianion with [ClP(l-NtBu)]2.
A particularly intriguing observation is the fact that the
reactions of 3 and 4 with [ClP(l-NtBu)]2 give 1·cis and 2·trans,
respectively, whereas the reactions of 3 or 4 with nBuLi/[ClP(l-
NtBu)]2 give the alternative isomers 1·trans and 2·cis. This switch
in the isomers 1 and 2 is not easily explained without obtaining
structures of the dilithiates of 3 and 4, although it seems likely
that the nature of the coordination of Li+ within these species will
be important in dictating the outcome of the reactions. Further
studies in this area are underway.
with a Belle Technology O2 and H2O internal recirculation
system. Melting points were determined using a conventional
apparatus and sealing samples in capillaries under N2. IR spectra
were recorded as Nujol mulls using NaCl plates and were run
on a Perkin-Elmer Paragon 1000 FTIR spectrophotometer.
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. 1H, 31P NMR spectra were recorded on a Bruker ATM
DRX500 spectrometer, in dry deuterated CDCl3 (using the
1
solvent resonances as the internal reference standard for H
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. The relative
amounts of cis and trans isomers prior to extraction with n-
Experimental
General experimental
pentane or toluene were determined from integration of the 31
NMR spectra.
P
Compounds 1–4 were prepared under dry, O2-free N2 on a
vacuum line. [ClP(l-NtBu)]2 was obtained from the 3 : 1
t
reaction of BuNH2 with PCl3, respectively, in THF. THF,
Synthesis of 1; Method A, 1·cis/trans (70 : 30). To a solution
of 2-aminophenol (0.38 g, 3.5 mmol) and Et3N (1.95 mL,
14.0 mmol) in THF (30 mL) at −78 ◦C was added dropwise
a solution of [ClP(l-NtBu)]2 (0.96 g, 3.5 mmol) in THF (30 mL)
toluene and n-pentane were dried by distillation over sodium or
sodium/benzophenone prior to the reactions. 1–4 were isolated
and characterised with the aid of an N2-filled glove box fitted
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3
1 7 7 1

View Article Online
over the course of 1 h. The mixture was allowed to warm to
room temperature and then brought to reflux (14 h). The solvent
was removed under vacuum and the residue extracted with n-
pentane (60 mL) and filtered (Celite, P3). The volume of the
resulting pale yellow solution was reduced under vacuum until
precipitation occurred. The precipitate was redissolved by the
addition of THF (2 mL) and heating. Storage (−30 ◦C, 24 h) gave
colourless crystals of 1. Yield 0.62 g (58%). Mp 298–300 ◦C. IR
(NaCl, Nujol), m/cm−1 = 3480 (w), 3345 (w) (N–H str.), 1595
NMR (500.20 MHz, +25 ◦C, CDCl3), d = 7.46 (d, trans isomer),
7.45 (d, J = 7.1 Hz, cis isomer), 7.21 (dd, JHH = 7.9, 1.3 Hz, cis
isomer), 6.90 (dt, JHH = 7.7, 1.3 Hz, cis isomer), 6.82 (dt, trans
isomer), 6.70 (dt, trans isomer), 6.68 (dt, JHH = 7.7, 1.5 Hz, cis
isomer), 6.53 (br. dd, JHH = 7.8 Hz, trans isomer), 6.42 (brs.,
N–H cis isomer), 4.89 [d, N–H (trans-isomer), 2JPH = ca. 30 Hz],
1.32 (s, tBu), 1.27 (s, tBu) (1 : 1, cis isomer), 1.09 (s, trans isomer)
(cis : trans isomers ca. 70 : 30 on the basis of integration of the
filtered (Celite, P3) and the solvent removed under vacuum to
1
give a white powder, identified as 2·cis by H and 31P NMR
spectroscopy (Yield 0.125 g, 40%). 1H NMR (500.20 MHz,
+25 ◦C, CDCl3), d = 7.60 (dd, J = 7.7, 1.5 Hz), 7.45 (br. d, J =
7.7 Hz), 7.25 (dt, 7.7, 1.5 Hz), 6.70 (dt, J = 7.7, 1.5 Hz), 1.40
(s, tBu), 1.37 (s, tBu) (the N–H proton could not be observed).
31P{ H} NMR (202.48 MHz, +25 ◦C, CDCl3), d = 231.1 (s),
1
128.0 (s).
Synthesis of 3. To a solution of 2-aminophenol (0.76 g,
7.0 mmol) and Et3N (0.98 mL, 7.0 mL) in toluene (20 mL)
at −78 ◦C was added dropwise to a solution of [ClP(l-NtBu)]2
(0.96 g, 3.5 mmol) in toluene (20 mL). The mixture was allowed
to warm to room temperature and was then stirred (16 h). The
mixture was then filtered (Celite, P3) and the filtrate evaporated
to dryness under vacuum. The residue was crystallised from n-
1
(w) (C–C str.), 1099 (s), 1043 (s), 894 (m), 798 (s), 746 (s). H
◦
pentane (15 mL)/THF (2 mL) at −20 C (20 h). Yield 0.84 g
◦
(57%). Mp 94–96 C. IR (NaCl, Nujol), m/cm−1 = 3487 (w),
◦
3390 (w) (sym., asym. N–H str.), other bands at 1300 (m), 1195
(w), 1097 (w), 1034 (s), 914 (w), 868 (m◦), 795 (s), 741 (s), 722 (w),
697 (m). 1H NMR (500.20 MHz, +25 C, CDCl3), d = 7.70 (dd,
1
tBu resonances). 31P{ H} NMR (202.48 MHz, +25 C, CDCl3),
2
2
d = 173.8 (d, P–O, JPNP = 30 Hz), 148.4 (d, P–NH, JPNP
=
30 Hz; splits into an apparent triplet in the H-coupled spectrum,
J = 29.0 Hz), 171.1 (s), 133.7 (s). Elemental analysis, found C
52.6, H 7.7, N 13.1, P 18.8, calc. for 1 C 54.0, H 7.4, N 13.5, P
19.9. (NB. Extraction of the crude reaction residue with toluene
(60 mL) instead of n-pentane gives pure 1·cis).
JHH = 8.0, 1.3 Hz), 6.97 (dt, JHH = 7.7, 1.3 Hz), 6.81 (dt, JHH
=
7.6, 1.6), 6.60 (dd, JHH = 7.8, 1.6), 3.61 (s, 4H, N–H), 1.39 (s,
18H, tBu). 31P NMR (202.48 MHz, +25 ◦C, CDCl3), d = 144.8
(s). Elemental analysis, found C 56.6, H 7.1, N 13.1, calc. for 3
C 57.1, H 7.2, N 13.3.
Method B, 1·trans. To a solution of 3 (0.84 g, 2.0 mmol) in
THF (20 mL) at −78 ◦C was added nBuLi (2.75 mL, 1.6 mol L−1,
4.4 mmol) over the course of 5 min and the resulting bright
yellow solution stirred for ca. 30 min at −78 ◦C. A solution of
[ClP(l-NtBu)]2 (0.55 g, 2.0 mmol) in THF (20 mL) was added
dropwise and the mixture allowed to warm to room temperature
and stirred overnight. The solvent was removed under vacuum
and the residue extracted with toluene (60 mL) and filtered
(Celite, P3). The solvent was removed under vacuum to give
Synthesis of 4. To
a solution of 2-aminothiophenol
(0.21 mL, 2.0 mmol) in toluene (30 mL) at −78 ◦C was
added Et3N (0.28 mL, 2.0 mmol). A solution of [ClP(l-NtBu)]2
(0.275 g, 1.0 mmol) in toluene was added dropwise. The reaction
was allowed to warm to room temperature and stirred (16 h)
before being filtered (Celite, P3) to remove the precipitate of
Et3NHCl. The volume of the filtrate was reduced under vacuum
to ca. 2 mL. Storage at −5 ◦C (24 h) gave colourless crystals of 4.
Yield 0.22 g (43%). Mp 129–130 ◦C. IR (NaCl, Nujol), m/cm−1
=
1
a pale yellow powder, identified as the 1·trans by H and 31P
3466 (w), 3368 (w) (sym., asym. N–H str.), other bands at 1308
(m), 1246 (W), 1200 (m), 1157 (w), 1076 (w)◦, 1024 (s), 886 (s), 798
(w), 749 (m). 1H NMR (500.20 MHz, +25 C, CDCl3), d = 7.75
[dd, 2H, JH–H = 7.7, 1.5 Hz], 7.09 [dt, 2H, JH–H = 7.64, 1.5 Hz],
6.80 [dt, 2H, JH–H = 7.5, 1.4 Hz], 6.50 [dd, 2H, JH–H = 8.0, 1.2 Hz],
NMR spectroscopy (Yield 0.41 g, 67%).
Synthesis of 2; Method A, 2·trans. To a solution of 2-
aminothiophenol (0.21 mL, 2.0 mmol) and Et3N (1.1 mL,
8.0 mmol) in THF (30 mL) at −78 ◦C was added dropwise a
solution of [ClP(l-NtBu)]2 (0.55 g, 2.0 mmol) in THF (30 mL)
over the course of 1 h. The mixture was allowed to warm to
room temperature and then brought to reflux (14 h). The solvent
was removed under vacuum and the residue extracted with n-
pentane (60 mL) and filtered (Celite, P3). The volume of the
resulting pale yellow solution was reduced under vacuum until
precipitation occurred. The precipitate was redissolved by the
3.95 (s, 4H, N–H), 1.46 (s, 18H, Bu). 31P NMR (202.48 MHz,
t
+25 ◦C, CDCl3), d = 221.2 (s.). Elemental analysis, found C 52.9,
H 6.6, N 12.3, P 14.0, calc. for 4 C 53.1; H 6.7; N 12.4; P 13.7.
X-Ray crystallographic studies of 1–4
Crystals of 1·cis/trans, 2·cis, 3 and 4 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 collections and structural refinements are
given in Table 1. The structures were solved by direct methods
and refined by full-matrix least squares on F2.10 The structure
of 1·cis/trans was solved in P-1 with half a molecule per
asymmetric unit, which would indicate a trans macrocycle.
However, when the oxygen and nitrogen atoms of the linker
[N(3) and O(1)] were refined, very unequal atomic displacement
parameters were observed indicating mixed atom types on these
positions. The occupancies of each position were adjusted so
that within the lattice there is a ratio of 50% trans-isomer with
25% each of two orientations of the cis-isomer. This produced
more reasonable atomic displacement parameters as well as
improved R-values. The hydrogen atom on N(3) (occupancy
◦
addition of THF (2 mL) and heating. Storage (−30 C, 24 h)
gave colourless crystals of 2. Yield 0.36 g (56%). Mp 187 ◦C. IR
(NaCl, Nujol), m/cm−1 = 3372 (w), 3286 (w) (N–H str.), 1582
(m) (C–C str.), 1087 (m), 1034 (s), 901 (s), 871 (s), 798 (s), 743 (s)
721 (m). 1H NMR (500.20 MHz, +25 ◦C, CDCl3), d = 7.85 (br.
d, J = 7.6 Hz), 6.93 (br. dt, J = 7.6, 1.3 Hz), 6.71 (br. dt, J = 7.6,
1.3 Hz), 6.52 (d, J = 7.6), 5.06 [d, N–H, 2JPH = ca. 32.9 Hz], 1.24
◦
1
(s, tBu). 31P{ H} NMR (202.48 MHz, +25 C, CDCl3), d = 185.1
(d, P–S, 2JPNP = 39.6 Hz), 149.3 (d, P–NH, 2JPNP = 39.6 Hz; splits
into an apparent triplet in the H-coupled spectrum, J = 35.7 Hz)
[trans isomer], 231.1 (s), 128.0 (s) [cis isomer] (>90% trans isomer
by integration of the 31P NMR spectrum). Elemental analysis,
found C 51.1, H 7.0, N 12.8, calc. for 2 C 51.4, H 7.1, N 12.8.
Method B, 2·cis. To a solution of 4 (0.226 g, 0.5 mmol) at
◦
n
0 C was added BuLi (0.62 mL, 1.0 mmol) in THF (15 mL).
The solution was warmed to room temperature and stirred
for 30 min. The resulting clear yellow solution was cooled to
−78 ◦C and [ClP(l-NtBu)]2 (0.138 g, 0.5 mmol) in THF (15 mL)
was added dropwise. The mixture was allowed to warm to
room temperature and stirred overnight. The solvent was then
removed under vacuum and the white powdered residue was
extracted with n-pentane (40 mL). The suspension was then
˚
0.75) was directly located and constrained to be 0.9 A from N(3).
All other hydrogen atoms were included in idealized positions,
including that on N(3ꢀ) (occupancy 0.25) which could not be
directly located. There is one molecule of compound 2 per
asymmetric unit and the calculated bond lengths, bond angles
and atomic displacement parameters indicate no disorder and
a cis-structure. The hydrogen atoms on N(5) and N(6) were
1 7 7 2
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3

View Article Online
directly located and included in the refinement. In compound
3 all eight hydrogen atoms on N(5), N(6), N(7) and N(8)
were directly located and included in the refinement (with
the bond lengths N(6)–H(6NA) and N(8)–H(8NB) constrained
A. A. Karasik, A. T. Gubaidullin, I. A. Litvinov, O. G. Sinyashin, P.
Loennecke and E. Hey-Hawkins, J. Chem. Soc., Dalton Trans., 2004,
442.
2 (a) A. Bashall, E. L. Doyle, C. Tubb, S. J. Kidd, M. McPartlin, A. D.
Wood and D. S. Wright, J. Chem., Soc. Chem. Commun., 2001, 2542;
(b) A. Bashall, A. D. Bond, E. L. Doyle, F. Garc´ıa, S. Kidd, G. T.
Lawson, M. McPartlin, M. C. Parry, A. D. Woods and D. S. Wright,
Chem. Eur. J., 2002, 8, 3377; (c) F. Garc´ıa, J. M. Goodman, R. A.
Kowenicki, I. Kuzu, M. McPartlin, M. A. Silva, L. Riera, A. D.
Woods and D. S. Wright, Chem. Eur. J., 2004, 10, 6066.
3 F. Garcia, R. A. Kowenicki, I. Kuzu, L. Riera, M. McPartlin and
D. S. Wright, J. Chem. Soc., Dalton Trans., 2004, 2904.
4 P. Kommana, K. V. P. P. Kumar and K. C. K. Swamy, Indian J. Chem.,
2003, 42A, 2371.
5 (a) See alsoS. S. Kumaravel, S. S. Krishnamurthy, T. S. Cameron
and A. Linden, Inorg. Chem., 1988, 27, 4546; (b) 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.
˚
to 0.9 A). In compound 4 the hydrogen atoms on N(3) and
N(4) were directly located, and were included in the refinement
with N–H lengths constrained to be equal, and displacement
parameters equal to 1.2 Ueq(N). In all four structures the
carbon-bonded H-atoms were included in idealized positions.
Anisotropic displacement parameters were assigned to all non-
hydrogen atoms in the final cycles of refinement.
CCDC reference numbers 259069 for 1·cis/trans; 259070 for
2·cis; 259071 for 3; 259072 for 4.
See http://www.rsc.org/suppdata/dt/b5/b502200b/ for cry-
stallographic data in CIF or other electronic format.
Acknowledgements
6 (a) W. J. Hehre, L. Radom, P. v. R. Schleyer, J. A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, New York, 1986; (b) A. D. Becke,
J. Chem. Phys., 1993, 98, 5648; (c) C. Lee, W. Yang and R. Parr, Phys.
Rev. B, 1988, 37, 785.
We thank the EPSRC (F.G., J.M.G., M.McP., D.S.W.), The
European Union (Fellowship for L.R., Erasmus grant for A.W.),
The States of Guernsey and The Domestic and Millenium Fund
(R.A.K.), Unilever (M.A.S.), and The Cambridge European
Trust (F.G.) for financial support. We also thank Dr J. E. Davies
(Cambridge) for collecting X-ray data for 1·cis/trans, 2·cis, 3
and 4.
7 P. L. Meunier, R. O. Day, J. R. Devillers and R. R. Holmes, Inorg.
Chem., 1978, 17, 3270; K. N. Gavrilov, A. V. Korostylev, A. I.
Polosukhin, O. G. Bondarev, A. Y. Kovalevsky and V. A. Davankov,
J. Organomet. Chem., 2000, 613, 148.
8 As part of these studies we have found that the reac-
tion of the dimer [ClP(l-NtBu)]2 with HOCH2CH2OH gives
a
mixture of the monomer [{P(l-NtBu)}2{OCH2CH2O}], the
dimer [{P(l-NtBu)} {OCH2CH2O}]2 and the P(V) compound
[P(H){OCH2CH2O}22].
References
1 For recent examples, seeR. N. Naumov, A. A. Karasik, O. G.
Sinyashin, P. Loennecke and E. Hey-Hawkins, J. Chem. Soc., Dalton
Trans., 2004, 357; A. S. Balueva, R. M. Kuznetsov, S. N. Ignat’eva,
9 J. K. Brask, T. Chivers, M. L. Krahn and M. Parvez, Inorg. Chem.,
1999, 38, 290.
10 G. M. Sheldrick, SHELX-97, Go¨ttingen, Germany, 1997.
D a l t o n T r a n s . , 2 0 0 5 , 1 7 6 4 – 1 7 7 3
1 7 7 3