Titanium complexes of bis(1°- amido)cyclodiphosph(III)azanes and bis(1°-amido)cyclodiphosph(V)azanes: Facial versus lateral coordination

Article abstract of DOI:10.1039/b007877h

Bis(1°-amino)cyclodiphosph(III)azanes and bis(1°-amino)cyclodiphosph(V)azanes show different coordination preferences with titanium(IV). The bis(tert-butylamino)cyclodiphosph(III)azane cis-[Bu^t(H)N(Bu^tNP)₂N(H)Bu^t

Full text of DOI:10.1039/b007877h

View Article Online / Journal Homepage / Table of Contents for this issue
Titanium complexes of bis(1Њ-amido)cyclodiphosph(III)azanes and
bis(1Њ-amido)cyclodiphosph(V)azanes: facial versus lateral
coordination†
Daniel F. Moser,a Christopher J. Carrow,a Lothar Stahl*a and Richard J. Staplesb
a Department of Chemistry, University of North Dakota, Grand Forks, ND 58202-9024, USA.
E-mail: Lothar.Stahl@mail.chem.und.nodak.edu
b Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA 02138,
USA
Received 28th September 2000, Accepted 12th February 2001
First published as an Advance Article on the web 28th March 2001
Bis(1Њ-amino)cyclodiphosph()azanes and bis(1Њ-amino)cyclodiphosph()azanes show different coordination
preferences with titanium(). The bis(tert-butylamino)cyclodiphosph()azane cis-[But(H)N(ButNP)2N(H)But]
(1) reacts with TiCl4 to afford {[(ButNP)2(NBut)2]TiCl2} (2), in which the ligand chelates the metal as a
diamide, above the heterocycle. Oxidations of 1 with phenyl- or p-tolyl azide yield the cyclodiphosph()azanes
cis-[But(H)N(ArN᎐PNBut) N(H)But] (Ar = phenyl (3), p-tolyl (4)). Single-crystal X-ray studies of 3 and 4
᎐
2
reveal structures similar to those of their parent cyclodiphosph()azanes, but with pendent arylimino
groups attached to the phosphorus() atoms. When 3 and 4 are allowed to react with TiCl4 the complexes
{[But(H)N(ArN᎐PNBut) NBut]TiCl } (Ar = phenyl (5), p-tolyl (6)) are isolated. In these compounds the
᎐
2
3
cyclodiphosph()azanes coordinate the metal laterally as monoanionic N–P–N ligands, similar to phosphoranates.
An analogous side-on coordination is also found for the dichalcogeno complexes {[But(H)N(E᎐PNBut) NBut]TiCl }
᎐
2
3
(E = S (9), Se (10)).
cannot be used for the synthesis of the corresponding titanium
complex, because the dianion reduces the metal to the trivalent
state.5 The titanium–chloride bond of TiCl4, however, is polar
enough to react with neutral 1,6 Scheme 1, to yield the titanium
Introduction
Because of the importance of early-transition metal complexes
in homogeneous polyolefin catalysis, there has been a resurgent
interest in bis(amido) complexes of these metals.1 We previously
reported syntheses and structures of bis(tert-butylamido)cyclo-
diphosph()azane compounds of zirconium and hafnium,
which, when activated with methylaluminoxane (MAO), poly-
merize ethylene.2 Unlike their cyclodisilazane analogues,3
however, these catalysts quickly deactivate during realistic
polymerization conditions. The strongly Lewis-acidic catalyst,
or the cocatalyst, presumably coordinates the lone-pair
electrons of the phosphorus() atoms to initiate the destructive
ring-opening of the ligand.
Cyclodiphosph()azanes are more robust than their phos-
phorus() counterparts, and many of the former are either
known or easy to prepare from the latter by stereospecific
chalcogen oxidations.4 The phosphorus() heterocycles can
also be oxidized with organic azides—a reaction that has not
been reported previously. We had hoped that, just like chal-
cogen oxidation, the oxidation with organic azides would leave
the bis(amido) moiety of the ligand essentially unchanged,
while significantly increasing the kinetic and thermodynamic
stabilities of these ligands to ring opening. Below we report
syntheses and solid-state structures of bis(1Њ-amino)cyclodi-
phosph()azanes and their titanium complexes.
Scheme 1
1
compound 2 as an orange–red solid. The very simple H, 13C
and 31P NMR spectra of 2 suggested that, just like its zirconium
and hafnium analogues, it was C2v-symmetric in solution.
A single-crystal X-ray analysis, Fig. 1, confirmed this
spectroscopically assigned structure. Crystal data and selected
bond parameters for 2 are listed in Tables 1 and 2, respectively.
In the solid state, the molecule is a seco-heterocube of Cs
symmetry, the titanium atom being the recipient of a donor
bond from one of the ring-nitrogen atoms. All P–N bonds of
the three-coordinate nitrogen atoms have essentially identical
lengths (avg. 1.716(2) Å) while those of the four-coordinate
nitrogen atom are expectedly longer, 1.779(2) Å. The titanium
atom has idealized trigonal-bipyramidal coordination geom-
etry and forms two normal, 1.925(2) Å, and one longer N–Ti
bond, 2.267(2) Å, with the ligand. The slightly inequivalent
titanium–chloride bonds of 2.2538(8) and 2.2780(7) Å have
similar lengths as those in related compounds.7,8
Results
Bis(amido)cyclodiphosph()azane complexes of zirconium
and hafnium are readily prepared by the interaction of [(But-
NP)2(NButLiؒTHF)2] with MCl4 (M = Zr, Hf).2 This procedure
Due to their instability, homogeneous titanium() catalysts
like 2, are unsuitable for commercial olefin polymerization, but
are useful model systems. For complex 2 the main mode of
† Communicated in part: D. F. Moser and L. Stahl, Abstracts of
Papers, 218th National Meeting of the American Chemical Society,
Anaheim, CA, March 1999; INOR 248.
1246
J. Chem. Soc., Dalton Trans., 2001, 1246–1252
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b007877h

View Article Online
Table 1 Crystal data for compounds 2–5 and 9–10
2
3
4
5
9
10
Formula
M
C16H36Cl2N4P2Ti
465.23
C32H56N6OP2 C30H52N6P2
C31.5H51Cl3N6P2Ti
729.91
C23H45Cl3N4P2S2Ti C23H45Cl3N4P2Se2Ti
602.77
558.72
657.94
750.93
Crystal system
Space group
T/K
Monoclinic
P21/c (no. 14)
213
Triclinic
P1 (no. 2)
293
Monoclinic
P21/n (no. 14) P21/n (no. 14)
293
Monoclinic
Monoclinic
P21/m (no. 11)
293
Monoclinic
P21/m (no. 11)
213
¯
213
a/Å
b/Å
c/Å
α/Њ
9.9454(1)
15.1386(1)
15.9841(1)
9.882(2)
10.931(2)
18.481(4)
89.14(3)
78.07(3)
68.41(3)
1812.2(6)
2
12.144(3)
20.535(2)
13.454(1)
10.594(1)
24.816(3)
14.361(2)
9.130(1)
13.450(1)
14.142(1)
9.108(1)
13.416(2)
14.113(2)
β/Њ
93.614(1)
90.84(1)
97.151(4)
94.488(9)
94.208(3)
γ/Њ
U/Å3
2401.77(3)
4
0.720
11105
3354.8(9)
4
0.157
7190
5887
3746.1(8)
4
0.557
23081
1731.2(3)
2
0.700
4142
1719.9(4)
2
2.710
11152
Z
µ/mmϪ1
0.152
10728
Reflections collected
Independent reflections 4838 (Rint = 0.0292) 7478
7630 (Rint = 0.0778) 3177 (Rint = 0.0848) 3895 (Rint = 0.0275)
(Rint = 0.0393) (Rint = 0.0180)
0.1051
0.2424
R(F)a [I > 2σ(I)]
0.0404
0.0886
0.0501
0.1526
0.0504
0.1065
0.0480
0.1390
0.0336
0.0975
wR2(F2)b [all data]
¹
2
2
2
2
2
2
2
a R = Σ|Fo Ϫ Fc|/Σ|Fo|. b Rw = {[Σw(Fo Ϫ Fc) ]/[Σw(Fo ) ]} ; w = 1/[σ (Fo) ϩ (xP) ϩ yP] where P = (Fo2 ϩ 2Fc2)/3.

²
Table 2 Selected bond lengths (Å) and angles (Њ) for 2
Ti–Cl(1)
Ti–Cl(2)
Ti–N(1)
Ti–N(3)
Ti–N(4)
P(1)–N(1)
2.2780(7)
2.2538(8)
2.267(2)
1.927(2)
1.923(2)
1.781(2)
P(1)–N(2)
P(1)–N(3)
P(2)–N(1)
P(2)–N(2)
P(2)–N(4)
1.718(2)
1.711(2)
1.776(2)
1.716(2)
1.719(2)
Cl(1)–Ti–Cl(2)
N(3)–Ti–N(4)
Cl(1)–Ti–N(3)
Cl(2)–Ti–N(3)
99.20(3)
116.35(8)
99.41(6)
Cl(1)–Ti–N(4)
Cl(1)–Ti–N(1)
N(1)–P(1)–N(2)
P(1)–N(1)–P(2)
99.48(6)
160.63(5)
81.33(9)
96.15(9)
119.07(6)
cyclodiphosph()azanes 3 and 4 in very good yields. The
attempted oxidations of cyclodiphosph()azanes with the less
reactive trimethylsilyl azide, however, gave intractable mixtures.
Compounds 3 and 4 are formal dimers of the aminobis-
(imino)phosphoranes [(ButN᎐)(ArN᎐)PN(H)But] (Ar = phenyl,
᎐
᎐
p-tolyl) respectively. The isolation of trans-{(Me3Si)2N-
[Me SiN᎐P(NSiMe ) ] N(SiMe ) }, the product of such a
Fig. 1 Solid-state structure of 2.
᎐
3
3
2 2
3 2
dimerization, had been communicated previously, but no
structural details were given.11
deactivation appears to be the cycloreversion of the cyclo-
diphosph()azane, as suggested by the isolation of ligand
recombination products from the reaction of asymmetrically
substituted analogues of 2 with MAO.9
The 2,4-bis(phenylimino)cyclodiphosph()azane was initially
isolated as an oil, but on addition of tetrahydrofuran, it crystal-
lized as a THF solvate, viz., cis-{[But(H)N(PhN᎐PNBut) -
᎐
2
N(H)But]ؒTHF} (3). Even these crystals, however, were only
of marginal quality, and we therefore also synthesized the
2,4-bis( p-tolylimino)cyclodiphosph()azane (4), which was free
of solvent. This latter compound has the additional advantages
of providing an NMR “handle” and having simpler NMR
spectra. Both compounds were subjected to single-crystal X-ray
studies. Selected crystal data and bond parameters of 3 and
4 appear in Tables 1 and 3, respectively. These bis(1Њ-amino)-
bis(imino)cyclodiphosph()azanes differ only in the presence
of two methyl groups, and have thus virtually identical bond
parameters. For this reason only the better-refined structure of
4 is discussed below.
Because cyclodiphosph()azanes are thermodynamically
more stable than their phosphorus() counterparts, we have
begun using the phosphorus() molecules as ligands. Oxid-
ations of phosphorus() compounds with organic azides
(Staudinger reaction) are well known,10 but these transform-
ations have, to our knowledge, not been reported for cyclo-
diphosph()azanes. The treatment of 1 with aryl azides
(Scheme 2) yielded the corresponding cis-2,4-bis(arylimino)-
The thermal-ellipsoid plot of 4, Fig. 2, shows that azide oxid-
ation is stereospecific, and that the top half of the molecule has
retained its conformation, as we had hoped. All P–N bonds,
however, are about 0.035(2) Å shorter than those in the parent
molecule. The tert-butylimino groups are parallel with the
perfectly planar heterocycle, but both the tert-butylamino and
the p-tolylimino groups form obtuse angles with the ring, giving
the molecule a letter “H” shape. As in 1 the N–H groups are
endo, i.e., they lie above the (P–N)2 ring, while the aromatic
rings are tucked below it. The exocyclic P–N(amino) and
Scheme 2
J. Chem. Soc., Dalton Trans., 2001, 1246–1252
1247

View Article Online
Table 3 Selected bond lengths (Å) and angles (Њ) for 3 and 4
Table 4 Selected bond lengths (Å) and angles (Њ) for 5, 9 and 10
3
4
5
9
10
P(1)–N(1)
P(1)–N(2)
P(1)–N(3)
P(1)–N(5)
P(2)–N(1)
P(2)–N(2)
P(2)–N(4)
P(2)–N(6)
1.690(4)
1.685(4)
1.621(5)
1.510(4)
1.674(4)
1.691(3)
1.621(4)
1.536(4)
1.689(2)
1.693(2)
1.630(2)
1.515(2)
1.688(2)
1.689(2)
1.636(3)
1.542(2)
Ti–Cl(1)
Ti–Cl(2)
Ti–Cl(3)
Ti–N(1)
Ti–N(2)
Ti–S(Se)
P(1)–S(1)(Se1)
P(2)–S(2)(Se2)
P(1)–N(1)
P(1)–N(2)
P(1)–N(3)
P(1)–N(4)
2.2519(11)
2.2205(11)
2.2540(12)
2.003(3)
2.2917(17)
2.2068(11)
2.2994(13)
2.2133(9)
2.037(4)
2.042(3)
2.014(2)
2.3938(14)
2.0097(13)
1.9210(16)
1.615(3)
2.5058(8)
2.1569(8)
2.0652(10)
1.613(3)
1.622(2)
1.621(3)
1.652(2)
1.655(2)
N(1)–P(1)–N(2)
N(3)–P(1)–N(5)
P(1)–N(1)–P(2)
P(1)–N(2)–P(2)
N(1)–P(2)–N(2)
N(4)–P(2)–N(6)
82.5(2)
106.4(2)
97.6(2)
97.1(2)
82.8(2)
105.6(2)
82.71(10)
106.41(13)
97.28(10)
97.09(10)
82.88(9)
1.664(2)
1.669(2)
Cl(1)–Ti–Cl(2)
Cl(2)–Ti–Cl(2)#
Cl(1)–Ti–Cl(3)
Cl(2)–Ti–Cl(3)
N(1)–Ti–N(2)
Cl(1)–Ti–N(1)
Cl(2)–Ti–N(1)
112.68(4)
95.03(5)
117.08(8)
94.64(4)
119.59(6)
105.81(13)
92.77(5)
96.23(5)
71.90(10)
88.29(8)
97.95(8)
156.37(11)
97.23(6)
156.45(12)
97.19(7)
Fig. 3 Solid-state structure of 5. The solvent molecule (toluene) is not
shown.
Fig. 2 Solid-state structure of 4. The proximal tert-butyl group has
been truncated to the quaternary carbon for clarity. Compound 3 has a
similar structure.
P᎐N(imino) are decidedly nonequivalent (1.633(3) vs. 1.529(2)
᎐
Å); the former being intermediate between single and double
bonds, while the latter are true double bonds.10
Treatment of 3 or 4 with TiCl4, Scheme 3, afforded red
crystals, whose NMR spectra revealed much lower symmetries
than required for seco-heterocubic molecules. The appearance
of two signals in the 31P NMR spectra provided convincing
evidence that the cyclodiphosph()azanes coordinated one
titanium atom laterally as monoanionic ligands. Spectroscopic
and X-ray crystallographic data also showed that 5 was a
hemi-toluene solvate, while 6 was solvent free.
A single-crystal X-ray study of 5 confirmed the preliminary
structure assignment. The thermal-ellipsoid plot, Fig. 3, shows
that the molecule consists of mutually perpendicular (P–N)2
and (P–N–Ti–N) heterocycles, connected by a common atom
(P1). Crystal data and bond parameters for 5 are collected
in Tables 1 and 4, respectively. The cyclodiphosph()azane
coordinates one titanium atom laterally as an aminoimino-
phosphoranate. Almost identical coordination geometries of
the titanium atoms had been found in complexes of the type
[XYP(NSiMe3)2TiCl3] [X = Cl, Y = N(SiMe3)2]7 (X = Cl, Y =
CCl3)8 and [Ph2P(NSiMe3)2TiCl3ؒMeCN],8 which have con-
ventional phosphoranate ligands. The equidistant phosphorus–
nitrogen (1.622(2) and 1.621(3) Å) and titanium–nitrogen bonds
(2.003(3) and 2.014(2) Å) of 5 demonstrate the highly sym-
metrical metal chelation by the ligand. The titanium atom has a
Scheme 3
1248
J. Chem. Soc., Dalton Trans., 2001, 1246–1252

View Article Online
more perfect trigonal-bipyramidal coordination environment
than that in 2. One nitrogen (N1) and one chlorine atom (Cl3)
define the axial direction of the bipyramid, while the second
nitrogen atom (N2) is equatorial. The nitrogen–titanium bonds
are slightly longer than those in 2, while the titanium–chloride
bonds are slightly shorter (2.2205(11)–2.2540(12) Å).
The isolation of 5 and 6 showed, to our initial disappoint-
ment, that arylimino-substituted bis(amino)cyclodiphosph()-
azanes chelate titanium as amino(imino)phosphoranates and
not as diamides. Ligands like 3 and 4 are apparently predisposed
to side-on coordination, due to the π-delocalization in the
N–P–N moiety on chelation. This unwanted result caused us to
investigatealternativephosphorus()ligands, namely2,4-dichal-
cogenocyclodiphosph()azanes. These molecules, especially
the selenium analogue, are less likely to exhibit π-delocaliz-
ation through their N–P–E groups. Moreover, the softer
chalcogen atoms are expected to bind the hard titanium() less
strongly than the nitrogen atoms, thereby encouraging the
coordination of the bis(amino)cyclodiphosph()azanes as
bis(amido) ligands.
Fig. 4 Solid-state structure of 10. The solvent molecule (toluene) is
not shown. The disulfur complex (9) is isostructural.
slightly shorter than those in 1,4-[(MeC5H4)2Ti]2Se4 (2.542(1)
and 2.563(1) Å).15
Dithio-, and diseleno-cyclodiphosph()azanes are well
Discussion
known,4 and structural studies on cis-[But(H)N(S᎐PNBut) -
᎐
2
N(H)But] (7),12 for example, had shown that its tert-butylamino
groups have a cis configuration, very similar to those in 3 and 4.
In contrast to the bis(arylimino)-substituted ligands, however,
one of the NH groups of 7 has an exo orientation, while the
Cyclodiphosph()azanes are thermally stable, easy to syn-
thesize heterocycles, whose central (P–N)2 moiety can be substi-
tuted with primary amines to yield chelating diamide ligands.
Their N–(P–N)2–N backbone encapsulates metals better than
conventional diamides, and this obviates the need for large
amino-nitrogen substituents to prevent dimerization of their
complexes.
second one is endo. The selenium analogue cis-[But(H)N(Se᎐
᎐
PNBut)2N(H)But] (8) can be readily prepared by refluxing 1
with grey selenium.13
When these 2,4-bis(1Њ-amino)-2,4-dichalcogenocyclodiphos-
ph()azanes were allowed to react with TiCl4 (Scheme 4), under
A typical example of such a bis(amido)cyclodiphosph()-
azane complex is the titanium compound 2, which is readily
synthesized (Scheme 1) from neutral cis-[But(H)N(ButNP)2-
N(H)But] 1, triethylamine and titanium tetrachloride. In this
complex the P–N ligand chelates the titanium atom as a
diamide, in a trihapto fashion, to form a seco-heterocubic
complex. The coordination geometry of the metal atom in 2 is
reminiscent of that of a corner position in α-TiCl3, which is still
the most widely used polymerization catalyst for isotactic
polypropylene.16
Like most trivalent phosphorus compounds, however, cyclo-
diphosph()azanes are both Lewis-basic and susceptible to
oxidation—properties which render them and their complexes
chemically reactive. Specifically, these ligands have a propensity
to undergo cycloreversion in the presence of strong Lewis
acids,17 such as those present during olefin polymerization.
The oxidative instability of cyclodiphosph()azanes has
previously been exploited for the rational synthesis of mono-
and di-chalcogeno cyclodiphosph()azanes.4 We have now used
such P() cyclodiphosphazanes as chelating ligands in place of
their P() analogues. By treating cyclodiphosph()azanes
with azides or chalcogens the more robust cyclodiphosph()-
azanes can be synthesized with minimal additional synthetic
effort. Because ligand design in polyolefin catalysis requires an
almost combinatorial approach, ease of conversion to cyclo-
diphosph()azanes was a consideration in this modification.
Oxidation removes the lone pairs as the most probable source
of facile Lewis-acid attack. Perhaps even more important is the
effect of the oxidation on the thermodynamic stability of the
P() ligands. This is demonstrated by the P–N bond lengths of
3, 4 and 7, which are 0.027(2)–0.045(2) Å shorter than those of
the P() species, indicative of stronger phosphorus–nitrogen
bonds. Somewhat unexpectedly, equivalent P–N bonds in the
P() molecules 3, 4 and 7 are isometric, despite significant
differences in the steric and electronic properties of their
pendent groups. This suggests that the P–N bonds of these lig-
ands are approaching their minimal values.
Scheme 4
identical conditions as for the syntheses of 5 and 6, similar
orange–red solids were isolated. The NMR spectra immediately
identified the products as structural analogues of 5 and 6,
rather than of 2. Both complexes, 9 and 10, crystallized as
mono-toluene solvates of the formula {[But(H)N(E᎐PNBut) -
᎐
2
NBut]TiCl3ؒC6H5Me} (E = S (9), Se (10)).
Crystal and bond parameters for the isostructural complexes
9 and 10 are collected in Tables 1 and 4, respectively. The
thermal ellipsoid plot of 10, Fig. 4, is representative for both
compounds. Like 5 these molecules are spirocycles, but in
contrast to 5, they have crystallographic m symmetry. The
mirror plane, which contains the (P–N–Ti–E) ring and both
phosphorus atoms, bisects the cyclodiphosph()azane ring. The
cyclodiphosph()azane and the three chloride ligands create
a similar, albeit more distorted, trigonal-bipyramidal coord-
ination geometry about titanium, N1 and Cl1 defining the axial
direction. Due to the presence of the larger sulfur atom the
Ti–N bond is slightly longer and the Ti–Cl bonds are more
dissimilar. The titanium–sulfur bond in 9 2.3938(14) Å, is
shorter than in [(MeC5H4)2Ti(OS5)], Ti–S = 2.467 Å,14 but this
may be the result of the greater steric crowding in the organo-
metallic compound. As expected, the P1–S1 bond (2.0097(13)
Å) has lost some of its double-bond character and is longer
than the free P2–S2 bond (1.9210(16) Å).
Contrary to our intentions bis(1Њ-amino)cyclodiphosph()-
azanes do not function as simple chelating bis(amido) ligands,
but are better considered dimeric aminobis(imino)phos-
phoranes. This discovery should have come as no surprise, of
The only structural difference between 9 and 10 is the bonds
to the chalcogen atoms, which are consistently ca. 0.1 Å longer
for the selenium analogue. The Ti–Se bond (2.5058(8) Å) is
J. Chem. Soc., Dalton Trans., 2001, 1246–1252
1249

View Article Online
course, as a comparison of bis(1Њ-amino) cyclodiphosph()-
azanes in their dianionic forms, A and B, shows. Because of
π-electron delocalization B is the more stable of the two
structures, and one may therefore expect these ligands to
chelate metal atoms as heteroallyls.
reason. Our inability to detect dititanium complexes is thus
likely due to kinetic factors (either the viscous reaction mixtures
or the insolubility of the monotitanium complexes prevented
proper mixing).
The title compounds, together with 11 and 12, thus demon-
strate the coordinative-flexibility of bis(1Њ-amido)cyclodi-
phosph()azanes. These heterocycles can chelate metals
as bis(amido) (12), aminoiminophosphoranato (11), or as
bis(amido)-bis(chalcogeno) ligands, depending on the size of
the metal fragment and the degree of covalency of the bonds.
Chemically these ligands behave as aminoiminophosphor-
anates, irrespective if the phosphorus() atoms bear the rel-
atively small and soft chalcogens sulfur and selenium, or the
harder and sterically more encumbered arylimino groups.
Despite their unexpected structures the title compounds may
still be expected to have applications in olefin catalysis. Nickel
complexes of aminoiminophosphoranates, for example, poly-
merize ethylene to short-chain branched polymers by an
unusual 2,ω coupling process.20 Catalytic tests on the titanium
complexes are in progress.
Although such resonance arguments may explain the side-on
binding of these ligands in complexes 5 and 6 and 9 and 10,
the true coordination behaviour of bis(amido)cyclodiphosph()-
azanes is more complex. Scherer et al. had discovered that
Re(CO) Cl reacts with [(Me Si) NP(᎐NSiMe ) ], Scheme 5,
᎐
5
3
2
3 2
Conclusion
Bis(1Њ-amino)cyclodiphosph()azanes react with titanium
tetrachloride to form seco-heterocubic bis(amido)complexes.
Their oxidation products, the 2,4-dichalcogeno- and 2,4-
diarylimino-substituted bis(1Њ-amino)cyclodiphosph()azanes,
however, chelate the titanium atom as aminoimino- or amino-
chalcogenophosphoranates. Given the extensive coordination
chemistry of phosphoranates21 (and of amidinates),22 a simi-
larly rich coordination chemistry can be predicted for bis(1Њ-
amino)cyclodiphosph()azanes. The unique feature of these
tetradentate ligands—their bifunctionality—may find appli-
cations in catalysis (tethering to surfaces) and materials
research (synthesis of metalla-oligomers and -polymers).
Experimental
CAUTION: Organic azides are potentially explosive com-
pounds. We encountered no problems while handling these
reagents or their reaction mixtures.
All experiments were performed under an atmosphere of
purified nitrogen or argon, using standard Schlenk techniques.
Solvents were dried and freed of molecular oxygen by distil-
lation under an atmosphere of nitrogen from sodium or
potassium benzophenone ketyl immediately before use. NMR
spectra were recorded on Varian VXR-300 and Bruker
AVANCE-500 NMR spectrometers. The 1H, 13C and 31P NMR
spectra are referenced relative to C6D5H (7.15 ppm), C6D6
(128.0 ppm) and P(OEt)3 (137.0 ppm), respectively. Melting
points were obtained on a Mel-Temp apparatus and are
uncorrected. Elemental analyses were performed by E and R
Microanalytical Services, Parsipanny, NJ, and Desert Analytics,
Tucson, AZ. Triethylamine and grey selenium were purchased
from Aldrich and used without further purification. Titanium
tetrachloride was purchased from Aldrich and used as a 2.0 M
stock solution in toluene. Phenyl and p-tolyl azide,23 and
cis-[But(H)N(ButNP)2N(H)But]6 were synthesized by published
procedures.
Scheme 5
to yield three products, of which only two (11 and 12) are
pertinent to our results.18 In 11 the bis(amido)cyclodiphosph-
()azane chelates two Re(CO)4 fragments as a bisphosphor-
anate to form a trispirocyclic complex, quite similar to 5 and
6. Compound 12, by contrast, is a heterocubic bimetallic
bis(amido) complex, which is structurally related to 2. It
displays the coordination mode (albeit a dinuclear one) of the
ligand we had hoped to see in 5 and 6 and 9 and 10. Additional
proof that bis(amino)cyclodiphosph()azanes have a diverse
coordination behaviour, which is highly dependent upon the
incorporated metal fragment, was recently provided by Chivers
et al.13 When they treated 7 or 8 with sodium or potassium
bis(trimethylsilyl)amide a still different binding mode of these
ligands was discovered. In these salts the ligand chelates one
alkali metal with both amido-nitrogen atoms and the second
one with both chalcogen atoms.
Syntheses
Despite using 2 : 1 metal/ligand ratios (Schemes 3 and 4), we
were unable to isolate dinuclear titanium complexes of 3, 4, 7
and 8. We initially thought this to be due to thermodynamic
factors (in bimetallic complexes the bulky tert-butyl groups
would presumably be within van der Waals distances of each
other). The recent isolations of trispirocyclic dinuclear main
group derivatives of 3, 4, 7 and 8,19 which exhibit coordination
modes analogous to 11, however, show that this cannot be the
{[(ButNP)2(NBut)2]TiCl2} (2). A two-neck 100-mL round
bottom flask, equipped with an inlet and stir bar, was charged
with a mixture of toluene (20 mL), triethylamine (0.48 mL, 3.5
mmol) and titanium tetrachloride (1.7 mL of a stock solution).
To the cooled (Ϫ78 ЊC) flask was added dropwise a toluene
solution of 1 (0.601 g, 1.72 mmol). The vigorously stirred
orange reaction mixture was allowed to warm and stirred at RT
for 12 h. It was then filtered through a medium-porosity frit
1250
J. Chem. Soc., Dalton Trans., 2001, 1246–1252

View Article Online
to remove triethylammonium chloride, and the orange filtrate
was concentrated in vacuo until crystals formed. These were
redissolved and the flask was kept at Ϫ12 ЊC until flaky orange
crystals (0.621 g, 77.6%) had formed. The crystals were
sublimed under a dynamic vacuum (70 ЊC, 0.03 Torr) to give
orange blocks (0.576 g, 72.0%), mp 200–203 ЊC (Found: C,
41.11; H, 7.97; N, 11.79. C16H36Cl2N4P2Ti requires C, 41.30; H,
7.80; N, 12.04%); δH (C6D6) 1.57 (s, 18 H, NBut), 1.15 (s, 18 H,
NBut); δC (C6D6) 60.2 (d, JPC = 19.3 Hz), 54.6 (t, JPC = 10.2 Hz),
33.7 (d, JPC = 12.2 Hz), 28.3 (t, JPC = 7.2 Hz); δP (C6D6) 115.5
(s).
combined in toluene. Upon cooling (Ϫ12 ЊC) the solution for
several days, purple needles formed (0.99 g, 74%), mp 200–
202 ЊC (Found: C, 50.58; H, 7.21; N, 11.52; C30H51Cl3N6P2Ti
requires C, 50.61; H, 7.22; N, 11.80%); δH (C6D6) 7.57 (dd, 2 H,
JHH = 6.5 Hz), 6.89 (t, 4 H, JHH = 8 Hz), 6.61 (d, 2 H, JHH = 7.8
Hz), 2.18 (s, 3 H, CH3), 2.10 (s, 3 H, CH3), 1.96 (s, 1H, NH),
1.71 (s, 9 H, NBut), 1.33 (s, 18 H, NBut), 1.20 (s, 9 H, N(H)But);
δC (C6D6) 129.8 (s), 129.3 (s), 125.6 (s), 125.0 (d, JPC = 9.5 Hz),
124.0 (d, JPC = 12.5 Hz), 56.3 (s), 54.0 (s), 32.7 (d, JPC = 8.4 Hz),
31.5 (d, JPC = 5.0 Hz), 30.8 (t, JPC = 4.0 Hz), 20.9 (s), 20.8 (s);
δP (C6D6) Ϫ3.04 (d, JPP = 47 Hz), Ϫ50.3 (d, JPP = 47 Hz).
{cis-[But(H)N(PhN᎐PNBut) N(H)But]ؒTHF} (3). A sample
cis-[But(H)N(Se᎐PNBut) N(H)But] (8). A mixture of 1 (1.01
᎐
2
᎐
2
of 1 (5.20 g, 14.9 mmol) was dissolved in THF (50 mL) in a
two-necked flask, equipped with an inlet, stir bar, and dropping
funnel. Phenyl azide (3.56 g, 29.8 mmol), dissolved in THF
(25 mL), was then added dropwise at RT to the clear solution,
which slowly evolved gas (N2). The orange solution was stirred
(12 h), then refluxed (2.5 h) and finally transferred to a 250-mL
flask, where it was concentrated in vacuo until crystals formed.
These were redissolved in a minimal amount of THF, and the
ensuing solution was refrigerated (Ϫ12 ЊC) to afford off-white
crystals (6.62 g, 74.0%), mp 152–154 ЊC (Found: C, 63.92; H,
9.20; N, 13.99. C32H56N6OP2 requires C, 63.76; H, 9.36; N,
13.94%); δH (C6D6) 7.37 (s, 4 H), 7.36 (s, 4 H), 6.97 (m, 2 H),
3.54 (m, 4 H, THF), 1.92 (s, 2 H, NH), 1.68 (m, 4 H, THF),
1.38 (s, 18 H, N(H)But), 1.34 (s, 18 H, NBut); δC (C6D6) 148.2
(s), 129.1 (s), 124.9 (t, JPC = 9.1 Hz), 119.2 (s), 67.8 (s), 54.0 (s),
53.4 (t, JPC = 2.5 Hz), 31.9 (s), 30.5 (t, JPC = 4.2 Hz), 25.8 (s);
δP (C6D6) Ϫ26.7 (s).
g, 2.89 mmol) and grey selenium (0.452 g, 5.72 mmol) was
dissolved/suspended in toluene (20 mL) and refluxed until it
was clear (ca. 6 h). The light yellow solution was concentrated
and placed in a freezer (Ϫ21 ЊC). After several days light-
yellow, rod-shaped crystals (1.26 g. 86.9%) formed, mp 144–
148 ЊC (Found: C, 37.92; H, 7.28; N, 10.88. C16H38N4P2Se2
requires C, 37.95; H, 7.56; N, 11.06%); δH (C6D6) 3.19 (s, 2 H,
NH), 1.78 (s, 18 H, N(H)But), 1.15 (s, 18 H, NBut); δC (C6D6)
59.2 (s), 55.9 (t, JPC = 2.8 Hz), 32.0 (s), 30.9 (t, J = 4.6 Hz);
δP (C6D6) 40.35 (s, J(77Se) satellites (7.5%) = 886 Hz).
{[But(H)N(S᎐PNBut) NBut]TiCl ؒC H Me} (9). In a manner
᎐
2
3
6
5
strictly analogous to that used for the syntheses of 5 and 6,
toluene (45 mL), triethylamine (1.4 mL, 10 mmol) and TiCl4
(5 mL of a stock solution) were combined. To this cooled
(Ϫ78 ЊC) solution was added dropwise 7 (2.06 g, 5.00 mmol),
dissolved in toluene (20 mL). The dark orange solution was
allowed to slowly warm to RT and stirred for 12 h. It was then
filtered through a medium-porosity frit, and the resulting dark
orange solution was concentrated in vacuo. The supersaturated
solution was stored in a freezer (Ϫ12 ЊC) until orange crystals
had formed (1.7 g, 57%), mp 173–175 ЊC (Found: C, 41.68; H,
7.27; N, 8.90; C23H45Cl3N4P2S2Ti requires C, 41.99; H, 6.89;
N, 8.51%); δH (C6D6) 7.52–7.05 (m, 5 H, C6H5Me), 2.83 (s, 1 H,
NH), 2.10 (s, 3 H, C6H5Me), 1.59 (s, 9 H, NBut), 1.37 (s, 18 H,
NtBu), 1.16 (s, 9 H, N(H)But); δC (C6D6) 59.6 (s), 34.1 (s), 32.9
cis-[But(H)N( p-tolylN᎐PNBut) N(H)But] (4). In a manner
᎐
2
identical to that used for the synthesis of 3, a sample of 1
(1.04 g, 3.00 mmol) was treated with p-tolyl azide (0.798 g,
6.00 mmol). After the solution had been refrigerated for several
days, colourless rectangular crystals of 4 (1.59 g, 95.0%) were
recovered, mp 140–143 ЊC (Found: C, 63.85; H, 9.31; N, 15.07.
C30H52N6P2 requires C, 64.49; H, 9.38; N, 15.03%); δH (C6D6)
7.32 (d, 4 H, JHH = 8 Hz), 7.20 (d, 4 H, JHH = 8 Hz), 2.28 (s, 6 H,
CH3), 1.73 (br t, 2 H, NH), 1.39 (s, 18 H, N(H)But), 1.37 (s,
18 H, NBut); δC (C6D6) 145.6 (s), 129.8 (s), 128.3 (s), 124.7
(t, JPC = 9 Hz), 54.0 (s), 53.3 (s), 31.9 (s), 30.5 (t, JPC = 4.1 Hz),
21.0 (s); δP (C6D6) Ϫ29.2 (s).
(d, JPC = 7.5 Hz), 31.6 (s), 25.8 (s); δP (C6D6) 36.2 (d, JPP
=
26 Hz); 25.7 (d, JPP = 26 Hz).
{[But(H)N(Se᎐PNBut) NBut]TiCl ؒC H Me} (10). In a man-
᎐
2
3
6
5
ner strictly analogous to that used for the synthesis of 9, TiCl4
(2.1 mL of a stock solution), NEt3 (0.6 mL, 4.2 mmol) and 8
(1.0 g, 2.0 mmol) were combined in toluene (20 mL). The dark
orange solution was allowed to warm to RT and then stirred for
12 h. Yield: 1.2 g, 93%, mp 170 ЊC (decomp.) (Found: C, 36.46;
H, 6.24; N, 7.65; C23H45Cl3N4P2Se2Ti requires C, 36.70; H, 6.03;
N, 7.44%); δH (C6D6) 7.53–7.05 (m, 5 H, C6H5Me), 3.21 (s, 1 H,
NH), 2.10 (s, 3 H, C6H5Me), 1.61 (s, 9 H, NBut), 1.42 (s, 18 H,
NBut), 1.20 (s, 9H, N(H)But); δC (C6D6) 59.3 (s), 32.9 (d,
JPC = 7.9 Hz), 31.6 (s), 30.7 (s), 25.8 (s); δP (C6D6) 22.5 (d,
JPP = 14 Hz), 10.1 (d, JPP = 14 Hz).
{[But(H)N(PhN᎐PNBut) NBut]TiCl ؒ0.5C H Me} (5). In a
᎐
2
3
6
5
cooled (Ϫ78 ЊC) three-neck flask, triethylamine (0.41 mL, 2.9
mmol) and titanium tetrachloride (1.4 mL of a stock solution)
were dissolved in toluene (10 mL), producing a dark orange
solution. A sample of 3 (0.884 g, 1.47 mmol), dissolved in
toluene (10 mL), was then added dropwise to the TiCl4/NEt3
solution. The dark-red reaction mixture was stirred for 12 h
while it warmed to RT. After filtration on a fine frit, the red
filtrate was concentrated in vacuo until crystals formed. These
were redissolved (35 ЊC), and the warm filtrate was allowed to
slowly cool to RT until dark-red crystals of 5 (0.77 g, 76%) had
formed, mp 208–210 ЊC (Found: C, 52.28; H, 7.58; N, 10.85.
C31.5H51Cl3N6P2Ti requires C, 51.83; H, 7.05; N, 11.50%);
δH (C6D6) 7.64 (d, 2 H, JHH = 7.6 Hz), 7.31 (s, 1 H), 7.14–7.00
(m, 5 H), 6.88 (m, 2 H), 6.71 (d, 2 H, JHH = 8.5 Hz), 2.10 (s,
1.5 H, C6H5Me), 1.92 (s, 1 H, NH), 1.70 (s, 9 H, NBut), 1.29
(s, 18 H, NBut), 1.17 (s, 9 H, N(H)But); δC (C6D6) 148.5 (s),
144.0 (s), 129.4 (s), 129.3 (s), 125.7 (s), 124.9 (d, JPC = 11 Hz),
124.1 (d, JPC = 11 Hz), 120.2 (s), 60.8 (d, JPC = 3.6 Hz), 56.4 (s),
54.2 (d, JPC = 4.1 Hz), 33.1 (d, JPC = 7.8 Hz), 31.4 (d, JPC = 4.9
Hz), 30.8 (t, JPC = 3.9 Hz), 21.4 (s, CH3); δP (C6D6) Ϫ1.42
(d, JPP = 48 Hz), Ϫ50.1 (d, JPP = 48 Hz).
X-Ray crystallography
Compounds 2, 5 and 10. Suitable, single crystals were coated
with oil, affixed to a glass capillary, and centred on the dif-
fractometer in a stream of cold nitrogen. Reflection intensities
were collected with a Bruker SMART CCD diffractometer,
equipped with an LT-2 low-temperature apparatus, operating at
213 K. Data were measured using ω scans of 0.3Њ per frame for
30 s until a complete hemisphere had been collected. The first
50 frames were recollected at the end of the data collection
to monitor for decay. Cell parameters were retrieved using
SMART24 software and refined with SAINT25 on all observed
reflections. Data were reduced with SAINT, which corrects for
Lorentz factor, polarization and decay. An empirical absorption
correction was applied with SADABS.26 The structures were
{[But(H)N( p-tolylN᎐PNBut) NBut]TiCl } (6). In a manner
᎐
2
3
identical to that used for the synthesis of 5, NEt3 (0.60 mL, 4.0
mmol), TiCl4 (4.0 mmol) and 4 (1.11 g, 1.97 mmol) were
J. Chem. Soc., Dalton Trans., 2001, 1246–1252
1251

View Article Online
solved by direct methods with the SHELXS-9027 program
and refined by full-matrix least squares methods on F2 with
SHELXL-97,28 incorporated in SHELXTL Version 5.10.29
9 D. F. Moser and L. Stahl, unpublished results.
10 D. E. C. Corbridge, Phosphorus: An Outline of its Chemistry,
Biochemistry and Technology, Elsevier, Amsterdam, 5th edn., 1995.
11 R. Appel and M. Halstenberg, J. Organomet. Chem., 1976, 121, C47.
12 T. G. Hill, R. C. Haltiwanger, M. L. Thompson, S. A. Katz and
A. D. Norman, Inorg. Chem., 1994, 33, 1770.
13 T. Chivers, M. Krahn and M. Parvez, Chem. Commun., 2000, 463.
14 R. Steudel, A. Prenzel and J. Pickardt, Angew. Chem., Int. Ed. Engl.,
1991, 30, 550.
15 D. M. Giolando, M. Papavassiliou, J. Pickardt, T. B. Rauchfuss and
R. Steudel, Inorg. Chem., 1988, 27, 2596.
16 M. Boero, M. Parrinello, S. Hüffler and H. Weiss, J. Am. Chem.
Soc., 2000, 122, 501; P. Pino and R. Mülhaupt, Angew. Chem., Int.
Ed. Engl., 1980, 20, 869.
Compounds 3, 4 and 9. The crystals were sealed inside argon-
filled glass capillaries, and the intensity data were collected
on a Bruker P4 diffractometer. The structures were solved by
direct methods with SHELXL-NT, Version 5.10,30 and refined
analogously to 2, 5 and 10.
CCDC reference numbers 158377–158382.
See http://www.rsc.org/suppdata/dt/b0/b007877h/ for crystal-
lographic data in CIF or other electronic format.
17 I. Schranz, D. F. Moser, L. Stahl and R. J. Staples, Inorg. Chem.,
1999, 38, 5814.
18 O. J. Scherer, P. Quintus and W. S. Sheldrick, Chem. Ber., 1987, 120,
1183; O. J. Scherer, P. Quintus, J. Kaub and W. S. Sheldrick, Chem.
Ber., 1987, 120, 1463.
19 T. Chivers, M. Krahn and M. Parvez, 9th International Conference
on Inorganic Ring Systems, Saarbrücken, July 2000, Book of
Abstracts, P-14; G. R. Lief, C. J. Carrow, R. J. Staples and L. Stahl,
Organometallics, 2001, 20, in press.
Acknowledgements
We thank the Phillips Petroleum Company for Research
Assistantships to D. F. M. and C. J .C.
References
1 J. D. Scollard, D. H. McConville and S. J. Rettig, Organometallics,
1997, 16, 1810; H. C. S. Clark, F. G. N. Cloke, P. B. Hitchcock, J. B.
Love and A. P. Wainwright, J. Organomet. Chem., 1995, 501, 333;
T. H. Warren, R. R. Schrock and W. M. Davis, Organometallics,
1996, 15, 562; K. Aoyagi, P. K. Gantzel and T. D. Tilley,
Organometallics, 1996, 15, 923; R. Minhas, L. Scoles, S. Wong and
S. Gambarotta, Organometallics, 1996, 15, 1113; B. Tsuie, D. C.
Swenson, R. F. Jordan and J. L. Petersen, Organometallics, 1997, 16,
1392; S. A. A. Shah, H. Dorn, A. Voigt, H. W. Roesky, E. Parisini,
H. G. Schmidt and M. Noltemeyer, Organometallics, 1996, 15, 3176;
H. V. R. Dias, W. Jin and Z. Wang, Inorg. Chem., 1996, 35, 6074.
2 L. Grocholl, I. Schranz, L. Stahl and R. J. Staples, Chem. Commun.,
1997, 1465.
20 W. Keim, R. Appel, A. Storeck, C. Krüger and R. Goddard, Angew.
Chem., Int. Ed. Engl., 1981, 20, 116; V. M. Möhring and G. Fink,
Angew. Chem., Int. Ed. Engl., 1985, 24, 116.
21 W. Kuchen, D. Langsch and W. Peters, Phosphorus, Sulfur, Silicon
Relat. Elem., 1990, 54, 55; O. J. Scherer, J. Kerth, B. K. Balbach
and M. L. Ziegler, Angew. Chem., Int. Ed. Engl., 1982, 21, 136;
T. Chivers, M. Parvez and M. A. Seay, Inorg. Chem., 1994, 33, 2147;
A. Steiner and D. Stalke, Inorg. Chem., 1993, 32, 1977.
22 P. J. Stewart, A. J. Blake and P. Mountford, Organometallics, 1998,
17, 3271; F. T. Edelmann, Coord. Chem. Rev., 1994, 137, 403; J. R.
Hagadorn and J. Arnold, J. Chem. Soc., Dalton Trans., 1997, 3087;
J. Richter, M. Belay, W. Brieser and F. Edelmann, Main Group Met.
Chem., 1997, 20, 191; E. A. C. Brussee, A. Meetsma, B. Hessen and
J. H. Teuben, Organometallics, 1998, 17, 4090; J. Barker, N. C.
Blacker, P. R. Phillips, N. W. Allcock, W. Errington and M. G. H.
Wallbridge, J. Chem. Soc., Dalton Trans., 1996, 431.
23 P. A. Smith and B. B. Brown, J. Am. Chem. Soc., 1951, 73, 2438.
24 SMART Version 4.043, Software for the CCD Detector System,
Bruker Analytical X-Ray Systems, Madison, WI, 1995.
25 SAINT Version 4.035, Software for the CCD Detector System,
Bruker Analytical X-Ray Systems, Madison, WI, 1995.
26 SADABS, Program for absorption corrections using the Bruker
CCD Detector System. Based on: R. Blessing, Acta Crystallogr.,
Sect. A, 1995, 51, 33.
3 L. Grocholl, V. Huch, L. Stahl, R. J. Staples, P. Steinhart and
A. Johnson, Inorg. Chem., 1997, 36, 4451.
4 R. Keat, K. W. Muir and D. G. Thompson, Tetrahedron Lett., 1977,
3087; R. Keat and D. G. Thompson, J. Chem. Soc., Dalton Trans.,
1980, 928; R. Keat, A. N. Keith, A. Macphee, K. W. Muir and D. G.
Thompson, J. Chem. Soc., Chem. Commun., 1978, 372; O. J. Scherer
and G. Schnabl, Angew. Chem., Int. Ed. Engl., 1976, 15, 772;
R. Jefferson, J. F. Nixon, T. M. Painter, R. Keat and L. Stobbs,
J. Chem. Soc., Dalton Trans., 1973, 1414; G. Bulloch and R. Keat,
J. Chem. Soc., Dalton Trans., 1974, 2010.
5 D. F. Moser and L. Stahl, unpublished results.
6 R. R. Holmes and J. A. Forstner, Inorg. Chem., 1963, 2, 380;
I. Schranz, L. Stahl and R. J. Staples, Inorg. Chem., 1998, 37, 1493.
7 E. Niecke, R. Kröher and S. Pohl, Angew. Chem., Int. Ed. Engl.,
1977, 16, 864.
8 M. Witt, M. Noltemeyer, H.-G. Schmidt, T. Lübben and H. W.
Roesky, J. Organomet. Chem., 1999, 591, 138; M. Witt, H. W.
Roesky, D. Stalke, F. Pauer, T. Henkel and G. M. Sheldrick,
J. Chem. Soc., Dalton Trans., 1989, 2173; M. Witt, H. W. Roesky and
M. Noltemeyer, Inorg. Chem., 1997, 36, 3476.
27 G. M. Sheldrick, SHELXS-90, Program for the Solution of Crystal
Structures, University of Göttingen, Göttingen, Germany, 1990.
28 G. M. Sheldrick, SHELXL-97, Program for the Solution of Crystal
Structures, University of Göttingen, Göttingen, Germany, 1997.
29 SHELXTL 5.10 (PC-Version), Siemens Analytical X-Ray
Instruments Inc., Madison, WI, 1998.
30 SHELXTL-NT Version 5.10, Program Library for Structure
Solution and Molecular Graphics, Bruker Analytical X-Ray
Systems, Madison, WI, 1999.
1252
J. Chem. Soc., Dalton Trans., 2001, 1246–1252