Synthesis, structure and polymerization behaviour of borane adducts of a phosphorus-bridged [1] ferrocenophane, [(η-C5H4)2FePPh]

Article abstract of DOI:10.1039/b000935k

Phosphorus(III)-bridged [1]ferrocenophanes cannot be polymerized via transition-metal catalyzed ring-opening polymerization, presumably due to the phosphorus lone pair ligating the catalyst. We have investigated a potential solution to this problem, which involves protection of the phosphorus lone pair through the formation of borane adducts. The adducts [(η- C₅H₄)₂FeP(Ph)BX₃] (2, X = H; 3, X = Cl) were synthesized in high yield from the reactions of appropriate boranes with P- phenylphospha[1]ferrocenophane, 1, and fully characterized by NMR, UV-Vis, MS, elemental analysis, and X-ray crystallography. Ring-opening polymerizations of 2 and 3 have been attempted both thermally and using transition-metal catalysts, resulting in insoluble polymeric products, [(η- C₅H₄)₂FeP(Ph)BX₃], (5, X = H; 6, X = Cl), which have been characterized by solid-state NMR and by pyrolysis-MS. Well-defined borane adducts of poly(ferrocenylphenylphosphine), [(η-C₅H₄)₂FeP(Ph)BX₃](n) (5b, X = H; 6b, X = Cl) were synthesized and characterized to provide comparative data for the insoluble products 5a, 6a and 6c. The data support the conclusion that both 2 and 3 may be ring-opened thermally to provide polymeric products consistent with the corresponding adducts of poly(ferrocenylphenylphosphine) 4. In contrast, transition-metal catalyzed ROP proceeds inefficiently, leading predominantly to products of unknown structure.

Full text of DOI:10.1039/b000935k

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis, structure and polymerization behaviour of borane adducts
of a phosphorus-bridged [1]ferrocenophane, [(g-C H ) FePPh]
5 4 2
Chris E. B. Evans, Alan J. Lough, Hiltrud Grondey and Ian Manners*
Department of Chemistry, University of T oronto, T oronto, ON, M5S 3H6, Canada.
E-mail: imanners=chem.utoronto.ca
Received (in New Haven, USA) 27th January 2000, Accepted 19th March 2000
Phosphorus(III)-bridged [1]ferrocenophanes cannot be polymerized via transition-metal catalyzed ring-opening
polymerization, presumably due to the phosphorus lone pair ligating the catalyst. We have investigated a potential
solution to this problem, which involves protection of the phosphorus lone pair through the formation of borane
adducts. The adducts [(g-C H ) FeP(Ph)BX ] (2, X \ H; 3, X \ Cl) were synthesized in high yield from the
5
4 2
3
reactions of appropriate boranes with P-phenylphospha[1]ferrocenophane, 1, and fully characterized by NMR,
UVÈVis, MS, elemental analysis, and X-ray crystallography. Ring-opening polymerizations of 2 and 3 have been
attempted both thermally and using transition-metal catalysts, resulting in insoluble polymeric products, [(g-
C H ) FeP(Ph)BX ] (5, X \ H; 6, X \ Cl), which have been characterized by solid-state NMR and by
5
4 2
3 n
pyrolysis-MS. Well-deÐned borane adducts of poly(ferrocenylphenylphosphine), [(g-C H ) FeP(Ph)BX ] (5b,
5
4 2
3 n
X \ H; 6b, X \ Cl) were synthesized and characterized to provide comparative data for the insoluble products 5a,
6a and 6c. The data support the conclusion that both 2 and 3 may be ring-opened thermally to provide polymeric
products consistent with the corresponding adducts of poly(ferrocenylphenylphosphine) 4. In contrast,
transition-metal catalyzed ROP proceeds inefficiently, leading predominantly to products of unknown structure.
Poly(ferrocenylphosphines) (e.g. 4) obtained from the ring-
opening polymerization (ROP) of [1]ferrocenophanes such as
P-phenylphospha[1]ferrocenophane, 1, have potential appli-
cations as catalytic materials based upon the reactivity of the
phosphorus lone pair.1,2,3 Unfortunately, thermally induced
ROP does not permit molecular weight control, and well-
deÐned polymers must be synthesized via living anionic poly-
CH Cl solution of 1 at [40 ¡C led to the formation of 3.
2
2
Reaction progress can be readily monitored via 31P NMR, as
reactant and product resonances may be distinguished both
from their positions and multiplicity.
The 1H and 13C NMR spectra of 1, 2 and 3 are featurally
very similar, the most striking exception being the broad
1 : 1 : 1 : 1 quartet of the borane hydrogens in the proton
spectrum of 2. In all the 1H NMR spectra, the Cp ligands give
rise to four resonances, because the phosphorus bridge is
asymmetric relative to a plane containing itself and the Cp
ipso-carbons. An atypical feature in the spectrum of 2 is the
coincidence of the resonances due to the phenyl groupÏs meta
and para position hydrogens, whereas in the other systems
merization,
a
painstaking synthetic technique. Ideally,
transition-metal catalyzed ROP reactions, using capping
agents to control molecular weight,4 would be used to gener-
ate suitably tractable polymers, a process presenting much less
of a synthetic challenge. The versatility of this approach has
been demonstrated for a number of silicon-bridged fer-
rocenophanes, where access to both block copolymers and
poly(silaferrocenes) with comb and star architectures has been
possible.5 As expected, however, the P(III)-bridged
[1]ferrocenophanes do not undergo transition-metal catalyzed
ROP, presumably because the lone pair on phosphorus binds
strongly to the catalystÏs metal centre.
We have previously reported that reaction of a P(III)-
bridged [1]ferrocenophane with methyl triÑate led to a
phosphonium-bridged system (7) that did undergo transition-
metal catalyzed ROP.6 The work presented herein is part of
an investigation into whether the phosphorus lone pair could
be protected to allow a formally P(III)-bridged ferrocenophane
to be polymerized using transition-metal catalysts. This
approach has the advantage that subsequent deborylation of
the poly(ferrocenylphosphine) boraneadduct should be possible,
for example by reaction with amines, thus providing a route
to controlled molecular weight poly(ferrocenylphosphines).
Results and discussion
Addition of boranes (BX , X \ H, Cl) to 1 at reduced tem-
3
peratures led to quantitative formation of the corresponding
borane adducts of 1. In particular, mixing THF solutions of
BH É THF and 1 at 0 ¡C led (see Scheme 1) to the formation
Fig. 1 31P solution NMR spectra: (A) 2 in CH Cl , (B) 3 in
3
2
2
of 2. Similarly, adding a heptane solution of BCl to a
CD Cl .
2
2
3
DOI: 10.1039/b000935k
New J. Chem., 2000, 24, 447È453
447
This journal is ( The Royal Society of Chemistry and the Centre National de la Recherche ScientiÐque 2000

View Article Online
three distinct resonances are seen for each phenyl groupÏs
hydrogens. A very clean doublet of doublet of doublets is
observed for the ortho position hydrogens of 2, exhibiting
3J \ 12 Hz, 3J \ 7.5 Hz and 4J \ 1.85 Hz.
In the 13C NMR spectra, the position of the ipso-carbons
was the most variable feature, becoming increasingly shielded
as the phosphorusÏ substituent became increasingly electron-
withdrawing. Thus, the position of the phenyl ipso-carbonÏs
resonance shifts from 138 ppm for 1 to 129 ppm for 2 to 121
ppm for 3. Similarly, the cyclopentadienyl ipso-carbonÏs reso-
nance shifts from 18.5 ppm for 1 to 16.3 ppm for 2 to near 10
ppm for 3.7
Crystals suitable for X-ray di†raction studies were obtained
for both adducts, and ORTEP diagrams of 2 and 3 É CH Cl
2
2
are provided as Fig. 3 and 4, respectively. Table 1 contains
pertinent crystal data and structural reÐnement information
for these experiments. Bond lengths and selected angles have
been provided in Tables 2 (for 2) and 3 (for 3 É CH Cl ). The
PH
HH
HH
2
2
structures of both adducts are consistent both with those of
other phosphorus-bridged [1]ferrocenophanes, and with those
of other borane adducts of tertiary phosphines. For example,
the BÈP bond length of 3 É CH Cl [1.976(4) A] is 0.059(6) A
2
2
longer than that of 2 [1.917(2) A]. This compares with a di†er-
ence of 0.056(12) A between the BÈP bond lengths of
Me P É BCl [1.957(5) A]8 and Me P É BH [1.901(7) A].9 The
Although the 1H and 13C NMR spectra were useful in
determining the identity of the products 2 and 3 as being
[1]metallocenophanes, the most useful nuclei for di†erentiat-
ing the various compounds were, as expected, 11B and 31P.
The singlet 31P NMR resonance of 1 (at ca. 10 ppm) became
(see Fig. 1) a poorly resolved quartet (pseudo-quartet at ca. 43
ppm) in the case of 2, and a well resolved quartet (at ca. 18
3
3
3
3
BÈP bond lengths of both 2 and 3 É CH Cl may be slightly
2
2
(ca. 0.02 A) longer than those of the corresponding Me P
adducts, but the BÈCl bond lengths in both 3 É CH Cl and
the Me P adduct are similar (i.e. 1.850 A).
3
2
2
3
A plane of symmetry in 3 É CH Cl contains the iron atom,
2
2
the bridging element (phosphorus), the boron atom, and the
ipso and para carbons of the phenyl ring. There is, thus,
no torsion angle between the Cp rings. This was also the
case for the diphenylgermane-bridged [1]ferrocenophane,
[(g-C H ) FeGePh ],10 but was not the case for
ppm with 1J \ 157 Hz) in the case of 3. The presence of
BP
resonances in the 11B NMR spectra of 2 and 3 (see Fig. 2)
provides additional evidence of the formation of the adducts.
The large boronÈphosphorus coupling constant of 3 caused its
11B NMR spectrum to consist of a well-resolved doublet
5
4 2
2
the
dimethylsilane-bridged
[1]ferrocenophane,
[(g-
whose value of 1J (154 Hz) is in close agreement with that
C H ) FeSiMe ], which had
4 2
[0.4(5)¡].11
a very small ring torsion
PB
5
2
obtained from the 31P NMR spectrum. For 2, in contrast, the
value of 1J had to be estimated from the pseudo-quartet in
Both structures reveal ring-tilt angles, a, (see Fig. 5 for the
deÐnitions of the various angles in [1]metallocenophanes)
consistent with other phosphorus-bridged [1]ferroceno-
phanes. The values of 24.38(10)¡ and 25.01(4)¡ for 2 and
3 É CH Cl , respectively, are somewhat smaller than the
BP
the 31P NMR spectrum, and does not agree with the value of
1J \ 39 Hz obtained from the 11B NMR spectrum, which
PB
consists of an incompletely resolved doublet. This latter value
should be considered to be closer to the correct magnitude of
the phosphorusÈboron coupling constant for 2.
2
2
26.7¡ observed in 1, and similar to the value of 24.4(5)¡
observed for the P(V)-bridged (and cationic) compound 7. If
Fig. 2 11B solution NMR spectra: (A) 2 in CDCl , (B) 3 in CD Cl .
3
2 2
Fig. 4 ORTEP molecular structure diagram of 3 É CH Cl with
2
2
thermal ellipsoids at 30% probability. The solvent molecule has been
omitted for clarity.
Fig. 3 ORTEP molecular structure diagram of 2 with thermal ellip-
soids at 30% probability.
Fig. 5 Schematic deÐning the various angles of signiÐcance in a
[1]ferrocenophane.
448
New J. Chem., 2000, 24, 447È453

View Article Online
Table 1 Crystal data and structural reÐnement information for the X-ray structural determination of 2 and 3 É CH Cl
2
2
2
3 É CH Cl
2
2
Empirical formula
Formula weight
C
305.92
H
BFeP
C
494.17
H
BCl FeP
16 16
17 15
5
T /K
100.0(2)
Triclinic
P1
7.4376(3)
8.9729(4)
11.9365(5)
73.754(2)
83.435(2)
66.728(2)
702.56(5)
2
1.167
6152
100.0(1)
Orthorhombic
Pnma
15.6327(3)
11.9939(3)
10.2391(6)
90
Crystal system
Space group
a/A
b/A
c/A
a/¡
b/¡
c/¡
90
90
U/A
Ž
3
1919.80(13)
4
1.563
Z
k/mm~1
ReÑections collected
Independent reÑections
Final R indices [I [ 2p(I)]
R indices (all data)
17120
2843 [R(int) \ 0.033]
2306 [R(int) \ 0.080]
R \ 0.0282, wR \ 0.0731
R \ 0.0308, wR \ 0.0672
1
2
1
2
R \ 0.0328, wR \ 0.0764
R \ 0.0530, wR \ 0.0721
1
2
1
2
the structures of these four compounds are compared, a trend
of) an increasing bonding interaction between Fe and the
bridging P as the positive character of the latter is increased.
Such a trend has been commented on in previous work
in which the FeÈP distance was observed before and
after quaternization of the phosphorus centre either with
sulfur2 or methyl triÑate6 (to form 7). In the former case, a
decrease in the FeÈP distance of 0.096(4) A was observed
of decreasing FeÈP separation and C ÈP bond length (see
ipso
Table 4) with increasing electron deÐciency at phosphorus is
apparent. The longest bonds are those in 1, with shorter bond
lengths observed for 2, 3 É CH Cl , and the shortest for 7. The
2
2
angles h, b and d are correspondingly smallest for 1 and
largest for 7. These trends are consistent with (but not proof
Table 2 Bond lengths (A) and selected angles (¡) for 2 (ESDs in parentheses)
Fe(1)ÈC(1)
Fe(1)ÈC(2)
Fe(1)ÈC(3)
Fe(1)ÈC(4)
Fe(1)ÈC(5)
Fe(1)ÈC(6)
Fe(1)ÈC(7)
Fe(1)ÈC(8)
Fe(1)ÈC(9)
Fe(1)ÈC(10)
Fe(1)ÈP(1)
P(1)ÈC(1)
2.0004(18)
2.0277(19)
2.0856(19)
2.0851(18)
2.0320(18)
1.9951(18)
2.0258(18)
2.0932(18)
2.0983(19)
2.0346(18)
2.6847(5)
1.8324(19)
1.8276(18)
1.8085(19)
1.917(2)
C(1)ÈC(5)
1.453(3)
1.430(3)
1.431(3)
1.424(3)
1.448(3)
1.452(3)
1.427(3)
1.430(3)
1.425(3)
1.398(3)
1.397(3)
1.390(3)
1.394(3)
1.390(3)
1.392(3)
C(2)ÈC(3)
C(3)ÈC(4)
C(4)ÈC(5)
C(6)ÈC(10)
C(6)ÈC(7)
C(7)ÈC(8)
C(8)ÈC(9)
C(9)ÈC(10)
C(11)ÈC(16)
C(11)ÈC(12)
C(12)ÈC(13)
C(13)ÈC(14)
C(14)ÈC(15)
C(15)ÈC(16)
P(1)ÈC(6)
P(1)ÈC(11)
P(1)ÈB(1)
C(1)ÈC(2)
C(1)ÈP(1)ÈC(6)
C(11)ÈP(1)ÈB(1)
C(1)ÈP(1)ÈC(11)
1.451(3)
96.06(8)
115.49(9)
107.12(8)
C(1)ÈP(1)ÈB(1)
C(5)ÈC(1)ÈC(2)
C(12)ÈC(11)ÈC(16)
114.11(9)
106.34(16)
119.91(17)
Table 3 Bond lengths (A) and selected angles (¡) for 3 É CH Cl (ESDs in parentheses)
2
2
Fe(1)ÈC(1)
1.997(2)
2.041(2)
2.103(2)
2.090(2)
2.024(2)
2.6464(9)
1.812(2)
1.796(3)
1.976(4)
1.838(4)
Cl(2)ÈB(1)
C(1)ÈC(5)
C(1)ÈC(2)
C(2)ÈC(3)
C(3)ÈC(4)
C(4)ÈC(5)
C(6)ÈC(7)
C(7)ÈC(8)
C(8) C(9)
1.850(2)
1.451(3)
1.454(3)
1.422(3)
1.428(4)
1.424(3)
1.397(3)
1.385(3)
1.389(3)
Fe(1)ÈC(2)
Fe(1)ÈC(3)
Fe(1)ÈC(4)
Fe(1)ÈC(5)
Fe(1)ÈP(1)
P(1)ÈC(1)
P(1)ÈC(6)
P(1)ÈB(1)
Cl(1)ÈB(1)
C(1)ÈP(1)È C(1)A
C(6)ÈP(1)ÈB(1)
C(6)ÈP(1)ÈC(1)
C(1)ÈP(1)ÈB(1)
97.83(13)
109.94(15)
108.81(9)
115.36(10)
C(5)ÈC(1)ÈC(2)
C(7)ÈC(6)ÈC(7)A
Cl(1)ÈB(1)ÈCl(2)
106.88(19)
119.9(3)
111.03(14)
New J. Chem., 2000, 24, 447È453
449

View Article Online
Scheme 1
between 1 and its sulfurized analogue (along with a shift in the
31P resonance from 10 to 46 ppm) and, in the latter case, a
decrease in FeÈP distance of 0.197(6) A was observed between
1 and 7 (along with a shift in the 31P resonance from 10 to 38
ppm). The changes in FeÈP distance between 1 and both 2
and 3 É CH Cl are intermediate to these, with a decrease of
of phosphorus. Such a conclusion must be treated with
caution, however, since it is known that the energies of the
signiÐcant orbitals are a†ected not only by the ringsÏ substit-
uents, but also by the magnitude of the tilt angle, a.15
The thermal behaviour of 2 was examined by di†erential
scanning calorimetry, and revealed a ROP exotherm centred
at 166 ¡C, which integrated to 60 kJ mol~1. This represents a
lower limit to the energy released during thermal ROP, since
no melt endotherm was apparent. Based on this information,
thermal polymerization of 2 was e†ected in a sealed, evac-
uated Pyrex tube at 170 ¡C over 30 min. This resulted in an
insoluble redÈbrown product, 5a, which was characterized by
pyrolysis-MS and solid state NMR. The pyrolysis-MS con-
tained peaks consistent with the structure expected for the
polymer. For example, peaks corresponding to ring-opened 1
2
2
0.085(4) A between 1 and 2 and of 0.128(4) A between 1 and
3 É CH Cl . Somewhat surprising is the fact that the 31P
2
2
NMR resonance of 3 occurs at a higher Ðeld (19 ppm) than
that of 2 (42 ppm). One would expect, using electronegativity
arguments, BH to be a weaker Lewis acid than BCl , and
3
3
thus to have a less deshielding e†ect upon the phosphorus.12
It has been previously noted, however, that BH can behave
as an anomalously strong acceptor towards phosphines
3
(relative to haloboranes) compared to its behaviour towards
amines.8,12,13
[FcP(Ph)H, Fc \ ferrocenyl] and cyclic dimers of
1
The UVÈVis data, however, do not seem to support BH
M[fcP(Ph)] , fc \ ferrocenediylN were observed. The MS of 2
3
2
behaving as a stronger acceptor towards phosphines. For 1,14
does not result in the same fragmentation pattern, exhibiting a
2 and 3 in CH Cl , the maxima of the lowest energy bands,
small molecular ion peak at 306, and peaks for 1 and the step-
wise loss of CpH and Fe from 1. This may be due in part to
the fact that ionization of 2 was e†ected at a considerably
lower temperature than that used for 5a.
2
2
similar to ferroceneÏs ““band IIÏÏ, occur at 507, 499 and 494
nm, respectively. It appears, therefore, that as phosphorus is
made more electron deÐcient, the band maximum shifts to
higher energy. This is supported by the position of the band
maximum for 7, which occurs at 484 nm (in THF). While this
datum cannot be exactly compared to the data obtained in
CH Cl , it does clearly support the conclusion that the band
Solid-state CP-MAS NMR of 5a revealed two broad peaks
in the 13C NMR spectrum, consistent with the Cp and phenyl
regions. A non-quaternary suppression (NQS)16 experiment
revealed a quaternary carbon within the band envelope of
each peak, also consistent with the presence of bound Cp and
phenyl moieties. Singlet resonances were observed in the 31P
and 11B solid-state NMR spectra at ca. 7 and ca. [40 ppm,
respectively. Elemental analysis of 5a could not, unfortunately,
conÐrm the productÏs identity. The %C determined was more
than 10% below the expected value, probably due to the for-
mation of carbon-containing, combustion-resistant ceramics.
The formation of such ceramics at elevated temperatures is
known for ferrocene-containing polymers.17
2
2
maximumÏs position is related to the electronic environment
Table 4 Selected bond lengths (A) and angles (¡) for a variety of
phosphorus-bridged [1]ferrocenophanes. The deÐnition of the various
angles is given in Fig. 5. ESDs (where available) are given in parenth-
eses after each value
1
2
3 É CH Cl
7
2
2
In order to provide conÐrmation of the identity of 5a,
the borane adduct of 4 was synthesized. A sample of 4
a
26.7
32.5
90.6(3)
159.8
2.774(3)
1.842(13)
24.38(10)
36.0(1)
96.06(8)
162.2(1)
2.6847(5)
1.8300(19)
1.917(2)
25.01(4)
36.6(2)
97.83(13)
162.2(1)
2.6464(9)
1.812(2)
1.976(4)
24.4(5)
37.9(4)
99.8(4)
163.6(4)
2.577(3)
1.772(10)
b
[M \ 19 000; polydispersity index (PDI) 1.14], obtained by
n
h
anionic polymerization of 1, was treated with BH in THF
3
d
at 0 ¡C in a manner similar to the synthesis of 2. The yellow
FeÈP
product, 5b, a†orded 1H and 13C NMR spectra consistent
with formation of a borane adduct of 4. For example, the
ipso-carbons in 5b are shielded by 8 (Ph) and 3 (Cp) ppm
C
ÈP
ipso
PÈB
È
È
450
New J. Chem., 2000, 24, 447È453

View Article Online
from their positions in 4.18 Considering the corresponding
monomers, the ipso-carbon resonances of 2 are shielded by 9
(Ph) and 1 (Cp) ppm relative to their positions in 1. In the 1H
NMR spectrum of 5b, the most signiÐcant feature is the broad
tation pattern to that observed for 5a. The pyrolysis-MS of 6c
contains many of the fragments observed for 6b, but two large
peaks unobserved in the MS of 6b dominate the fragmenta-
tion pattern. The peaks at 376 and 309 may arise from frag-
mentation of the unidentiÐed species responsible for the 31P
NMR resonance at 35 ppm. From the intensity of the frag-
ments in the MS of 6c associated with fragments observed for
6b, we tentatively suggest that the product formed in the
peak associated with BH . The remainder of the proton spec-
3
trum is very similar in appearance to that of the sulfurized
derivative of 4,2 as might be expected. SigniÐcantly, the 31P
NMR spectrum showed a singlet peak at 8 ppm, and the 11B
NMR spectrum a singlet peak at [40 ppm, as were observed
for 5a. The pyrolysis-MS of 5b contains fragments such as
transition-metal catalyzed ROP of
3 is perhaps 20%
poly(ferrocenylphenylphosphine trichloroborane). While the
amount of product generated certainly indicates that a cata-
lytic process is at work, it is clear that the reaction mechanism
is dominated by processes other than the desired chain-
growth polymerization. These may involve, for example, reac-
tions with BÈH or BÈCl bonds in the case of 2 and 3
respectively.
FcP(Ph) and [fcP(Ph)] , n \ 2È4.
n
The correspondence of the solid-state NMR and pyrolysis-
MS data is highly suggestive that polymers 5a and 5b are the
same species, and that the insolubility of 5a is due either to a
very high molecular weight or to some degree of crosslinking
or to a combination of both these factors. Attempts to deter-
mine the molecular weight of 5b by gel permeation chroma-
tography (GPC) resulted in estimates of M of ca. 9000. This is
n
about half the expected value, and may be due to interactions
Conclusions
between the polymer and the column material. Polymer 4
itself does not elute from a GPC column due to such inter-
actions, unless it has Ðrst been derivatized with sulfur to a†ord
the analogous phosphine sulÐde polymer.2 The tail in the
The phosphorus-bridged [1]ferrocenophane 1, and the corre-
sponding ring-opened polymer, 4, readily form adducts with
boranes BX (X \ H, Cl). The monomeric borane adducts (2
3
and 3) undergo thermal polymerization to yield products (5a
GPC trace of 5b (which extends to M ca. 19 000), and the
n
and 6a) spectroscopically consistent with 5b and 6b, the
borane adducts of 4 with BH and BCl , respectively.
broad PDI (about 2), are suggestive of some polymerÈcolumn
interactions that retard elution.
3
3
Attempts at transition-metal catalyzed ROP of 2 and 3 led to
insoluble, poorly deÐned products. The reaction of 3 with
Attempts to polymerize 2 via transition-metal catalyzed
ROP using PtCl , KarstedtÏs catalyst, [Rh(cod) ]OTf
2
2
Pt(cod) led to the formation of a mixture of products, 6c, due
to a catalytic process. Comparison of spectral data of 6c with
(cod \ 1,5-cyclooctadiene), and Pt(cod) met with limited
2
2
success, typeÐed by extremely poor yields of dark, insoluble
that for 6b suggests the presence of the ring-opened polymer
derived from 3 in low yield.
product. The transition-metal catalyzed ROP attempt on 3
with Pt(cod) , in contrast, resulted in a bright orange precipi-
2
tate that was extracted with CH Cl , resulting in a tan solid
2
2
(6c) which was further analyzed by solid-state NMR and
pyrolysis-MS. As with 5a, 6c possessed two broad peaks in its
13C NMR spectrum, corresponding to the Cp (70 ppm) and
phenyl (130 ppm) carbons. The NQS experiment indicated
quaternary carbons in the same regions, as expected. The 11B
NMR spectrum of 6c consisted of a singlet at 0 ppm with a
broad shoulder to high Ðeld, while the 31P NMR spectrum
contained two broad singlets at [15 and 35 ppm, with the
latter being of much greater intensity. As in the case of 5a, we
wished to synthesize a soluble adduct of polymer 4 to provide
Experimental
Materials and methods
All reactions involving monomer were carried out under
nitrogen atmospheres using standard Schlenk techniques in
conjunction with an Innovative Technologies drybox. THF
and hexanes were distilled from Na, and CH Cl was dried on
2
2
a
Grubbs-type19 solvent puriÐcation system using A-2
comparative NMR data for 6c. The BCl adduct of 4, 6b, is,
however, an insoluble yellowÈorange solid. It was examined
alumina under a nitrogen atmosphere, and deoxygenated by
three freeze-pump-thaw cycles. BH (1 M in THF) and BCl
(1 M in heptane) were purchased from Aldrich and used as
3
3
3
by solid-state NMR, and was found to exhibit two broad
peaks in its 13C NMR spectrum, at 75 and 135 ppm, which
compare favourably to the positions of the analogous peaks
observed for 6c (at 70 and 130 ppm, respectively), considering
the peak widths. In the NQS experiment on 6b, a single quat-
ernary carbon resonance was observed in the Cp region (at ca.
70 ppm), and two resonances due to quaternary carbons in the
phenyl region (at ca. 128 and 133 ppm). Its 31P NMR spec-
trum contains a single peak (at [15 ppm), but two peaks are
seen in the 11B NMR spectrum, at 2 and 8 ppm. The peak at
2 ppm shows a pronounced shoulder (at ca. 4 ppm), which is
roughly consistent with the expected magnitude of BÈP coup-
ling. The reason why multiple boron and phenyl ipso-carbon
resonances are observed (but a single phosphorus resonance)
is not clear. Polymer 6a, an insoluble purpleÈbrown solid
obtained in low yield from the thermal ROP of 3, exhibited
peaks in its 11B and 31P NMR spectra at 4 and [20 ppm,
respectively, in excellent agreement with the data for 6b.
received. Compound 1 was prepared using a published pro-
cedure.20 Polymer 4 [M \ 19 000 (calcd 18 800); PDI 1.14]
n
was prepared by living anionic ROP in triply distilled THF
using n-BuLi (1.6 M in hexanes) as the initiator and degassed
H O as the terminator.1b
2
GPC were obtained on a Waters 2690 liquid chromato-
graph equipped with a model 510 HPLC pump, a model U6K
injector, Ultrastyragel columns with pore sizes of 103È105 A,
and a Waters 410 di†erential refractometer. The eluent was
THF with a Ñow rate of 1.0 mL min~1, and results are report-
ed vs. polystyrene standards. The sample of 4 was Ðrst deriva-
tized with sulfur in order to allow it to pass through the
column. This methodology has been discussed elsewhere.2
Solution NMR spectra were obtained on a Varian Unity
500 spectrometer (1H at 500.0 MHz, 11B at 160.5 MHz), a
Varian VXR 400 spectrometer (1H at 400.0 MHz, 13C at 100.6
MHz), or a Varian Gemini-300 spectrometer (31P at 121.6
MHz). 1H, 13C and 31P spectra were referenced to solvent.
11B spectra were run unlocked, and the glass peak was
removed using a 32 point backwards linear prediction on a
basis of 512 points, using 16 coefficients. Chemical shifts are
reported relative to TMS (1H and 13C), 85% H PO (31P), or
The pyrolysis-MS data of 6a, 6b and 6c were consistent
with the NMR data. The fragmentation pattern of 6b at
300 ¡C revealed peaks corresponding to BCl (116) and BCl
3
2
(81), with no fragmentation of the polymer backbone being
observed. At 550 ¡C, however, fragmentation of the backbone
3
4
results in peaks due to CpH, P(Ph), FcH and/or Ph PH,
BF É O(Et) (11B). Solid-state CP-MAS NMR spectra were
2
3
2
FcP(Ph), fcPFc, Fc P(Ph), and [fcP(Ph)] ; a similar fragmen-
obtained on a Bruker DSX-400 (13C at 100.6 MHz) or Bruker
2
2
New J. Chem., 2000, 24, 447È453
451

View Article Online
DSX-200 (11B at 64.2 MHz, 31P at 81.0 MHz) spectrometers
and were referenced to adamantane (13C, 38.4 ppm vs. TMS),
85% H PO (31P), or NaBH [11B, [42.0 ppm vs.
fcP(Ph) É BH , 584 (32) [fcP(Ph)] . Elem. anal. calcd (%): C,
62.82; H, 5.27; Found (%): C, 62.21; H, 5.19.
3
2
3
4
4
BF É O(Et) ].21 The 11B NMR spectra of the samples con-
3
2
taining BCl were run as 90¡-q-180¡ spin-echo experiments,
P-Phenylphospha[1]ferrocenophane trichloroborane (3).
Compound 1 (497 mg, 1.70 mmol, FW 292.10) was dissolved
in CH Cl (10 mL), cooled to [40 ¡C, and trichloroborane
3
where q is one rotor period.
Mass spectrometry was performed on a Micro Mass
70S-250 mass spectrometer in electron impact mode at 70 eV.
Ionization was achieved thermally at ca. 200 ¡C for 2 and 3,
using a temperature ramp of 0.5 ¡C s~1 from ca. 40 to 533 ¡C
for 5a, discretely at 300 and 550 ¡C for 6a and 6b, and dis-
cretely at 200 and 550 ¡C for 5b and 6c. Di†erential scanning
calorimetry was performed on a PerkinÈElmer DSC-7 cali-
brated to cyclohexane and indium. A heating rate of 10 ¡C
min~1 was used between 40 and 200 ¡C. UVÈVis spectroscopy
was obtained in CH Cl on a PerkinÈElmer Lambda900 UV/
2
2
(1.8 mL, 1 M in heptane, 1.06 equiv.), also at [40 ¡C, added
quickly. Precipitate formation was noted immediately, and
the reaction mixture was left overnight at [40 ¡C. The
supernatant liquor was decanted, the product dissolved in a
minimum of CH Cl (15È20 mL) and recrystallized overnight
2
2
(again at [40 ¡C), resulting in the formation of orangeÈred
crystals (3 É CH Cl ) suitable for X-ray di†raction analysis.
2
2
The product was dried in vacuo at ambient temperature for
3 h and was isolated in 84% yield (587 mg, 1.43 mmol, FW
409.27). 1H (500.0 MHz, CD Cl , 293 K) d 4.38 (s, Cp, 2 H),
2
2
Vis/NIR spectrometer. Elemental analysis was performed by
Quantitative Technologies Inc., Whitehouse, NJ.
2
2
4.74 (s, Cp, 2 H), 4.84 (s, Cp, 2 H), 5.43 (s, Cp, 2 H), 7.66 (s, Ph,
2 H), 7.78 (s, Ph, 1 H), 7.99 (s, Ph, 2 H). 13C (100.6 MHz,
CD Cl , 293 K) d 10.6 (d, 1J \ 33 Hz, ipso-Cp), 77.5 (d,
X-Ray crystallography data were collected on a Nonius
Kappa CCD di†ractometer using graphite monochromated
Mo-Ka radiation (j \ 0.710 73 A). A combination of 1¡ phi
and omega (with kappa o†sets) scans were used to collect suf-
Ðcient data. The data frames were integrated and scaled using
the Denzo-SMN package.22 The structures were solved and
reÐned using the SHELXTLCPC v5.1 package.23 ReÐnement
was by full-matrix least-squares on F2 using all data (negative
intensities included). Hydrogen atoms were included in calcu-
lated positions.
2
2
PC
J
J
\ 14 Hz, Cp), 77.7 (s, Cp), 81.1 (d, J \ 7 Hz, Cp), 82.4 (d,
PC
PC
PC
\ 8 Hz, Cp), 121.2 (d, 1J \ 69 Hz, ipso-Ph), 129.7 (d,
PC
PC
2J \ 11 Hz, Ph), 133.8 (d, 3J \ 9 Hz, Ph), 134.1 (d, 4J
\
PC
PC
2 Hz, Ph). 11B (160.5 MHz, CD Cl , 293 K) d [8.0 (d,
2
2
1J \ 154 Hz). 31P (121.6 MHz, CD Cl , 293 K) d 18.5 (q,
PB
2
2
1J \ 157 Hz). UVÈVis (CH Cl ) j
494 nm, e \ 4.33
BP
2
2
max
494
^ 0.06 ] 102 L mol~1 cm~1. Pyrolysis-MS (ca. 200 ¡C, EI, 70
eV) m/z (%) 56 (18) Fe, 81 (26) BCl , 170 (22) [226 [ Fe], 226
2
CCDC reference number 440/176. See http://www.rsc.org/
suppdata/nj/b0/b000935k/ for crystallographic Ðles in .cif
format.
(42) [292 [ CpH], 292 (100) fcP(Ph), 328 (23) FcP(Ph)Cl
(Fc \ ferrocenyl). Elem. anal. cald (%): C, 46.96; H, 3.20.
Found (%): C, 46.99; H, 3.16.
Syntheses
Thermal polymerization of
2 to give poly-P-phenyl-
phospha[1]ferrocenophane borane (5a). Compound 2 (150 mg,
0.490 mmol, FW 305.93) was sealed into an evacuated Pyrex
tube and placed in a polymerization oven at 170 ¡C for 30
min. The cooled tube was opened under inert atmosphere.
The redÈbrown product, 5a, was obtained quantitatively, and
found to be essentially insoluble in all solvents. (In CH Cl ,
P-Phenylphospha[1]ferrocenophane borane (2). Compound
1 (222 mg, 0.760 mmol, FW 292.10) was dissolved in THF (3
mL), cooled in an ice bath, and borane (0.84 mL, 1 M in THF,
1.1 equiv.) was added slowly. The solution was allowed to stir
for several minutes, and an aliquot was removed to determine
reaction progress. 100% conversion (by 31P NMR) was
observed with the Ðrst aliquot (removed after 25 min.);
hexanes (5 mL, ca. 2 equiv. v/v with THF) were added to the
solution, which was then cooled to [20 ¡C for 2 h, resulting
in orangeÈred crystals suitable for X-ray di†raction analysis.
An isolated yield of 89% was obtained after drying overnight
in vacuo at ambient temperature (206 mg, 0.673 mmol, FW
2
2
CHCl and THF a slight colouration of the solvent was
3
observed.) Solid-state NMR: 13C (100.6 MHz, 293 K) d 70 (s,
Cp, 1.7 kHz), 130 (s, Ph, 2 kHz). 11B (64.2 MHz, 293 K) d [42
(s, 2.2 kHz). 31P (81.0 MHz, 293 K) d 7.9 (s, 350 Hz).
Pyrolysis-MS (T \ 533 ¡C, EI, 70 eV) m/z (%) 66 (35) CpH,
max
121 (45) CpFe, 186 (100) ferrocene and/or Ph PH, 294 (52)
2
FcP(Ph)H, 584 (32) [fcP(Ph)] . Elem. anal. calcd (%): C,
2
305.93). 1H (400.0 MHz, CDCl , 293 K) d 1.0 (br 1 : 1 : 1 : 1 q,
62.82; H, 5.27. Found (%): C, 52.05; H, 4.97.
3
1J \ ca. 100 Hz, BH ), 4.33 (s, Cp, 2 H), 4.56 (s, Cp, 2 H),
BH
3
4.60 (s, Cp, 2 H), 4.92 (s, Cp, 2 H), 7.55 (m, Ph, 3 H), 7.82 (m,
Ph, 2 H); 1H (400.0, MHz, C D , 293 K) d 1.95 (br q, 1J
\
Poly-P-phenylphospha[1]ferrocenophane
borane
(5b).
6
6
BH
97 Hz, av., BH ), 3.81 (s, Cp, 2 H), 4.09 (s, Cp, 2 H), 4.15 (s, Cp,
Polymer 4 (100 mg, M \ 19 000, PDI \ 1.14) was dissolved
3
n
2 H), 4.88 (s, Cp, 2 H), 7.32 (m, Ph, 3 H), 7.60 (ddd, 3J \ 12
in THF (7 mL), cooled in an ice bath and borane (0.6 mL, 1 M
PH
Hz, 3J \ 7.5 Hz, 4J \ 1.9 Hz, Ph-ortho, 2 H). Note: the
“singletsÏ observed for the Cp resonances in C D were
in THF, 1.8 equiv.) was added slowly. The solution was
warmed to room temperature, allowed to stir overnight, con-
centrated to ca. 4 mL, and was added dropwise to hexanes
(100 mL) with vigorous stirring. The solid was redissolved in
THF (1 mL), reprecipitated by dropwise addition into rapidly
stirred hexanes, collected by vacuum Ðltration and dried over-
night in vacuo at 35 ¡C. While the reaction appeared quanti-
tative by 31P NMR, the isolated yield of yellow polymer was
78 mg (74%). 1H (400.0 MHz, CD Cl , 293 K) d 1.12 (br s
HH
HH
6
6
actually Ðnely split multiplets (J O 1 Hz). 13C (100.6 MHz,
CDCl , 293 K) d 16.3 (d, 1J \ 32 Hz, ipso-Cp), 76.0 (s, Cp),
3
PC
76.4 (d, 2J \ 18 Hz, Cp), 79.4 (d, 3J \ 5 Hz, Cp), 79.7 (d,
PC
2J \ 9 Hz, Cp), 128.5 (d, 1J \ 61 Hz, ipso-Ph), 129.2 (d,
PC PC
2J \ 10 Hz, Ph), 130.6 (d, 3J \ 10 Hz, Ph), 131.5 (d,
PC PC
PC
4J \ 2 Hz, Ph). Note: n of the various nJ were assigned
PC
PC
with the assumption that the magnitude of J decreases as n
2
2
increases. Quaternary carbons were assigned on the basis of
BH ), 3.8È4.5 (m, Cp, 8 H), 7.43 (s, Ph, 2 H), 7.48 (s, Ph, 1 H),
3
intensity. 11B (160.5 MHz, CDCl , 293 K) d [37.5 (d, 1J
\
7.66 (s, Ph, 2 H). 13C (75.5 MHz, CD Cl , 293 K) d 73.0 (d,
3
PB
2 2
39 Hz). 31P (121.55 MHz, CDCl , 293 K) d 42.6 (d, 1J \ 58
1J \ 68 Hz, ipso-Cp), 73.2 (d, J \ 8 Hz, Cp), 73.8 (d,
3
BP
PC
PC
PC
Hz). DSC (10 ¡C min~1) 60 kJ mol~1 ROP exotherm (onset
J \ 12 Hz, Cp), 73.9 (d, J \ 5 Hz, Cp), 74.8 (d, J \ 6 Hz,
PC
PC
164 ¡C, peak 166 ¡C). UVÈVis (CH Cl ) j
499 nm, e
\
Cp), 128.8 (d, J \ 9 Hz, Ph), 130.8 (s, ipso-Ph), 131.6 (s, Ph),
2
2
max
499
PC
5.04 ^ 0.08 ] 102 L mol~1 cm~1. Pyrolysis-MS (ca. 200 ¡C,
EI, 70 eV) m/z (%) 56 (15) Fe, 170 (19) [226 [ Fe], 226 (37)
[292 [ CpH], 292 (100) fcP(Ph) (fc \ ferrocenediyl), 306 (2)
132.5 (d, J \ 9 Hz, Ph). 11B (160.5 MHz, CD Cl , 293 K)
PC
2 2
d [40 (s). 31P (121.6 MHz, CD Cl , 293 K) d 7.7 (s). GPC
2
2
M \ 9000; PDI \ 1.94. These GPC data should be
n
452
New J. Chem., 2000, 24, 447È453

View Article Online
considered to be inaccurate, as interactions between the
analyte and the column material have likely skewed the
results. Pyrolysis-MS (550 ¡C, EI, 70 eV) m/z (%) 78 (72) PhH,
Acknowledgements
C. E. B. E. and I. M. thank the Natural Sciences and Engin-
eering Research Council of Canada for Ðnancial support. We
are also grateful to Alex Young for the mass spectral measure-
ments and to Karen Temple for the DSC data.
108 (66) P(Ph), 186 (62) FcH and/or Ph PH, 293 (29) FcP(Ph),
2
400 (8) fcPFc, 478 (12) Fc P(Ph), 584 (31) [fcP(Ph)] , 876 (4)
2
2
[fcP(Ph)] , 1168 (0.5) [fcP(Ph)] .
3
4
Thermal polymerization of
3
to give poly-P-phenyl-
phospha[1]ferrocenophane trichloroborane (6a). Compound 3
(200 mg, 0.489 mmol, FW 409.27) was heated to 170 ¡C for 30
min. under a nitrogen atmosphere. No melting of the solid
was observed, although a colour change from orangeÈred to
purpleÈbrown occurred. The resultant solid was extracted
with 1.5 mL aliquots of CH Cl until the extracts were
colourless. After drying in vacuo the product 6a (36 mg, 18%)
was characterized by 11B and 31P NMR spectroscopy. Solid-
state NMR: 11B (64.2 MHz, 293 K) d 4 (s, 1.1 kHz). 31P (81.0
MHz, 293 K) d [20 (s, 2.5 kHz). Pyrolysis-MS (550 ¡C, EI, 70
eV) m/z (%) 66 (64) CpH, 78 (100) PhH, 108 (52) P(Ph), 186
(75) FcH, 293 (1) FcP(Ph).
Notes and references
1
(a) C. H. Honeyman, T. J. Peckham, J. A. Massey and I.
Manners, Chem. Commun., 1996, 2589; (b) T. J. Peckham, J. A.
Massey, C. H. Honeyman and I. Manners, Macromolecules, 1999,
32, 2830.
2
2
2
3
C. H. Honeyman, D. A. Foucher, F. Y. Dahmen, R. Rulkens, A.
J. Lough and I. Manners, Organometallics, 1995, 14, 5503.
For previous studies of the catalytic applications of
poly(ferrocenylphosphines) formed by condensation routes, see J.
D. Fellman, P. E. Garrou, H. P. Withers, D. Seyferth and D. D.
TraÐcante, Organometallics, 1983, 2, 818.
4
5
6
7
P. GomezÈElipe, P. M. Macdonald and I. Manners, Angew.
Chem., Int. Ed. Engl., 1997, 36, 762.
Poly-P-phenylphospha[1]ferrocenophane
trichloroborane
P. GomezÈElipe, R. Resendes, P. M. Macdonald and I. Manners,
J. Am. Chem. Soc., 1998, 120, 8348.
T. J. Peckham, A. J. Lough and I. Manners, Organometallics,
1999, 18, 1030.
The numbers reported for the Cp ipso-carbon must be treated
with caution as the peaks were ill-resolved, with an extremely
poor signal-to-noise ratio. The peaks observed at ca. 10 ppm do
not, however, appear to be artifacts.
(6b). Polymer 4 (67 mg, M \ 19 000, PDI \ 1.14) was dis-
n
solved in CH Cl (7 mL), cooled to [40 ¡C, and trichloro-
2
2
borane (0.25 mL, 1 M in heptane, 1.1 equiv., also at [40 ¡C)
was added. A bright yellowÈorange precipitate formed imme-
diately. The colourless supernatant was decanted, and the pre-
cipitate was rinsed with a small volume of CH Cl (resulting
2
2
in a colourless supernatant liquor) prior to being dried in
vacuo for 2 h at 35 ¡C, to yield 62 mg (66%) of a bright
yellowÈorange powder (insoluble in hexanes, toluene,
dichloromethane and THF). Solid-state NMR: 13C (50.3
MHz, 293 K) d 75 (s, Cp, 1.3 kHz), 135 (s, Ph, 1.1 kHz). 11B
(64.2 MHz, 293 K) d 2 (s, 450 Hz), 8 (s, 350 Hz). 31P (81.0
MHz, 293 K) d [15 (s, 1.1 kHz). Pyrolysis-MS (300 ¡C, EI, 70
eV) m/z (%) 81 (100) BCl , 116 (35) BCl ; (550 ¡C, EI, 70 eV)
8
9
D. L. Black and R. C. Taylor, Acta Crystallogr., Sect. B, 1975, 31,
1116.
P. S. Bryan and R. L. Kuczkowski, Inorg. Chem., 1972, 11, 553.
10 H. StoeckliÈEvans, A. G. Osborne and R. H. Whiteley, J.
Organomet. Chem., 1980, 194, 91.
11 W. Finckh, B.-Z. Tang, D. A. Foucher, D. B. Zamble, R. Ziem-
binski, A. Lough and I. Manners, Organametallics, 1993, 12, 823.
12 Comprehensive Coordination Chemistry, ed. G. Wilkinson,
Pergamon Press, Oxford, UK, 1987, vol. 3, ch. 24.
13 J. Apel and J. Grobe, Z. Anorg. Allg. Chem., 1979, 453, 28.
14 UVÈVis data for 1 in CH Cl has not been published elsewhere:
2
3
m/z (%) 66 (100) CpH, 108 (54) P(Ph), 186 (91) FcH, 293 (25)
FcP(Ph), 400 (19) fcPFc, 478 (13) Fc P(Ph), 584 (9) [fcP(Ph)] .
2
2
2
2
j
507 nm, e \ 4.38 ^ 0.05 ] 102 L mol~1 cm~1.
max
494
Reaction of
3 with Pt(cod) to give poly-P-phenyl-
15 (a) S. Barlow, M. J. Drewitt, T. Dijkstra, J. C. Green, D. OÏHara,
C. Whittingham, H. H. Wynn, D. P. Gates, I. Manners, J. M.
Nelson and J. K. Pudelski, Organometallics, 199817, 2113; (b) J.
C. Green, Chem Soc. Rev., 1998, 27, 263; (c) A. Berenbaum, H.
Braunschweig, R. Dirk, U. Englert, J. C. Green, F. Jakle, A. J.
Lough and I. Manners, J. Am. Chem. Soc., in press
2
phospha[1]ferrocenophane trichloroborane (6c). A solution of
compound 3 (200 mg) in CH Cl (12 mL) was treated with
Pt(cod) (12 mg, 6 mol %), and stirred for 48 h, resulting in
2
2
2
the precipitation of a bright orange product. The yellowÈ
orange supernatant was decanted, and the product was rinsed
with hexanes and dried in vacuo for 2 h. The solid residue was
extracted 5 times with 1.5 mL aliquots of CH Cl (until the
16 S. J. Opella and M. H. Frey, J. Am. Chem. Soc., 1979, 101, 5854.
17 B.-Z. Tang, R. Petersen, D. A. Foucher, A. Lough, N. Coombs, R.
Sodhi and I. Manners, J. Chem. Soc., Chem. Commun., 1993, 523.
18 13C (75.5 MHz, CD Cl , 293 K) data for 4 has not been
2
2
supernatant liquors were colourless), and the insoluble tan
product (147 mg, 74%) was dried in vacuo for 2 h at 35 ¡C.
Solid-state NMR: 13C (100.6 MHz, 293 K) d 70 (s, Cp, 2 kHz),
130 (s, Ph, 2 kHz). 11B (64.2 MHz, 293 K) d 0 (s, 1.5 kHz). 31P
(81.0 MHz, 293 K) d 35 (s, 2 kHz), [15 (s, 2 kHz). Pyrolysis-
MS (200 ¡C, EI, 70 eV) m/z (%) 173 (100) CpPhPH, 238 (54)
2
2
published elsewhere: d 71.8 (s, Cp), 72.5 (s, Cp), 73.0 (d, J \ 12
PC
Hz, Cp), 74.1 (d, J \ 12 Hz, Cp), 78.9 (d, 1J \ 5 Hz, ipso-Cp),
PC
PC
128.2 (d, J \ 7 Hz, Ph), 129.2 (s, Ph), 134.3 (d, J \ 22 Hz, Ph),
PC
PC
139.5 (d, 1J \ 10 Hz, ipso-Ph).
PC
19 A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen and
F. J. Timmers, Organometallics, 1996, 15, 1518.
20 D. Seyferth and H. P. Withers, Jr., Organometallics, 1982, 1, 1275.
21 R. K. Harris, J. Bowles, I. R. Stephenson and E. H. Wong,
Spectrochim. Acta, Part A, 1988, 44, 273.
22 Z. Otwinowski and W. Minor, Methods Enzymol., 1997, 276, 307.
23 G. M. Sheldrick, SHEL XT L CPC, v5.1, Bruker Analytical X-ray
Systems, Madison, WI, USA, 1997.
Cp P(Ph), 294 (9) FcP(Ph)H, 328 (31) FcP(Ph)Cl; (550 ¡C, EI,
2
70 eV) m/z (%) 66 (30) CpH, 121 (34) CpFe, 186 (100) FcH
and/or Ph PH, 262 (21) unassigned, 294 (13) FcP(Ph)H, 309
2
(53) unassigned, 376 (41) unassigned, 400 (3) fcPFc, 478 (4)
Fc P(Ph).
2
New J. Chem., 2000, 24, 447È453
453