Synthesis, structure, and ring-opening polymerization (ROP) of a phosphonium-bridged [1]ferrocenophane

Article abstract of DOI:10.1021/om980696t

The stable phosphonium-bridged [1]ferrocenophane [(η-C₅H₄)₂FePPhMe][OTf] (1Ob) was synthesized by the reaction of the phosphorus-bridged [1]ferrocenophane (η-C₅H₄)₂FePPh (3a) with methyl triflate (MeOTf). A single-crystal X-ray diffraction study of 10b revealed an angle of 24.4(5)° between the planes of the cyclopentadienyl rings, less than the respective angle (26.7°) for 3a. Compound 10b and a number of other tetracoordinate, phosphorus-bridged [1]ferrocenophanes, (η-C₅H₄)₂FeP(S)Ph (5a), (η-C₅H₄)₂FeP[Fe(CO)4]Ph (6), and [(η-C₅H₄)₂FePFpPh]tPF₆] (7) (where Fp = (η-C₅H₅)Fe(CO)₂), were investigated with respect to their ring-opening polymerization (ROP) behavior. Only compound 10b was found to undergo ROP, which occurred both thermally and in the presence of a transition-metal catalyst (PtCl₂). The resultant ionomeric polymer {[(η-C₅H₄)₂FePPhMe][OTf]}_η (11) was found to be soluble in dimethylformamide (DMF), dimethyl sulfoxide (DMSO), methanol, and acetone but displayed only limited stability in these solvents. The thermally ring-opened polymer was found to possess a glass transition temperature of 176 °C and was thermally stable to weight loss up to ca. 400 °C. Analysis by wide-angle X-ray scattering (WAXS) revealed that the polymer was amorphous. A study on the partial to full methylation of the polymer [(η-C₅H₄)₂-FePPh]_n (14) gave results that were consistent with those from the ROP of 10b. Dynamic light scattering studies on polymer 11 produced via thermal ROP and transition-metal-catalyzed ROP gave hydrodynamic radii in the range of 30-45 nm, which suggested that the compounds were polymeric rather than oligomeric in nature. On the basis of the glass transition temperatures for a series of samples of polymers 11 of known molecular weight with varying numbers of repeat units (from 20 to 100), the molecular weight (M_n) of the transition-metal-catalyzed ROP product was greater than 46 000 (ca. 100 repeat units) and the molecular weight of the thermally produced polymer was even higher.

Full text of DOI:10.1021/om980696t

1030
Organometallics 1999, 18, 1030-1040
Syn th esis, Str u ctu r e, a n d Rin g-Op en in g P olym er iza tion
(ROP ) of a P h osp h on iu m -Br id ged [1]F er r ocen op h a n e
Timothy J . Peckham, Alan J . Lough, and Ian Manners*
Department of Chemistry, University of Toronto, 80 St. George Street,
Toronto M5S 3H6, Ontario, Canada
Received August 17, 1998
The stable phosphonium-bridged [1]ferrocenophane [(η-C5H4)2FePPhMe][OTf] (10b) was
synthesized by the reaction of the phosphorus-bridged [1]ferrocenophane (η-C5H4)2FePPh
(3a ) with methyl triflate (MeOTf). A single-crystal X-ray diffraction study of 10b revealed
an angle of 24.4(5)° between the planes of the cyclopentadienyl rings, less than the respective
angle (26.7°) for 3a . Compound 10b and a number of other tetracoordinate, phosphorus-
bridged [1]ferrocenophanes, (η-C5H4)2FeP(S)Ph (5a ), (η-C5H4)2FeP[Fe(CO)4]Ph (6), and [(η-
C5H4)2FePFpPh][PF6] (7) (where Fp ) (η-C5H5)Fe(CO)2), were investigated with respect to
their ring-opening polymerization (ROP) behavior. Only compound 10b was found to undergo
ROP, which occurred both thermally and in the presence of a transition-metal catalyst (PtCl2).
The resultant ionomeric polymer {[(η-C5H4)2FePPhMe][OTf]}n (11) was found to be soluble
in dimethylformamide (DMF), dimethyl sulfoxide (DMSO), methanol, and acetone but
displayed only limited stability in these solvents. The thermally ring-opened polymer was
found to possess a glass transition temperature of 176 °C and was thermally stable to weight
loss up to ca. 400 °C. Analysis by wide-angle X-ray scattering (WAXS) revealed that the
polymer was amorphous. A study on the partial to full methylation of the polymer [(η-C5H4)2-
FePPh]n (14) gave results that were consistent with those from the ROP of 10b. Dynamic
light scattering studies on polymer 11 produced via thermal ROP and transition-metal-
catalyzed ROP gave hydrodynamic radii in the range of 30-45 nm, which suggested that
the compounds were polymeric rather than oligomeric in nature. On the basis of the glass
transition temperatures for a series of samples of polymers 11 of known molecular weight
with varying numbers of repeat units (from 20 to 100), the molecular weight (Mn) of the
transition-metal-catalyzed ROP product was greater than 46 000 (ca. 100 repeat units) and
the molecular weight of the thermally produced polymer was even higher.
In tr od u ction
Transition-metal-containing polymers remain an area
of active research due to the potential for obtaining
materials with possibly useful physical or catalytic
properties that are not easily obtainable with most
organic polymer systems.1-3 We have previously shown
that ring-opening polymerization (ROP) of strained [1]-
ferrocenophanes (1) represents a general route to high-
molecular-weight poly(ferrocenes) (2).4-11 The polym-
erization behavior of the silicon-bridged monomers 1
(ERx ) SiRR′) is particularly well-developed with a
variety of possible substituents at silicon, and thermal,
anionic, and transition-metal-catalyzed ROP routes to
ROP of 3a ,b to afford polymers 4a ,b was reported by
our group in 1995.10 More recently, we have also
(2) (a) Wright, M. E.; Sigman, M. S. Macromolecules 1992, 25, 6055.
(b) Davies, S. J .; J ohnson, B. F. G.; Khan, M. S.; Lewis, J . J . Chem.
Soc., Chem. Commun. 1991, 187. (c) Tenhaeff, S. C.; Tyler, D. R. J .
Chem. Soc., Chem. Commun. 1989, 1459. (d) Neuse, E. W.; Bednarik,
L. Macromolecules 1979, 12, 187. (e) Brandt, P. F.; Rauchfuss, T. B.
J . Am. Chem. Soc. 1992, 114, 1926. (f) Sturge, K. C.; Hunter, A. D.;
McDonald, R.; Santarsiero, B. D. Organometallics 1992, 11, 3056. (g)
Gonsalves, K.; Zhan-ru, L.; Rausch, M. D. J . Am. Chem. Soc. 1984,
106, 3862. (h) Katz, T. J .; Sudhakar, A.; Teasley, M. F.; Gilbert, A. M.;
Geiger, W. E.; Robben, M. P.; Wuensch, M.; Ward, M. D. J . Am. Chem.
Soc. 1993, 115, 3182. (i) Nugent, H. M.; Rosenblum, M. J . Am. Chem.
Soc. 1993, 115, 3848. (j) Patterson, W. J .; McManus, S. P.; Pittman,
C. U. J . Polym. Sci.: Polym. Chem. Ed. 1974, 12, 837. (k) Buretea, M.
A.; Tilley, T. D. Organometallics 1997, 16, 1507. (l) Stanton, C. E.;
Lee, T. R.; Grubbs, R. H.; Lewis, N. S.; Pudelski, J . K.; Callstrom, M.
R.; Erickson, M. S.; McLaughlin, M. L. Macromolecules 1995, 28, 8713.
(m) Heo, R. W.; Somoza, F. B.; Lee, T. R. J . Am. Chem. Soc. 1998, 120,
1621.
the interesting polymers 2 (ERx ) SiRR′) are all known.
4-6,11-13
The ROP behavior of phosphorus-bridged [1]ferro-
cenophanes (3a ) has also been studied. Thus, thermal
(1) (a) Pittman, C. U.; Carraher, C. E.; Reynolds, J . R In Encyclo-
pedia of Polymer Science and Engineering; Mark, H. F., Bikales, N.
M., Overberger, C. G., Menges, G., Eds.; Wiley: New York, 1989; Vol.
10, p 541. (b) Sheats, J . E.; Carraher, C. E.; Pittman, C. U.; Zeldin,
M.; Currell, B. Inorganic and Metal-Containing Polymeric Materials;
Plenum: New York, 1989. (c) Gonsalves, K. E.; Rausch, M. D. In
Inorganic and Organometallic Polymers; ACS Symposium Series 360;
Zeldin, M., Wynne, K., Allcock, H. R., Eds.; American Chemical
Society: Washington, DC, 1988. (d) Allcock, H. R. Adv. Mater. 1994,
6, 106. (e) Manners, I. Chem. Br. 1996, 32, 46.
10.1021/om980696t CCC: $18.00 © 1999 American Chemical Society
Publication on Web 02/24/1999

A Phosphonium-Bridged [1]Ferrocenophane
Organometallics, Vol. 18, No. 6, 1999 1031
reported that high-molecular-weight poly(ferrocenylphos-
phines) and related block copolymers can be obtained
opening and formation of 8 in the presence of [(n-
Bu)4N]F with the concurrent loss of CO and formation
of [(n-Bu)4N][PF6].17
via living anionic ROP of 3a .14,15 To date, no transition-
metal-catalyzed routes have been reported, as all known
catalysts simply coordinate to the P(III) center of 3a
without leading to ROP.
The interesting chemistry of compound 3a has been
explored by a number of groups, particularly by Seyferth
et al., who were able to coordinate a number of different
moieties to the phosphorus, including sulfur to give 5a
and Fe(CO)4 moieties to give 6.16
Ring-opening of 3a in the presence of a transition
metal has also been reported by Cullen et al.18 This
involved the thermal reaction (125 °C) of 3a with Os3-
(CO)12, which resulted in the major product Os3(CO)9-
[µ3-(C5H4PPh)Fe(C5H4)] (9).
Nakazawa, Miyoshi, et al. found that upon coordina-
tion of Fe(η5-C5H5)(CO)2 to the bridging phosphorus
atom (to give 7), irradiation with UV light led to ring-
(3) Manners, I. Angew. Chem., Int. Ed. Engl. 1996, 35, 1602.
(4) Foucher, D. A.; Tang, B.-Z.; Manners, I. J . Am. Chem. Soc. 1992,
114, 6246.
(5) Manners, I. Adv. Organomet. Chem. 1995, 37, 131.
(6) Manners, I. Can. J . Chem. 1998, 76, 371.
(7) Rulkens, R.; Lough, A. J .; Manners, I. Angew. Chem., Int. Ed.
Engl. 1996, 35, 1805.
(8) Pudelski, J . K.; Gates, D. P.; Rulkens, R.; Lough, A. J .; Manners,
I. Angew. Chem., Int. Ed. Engl. 1995, 34, 1506.
(9) Braunschweig, H.; Dirk, R.; Mu¨ller, M.; Nguyen, P.; Resendes,
R.; Gates, D. P.; Manners, I. Angew. Chem., Int. Ed. Engl. 1997, 36,
2338.
Another potentially interesting tetracoordinate phos-
phorus-bridged [1]ferrocenophane, 10a , was described
by Seyferth et al.16 This compound was synthesized by
(10) Honeyman, C. H.; Foucher, D. A.; Dahmen, F. Y.; Rulkens, R.;
Lough, A. J .; Manners, I. Organometallics 1995, 14, 5503. Note that
poly(ferrocenylphosphine) 4a has been previously reported by
a
condensation route: Withers, H. P.; Seyferth, D.; Fellmann, J . D.;
Garrou, P. E.; Martin, S. Organometallics 1982, 1, 1283.
(11) For the work of other groups on poly(ferrocenylsilanes) see: (a)
Nguyen, M. T.; Diaz, A. F.; Dement’ev, V. V.; Pannell, K. H. Chem.
Mater. 1993, 5, 1389. (b) Hmyene, M.; Yasser, A.; Escorne, M.;
Percheron-Guegan, A.; Garnier, F. Adv. Mater. 1994, 6, 564. (c) Reddy,
N. P.; Yamashita, H.; Tanaka, M. J . Chem. Soc., Chem. Commun. 1995,
2263. (d) Barlow, S.; Rohl, A. L.; Shi, S.; Freeman, C. M.; O’Hare, D.
J . Am. Chem. Soc. 1996, 118, 7578. (e) Park, J .; Seo, Y.; Cho, S.; Whang,
D.; Kim, K.; Chang, T. J . Organomet. Chem. 1995, 489, 23.
(12) (a) Ni, Y.; Rulkens, R.; Manners, I. J . Am. Chem. Soc. 1996,
118, 4102. (b) Massey, J . A.; Power, K. N.; Winnik, M.; Manners, I. J .
Am. Chem. Soc. 1998, 120, 9533. (c) Massey, J . A.; Power, K. N.;
Winnik, M.; Manners, I. Adv. Mater. 1998, 10, 1559.
(13) (a) Ni, Y.; Rulkens, R.; Pudelski, J . K.; Manners, I. Makromol.
Chem. Rapid Commun. 1995, 16, 637. (b) Go´mez-Elipe, P.; Resendes,
R.; Macdonald, P. M.; Manners, I. J . Am. Chem. Soc. 1998, 120, 8348.
(14) Honeyman, C. H.; Peckham, T. J .; Massey, J . A.; Manners, I.
J . Chem. Soc., Chem. Commun. 1996, 2589.
the reaction of 3a with MeI. Although the compound
was characterized by 31P and 1H NMR, no crystals
suitable for X-ray analysis could be grown and the
compound was reported to be unstable in solution.
In this paper we report on our studies of the ROP
behavior, particularly by transition-metal catalysts, of
several tetracoordinate phosphorus-bridged [1]ferro-
cenophanes (5a , 6, and 7) as well as our preparation of
(15) Peckham, T. J .; Massey, J . A.; Honeyman, C. H.; Manners, I.
Macromolecules, in press.
(16) Seyferth, D.; Withers, H. P. Organometallics 1982, 1, 1275.
Recently Brinzinger and co-workers have reported phosphonium-
bridged group 1 and 2 bis(cyclopentadienyl) complexes: Leyser, N.;
Schmidt, K.; Brintzinger, H. H. Organometallics 1998, 17, 2155.
(17) Mizuta, T.; Yamasaki, T.; Nakazawa, H.; Miyoshi, K. Organo-
metallics 1996, 15, 1093.
(18) Cullen, W. R.; Retting, S. J .; Zheng, T. C. J . Organomet. Chem.
1993, 452, 97.

1032 Organometallics, Vol. 18, No. 6, 1999
Peckham et al.
a more stable analogue of 10a , with a triflate counterion,
and our ROP studies on this species.
Resu lts a n d Discu ssion
Attem p ted Tr a n sition -Meta l-Ca ta lyzed ROP of
[1]F er r ocen op h a n es 5a , 6, a n d 7. Although thermal
ROP of phosphorus-bridged [1]ferrocenophanes has been
reported previously by our group, mainly bis(trimeth-
ylsilyl) sulfide and low-molecular-weight material were
isolated from an otherwise insoluble product from the
thermal ROP for 5b.10 In light of this evidence, it was
thought that an attempt to thermally ROP 5a would
lead to the formation of H2S, a possible safety hazard.
As expected, no formation of polymer was observed on
an attempted transition-metal-catalyzed polymerization
in solution using either Pt(II) (PtCl2) or Pt(0) (Karstedt’s
catalyst) initiators.
F igu r e 1. Molecular structure of 10b (vibrational el-
lipsoids at 25% probability level).
Ta ble 1. Su m m a r y of Cr ysta l Da ta a n d In ten sity
Collection P a r a m eter s for 10b
Similarly, the thermal ROPs of 6 and 7 were avoided
due to the possibility for the release of CO. Transition-
metal-catalyzed experiments with both PtCl2 and
Karstedt’s catalyst also did not lead to the formation of
polymer in the case of either compound. We then
decided to investigate the ROP behavior of phospho-
nium-bridged [1]ferrocenophanes.
Syn th esis a n d Ch a r a cter iza tion of 10b. As men-
tioned above, Seyferth reported that the treatment of
3a with MeI led to the formation of 10a which was
determined to be unstable in solution.16 In light of the
sensitivity of 3a to nucleophiles, we attempted to
synthesize a derivative of 10a with a less nucleophilic
counteranion.
empirical formula
fw
wavelength (Å)
cryst syst
space group
a (Å)
C18.5H16ClF3FeO3PS
497.64
0.710 73
monoclinic
C2/c
23.380(1)
13.339(1)
12.715(1)
90
91.79(1)
90
3963.4(14)
8
1.668
1.127
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
F(calcd) (g cm-3
)
)
abs coeff (mm-1
F(000)
2016
cryst size (mm)
θ range for data collecn (deg)
no. of rflns collected
0.18 × 0.04 × 0.03
Our choice of methyl triflate as a methylating reagent
proved to be successful. Compound 10b was isolated in
high yield and characterized by 1H, 13C, 19F, and 31P
NMR as well as elemental analysis, which were all
consistent with the proposed structure. The compound
was also found to be stable in solution under N2 and
could be isolated as dark red microcrystals suitable for
single-crystal X-ray analysis by recrystallization from
a THF/CH2Cl2 mixture. Additionally, the compound was
found to be stable to air and moisture in the solid state
but decomposed slowly over several hours in solution.
An examination of the NMR data revealed some
2.40-20.90
7331
no. of indep rflns
2052 (Rint ) 0.1013)
2052/0/268
1.077
R1 ) 0.0569, wR2 ) 0.1421
R1 ) 0.1023, wR2 ) 0.1640
0.0007(2)
no. of data/restraints/params
goodness of fit on F2
final R indices [I>2σ(I)]
R indices (all data)
extinction coeff
largest diff peak and hole (e Å-3
)
0.510 and -0.411
diffraction study were obtained from a THF/CH2Cl2
mixture at -30 °C. Only very small crystals of the
sample (0.18 × 0.04 × 0.03 mm) were available for data
collection. Data were collected on a Nonius Kappa-CCD
using graphite-monochromated Mo KR radiation (λ )
0.710 73 Å). A total of 180 frames, of 1° rotations of æ,
were exposed for 90 s each. There were no measurable
data higher than 21° in θ. The data were integrated and
scaled using the DENZO package.19 The structure was
solved and refined using the SHELXTL/PC package.20
Refinement was by full-matrix least squares on F2 using
all data (negative intensities included). The weighting
1
interesting features. The H NMR spectrum of 10b has
four distinct resonances due to four inequivalent cyclo-
pentadienyl (Cp) protons. This is in contrast to the
corresponding spectrum of 3a , in which only three
resonances are seen in the Cp region. The phenyl region
for 10b is also more distinct than that for 3a and, in
addition, displays a coupling pattern consistent with a
monosubstituted benzene ring attached to a phospho-
nium center. Perhaps most revealing in the 1H NMR
spectrum is the doublet at 2.66 ppm due to a methyl
group attached directly to phosphorus (2J PH ) 13.9 Hz).
The resonance at 37.5 ppm in the 31P NMR spectrum is
in agreement with a tetrasubstituted phosphorus center
(cf. the resonance for 3a at 13 ppm) and has almost the
same shift (37 ppm) found for 10a .16 The data provided
by the 13C NMR spectrum is similarly consistent with
the strained [1]ferrocenophane structure proposed for
10b, especially the dramatically upfield-shifted ipso
carbon of the cyclopentadienyl rings at 4.2 ppm.5
Sin gle-Cr yst a l X-r a y Diffr a ct ion St u d y of 10b.
Crystals of 10b suitable for a single-crystal X-ray
2
scheme was w ) 1/[σ2(Fo ) + (0.0699)2 + 25.07P], where
2
2
P ) (Fo + 2Fc )/3. Hydrogen atoms were included in
calculated positions. A view of the molecular structure
is shown in Figure 1. Table 2 gives selected structural
data for 10b as well as for related phosphorus-bridged
[1]ferrocenophanes. The angles R, â, θ, and δ are defined
in Figure 2. A summary of the crystal data and collec-
tion parameters can be found in Table 1. A table of the
fractional coordinates is given in the Supporting Infor-
(19) DENZO-SMN; Nonius Co., Delft, Holland, 1997.
(20) Sheldrick, G. M. SHELXTL[\ ]PC; Siemens Analytical X-ray
Instruments Inc., Madison, WI, 1994.

A Phosphonium-Bridged [1]Ferrocenophane
Organometallics, Vol. 18, No. 6, 1999 1033
Ta ble 2. Selected Str u ctu r a l P a r a m eter s for 10b a n d Rela ted P h osp h or u s-Br id ged [1]F er r ocen op h a n es,
w ith Esd ’s in P a r en th eses (Wh er e Ava ila ble)
compd
3a
3a a
3c
5b
10b
Fe-P dist (Å)
Fe displacement (Å)
P-C (Cpipso) (av) (Å)
ring tilt R (deg)
â (av) (deg)
2.774(3)
2.784(2)
0.291(7)
1.857(8)
27.5(6)
32.0(3)
95.7(4)
159.5(3)
10
2.755(5)
0.277(8)
1.841(20)
27.0(6)
31.9(7)
90.1(7)
160.4(6)
10
2.688(2)
0.259(2)
1.812(5)
25.3(3)
35.0(4)
95.0(2)
161.8(2)
10
2.577(3)
0.232(9)
1.772(10)
24.4(5)
37.9(4)
99.8(4)
1.842(13)
26.7
32.5
90.6(3)
159.8
43
θ (av) (deg)
δ (av) (deg)
ref
163.6(4)
this work
a
Trimethylsilyl substituents in the 3,3′-positions on the cyclopentadienyl rings.
shorter even than that for 5b (2.688(2) Å). This may
indicate that there is a greater iron-phosphorus inter-
action occurring in this compound than in previous
examples of phosphorus-bridged [1]ferrocenophanes.
This is not particularly surprising, as the electrophilicity
of the phosphorus atom in 10b is probably greater than
in any of the other examples and thus possibly a
stronger dative bond from the relatively electron-rich
iron atom exists in this situation. This may be partially
responsible for the large â angle for this compound
(37.9(4)°). However, it should also be noted that the ipso
Cp carbon-phosphorus bond length is also shorter in
10b (1.772(10) Å) than in the other examples. This
might also at least be partially responsible for the
shorter Fe-P distance as well as the larger â angle.
In comparison to other phosphonium compounds, the
geometry and bond lengths in the case of 10b are fairly
typical with expected values in the area of 1.78-1.83 Å
for P-C(sp2) bonds and bond angles of 107-114°.21 The
P-C(sp3) bond length (1.765(9) Å) is somewhat shorter
than the expected value (1.80 Å), and the θ angle of
99.8(4)° is also smaller than expected for a phosphonium
compound, although it is a larger angle than is found
for other phosphorus-bridged [1]ferrocenophanes. These
results may be the effect of the strained nature of this
compound. In the case of the θ angle, this is probably
due to the combined effect of the shorter P-C (Cp-ipso)
bond lengths and the possible dative interaction be-
tween iron and phosphorus.
Electr on ic Sp ectr u m a n d Electr och em ica l Be-
h a vior of 10b. The electronic structure of 10b was
examined by UV/vis spectroscopy. The band II λmax
value of 484 nm (in THF; the value in DMF is the same)
is significantly blue-shifted in comparison to the value
of 498 nm (in hexanes) for 3a . Both values are much
higher than was found for ferrocene (440 nm in hex-
anes). This red shift has been observed previously in
other strained [1]ferrocenophanes and has been at least
partially attributed to the tilting of the cyclopentadienyl
rings from the planarity of ferrocene and unstrained,
bridged ferrocenophanes. EHMO calculations on silicon-
and sulfur-bridged [1]ferrocenophanes as well as on
ferrocene in which the rings are tilted by the same angle
(31.05°) as for the sulfur-bridged [1]ferrocenophane
illustrate that the HOMO-LUMO gap grows smaller
as the tilt angle is increased due to a lowering of the
LUMO energy level.22,23 The smaller tilt angle of 10b
F igu r e 2. Distortions in group 8 metallocenophanes:
defining angles R, â, θ, and δ.
Ta ble 3. Bon d Len gth s (Å) for 10b w ith Esd ’s in
P a r en th eses
Fe(1)-C(6)
Fe(1)-C(1)
Fe(1)-C(2)
Fe(1)-C(3)
Fe(1)-C(8)
Fe(1)-P(1)
P(1)-C(6)
1.968(8)
1.979(9)
2.015(9)
2.062(9)
2.083(11)
2.577(3)
1.768(9)
1.785(9)
1.437(12)
1.389(13)
1.465(12)
1.384(13)
1.395(13)
1.380(13)
1.36(2)
Fe(1)-C(10)
Fe(1)-C(7)
Fe(1)-C(5)
Fe(1)-C(4)
Fe(1)-C(9)
P(1)-C(17)
P(1)-C(6)
1.979(9)
1.997(9)
2.020(9)
2.062(9)
2.090(10)
1.765(9)
1.777(9)
1.432(12)
1.388(13)
1.409(13)
1.465(12)
1.414(14)
1.374(13)
1.405(14)
1.35(2)
P(1)-C(11)
C(1)-C(5)
C(1)-C(2)
C(2)-C(3)
C(3)-C(4)
C(6)-C(10)
C(7)-C(8)
C(4)-C(5)
C(6)-C(7)
C(8)-C(9)
C(9)-C(10)
C(11)-C(16)
C(13)-C(14)
C(15)-C(16)
C(1S)-F(1)
C(1S)-S(1)
S(1)-O(1)
C(11)-C(12)
C(12)-C(13)
C(14)-C(15)
C(1S)-F(3)
C(1S)-F(2)
O(2)-S(1)
1.372(13)
1.311(13)
1.770(13)
1.371(9)
1.706(12)
1.246(14)
1.319(14)
1.413(7)
1.420(9)
1.706(12)
S(1)-O(3)
C(2S)-Cl(1) #1
C(2S)-Cl(1)
Ta ble 4. Selected Bon d An gles for 10b w ith Esd ’s
in P a r en th eses
C(1)-P(1)-C(6)
C(1)-P(1)-C(11)
C(6)-P(1)-C(11)
C(17)-P(1)-C(1)
C(17)-P(1)-C(6)
C(17)-P(1)-C(11)
99.8(4)
109.4(4)
110.4(4)
112.2(4)
111.5(4)
112.9(4)
C(2)-C(1)-P(1)
C(5)-C(1)-P(1)
C(2)-C(1)-C(5)
C(2)-C(3)-C(4)
C(3)-C(2)-C(1)
117.4(7)
118.5(6)
107.2(8)
110.5(9)
106.9(8)
mation; bond lengths and selected angles are found in
Tables 3 and 4, respectively.
The single-crystal X-ray diffraction study agrees with
the structure assigned to 10b. Disordered CH2Cl2
molecules were found to be included within the unit cell.
However, no significant interaction was found between
the ferrocenophane moiety and either CH2Cl2 or the
triflate counterion. In comparison to other phosphorus-
bridged [1]ferrocenophanes, 10b posseses the smallest
tilt angle yet known for such a species (24.4(5)°). This
value is even smaller than for the phosphorus(V)-
bridged species 5b (25.3(3)°).10 Perhaps most interesting
to note is the very short Fe-P distance of 2.577(3) Å,
(21) Cristau, H. J .; Ple´nat, F. In The Chemistry of Organophosphorus
Compounds; Hartley, F. R., Ed.; Wiley: Toronto, 1994; Vol. 3, p 45.
(22) Barlow, S. D., M. J .; Dijkstra, T.; Green, J . C.; O’Hare, D.;
Whittingham, C.; Wynn, H. H.; Gates, D. P.; Manners, I.; Nelson, J .
M.; Pudelski, J . K. Organometallics 1998, 17, 2113.

1034 Organometallics, Vol. 18, No. 6, 1999
Peckham et al.
upon heating, and no further thermal behavior was seen
during the cooling cycle. The energy of this exotherm
(enthalpy of polymerization) was found to be 61 ( 5 kJ
mol-1. By comparison, the enthalpy of polymerization
for 3a was found to be 68 ( 5 kJ mol-1
.
This behavior led us to believe that thermal ROP of
10b had taken place, leading to the formation of 11. This
F igu r e 3. DSC thermogram of 10b.
ROP experiment was conducted on a preparative scale
by heating 10b in a sealed, evacuated Pyrex tube at 145
°C for 30 min. This led to the formation of a product
which was insoluble in ethanol, water, and organic
solvents such as CH2Cl2, THF, and hexanes but that
was soluble in the highly polar solvents DMF, DMSO,
methanol, and acetone (the product was found to be
unstable in all the solvents but less so in DMF). Analysis
of the product by NMR was consistent with a ring-
opened structure. In particular, the 31P NMR spectrum
in DMSO-d6 showed only one resonance at 23.8 ppm
with no sign of any residual monomer (δ 37.5 ppm). The
1H NMR spectrum contained only broad resonances,
with none of the definition that had been seen for the
monomer but which were nonetheless consistent with
the assigned structure. For example, the four reso-
nances in the Cp region of 10b were reduced to three
broad resonances. Similarly, the coupling between the
methyl protons and phosphorus could no longer be
resolved for 11. In the case of the 13C NMR spectrum,
the ipso carbon had shifted downfield to 66.5 ppm in
comparison to 10b and was in the region normally
associated with Cp carbons in an unstrained system.
Tr a n sition -Meta l-Ca ta lyzed ROP Beh a vior of
10b. It has been previously noted by other research
groups that coordination to the lone pair of the P(III)
site or oxidation to P(V) permits facile cleavage of P-C
bonds.17,25-28 With this in mind, we attempted the
transition-metal-catalyzed ROP of 10b. In the presence
of approximately 10 mol % PtCl2, 10b was indeed found
to undergo ROP and the polymer formed possessed the
same NMR characteristics as that formed via thermal
ROP. To our knowledge, this represents the first ex-
ample of a phosphorus-containing ring that undergoes
transition-metal-catalyzed ROP. However, the yields
(ca. 56%) were significantly lower than those obtainable
via thermal ROP (ca. 70%). We have observed the
formation of other products in this reaction which are
as of yet unidentified (31P NMR spectrum: 25.8 and 25.5
ppm in DMSO-d6) and appear to be formed in greater
amounts with increasing dilution of the reaction solu-
(24.5°) in comparison to 3a (26.7°) would appear to agree
with these calculations. However, it should also be noted
that the iron-phosphorus distance in 10b (2.577 Å) is
significantly shorter than in 3a (2.774 Å) and there may
be also some degree of iron-phosphorus interaction
which is influencing the size of the HOMO-LUMO gap
in addition to the difference in tilt angle. Some evidence
for a dative iron-bridging element bond in ferro-
cenophanes has been provided in the past by 57Fe
Mo¨ssbauer spectroscopy.22,24
Cyclic voltammetry was used to examine the electro-
chemical behavior of 10b. Analysis (in CH2Cl2) revealed
one irreversible oxidation wave at (Ep(ox)) +0.72 V (vs
ferrocene) at a scan rate of 250 mV/s. For comparison,
the E1/2 value for 1 (ERx ) SiMe2) is only 0.00 V. This
suggests that the phosphonium center exhibits a strong
electron-withdrawing effect on the iron center. The
value of +0.72 V also represents the highest oxidation
potential found to date for a [1]ferrocenophane. Mea-
surements were also carried out in DMF. Again, there
was only one irreversible wave (Ep(ox) ) +0.26 V vs
ferrocene) at a scan rate of 250 mV s-1 with the smaller
difference in oxidation potential versus ferrocene prob-
ably attributable to the donating nature of the solvent
(DMF). In contrast to the scan carried out in CH2Cl2 in
which no other compound could be observed other than
that for 10b, the second scan of the solution of 10b in
DMF revealed that one or more uncharacterized prod-
ucts had been formed during the first scan which then
subsequently underwent two partially reversible oxida-
tion waves at -0.41 and -0.10 V (the wave due to 10b
was still visible) at a scan rate of 250 mV s-1
.
31P NMR
1
and H NMR analyses of the solution before and after
the cyclic voltammetric experiment indicated the pres-
ence of only 10b. These results, as expected, indicated
that the compounds formed during the experiment are
formed only in the vicinity of the electrode and not in
the bulk solution, in which 10b is relatively stable.
Th er m a l ROP Beh a vior of 10b. The thermal ROP
behavior of 10b was first examined by heating a sample
under N2 using a DSC (see Figure 3). No melting
endotherm was observed. However, a strong exotherm
was seen to occur at an onset temperature of 145 °C
(25) Nakazawa, H.; Matsuoka, Y.; Nakagawa, I.; Miyoshi, K. Or-
ganometallics 1992, 11, 1385.
(26) Arce, A. J .; De Sanctis, Y.; Machado, R.; Capparelli, M. V.;
Manzur, J .; Deeming, A. J . Organometallics 1995, 14, 3592.
(27) Wicht, D. K.; Kourkine, I. V.; Lew, B. M.; Nthenge, J . M.;
Glueck, D. S. J . Am. Chem. Soc. 1997, 119, 5039.
(23) Green, J . C. Chem. Soc. Rev. 1998, 263.
(24) Rulkens, R.; Gates, D. P.; Balaishis, D.; Pudelski, J . K.;
McIntosh, D. F.; Lough, A. J .; Manners, I. J . Am. Chem. Soc. 1997,
119, 10976.
(28) Carmichael, D.; Hitchcock, P. B.; Nixon, J . F.; Mathey, F.;
Pidcock, A. J . Chem. Soc., Chem. Commun. 1986, 762.

A Phosphonium-Bridged [1]Ferrocenophane
Organometallics, Vol. 18, No. 6, 1999 1035
tion. In the transition-metal-catalyzed ROP of 1 (ERx
) SiMe2), cyclic dimers analogous to 1 have been
isolated from the reaction mixtures.13 A similar situa-
tion may also occur with the thermal and transition-
metal-catalyzed ROP of 10b. The presence of two
resonances for the potential dimer may arise from the
possibility of cis and trans isomers, 12 and 13. Signifi-
cantly, in the case of the transition-metal-catalyzed ROP
of 1 (ERx ) SiMeCl), a cyclic dimer was proposed as a
byproduct of the reaction and the two resonances for
this dimer observed in the 29Si NMR spectrum were
attributed to cis and trans isomers.29
In an attempt to synthesize a more generally soluble
product, we attempted copolymerizations of 10b with
silicon-bridged [1]ferrocenophanes 1 (ERx ) SiMe2)
using both thermal and transition-metal-catalyzed routes.
In the case of attempted thermal copolymerizations,
only the homopolymer 2 (ERx ) SiMe2) and 11 were
isolated. For the transition-metal-catalyzed polymeriza-
tions, only 2 (ERx ) SiMe2) and unreacted 10b were
observed. In each case, the absence of copolymerization
was shown by the lack of crossover peaks in the 31P and
1H NMR spectra of the isolated products.
F igu r e 4. 31P NMR spectra (in CD2Cl2 for a-d and DMSO-
d6 for e) of the methylation of polymer 14 (n ) 100): (a)
0%, (b) 17%, (c) 33%, (d) 50%, and (e) 100% methylation.
Meth ylation of P oly(fer r ocen ylph en ylph osph in e)
14 a s a n Alter n a tive Rou te to 11. To explore an
alternative route to 11, we studied partial to complete
methylation of the well-defined poly(ferrocenylphos-
phine) 14, synthesized via anionic ROP of 3a using
n-BuLi as an initiator and H2O as a termination
reagent. Partial methylations ranging from 17% to 50%
of phosphorus sites in the polymer were easily achieved
by addition of the appropriate amount of MeOTf to a
CH2Cl2 solution of 14. At methylations higher than 50%,
it was found that the polymer 15 would precipitate from
solution and would not undergo further methylations
readily. To achieve 100% methylation, it was necessary
to first suspend 14 in MeOTf and to then add CH2Cl2.
This resulted in complete methylation for both the high-
(n ) 100) and low-molecular-weight (n ) 11) samples
of 14.
Even at a low degree of methylation, the presence of
the methyl groups at presumably random phosphorus
sites along the polymer main chain leads to different
environments that were readily distinguishable by
NMR. Thus, unmethylated phosphorus resonances (at
ca. -33 ppm) with (i) no Me groups on adjacent P sites
(-P-fc-P-fc-P-) (where fc ) (η-C5H4)2Fe and P )
PPh), (ii) one Me group on one adjacent P site on either
side (-P+-fc-P-fc-P-) (where P+ ) PMePh+), and (iii)
Me groups on both of the adjacent P sites (-P+-fc-P-
fc-P+-) were detected. Analogous resonances were
Figure 4 illustrates the effect of increasing methyla-
tion of 14 on the 31P NMR spectrum of the polymer.
(29) Zechel, D. L.; Hultzsch, K. C.; Rulkens, R.; Balaishis, D.; Ni,
Y.; Pudelski, J . P.; Lough, A. J .; Manners, I. Organometallics 1996,
15, 1972.

1036 Organometallics, Vol. 18, No. 6, 1999
Peckham et al.
observed for the methylated P sites at ca. 24 ppm. At
0% methylation, as expected, only one resonance was
detected. At 17% methylation, there is one resonance
for the methylated P sites as well as three resonances
for the free P sites corresponding to situations i-iii with
situation i being the most prevalent and situation iii
being the least. However, as the degree of methylation
increased, the resonance for -P-fc-P-fc-P- sites
decreased in intensity until, at 50% methylation, only
two resonances, due to -P+-fc-P-fc-P- and -P+-
fc-P-fc-P+- sites, were observed. Also, two resonances
due to methylated phosphorus sites can be seen in the
spectrum. At 100% methylation, only one resonance is
visible with a shift of 23.8 ppm, the same shift found
for both the thermal and transition-metal-catalyzed
polymerizations of 10b and, therefore, fully supports our
conclusion that 11 can be obtained from 10b via ROP.
is at a lower oxidation potential than ferrocene. To
provide a simple model for 11, we synthesized compound
16 by reaction of bis(ferrocenyl)phenylphosphine with
methyl triflate.
The dimer 16 was found to be soluble in both CH2Cl2
and DMF, and thus the electrochemical behavior of this
species was examined in each solvent. In CH2Cl2, there
were two oxidation waves at +0.65 and +0.97 V and a
single reduction wave at +0.39 V (vs ferrocene at 250
mV s-1) with the shape of the reduction wave possibly
due to a two-electron reduction as well as absorption.
Similar behavior has been reported for the analogous
species with an I- counterion.30 The high oxidation
potentials observed for 16 in CH2Cl2 are consistent with
the high value that was detected for the monomer 10b
in the same solvent.
In DMF, the cyclic voltammogram of 16 revealed only
one irreversible oxidation wave at -0.14 V (vs ferrocene
at 250 mV s-1). This is at lower potential than for the
monomer 10b, for which the oxidation wave was found
at +0.26 V but is at a similar value for polymer 11 in
the same solvent (-0.16 V vs ferrocene at 250 mV s-1).
The unusually low oxidation potentials for dimer 16 and
polymer 11 in DMF may be the result of strong solvation
of the cationic main chain by the donor solvent. Indeed,
on the basis of the solubility properties of polymer 11,
such interactions are probably vital to effect dissolution
of the material.
Th er m a l Tr a n sition Beh a vior , Th er m a l Sta bil-
ity, a n d Mor p h ology of 11. The thermal transition
behavior and the thermal stability of 11 (produced via
thermal ROP and transition-metal-catalyzed ROP) were
studied by differential scanning calorimetry (DSC) and
thermogravimetric analysis (TGA) and compared to
those of the unmethylated polymer 14. The morpholo-
gies of the two polymers were examined using wide-
angle X-ray scattering (WAXS). By DSC, both 11 and
14 were found to be amorphous, with no evidence for
melt transitions for either polymer. However, glass
transitions were observed in both cases at 176 and 126
°C, for 11 (via thermal ROP) (Figure 6) and 14,15
respectively. The higher value for 11 is reasonable,
considering that the addition of a methyl group to the
backbone of 14 probably reduces the flexibility of the
polymer main chain. For comparison, there is an
increase in the Tg value from 9 to 33 °C for 2 (ERx )
SiHMe) and 2 (ERx ) SiMe2), respectively. Also, the Tg
value for 11 is significantly higher than the value of 54
°C for the analogous poly(ferrocenylsilane) 2 (R ) Me,
R′ ) Ph), although the value for 14 is also higher (to
date, no analogous silicon-bridged species 2 has been
synthesized where R ) Ph, R′ ) H). However, the
triflate counterion no doubt plays a role in the thermal
In the case of the complete methylation of the lower
(n ) 11) molecular weight sample of 14, end groups can
also be seen at 26.9 and 24.1 ppm corresponding to the
initiated and terminated ends of the polymer chain. End
1
groups can also be detected by H NMR. For all cases,
however, no sample would elute from a GPC column
using THF for any partially or fully methylated poly-
mers, even after sulfurization of the remaining unre-
acted P sites and with only 17% methylation. It is
unclear at this point in time why this is the case. The
reason may involve the presence of interactions between
the polymer and the column packing material or simply
the decrease in solubility of the methylated polymer in
comparison to that of the unmethylated polymer. It
should be noted that polymers 4a ,b also do not elute
from a GPC column until all the phosphorus sites have
been sulfurized.10
Electr on ic a n d Electr och em ica l P r op er ties of
11. The λmax value obtained for polymer 11 is 452 nm
(in DMF), considerably blue-shifted in comparison to the
[1]ferrocenophane 10b (λmax ) 484 nm in THF and
DMF). This is not unusual for these systems and has
also been observed for poly(ferrocenylsilanes) and poly-
(ferrocenylgermanes) relative to their respective mono-
mers.5 The λmax value of 444 nm (in CH2Cl2) for polymer
4a is also blue-shifted relative to 11. Certainly, in this
situation, it is unlikely that this effect is due to a
difference in tilt angles for the Cp rings; the rings are
probably close to being parallel in the polymer. More
likely, it is an example of the effect that the bridging
element has upon the HOMO-LUMO gap. In this case,
it appears that the presence of phosphonium centers in
close proximity to iron leads to a red shift.
Cyclic voltammetry of 11 in DMF revealed the pres-
ence of only one redox wave at -0.16 V, which is
irreversible at low scan rates (Ired/Iox ) 0.91 and 0.82
at 250 and 100 mV s-1, respectively) but reversible at
higher scan rates. The reduction wave is split into two
components, which may be a consequence of adsorption
to the electrode (Figure 5). The presence of only one
wave is unusual and contrasts with the two waves
normally detected for poly(ferrocenes) with bridging
elements such as silicon, germanium, and phosphorus.5
However, it is possible that the presence of a cationic
phosphonium site as the spacer unit between iron sites
decreases the interaction between the metal atoms.
Even more unusual is the observation that the E1/2 value
(30) Kotz, J . C.; Nivert, C. L.; Lieber, J . M.; Reed, R. C. J .
Organomet. Chem. 1975, 91, 87.

A Phosphonium-Bridged [1]Ferrocenophane
Organometallics, Vol. 18, No. 6, 1999 1037
-2/3
F igu r e 7. Tg as a function of Mn
(temperature (2 °C).
for 11 (n ) 20-100)
F igu r e 5. Cyclic voltammogram of polymer 11 in DMF
at a scan rate of 250 mV s-1 (referenced vs ferrocene).
Ta ble 5. Th er m a l Da ta for P olym er 11 (w h er e n )
Nu m ber of Rep ea t Un its)
n
Mn
Tg (°C)
∆Cp (J /(g K))
20
55
70
9.1 × 103
2.5 × 104
3.2 × 104
4.6 × 104
135
154
158
162
0.18
0.18
0.19
0.19
100
of 11 is cationic, whereas the column possessed surface
negative charges, this is not particularly surprising. In
the case of viscometry, the polymer was not found to be
stable in DMF or methanol long enough for the rela-
tively time-consuming measurements to be completed
and reproducible results to be obtained. Analysis by
electrospray mass spectrometry of 11 produced by both
thermal transition-metal-catalyzed ROP detected only
oligomers (dimer, trimer, and tetramer).
F igu r e 6. DSC thermogram of polymer 11 (produced via
With the problems of polymer stability in mind,
dynamic light scattering analyses, which could be
performed rapidly, were conducted on samples of 11
obtained via both thermal ROP and transition-metal-
catalyzed ROP in methanol. The hydrodynamic radii
were in the range of 30-45 nm, which suggested that
the compound was polymeric rather than oligomeric in
nature. However, due to the instability of the polymer
in the solvent as well as the possibility for aggregation,
it was not possible to get an accurate value.
thermal ROP).
transition behavior of this polymer as well and may also
contribute to its fairly high Tg value. Interestingly, the
Tg value for 11 derived from the transition-metal-
catalyzed ROP of 10b was found to be 164 °C (see next
section). WAXS analysis of these polymers is also
consistent with a generally amorphous nature (one
broad amorphous halo was detected at 5.5 and 5.2 Å
for 11 and 14, respectively).
TGA results reveal similar thermal stabilities for 11
and 14, and the polymers started to lose weight at 385
and 410 °C, respectively. However, the weight loss for
11 is significantly higher than that for 14 with weight
losses of 50% and 30%, respectively, by 600 °C. This may
possibly coincide with the loss of all groups, leaving
behind an unsubstituted poly(ferrocenylphosphine) back-
bone: i.e., loss of phenyl, methyl, and triflate groups
from 11, while 14 only loses phenyl groups.
As solution methods were hindered by polymer in-
stability in solvents in which the material was soluble,
we determined the glass transition temperatures by
DSC for a range of samples of polymer 11 of known
molecular weight (n ) 20, 55, 70, and 100) that had been
prepared via the 100% methylation of samples of 14
prepared by anionic ROP. The results are shown in
Table 5, and the Tg values were plotted as a function of
-2/3
Mn
(Figure 7).
Molecu la r Weigh t Deter m in a tion for 11. We have
attempted to determine the molecular weight for 11 by
a number of methods, with limited success. In the case
of GPC, no polymer was detected using polystyrene
sulfonate columns and either N-methylpyrrolidinone
(NMP) or 0.2 M sodium nitrate in methanol/water
(75/25) as an elution solvent. As the polymer backbone
We have previously found for polymer 2 (where ERx
) SiMe2) that the Tg value reached a maxiumum of 33
°C at a length of approximately 90 repeat units31,32 and
(31) Ni, Y. Z.; Rulkens, R.; Manners, I. J . Am. Chem. Soc. 1996,
118, 4102.
(32) Lammertink, R. G. H.; Hempenius, M. A.; Manners, I.; Vancso,
G. J . Macromolecules 1998, 31, 795.

1038 Organometallics, Vol. 18, No. 6, 1999
Peckham et al.
phenylphosphine were synthesized according to the litera-
ture.16,17,41 Samples of polymer 14 were prepared by living
anionic polymerization as described previously by our group.14,15
Equ ip m en t. All reactions and manipulations were carried
out under an atmosphere of prepurified nitrogen using either
Schlenk techniques or an inert-atmosphere glovebox (Vacuum
Atmospheres) unless otherwise stated. Solvents were dried by
standard methods, distilled, and stored under nitrogen over
activated molecular sieves. 1H NMR spectra (300 MHz), 13C
NMR spectra (75.5 MHz), 19F NMR spectra (282.3 MHz), and
31P NMR spectra (121.4 MHz) were recorded on a Varian
Gemini 300 spectrometer. 1H NMR spectra (400 MHz) were
also recorded on a Varian Unity 400 spectrometer. Chemical
shifts are reported relative to residual protonated solvent (1H
or 13C), external H3PO4 (31P), or external CFCl3 (19F). Melting
points and enthalpies of polymerization were obtained with a
Perkin-Elmer DSC 7 differential scanning calorimeter operat-
ing at a heating rate of 10 °C min-1 under N2. TGA analyses
were carried out with a Perkin-Elmer TGA-7 thermogravi-
metric analyzer under nitrogen at a heating rate of 10 °C
min-1. UV/vis spectra were recorded on a Perkin-Elmer UV/
vis/near-IR Lambda 900 spectrometer in either anhydrous
THF or DMF. Elemental analyses were performed by Quan-
titative Technologies, Inc., Whitehouse, NJ .
Dynamic light scattering experiments were carried out on
a wide angle laser light scattering photometer from Brookhaven
Instruments Corp. A 5 mW vertically polarized He-Ne laser
from Spectra Physics was the light source. The solutions were
filtered through disposable 0.2 µm filters from Millipore into
glass scattering cells with a diameter of 12.3 nm. The cells
were placed into the BI-200SM goniometer and sat in a vat of
thermostated toluene which matched the index of refraction
of the glass cells. The angle of measurement for the goniometer
was 90°. The scattered light was detected by a photomultiplier
interfaced to the BI-2030AT digital correlator with 136 chan-
nels and measured the correlation function in real time. The
instrument was controlled by a 486AT computer. The data
were analyzed by software supplied by Brookhaven.
Electrochemical experiments were carried out using a PAR
Model 273 potentiostat with a Pt working electrode, a W
secondary electrode, and an Ag-wire reference electrode in a
Luggin capillary. Polymer solutions were 1 × 10-3 M in CH2-
Cl2 with 0.1 M [Bu4N][PF6] as a supporting electrolyte.
Powder diffraction data were obtained on a Siemens D5000
diffractometer using Ni-filtered Cu KR (λ ) 1.541 78 Å)
radiation. The samples were scanned at step widths of 0.02°
with 1.0 s per step in the range of 3-40° in 2θ. Samples were
prepared by spreading the finely ground polymer on grooved
glass slides.
Syn th esis of th e Meth yla ted P h osp h or u s-Br id ged [1]-
F er r ocen op h a n e 10b. A 1.88 g (6.44 mmol) portion of 3a 14,42
was dissolved in 50 mL of toluene, giving a dark red solution.
A 1.03 g (6.30 mmol) amount of MeOTf was added to this
solution dropwise with stirring. A bright red powder was
immediately observed to precipitate during the addition. The
reaction mixture was stirred for a further 30 min after the
addition was complete. The bright orange-red powder was then
isolated under N2 on a filtration frit or under air using a water
aspirator and washed with toluene until the washings were
colorless. Dark red microcrystals of 10b could be grown via
recrystallization from THF. Yield: 2.54 g (88%). 31P{1H} NMR
(CDCl3): δ 37.5 ppm. 19F NMR (CDCl3): δ - 78.6 ppm. 13C
NMR (in CDCl3): δ 136.3 (s, para Ph), 131.2 (d, meta Ph, 3J PC
the data fit well to the O’Driscoll equation, where Tg,∞
is the glass transition temperature of the polymer with
infinite molar mass:32,33
-2/3
Tg ) Tg,∞ - KMn
(1)
Similarly, we found that the data for 11 over the
range of 20-100 repeat units fits well to the O’Driscoll
equation with a regression coefficient of 0.998. The
predicted Tg,∞ value is 175.5 °C. On the basis of these
results, this would suggest the Mn value for 11 derived
by transition-metal-catalyzed ROP (Tg ) 164 °C) is ca.
4.6 × 104. It is more difficult to determine an ap-
proximate molecular weight for the thermal-ROP-
derived polymer, as the experimental Tg value (176 °C)
is close to the theoretical Tg,∞ value. The higher Tg value
for the thermal-ROP-derived product in comparison to
the value for the transition-metal-catalyzed product
does suggest, however, that the former possesses sig-
nificantly higher molecular weight than the latter. This
is not surprising, as the latter precipitates from solution
during the polymerization and thus probably cannot
reach such a long chain length before becoming com-
pletely insoluble in the reaction solvent.
Su m m a r y
The synthesis of a stable, phosphonium-bridged [1]-
ferrocenophane (10b) has been achieved. The compound
was characterized by a variety of spectroscopic tech-
niques as well as by single-crystal X-ray diffraction. In
contrast to compounds 5-7, compound 10b was found
to undergo both thermal and transition-metal-catalyzed
ROP. Although a number of phosphonium-containing
linear34-39 and cyclic polymers40 have been synthesized
in the past, the results represent the first synthesis of
a phosphonium-containing polymer via ROP as well as
the first transition-metal-catalyzed ROP of a phospho-
rus(III)-containing monomer. The resultant polymer 11
was found to be amorphous and displayed thermal
stability similar to that of poly(ferrocenylphosphine) 14.
On the basis of glass transition temperature measure-
ments for a series of samples of polymer 11 of known
molecular weight, the number-average molecular weight
of 11 produced via transition-metal-catalyzed ROP was
ca. 46 000 (ca. 100 repeat units) and the thermally
produced polymer was of even higher molecular weight.
In contrast to 10b, a series of tetracoordinate, phospho-
rus-bridged [1]ferrocenophanes (5-7) were found to be
resistant to undergo transition-metal-catalyzed ROP in
the presence of either Pt(II) or Pt(0) catalysts.
Exp er im en ta l Section
Ma ter ia ls. All chemicals, unless otherwise noted, were
purchased from Aldrich. Compounds 5-7 and bis(ferrocenyl)-
(33) O’Driscoll, K.; Sanayei, A. R. Macromolecules 1991, 24, 4479.
(34) Kanazawa, A.; Ikeda, T.; Endo, T. J . Polym. Sci.: Polym. Chem.
1994, 52, 641.
(35) Kanazawa, A.; Ikeda, T.; Endo, T. J . Appl. Polym. Sci. 1994,
52, 1237.
(36) Kanazawa, A.; Ikeda, T.; Endo, T. J . Appl. Polym. Sci. 1994,
52, 1305.
(37) Kanazawa, A.; Ikeda, T.; Endo, T. J . Appl. Polym. Sci. 1994,
53, 1245.
(38) Ilzawa, T.; Yamada, Y.; Ogura, Y.; Sato, Y. J . Polym. Sci.:
Polym. Chem. 1994, 32, 2057.
(41) Sollott, G. P.; Mertwoy, H. E.; Portnoy, S.; Snead, J . L. J .
Organomet. Chem. 1963, 28, 1090.
(42) The synthesis of 3a has been previously described. See: (a)
Osborne, A. G.; Whiteley, R. H.; Meads, R. E. J . Organomet. Chem.
1980, 193, 345. (b) Seyferth, D.; Withers, H. P., J r. J . Organomet.
Chem. 1980, 185, C1.
(39) Hughes, I. Tetrahedron Lett. 1996, 37, 7595.
(40) Seyferth, D.; Masterman, T. C. Macromolecules 1995, 28, 3055.
(43) Stoeckli-Evans, H.; Osborne, A. G.; Whiteley, R. H. J . Orga-
nomet. Chem. 1980, 134, 91.

A Phosphonium-Bridged [1]Ferrocenophane
Organometallics, Vol. 18, No. 6, 1999 1039
2
) 13 Hz), 130.8 (d, ortho Ph, J PC ) 14 Hz), 118.9 (d, ipso Ph,
and the remaining insoluble material was dried under vacuum.
1J PC ) 90 Hz), 84.5 (d, Cp, J PC ) 11 Hz), 83.5 (d, Cp, J PC
)
The product was analyzed by H and 31P NMR, and the data
3
3
1
2
2
10.4 Hz), 76.4 (d, Cp, J PC ) 11.7 Hz), 75.2 (d, Cp, J PC ) 14.9
were consistent with the structure proposed for the products
of thermal and transition-metal-catalyzed ROP of 10b. Ad-
ditionally, end groups were observed by 31P NMR (in DMSO-
d6) at 26.9 and 24.1 ppm.
1
1
Hz), 11.3 (d, Me, J PC ) 55 Hz), 4.2 (d, ipso Cp, J PC ) 71.9
1
Hz) ppm. H NMR (400 MHz, in CDCl3): δ 8.80 (m, 2H, Ph),
7.89 (m, 1H, Ph), 7.79 (m, 2H, Ph), 5.45 (m, 2H, Cp), 5.02 (m,
2H, Cp), 4.93 (m, 2H, Cp), 4.41 (m, 2H, Cp), 2.66 (d, 3H, Me,
2J PH ) 13.9 Hz) ppm. UV-visible (THF): λmax(band II) 484
nm (ꢀ ) 307 M-1 cm-1). Anal. Calcd for C18H16F3FeO3PS: C,
47.39; H, 3.54. Found: C, 47.11; H, 3.36.
Th er m a l ROP of 10b. A 0.75 g (1.64 mmol) portion of 10b
was sealed in an evacuated Pyrex tube and heated at 145 °C
for 30 min. During this time, the contents were observed to
undergo a change to a dark red-orange color. After this time,
the tube was broken and the polymer washed with CH2Cl2.
The product (11) was then dried under vacuum and isolated
as an orange powder. Yield: 0.65 g (87%). 31P NMR (DMSO-
d6): δ 23.8 ppm. 19F NMR (DMSO-d6): δ -78.2 ppm. 13C NMR
(DMSO-d6): δ 134.8 (s, para Ph), 131.6 (d, meta Ph, 3J PC ) 10
Attem p ted Th er m a l Cop olym er iza tion of 1 (ER x
)
SiMe2) a n d 10b. A 100 mg (0.413 mmol) portion of 1 (ERx )
SiMe2) and 53 mg (0.116 mmol) of 10b were sealed in an
evacuated Pyrex tube at 140 °C for 15 min. The THF-soluble
fraction of the contents of the tube was dissolved in THF and
then precipitated into hexanes. The insoluble and soluble
fractions were then dissolved in DMSO-d6 and C6D6, respec-
tively, and analyzed by 1H and 31P NMR. This analysis
revealed only the presence of the respective homopolymers 11
and 2 (ERx ) SiMe2), with no evidence for switching groups.
Attem p ted Tr a n sition -Meta l-Ca ta lyzed Cop olym er i-
za tion of 1 (ERx ) SiMe2) a n d 10b. A 193 mg (0.798 mmol)
portion of 1 (ERx ) SiMe2) and 109 mg (0.239 mmol) of 10b
were dissolved in 25 mL of CH2Cl2. To this was added 4 mg
(6.3 mol %) of PtCl2. The reaction mixture was stirred
overnight. Analysis as described above revealed only the
presence of homopolymers 2 (ERx ) SiMe2) and 11.
2
1
Hz), 130.0 (d, ortho Ph, J PC ) 13 Hz), 121.7 (d, ipso Ph, J PC
) 94 Hz), 76.3-72.8 (Cp, multiplet, broad), 66.5 (ipso Cp, 1J PC
) 102 Hz), 5.2 (d, Me, 1J PC ) 62 Hz) ppm. 1H NMR (300 MHz,
in DMSO-d6): δ 7.85 (broad, 3 H, phenyl), 7.68 (broad, 2 H,
Ph), 4.88 (broad, 2 H, Cp), 4.65 (broad, 4 H, Cp), 3.97 (broad,
Attem p ted Th er m a l Cop olym er iza tion of 1 (ER x
)
2 H, Cp), 2.93 (broad, 3H, Me) ppm. UV-visible (DMF): λmax
-
SiMeP h ) a n d 10b. A 151 mg (0.497 mmol) portion of 1 (ERx
) SiMePh) and 45 mg (0.099 mmol) of 10b were sealed in an
evacuated Pyrex tube at 140 °C for 15 min. Analysis as
described above revealed only the presence of homopolymers
2 (ERx ) SiMePh) and 11.
(band II) 452 nm (ꢀ ) 137 M-1 cm-1). Anal. Calcd for C18H16F3-
FeO3PS: C, 47.39; H, 3.54. Found: C, 46.65; H, 3.55. Electro-
spray MS: m/e 1676 ([fcPPhMe]4[OTf]3, 10%), 1219 ([fcPPhMe]3-
[OTf]2, 100%), 763 ([fcPPhMe]2[OTf], 10%), 307 ([fcPPhMe]+,
50%).
Attem p ted Tr a n sition -Meta l-Ca ta lyzed Cop olym er i-
za tion of 1 (ERx ) SiMeP h ) a n d 10b. A 160 mg (0.526 mmol)
portion of 1 (ERx ) SiMePh) and 50 mg (0.110 mmol) of 10b
were dissolved in 25 mL of CH2Cl2. To this was added 4 mg
(13.6 mol %) of PtCl2. The reaction mixture was stirred
overnight. Examination of the products by 1H and 31P NMR
showed evidence only for homopolymer 2 (ERx ) SiMePh) and
unreacted 10b.
Tr a n sition -Meta l-Ca ta lyzed ROP of 10b. A 0.196 g
(0.430 mmol) portion of 10b was dissolved in 5 mL of CH2Cl2.
To this solution was added 9 mg (7.9 mol %) of PtCl2. The
solution was then stirred overnight. A fine light orange powder
precipitated out of solution during this time. The precipitate
was allowed to settle, and then the solvent was removed via
cannulation. The product (11) was washed with CH2Cl2 and
then dried under vacuum. 1H, 13C, 19F, and 31P NMR data were
consistent with the same product obtained via thermal ROP.
Yield: 0.105 g (54%). Electrospray MS: m/e 1676 ([fcPPhMe]4-
[OTf]3, 10%), 1219 ([fcPPhMe]3[OTf]2, 100%), 763 ([fcPPhMe]2-
[OTf], 65%), 307 ([fcPPhMe]+, 100%).
P a r tia l Meth yla tion of P oly(fer r ocen ylp h en ylp h os-
p h in e) 14 (n ) 100). The procedure for various percent
methylations of 14 was the same for all samples: (a) 124 mg,
(b) 107 mg, or (c) 105 mg of 14 (n ) 100) (Mw of sulfurized 14
38 000, Mw/Mn ) 1.30) was dissolved in CD2Cl2 in a 5 mm NMR
tube. To this solution was added, via a microsyringe, (a) 7 µL,
(b) 14 µL, or (c) 21 µL of MeOTf. 31P NMR spectra were run.
Samples of the solution were then treated with S8 in an
attempt to produce a GPC-analyzable polymer. However, no
polymer was found to elute from the Ultrastyragel column
using THF as eluent, even for the polymer sample treated with
only 7 µL of MeOTf. Similarly, a 112 mg sample of 14 (n )
100) was treated sequentially with 7 µL portions of MeOTf
and a 31P NMR spectrum was obtained after each addition.
The spectra were essentially the same.
Com p lete Meth yla tion of P oly(fer r ocen ylp h en ylp h os-
p h in e) 14 (n ) 100). A 65 mg portion of 14 (n ) 100) (Mw of
sulfurized 14 ) 38 000, Mw/Mn ) 1.30) was suspended in an
excess of MeOTf. A 2 mL amount of CH2Cl2 was added and
the solution stirred for 10 min. The solution was removed and
the remaining insoluble material dried under vacuum. The
product was analyzed by NMR, and the data were consistent
with the structure proposed for the products of thermal and
transition-metal-catalyzed ROP of 10b.
Com p lete Meth yla tion of P oly(fer r ocen ylp h en ylp h os-
p h in e) 14 (n ) 11). A 60 mg portion of 14 (n ) 11) (Mw of
sulfurized 14 2600, Mw/Mn ) 1.08) were suspended in an excess
of MeOTf. A 2 mL amount of CH2Cl2 was added and the
solution stirred for 10 min. The mother liquors were removed,
Attem p ted Tr a n sition -Meta l-Ca ta lyzed Cop olym er i-
za tion of 5a . A 60 mg portion (0.19 mmol) of 5a was dissolved
in 0.7 mL of C6D6 in a 5 mm NMR tube. To this solution was
added 8 mol % of PtCl2. The contents of the tube were shaken
at intermittent intervals over 24 h. No change in the 31P NMR
spectrum was observed during this time. The contents were
also heated at 60 °C overnight. No changes were observed by
31P NMR.
Attem p ted Tr a n sition -Meta l-Ca ta lyzed ROP of 6. A 60
mg (0.13 mmol) portion of 6 was dissolved in 0.7 mL of C6D6
in a 5 mm NMR tube. To this solution was added 8 mol % of
PtCl2. The contents of the tube were shaken at intermittent
intervals over 24 h. No change in the 31P NMR spectrum was
observed during this time. The contents were also heated at
60 °C overnight. No changes were observed by 31P NMR.
Syn th esis of Mod el Com p ou n d 16. A 0.362 g (0.757
mmol) portion of bis(ferrocenyl)phenylphosphine was dissolved
in 5 mL of toluene. To this stirred solution was added 0.128 g
(0.780 mmol) of MeOTf. A reddish orange oil was immediately
observed to precipitate from solution. The solution was re-
moved by decantation. The oil was dissolved in CH2Cl2 and
precipitated into hexanes. The solution was removed by
decantation. The product was dried overnight under vacuum
and isolated as a reddish orange semisolid. Yield: 0.406 g
(84%). 31P{1H} NMR (CDCl3): δ 24.8 ppm. 19F NMR (CDCl3):
δ -78.5 ppm. 13C NMR (in CDCl3): δ 134.4 (s, para Ph), 131.6
3
2
(d, meta Ph, J PC ) 9 Hz), 129.9 (d, ortho Ph, J PC ) 12 Hz),
1
3
122.9 (d, ipso Ph, J PC ) 92 Hz), 74.3 (d, Cp, J PC ) 9.9 Hz),
3
2
73.9 (d, Cp, J PC ) 9.7 Hz), 72.6 (d, Cp, J PC ) 12.6 Hz), 71.8
2 1
(d, Cp, J PC ) 13.5 Hz), 70.8 (s, Cp), 63.6 (d, ipso Cp, J PC
)
1
1
105 Hz), 10.8 (d, Me, J PC )64 Hz) ppm. H NMR (400 MHz,
in CDCl3): δ 7.5-7.9 (m, 5 H, Ph), 4.79 (m, 2H, Cp), 4.73 (m,
2H, Cp), 4.63 (m, 2 H, Cp), 4.41 (m, 2 H, Cp), 4.20 (s, 10 H,

1040 Organometallics, Vol. 18, No. 6, 1999
Peckham et al.
2
Cp), 2.93 (d, 3 H, Me, J PH ) 13.9 Hz) ppm. Anal. Calcd for
light scattering studies on polymer 11 and Mark
MacLachlan for obtaining WAXS profiles of polymers
11 and 14. We are also grateful to Kristina Barr for help
in the preparation of the manuscript.
C
28H26F3Fe2O3PS: C, 52.37; H, 4.08. Found: C, 52.62; H, 4.20.
Ack n ow led gm en t. This work was funded by the
Natural Sciences and Engineering Research Council
(NSERC). T.J .P. thanks the NSERC for a graduate
scholarship, and I.M. thanks the Alfred P. Sloan Foun-
dation for a Research Fellowship (1994-1998), the
NSERC for an E. W. R. Steacie Fellowship (1997-1998),
and the University of Toronto for a McLean Fellowship.
We also thank J ason Massey for performing dynamic
Su p p or tin g In for m a tion Ava ila ble: A table giving po-
sitional parameters and isotropic thermal parameters for 10b.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OM980696T