Acrylonitrile Polymerization by Cy3PCuMe and (Bipy) 2FeEt2

Article abstract of DOI:10.1021/ja039018a

Cy₃PCuMe (1) undergoes reversible ligand redistribution at low temperature in solution to form the tight ion pair [Cu(PCy₃) ₂][CuMe₂] (3). The structure of 3 was assigned on the basis of (i) the stoichiometry of the 1 = 3 equilibrium, (ii) the observation of a triplet for the PCy₃ C1 ¹³C NMR resonance due to virtual coupling to two ³¹P nuclei, and (iii) reverse synthesis of 1 by combining separately generated Cu(PCy₃)₂⁺ and CuMe₂^- ions. Complex 1 and [Cu(PCy₃) ₂][PF₆] (5) coordinate additional PCy₃ to form (Cy₃P)₂CuMe and [Cu(PCy₃)₃][PF ₆], respectively, while 3 does not. Complex 1, free PCy₃, and (bipy)₂FeEt₂ (2) each initiate the polymerization of acrylonitrile. In each case, the polyacrylonitrile contains branches that are characteristic of an anionic polymerization mechanism. The major initiator in acrylonitrile polymerization by 1 is PCy₃, which is liberated from 1. A transient iron hydride complex is proposed to initiate acrylonitrile polymerization by 2.

Full text of DOI:10.1021/ja039018a

Published on Web 01/23/2004
Acrylonitrile Polymerization by Cy3PCuMe and (Bipy)2FeEt2
Frank Schaper, Stephen R. Foley, and Richard F. Jordan*
Contribution from the Department of Chemistry, The UniVersity of Chicago,
5735 S. Ellis AVenue, Chicago, Illinois 60637
Received October 13, 2003; E-mail: rfjordan@uchicago.edu
Abstract: Cy3PCuMe (1) undergoes reversible ligand redistribution at low temperature in solution to form
the tight ion pair [Cu(PCy3)2][CuMe2] (3). The structure of 3 was assigned on the basis of (i) the stoichiometry
of the 1 ) 3 equilibrium, (ii) the observation of a triplet for the PCy3 C1 13C NMR resonance due to virtual
coupling to two 31P nuclei, and (iii) reverse synthesis of 1 by combining separately generated Cu(PCy3)2
+
-
and CuMe2 ions. Complex 1 and [Cu(PCy3)2][PF6] (5) coordinate additional PCy3 to form (Cy3P)2CuMe
and [Cu(PCy3)3][PF6], respectively, while 3 does not. Complex 1, free PCy3, and (bipy)2FeEt2 (2) each
initiate the polymerization of acrylonitrile. In each case, the polyacrylonitrile contains branches that are
characteristic of an anionic polymerization mechanism. The major initiator in acrylonitrile polymerization
by 1 is PCy3, which is liberated from 1. A transient iron hydride complex is proposed to initiate acrylonitrile
polymerization by 2.
L ) substituted or unsubstituted 2,2′-bipyridine),4 dithioesters
Introduction
(RAFT),5 or N-oxides6 have been reported. Radical PAN is
The polymerization of olefins by insertion chemistry with
Ziegler-Natta, Cr-based, or single-site metal catalysts provides
a powerful approach to the synthesis of polyolefins, often with
a high degree of control of polymer composition and structure.1
A major current challenge in this area is to develop catalysts
that are capable of polymerizing or copolymerizing polar olefins,
particularly CH2dCHX monomers with functional groups
directly bonded to the olefin unit, by insertion mechanisms.2
While some success has been achieved in the copolymerization
of acrylates and vinyl ketones with ethylene and propene,3 a
general solution to this problem is lacking. Here we report
studies directed to the development of metal catalysts for the
insertion polymerization/copolymerization of acrylonitrile (AN).
Polyacrylonitrile (PAN) and related copolymers are prepared
commercially by radical polymerization of AN. Controlled/living
radical AN polymerizations in the presence of LCuX (ATRP,
normally completely linear. AN is also readily polymerized by
anionic polymerization initiators, in which case lower molecular
weight (vs typical radical PAN), branched PAN is formed.
Branching arises by inter- or intramolecular abstraction of
-CH2CH(CN)- methine protons from the polymer backbone by
the propagating carbanion.7
Numerous metal complexes have been reported to polymerize
AN.8-10 While many of these metal-initiated polymerizations
are clearly anionic, the mechanism is less clear for others.8g-j
In particular, Yamamoto and co-workers proposed that the
neutral metal alkyl complexes Cy3PCuMe (1)9 and (bipy)2FeEt2
(4) (a) Matyjaszewski, K.; Jo, S. M.; Paik, H.-J.; Gaynor, S. G. Macromolecules
1997, 30, 6398. (b) Matyjaszewski, K.; Jo, S. M.; Paik, H.-J.; Shipp, D. A.
Macromolecules 1999, 32, 6431. (c) Barboiu, B.; Percec, V. Macromol-
ecules 2001, 34, 8626. (d) Patten, T. E.; Matyjaszewski, K. AdV. Mater.
1998, 10, 901.
(5) Chiefari, J.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.; Le, T. P. T.;
Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.; Moad, G.; Rizzardo, E.;
Thang, S. H. Macromolecules 1998, 31, 5559.
(6) (a) Moad, G.; Rizzardo, E. Macromolecules 1995, 28, 8722. (b) Benoit,
D.; Chaplinski, V.; Braslau, R.; Hawker, C. J. J. Am. Chem. Soc. 1999,
121, 3904.
(7) (a) Malsch, G.; Dautzenberg, H.; Krippner, W.; Scheller, D.; Fritzsche, P.;
Berger, W. Acta Polym. 1982, 32, 626. (b) Scheller, D.; Krippner, W.;
Grossmann, G.; Berger, W. Acta Polym. 1982, 33, 632. (c) Ono, H.;
Hisatani, K.; Kamide, K. Polym. J. 1993, 25, 245. (d) Verneker, V. R. P.;
Shaha, B. Macromolecules 1986, 19, 1851.
(8) (a) Jenkins, A. D.; Lappert, M. F.; Srivastava, R. C. Polym. Lett. 1968, 6,
865. (b) Saegusa, T.; Horiguchi, S.; Tsuda, T. Macromolecules 1975, 8,
112. (c) Gandhi, V. G.; Sivaram, S.; Bhardwaj, I. S. J. Macromol. Sci.,
Chem. 1983, A19, 147. (d) Billingham, N. C.; Lees, P. D. Makromol. Chem.
1993, 194, 1445. (e) Tsuchihara, K.; Suzuki, Y.; Asai, M.; Soga, K. Chem.
Lett. 1999, 891. (f) Siemeling, U.; Ko¨lling, L.; Stammler, A.; Stammler,
H.-G.; Kaminski, E.; Fink, G. Chem. Commun. 2000, 1177. (g) Suh, M.
P.; Oh, Y. H.; Kwak, C. H. Organometallics 1987, 6, 411. (h) Mudalige,
D. C.; Rempel, G. L. J. Macromol. Sci., Chem. 1997, A34, 361. (i) Arndt,
S.; Beckerle, K.; Hultzsch, K. C.; Sinnema, P.-J.; Voth, P.; Spaniol, T. P.;
Okuda, J. J. Mol. Catal. A: Chem. 2002, 190, 215. (j) Yamamoto, A.;
Ikeda, S. J. Am. Chem. Soc. 1967, 89, 5989.
(1) (a) Gibson, V. C.; Spitzmesser, S. K. Chem. ReV. 2003, 103, 283. (b)
Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Chem. ReV. 2000, 100,
1253. (c) Kaminsky, W.; Arndt, M. AdV. Polym. Sci. 1997, 127, 143. (d)
Brintzinger, H.-H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.; Waymouth, R.
M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (e) McDaniel, M. P.
AdV. Catal. 1985, 33, 48.
(2) Boffa, L. S.; Novak, B. M. Chem. ReV. 2000, 100, 1479.
(3) (a) Johnson, L. K.; Mecking, S.; Brookhart, M. J. Am. Chem. Soc. 1996,
118, 267. (b) Mecking, S.; Johnson, L. K.; Wang, L.; Brookhart, M. J.
Am. Chem. Soc. 1998, 120, 888. (c) Stibrany, R. T.; Schulz, D. N.; Kacker,
S.; Patil, A. O.; Baugh, L. S.; Rucker, S. P.; Zushma, S.; Berluche, E.;
Sissano, J. A. Polym. Mater. Sci. Eng. 2002, 86, 325. (d) Heinemann, J.;
Mu¨lhaupt, R.; Brinkmann, P.; Luinstra, G. Macromol. Chem. Phys. 1999,
200, 384. (e) Yasuda, H. J. Polym. Sci., Polym. Chem. 2001, 39, 1955. (f)
Wang, L.; Hauptman, E.; Johnson, L. K.; Marshall, W. J.; McCord, E. F.;
Wang, Y.; Ittel, S. D.; Radzewich, C. E.; Kunitsky, K.; Ionkin, A. S. Polym.
Mater. Sci. Eng. 2002, 86, 322. (g) Johnson, L.; Bennett, A.; Dobbs, K.;
Hauptman, E.; Ionkin, A.; Ittel, S.; McCord, E.; McLain, S.; Radzewich,
C.; Yin, Z.; Wang, L.; Wang, Y.; Brookhart, M. Polym. Mater. Sci. Eng.
2002, 86, 319. (h) Tian, G.; Boone, H. W.; Novak, B. M. Macromolecules
2001, 34, 7656. (i) Drent, E.; van Dijk, R.; van Ginkel, R.; van Oort, B.;
Pugh, R. I. Chem. Commun. 2002, 744. (j) Chien, J. C. W.; Fernandes, S.;
Correia, S. G.; Rausch, M. D.; Dickson, L. C.; Marques, M. M. Polym.
Int. 2002, 51, 729.
(9) (a) Ikariya, T.; Yamamoto, A. J. Organomet. Chem. 1974, 72, 145. (b)
Miyashita, A.; Yamamoto, T.; Yamamoto, A. Bull. Chem. Soc. Jpn. 1977,
50, 1109.
9
2114
J. AM. CHEM. SOC. 2004, 126, 2114-2124
10.1021/ja039018a CCC: $27.50 © 2004 American Chemical Society

Acrylonitrile Polymerization
A R T I C L E S
Table 1. Selected Bond Lengths and Angles for 1a
Cu(1)-P(1)
Cu(1)-C(19)
P(1)-C(1)
P(1)-C(7)
P(1)-C(13)
2.1953(6)
1.916(2)
1.850(1)
1.855(2)
1.842(1)
P(1)-Cu(1)-C(19)
Cu(1)-P(1)-C(1)
Cu(1)-P(1)-C(7)
Cu(1)-P(1)-C(13)
175.55(5)
109.24(5)
112.60(5)
116.05(5)
a Bond lengths are given in angstroms; bond angles are given in degrees.
in 1 [1.916(2) Å] is shorter than those in the tetrahedral complex
(Ph3P)3CuMe [2.04(1) Å]15 and the trigonal planar “pincer”
complex (Me3SiC5H4)2Ti(C2SiMe3)2CuMe [1.966(2) Å]16 and
is also shorter than Cu-C distances in monomeric anionic Cu
alkyls (1.93-1.96 Å).17,18
Room-Temperature NMR Spectra of 1. Solutions of 1 in
benzene-d6, toluene-d8, CD2Cl2, THF-d8, or THF-d8/CD3CN
1
mixtures display simple H NMR spectra with multiplets for
Figure 1. Molecular structure of 1. Hydrogen atoms are omitted for clarity.
Thermal ellipsoids are drawn at the 50% probability level.
the cyclohexyl rings and a slightly broadened singlet for the
CuMe group.19 The 31P NMR spectra of 1 in these solvents
contain one broad singlet in the region δ 14-17. The 13C NMR
spectrum of 1 at room temperature contains four PCy3 signals,
three of which show 31P coupling with JCP ranging from 6 to
18 Hz. The CuMe 13C resonance appears as a singlet at δ -7
to -9, depending on the solvent, and is significantly broadened
by the 63,65Cu quadrupole. No 63Cu NMR signals are observed
for 1. With the exception of symmetrically substituted tetrahedral
complexes, quadrupolar broadening usually prevents the ob-
servation of 63Cu NMR spectra. In 63Cu spectra of [Cu(PR3)4]-
[CuR′2] ion pairs, for example, a signal was observed for the
tetrahedral cation but not for the linear anion.17b No 63Cu NMR
signals were found for (R3P)3CuR′ complexes.17b
(2; bipy ) 2,2′-bipyridine)10 polymerize AN by insertion
mechanisms. These reports, and the recent discovery of copper3c,11
and iron12 catalysts for ethylene polymerization, prompted us
to investigate the structure and solution behavior of 1 and the
reactions of 1 and 2 with AN in more detail, to determine the
mechanisms by which these complexes polymerize AN.
Results and Discussion
Cy3PCuMe (1). Complex 1 was synthesized by Yamamoto
and co-workers9 by reductive alkylation of Cu(acac)2 with Al-
(OEt)Me2 in the presence of PCy3. This copper alkyl complex
is quite stable, decomposing at 105 °C in the solid state with
evolution of ethane.9a,13 Complex 1 polymerizes acrylonitrile
and copolymerizes acrylonitrile and acrylates.9a Complex 1 does
not copolymerize AN and styrene, which the authors took as
evidence for an insertion mechanism for AN homopolymeriza-
tion.9a
We prepared 1 by the reaction of CuI with MeLi in Et2O in
the presence of PCy3 at -35 °C. After evaporation of the
solvent, the residue was extracted with a toluene/pentane mixture
and crystallized at -40 °C, to afford 1 in 46% yield. While the
overall yield is lower than in the original synthesis,9a isolation
problems due to the presence of aluminum byproducts are
avoided and crystalline material is obtained.
Variable-Temperature NMR Spectra of 1. When the
temperature is lowered to below ca. 260 K, 1 is partially
converted into a new species 3. Equilibrium mixtures of 1 and
3 are observed in the temperature range 180-260 K. The
transformation is reversible, and raising the temperature yields
1 quantitatively.
1
The H and 31P NMR spectra of 3 are similar to those of 1
(Figures 2 and 3).20 The 13C NMR spectrum of 3 contains four
signals for the PCy3 ligand, most of which are shifted slightly
downfield from the signals of 1 (Figure 4). Significantly, the
(14) Holloway, C. E.; Melnik, M. ReV. Inorg. Chem. 1995, 15, 147.
(15) Coan, P. S.; Folting, K.; Huffmann, J. C.; Caulton, K. G. Organometallics
1989, 8, 2724.
(16) Janssen, M. D.; Koehler, K.; Herres, M.; Dedieu, A.; Smeets, W. J. J.;
Spek, A. L.; Grove, D. M.; Lang, H.; van Koten, G. J. Am. Chem. Soc.
1996, 118, 4817.
(17) (a) Hope, H.; Olmstead, M. M.; Power, P. P.; Sandell, J.; Xu, X. J. Am.
Chem. Soc. 1985, 107, 4337. (b) Dempsey, D. F.; Girolami, G. S.
Organometallics 1988, 7, 1208. (c) John, M.; Auel, C.; Behrens, C.; Marsch,
M.; Harms, K.; Bosold, F.; Gschwind, R. M.; Rajamohanan, P. R.; Boche,
G. Chem. Eur. J. 2000, 6, 3060. (d) Martin, S. F.; Fishpaugh, J. R.; Power,
J. M.; Giolando, D. M.; Jones, R. A.; Nunn, C. M.; Cowley, A. H. J. Am.
Chem. Soc. 1988, 110, 7226.
Complex 1 is monomeric in the solid state with a linear
coordination geometry at Cu ( P(1)-Cu(1)-C(19) ) 175.6(1)°;
Figure 1, Table 1). No intermolecular contacts can be detected
in the packing diagram, in accordance with the preference of
Cu(I) for linear or tetrahedral geometries.14 The Cu-Me distance
(10) (a) Yamamoto, T.; Yamamoto, A.; Ikeda, S. Bull. Chem. Soc. Jpn. 1972,
45, 1111. (b) Yamamoto, T.; Yamamoto, A.; Ikeda, S. Bull. Chem. Soc.
Jpn. 1972, 45, 1104. (c) Yamamoto, A. J. Chem. Soc., Dalton Trans. 1999,
1027. (d) Yamamoto, T.; Yamamoto, A.; Ikeda, S. Polym. Lett. 1971, 9,
281. (e) Yamamoto, A.; Shimizu, T.; Ikeda, S. Makromol. Chem. 1970,
136, 297.
(11) (a) Stibrany, R. T.; Schulz, D. N.; Kacker, S.; Patil, A. O. (Exxon Research
and Engineering) WO 99/30822, 1999. (b) Stibrany, R. T.; Patil, A. O.;
Zushma, S. Polym. Mater. Sci. Eng. 2002, 86, 323. (c) Baugh, L. S.; Patil,
A. O.; Schulz, D. N.; Stibrany, R. T.; Sissano, J. A.; Zushma, S.
(ExxonMobil Research and Engineering Co.) WO 0320778, 2003. (d)
Gibson, V. C.; Tomov, A.; Wass, D. F.; White, A. J. P.; Williams, D. J. J.
Chem. Soc., Dalton Trans. 2002, 2261. (e) Stibrany, R. T.; Schulz, D. N.;
Kacker, S.; Patil, A. O.; Baugh, L. S.; Rucker, S. P.; Zushma, S.; Berluche,
E.; Sissano, J. A. Macromolecules 2003, 36, 8584.
(12) (a) Small, B. L.; Brookhart, M.; Bennett, A. M. A. J. Am. Chem. Soc.
1998, 120, 4049. (b) Britovsek, G. J. P.; Gibson, V. C.; Kimberley, B. S.;
Maddox, P. J.; McTavish, S. J.; Solan, G. A.; White, A. J. P.; Williams,
D. J. Chem. Commun. 1998, 849.
(13) Pasynkiewicz et al. observed a mixture of methane and ethane in the thermal
decomposition of 1. Pasynkiewicz, S.; Poplawska, J. J. Organomet. Chem.
1985, 282, 427.
(18) For comparison, Cu-aryl bond distances range from 1.89 to 1.94 Å. (a)
Hwang, C.-S.; Power, P. P. Organometallics 1999, 18, 697. (b) Wehman,
E.; van Koten, G.; Jastrzebski, J. T. B. H.; Rotteveel, M. A.; Stam, C. H.
Organometallics 1988, 7, 1477. (c) Wingerter, S.; Gornitzka, H.; Bertrand,
G.; Stalke, D. Eur. J. Inorg. Chem. 1999, 173. (d) Hitchcock, P. B.; Lappert,
M. F.; Layh, M. J. Chem. Soc., Dalton Trans. 1998, 1619. (e) He, X.;
Olmstead, M. M.; Power, P. P. J. Am. Chem. Soc. 1992, 114, 9668.
(19) The absence of 31P coupling for the Cu-Me resonance may result from
quadrupolar broadening by 63,65Cu or from exchange of free and coordinated
PCy3. Since PCy3 exchange in 1 is fast on the NMR time scale even at
190 K, even otherwise undetectably small amounts of free PCy3 (from
partial decomposition, hydrolysis, or impurities) would wash out the
coupling.
(20) It was reported previously that the 31P NMR resonance of 1 shifts upfield
at lower temperatures; however, only one broadened signal was observed.
These results are consistent with the 1 ) 3 equilibrium described here; the
lower field strength used in previous work did not enable resolution of the
resonances for 1 and 3. See Pasynkiewicz, S.; Poplawska, J. J. Organomet.
Chem. 1986, 302, 269.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2115

A R T I C L E S
Schaper et al.
Table 2. Equilibrium Constantsa and Thermodynamic Parameters
for the 1 ) 3 Equilibrium (eq 1) in Different Solvents
-1
solvent
K , 235K(M
)
∆H° (kJ/mol) ∆S° (J/(mol‚K)) T rangeb (K)
1
toluene-d8
CD2Cl2
2.8(2)c
0.73(3)
2.0(2)
-20.8(8)
-20.5(2)
-21.7(3)
-24.7(8)
-81(4)
-90(1)
-87(2)
-92(4)
180-260
190-250
183-250
210-230
THF-d8
THF-d8/CD3CNd
4.6(3)c
a Equilibrium constants were determined from 31P NMR spectra for
toluene-d8 and from 1H and 31P NMR spectra for all other solvents.
b Temperature range of K1 measurements used to determine ∆H° and ∆S°.
c Standard deviation is estimated. d 41/59 ratio.
Figure 2. Variable-temperature 1H NMR spectra of the 1 ) 3 equilibrium
in CD2Cl2.
Figure 5. van’t Hoff plots for the equilibrium constant K1 ) [3]/[1]2 in
toluene-d8 solution with [Cu]total ) 12 mM (O), 28 mM (0), 120 mM ([),
and 170 mM (4).
THF-d8, and THF-d8/CD3CN solution. The implications of this
observation will be discussed below. No 63Cu NMR signals are
observed for 3 over the temperature range of 180-290 K,
indicating that a symmetrical tetrahedral Cu species is not
formed.
Determination of Equilibrium Constants and Molecular
Formula of 3. The 1:3 ratio depends on the initial concentration
of 1, with higher concentrations favoring the formation of 3.
Measured values of the [1]/[3] ratio for solutions with [1]initial
) 12-170 mM, over the temperature range 180-260 K, are
uniquely consistent with the equilibrium shown in eq 1 and show
that 3 is formed from two molecules of 1. Values for K1 )
[3]/[1]2 are given in Table 2, and van’t Hoff plots for different
initial concentrations of 1 are shown in Figure 5. In accordance
with the observed stoichiometry, reaction entropies of ca. -90
J/(mol‚K) were obtained from these plots (Table 2). The
equilibrium in eq 1 is rather insensitive to solvent polarity (Table
2).
Figure 3. Variable-temperature 31P{1H} NMR spectra of the 1 ) 3
equilibrium in CD2Cl2.
Cy PCuMe
(Cy PCuMe)
2
h
(1)
3
3
2
1
3
Solution Structures of 1 and 3. Since Cy3PCuiBu22 and nBu3-
PCuMe23 are monomeric in benzene and 1 is monomeric in the
solid state, 1 is most likely monomeric in solution at room
temperature. The key observations that provide information
about the structure of 3 are (i) 3 has the molecular formula (1)2
and (ii) the PCy3 C1 13C NMR resonance of 3 is a triplet with
JCP ) 7 Hz. The triplet structure is most likely due to virtual
coupling to two phosphorus atoms.24 To determine the influence
Figure 4. 13C{1H} NMR spectrum (PCy3 region) of an equilibrium mixture
of 1 and 3 in CD2Cl2 solution at 190 K.
PCy3 C1 13C resonance of 3 in CD2Cl2 appears as triplet (δ
31.2) with JCP ) 7 Hz, in contrast to the doublet (δ 30.0) with
JCP ) 19 Hz observed for 1. The PCy3 C2 and C3 13C
resonances of 3 appear as broadened singlets, though the triplet
structure can be seen qualitatively for the C2 resonance (δ
27.1).21 Analogous results were obtained in toluene-d8,
(22) Miyashita, A.; Yamamoto, A. Bull. Chem. Soc. Jpn. 1977, 50, 1102.
(23) Pasynkiewicz, S.; Pikul, S.; Poplawska, J. J. Organomet. Chem. 1985, 293,
125.
(24) (a) Drago, R. S. Physical Methods in Chemistry; W. B. Saunders
Company: Philadelphia, PA, 1977; p 223. (b) Harris, R. K. Can. J. Chem.
1964, 42, 2275.
(21) The C1-C4 13C NMR signals of the PCy3 ligand were assigned in the
order of decreasing 31P coupling constants.
9
2116 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

Acrylonitrile Polymerization
A R T I C L E S
Table 3. NMR Data of Neutral and Cationic Cu(I) Phosphine
Scheme 1
Complexes in CD2Cl2 Solution at 190 K
31P(δ)
13Cof C1(δ)
Cy3PCuMe (1)
17.4
11.8
10.6a
25.1
15.6a,b
3.8
30.0 (d, J ) 19 Hz)
31.2 (t, J ) 7 Hz)
32.4 (d, J ) 16 Hz)a
29.8 (t, J ) 11 Hz)
32.3 (t, J ) 9 Hz)a,b
not resolved
[Cu(PCy3)2][CuMe2] (3)
{Cy3PCu(µ-I)}2 (4)
[Cu(PCy3)2][PF6] (5)
[Cu(PCy3)2][OTf] (6)
[Cu(PCy3)3][PF6]
PCy3
10.8
30.1 (d, J ) 15 Hz)
a At 290 K; 13C C1 chemical shifts at 190 K are usually 1-2 ppm
downfield from the position at 290 K. b In CDCl3; see ref 26.
of coordination geometry on 31P coupling in systems of this
type, we recorded 13C NMR spectra of the iodide-bridged dimer
results indicate that 1 is in fast equilibrium with a species
containing more than one phosphine ligand, most likely (Cy3P)2-
CuMe (7), as shown in eq 2.29 Trigonal planar (Cy3P)2CuX
complexes (X ) Cl, Br, I) are known.25,30
{Cy3PCu(µ-I)}2 (4)25 and the salt [Cu(PCy3)2][PF6] (5). The 13
C
NMR spectrum of 4 in CD2Cl2 contains a doublet for C1 (Table
3). Apparently 4JPP′ is smaller than 1JPC in this case and virtual
coupling is not observed.24 We thus exclude methyl-bridged
dimers, e.g., {Cy3PCu(µ-Me)}2, as possible structures for 3. In
Cy PCuMe
(Cy P) CuMe
+ PCy3 h
(2)
3
3
2
contrast, the PCy3 C1 13C resonance of 5 is a triplet with JCP
)
1
7
11 Hz, similar to that observed for 3. Analogous results have
been reported for [Cu(PCy3)2][OTf] (6, Table 3).26 The similarity
of the 13C NMR spectra of 3, 5, and 6 strongly implies that 3
contains a Cu(PCy3)2+ unit, and therefore that 3 is the tight ion
pair [Cu(PCy3)2][CuMe2].
Reaction of 5 with PCy3. The 31P NMR spectrum of a
solution of 5 and 0.5 equiv of PCy3 in CD2Cl2 at room
temperature contains only one PCy3 resonance, which is shifted
upfield from that of 5 toward the free PCy3 position. At 190 K,
this signal is split into two sharp resonances at δ 25.1
(corresponding to 5) and δ 3.8 in a ca. 2/3 intensity ratio. The
31P NMR spectra of a mixture of 5 and 1 equiv of PCy3 at 190
To support the proposed structure of 3, we investigated
whether 1 can be obtained by combining separately generated
Cu(PCy3)2+ and CuMe2- ions. Indeed, the methyl region of the
1H NMR spectrum of a suspension of CuMe2Li‚LiI and an
excess of 5 in THF-d8 at 190 K contains only signals for 1 and
3, in the ratio expected from the equilibrium constant K1
determined in THF-d8 at this temperature. The 31P NMR
spectrum also contains characteristic signals for 1 and 3.27
Syntheses of copper phosphine alkyl compounds generally
yield either monomeric (R3P)3CuR′ complexes or [Cu(PR3)4]-
[CuR′2] ion pairs, depending on the size of the alkyl and
phosphine ligands.14,15,17b,28 Interconversion or ligand redistribu-
tion reactions have not been reported for these species.
Complexes 1 and 3 thus represent a special case, where the
neutral monomeric and ion-paired forms are of comparable
energy.
K contains only the δ 3.8 resonance. Accordingly, the δ 3.8
+
resonance is assigned to the trisphosphine species Cu(PCy3)3
,
and we conclude that 5 coordinates one additional PCy3 ligand.
In summary, coordination of additional PCy3 takes place in
the order 5 > 1 . 3. The resistance of 3 to reaction with PCy3
+
reflects the strong interaction between the Cu(PCy3)2 and
-
CuMe2 units.
1
Coordination of Acrylonitrile to 1. Room-temperature H
NMR spectra of toluene-d8 solutions of 1 (0.1 M) containing
0.8-5 equiv of acrylonitrile (AN) contain only signals for 1
and free AN. However, as the temperature is lowered, the AN
signals broaden and shift slightly upfield, the shift being more
pronounced at lower AN:1 ratios. As the AN:1 ratio is increased,
Reaction of 1 with PCy3. The 31P spectra of mixtures of 1
and 3 in CD2Cl2 or toluene-d8 that contain 0.5-3 equiv of excess
PCy3 do not exhibit a free PCy3 resonance between 180 and
290 K. Instead, the 31P resonance of 1 is shifted upfield toward
the free PCy3 position. The 1H CuMe resonance of 1 also shifts
the CuMe 1H resonance of 1 shifts slightly upfield and the 31
P
NMR resonance of 1 shifts downfield (away from the position
of free PCy3). These changes are reversed by raising the
temperature back to ambient. Also, the formation of 3 is reduced
in the presence of AN. These observations indicate that AN
reacts with 1 to form (Cy3P)(AN)xCuMe in a fast equilibrium.
The reactivity of 1 in solution is summarized in Scheme 1.
Acrylonitrile Polymerization with 1. As originally reported
by Yamamoto and co-workers,9 solutions of 1 in toluene are
1
slightly in the presence of PCy3. In contrast, the 31P and H
CuMe resonances of 3 are unaffected by addition of PCy3.
However, the concentration of 3 is smaller than that in the
absence of PCy3 under otherwise identical conditions. These
(25) Bowmaker, G. A.; Boyd, S. E.; Hanna, J. V.; Hart, R. D.; Healy, P. C.;
Skelton, B. W.; White, A. H. J. Chem. Soc., Dalton Trans. 2002, 2722.
(26) Green, J.; Sinn, E.; Woodward, S. Inorg. Chim. Acta 1995, 230, 231.
(27) (a) An additional signal at δ 9.9 is observed in the 31P NMR spectra of
THF-d8 solutions of CuMe2Li‚LiI and 5. This signal is assigned to [Cu-
(PCy3)2][PF6]‚LiI (5‚LiI) since the same chemical shift was observed for a
1:1 mixture of 5 and LiI in THF-d8. At room temperature, only 1 and the
putative 5‚LiI species were observed. The formation of [Cu(PCy3)2][PF6]‚
LiI can be suppressed by use of excess CuMe2Li‚LiI. In this case, signals
assigned to CuMe2Li‚LiI (see ref 27b) can be detected at 190 K in addition
to the methyl signals for 1 and 3. At room temperature, fast methyl exchange
occurs and only one broadened signal is observed in the methyl region of
the 1H spectrum. (b) Ashby, E. C.; Watkins, J. J. J. Am. Chem. Soc. 1977,
99, 5312.
(29) The equilibrium constant for eq 2, K2 ) [7]/([1]‚[PCy3]), was determined
as follows. Since 3 does not coordinate additional PCy3, its concentration
can be obtained from 1H NMR spectra. This value, along with the
independently determined equilibrium constant for the 1 ) 3 equilibrium,
enables determination of the concentrations of 1, 7, and PCy3 (even though
these species are in fast exchange with each other) and determination of
K2. In CD2Cl2, K2 ) [7]/([1][PCy3]) ) 21(3) M-1 at 190 K, ∆H° ) -22(1)
kJ/mol, and ∆S° ) -91(6) J/(mol‚K). The ∆H° and ∆S° values were
determined from K2 values measured over the temperature range 190-
230 K. The standard deviation for K2 is a conservative estimate based on
likely errors and observed standard deviations in similar systems. Standard
deviations of ∆H° and ∆S° values are derived from the linear regression
of the van’t Hoff plot.
(30) Soloveichik, G. L.; Eisenstein, O.; Poulton, J. T.; Streib, W. E.; Huffman,
J. C.; Caulton, K. G. Inorg. Chem. 1992, 31, 3306.
(28) Leoni, P.; Pasquali, M.; Ghilardi, C. A. Chem. Commun. 1983, 240.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2117

A R T I C L E S
Schaper et al.
Table 4. Acrylonitrile Polymerizations with Different Initiators in Toluene Solution
% Et endsa
entry
initiator
T (°C)
[AN] (M)
[AN]/[initiator]
time (min)
% conversion
activityb
mm (%)
branch numberc
observed
theory
1
2
1
1
1
1
1
1
1
1
1g
none
PCy3
PCy3
PCy3
80
23
23
0
2.3
2.3
6.5
3.9
3.4
6.4
7.9
6.6
3.9
3.9
3.9
3.9
3.9
3.9
5.4
15.2j
8.9k
440
440
120
355
3
1
5
0.3
0.5
3
4400
1100
1100
4100
430
26
19
40
20
61
17
11
0
53
54
1
9
0
1700
32
34
32
32
30
31
35
16
34
19
26
0.03
0.1
0.05
0.04
0.1
0.5
0.3
0.1
4d
5
40
45
60
80
135
40
45
40
45
75
45
120
15
20
39 ( 18e
0f
56
60
2
25
6
7
8
9
10
11d
12
13
14d
15
16d
17d
0
-37
-37
0
4500
1000
15
44
0.05
0.2
0.1
1.0
15
0
0
0
0
0
0
23
23
0
1100
1100
1100
600
1000
240
75 ( 15e
32
33
34
31
0
0
h
i
86
4
8
0
8
Me2CuLi‚LiI
32
18
1.7
1.8
5
2
2
95
80
29
32
45
21
0.2
0.3
0.45
0.75
170
3
a The number of Et (Me in the case of 2) end groups per 100 monomer units was determined from the ratio of the intensities of the 1H -CH2CH3 resonance
at δ 1.03 vs the main-chain -CH2- resonances at δ 2.0-2.6. For 2, the 1H -CH(CN)CH3 resonance at δ 1.31 was used. The theoretical percentage of Et end
groups (Me in the case of 2) is the ratio that would be observed if exactly one Me group per 1 or Me2CuLi (or one hydride per 2) is utilized in initiation.
b Defined as grams of PAN/([initiator][AN]hour). c The branch number is the number of branches per 500 monomer units and was determined from NMR
1
spectra from the ratio of the intensities of the branch end -CH2CN resonances (δ 13.9 and 12.2 in 13C, δ 2.7 in H) to the main-chain -CH2- resonances (δ
1
32-39 in 13C, δ 2.0-2.6 in H). d Polymer molecular weights were determined for selected samples: entry 4, Mw ) 1.0 × 104 g/mol and Mw/Mn ) 1.9;
entry 11, Mw ) 1.2 × 104 g/mol and Mw/Mn ) 2.1; entry 14, Mw ) 4.4 × 103 g/mol and Mw/Mn ) 1.4; entry 16, Mw ) 7.3 × 104 g/mol and Mw/Mn ) 7.0;
entry 17, Mw ) 1.5 × 104 g/mol and Mw/Mn ) 3.8. e Average of three experiments. f Toluene/AN mixture stirred with 10 mg of 1 at room temperature for
3 min and then vacuum-transferred to 10 mg of 1 at -196 °C and warmed to 0 °C (see ref 43). g Toluene partly replaced by 1-hexene; hexene:AN ) 0.3.
h 5 equiv of TEMPO added. i 0.6 equiv of [Cu(NCMe)4][PF6] added. j Reaction conducted in neat AN. k Reaction conducted in a mixture of 1.5 g (28 mmol)
of AN and 2.0 g (32 mmol) of vinyl chloride.
active for AN polymerization. As summarized in Table 4, AN
(2-8 M in toluene) is polymerized by 1 to white PAN at
temperatures ranging from -37 to 23 °C. The 13C NMR
spectrum of a representative PAN sample produced by 1 is
shown in Figures 6a and 7a. The tacticity of PAN can be
determined from the -CH2C(CN)H- methine resonance at δ 27
or the nitrile resonance at δ 120.31 The PANs produced by 1
between -37 and 23 °C are all slightly isotactic (mm ) 30-
35%, Table 4). PAN obtained by radical polymerization is
usually atactic,32 while anionic polymerizations yield slightly
to highly isotactic polymer.32a,33 In previous work only small
variations in tacticity were found in anionic polymerizations
between -80 and 40 °C.32a
In addition to the main-chain -CH2C(CN)H- resonances,
several other peaks are present in the 13C NMR spectra of PAN
produced by 1, which are labeled with lowercase letters in Figure
6 or asterisks in Figure 7. These peaks are characteristic of
branches formed in anionic polymerizations,7b-d,32a,33a and their
Figure 6. 13C{1H} NMR spectra (alkyl region) in DMSO-d6 at 60 °C of
assignments by Kamide and co-workers7c are shown in Figure
PAN obtained with different initiators: (a) 1 (entry 4), (b) PCy3 (entry
8. In contrast, signals for branched structures are absent from
NMR spectra of PAN obtained by radical polymerization
(Figures 6d and 7d).4a-c,7c,32a,34 Approximately 30 branches per
11), (c) Me2CuLi‚LiI (entry 14), (d) commercial PAN prepared by radical
polymerization. Entry numbers refer to experiments in Table 4. The vertical
scales are adjusted so that the intensity of the -CH2- resonance is normalized.
Assignments are based on the labeling scheme shown in Figure 8. Small
arrows indicate peaks that are only observed when Me2CuLi‚LiI is used as
the initiator (see text).
(31) (a) Balard, H.; Fritz, H.; Meybeck, J. Makromol. Chem. 1977, 178, 2393.
(b) Kamide, K.; Yamazaki, H.; Okajima, K.; Hikichi, K. Polym J. (Tokyo)
1985, 17, 1233. (c) Kamide, K.; Yamazaki, H.; Okajima, K.; Hikichi, K.
Polym. J. (Tokyo) 1985, 17, 1291. (d) Katsuraya, K.; Hatanaka, K.;
Matsuzaki, K.; Minagawa, M. Polymer 2001, 42, 6323.
(32) (a) Kamide, K.; Ono, H.; Hisatani, K. Polym. J. (Tokyo) 1992, 24, 917. (b)
Matsuzaki, K.; Uryu, T.; Okada, M.; Shiroki, H. J. Polym. Sci., Part A
1968, 6, 1475.
(33) (a) Opitz, G.; Heller, A.; Scheller, D.; Berger, W. Acta Polym. 1996, 47,
67. (b) Zheng, H.; Zhang, Y.; Shen, Z. Polym. Int. 2002, 51, 622. (c)
Nakano, Y.; Hisatani, K.; Kamide, K. Polym. Int. 1994, 35, 207. (d) Nakano,
Y.; Hisatani, K.; Kamide, K. Polym. Int. 1994, 35, 249.
(34) (a) Malsch, G.; Fritzsche, P.; Dautzenberg, H. Acta Polym. 1981, 32, 758.
(b) Bajaj, P.; Sreekumar, T. V.; Sen, K.; Kumar, R.; Brar, A. S. J. Appl.
Polym. Sci. 2003, 88, 1211. (c) Bajaj, P.; Sreekumar, T. V.; Sen, K. J.
Appl. Polym. Sci. 2001, 79, 1640.
500 monomer units (which includes monomer units in the
branches) were found for AN:initiator ratios of ca. 1000 (Table
4).35
Compound 1 does not polymerize ethylene or hexene. When
the polymerization of AN was carried out in the presence of
hexene, no hexene incorporation in the PAN was observed and
the microstructure of the PAN was nearly identical to that
formed in the absence of hexene. The slight isotacticity, lack
of temperature dependence of the tacticity, lack of activity in
9
2118 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

Acrylonitrile Polymerization
A R T I C L E S
Scheme 2
Free PCy3 is slightly more active than 1 for AN polymeri-
zation under identical conditions (Table 4, entry 11 vs entry
4). The presence of TEMPO does not influence activity or
polymer microstructure in PCy3-initiated AN polymerizations,
in agreement with an anionic mechanism (Table 4, entry 12).
In contrast, addition of 0.6 equiv of [Cu(NCMe)4][PF6] strongly
inhibits polymerization, probably by trapping of PCy3 to form
5 (Table 4, entry 13). Control experiments show that 5 is inactive
for AN polymerization (Table 4, entry 15). The branching
patterns are practically identical for PANs obtained with PCy3
or 1. Indeed, the 13C NMR spectra of these PANs are virtually
superimposable (Figures 6 and 7, compare spectra a and b in
each figure).
To further probe the possible role of PCy3 in AN polymer-
ization initiated by 1, we investigated the presence of Cy3PR+
cations in the PANs by 31P NMR. 31P spectra of PAN obtained
with 1 at high initiator concentrations (AN/1 ) 40-50) and
washed with HCl/H2O and acetone contain three resonances
between δ 34.1 and 34.4 (Figure 9a). These resonances are in
the range expected for Cy3P+CH2CR(CN)‚‚‚ species (R ) H
or branch). For example, the 31P resonance of [Cy3PCH2CH2-
CN]Cl appears at δ 34.6 (Figure 9d), and chemical shifts of δ
28-33 were reported for [Cy3PR]Br salts in CDCl3 solution.37,38
Repetitive (2×) dissolution of the PAN in DMF and reprecipi-
tation in methanol, which was reported to remove noncovalently
attached phosphonium species,36d-f does not significantly reduce
the intensities of these resonances (Figure 9b). The 31P spectrum
of PAN produced with PCy3 at low AN/PCy3 ratios (and washed
with HCl/H2O and acetone) contains the same three δ 34.1-
34.4 resonances and an additional resonance at δ 34.6 (Figure
9c). Reprecipitation of this PAN from DMF/MeOH removes
the δ 34.6 resonance but not the δ 34.1-34.4 resonances. On
the basis of these results, we assign the δ 34.1-34.4 resonances
to Cy3P+CH2CR(CN)‚‚‚ species that are covalently attached to
the polymer chain and the δ 34.6 resonance to the [Cy3PCH2-
CH2CN]+cation,whichisderivedfromtheinitialCy3P+CH2CH(CN)-
zwitterion by proton abstraction from AN (Scheme 2) or by
protonation during workup.36 The small chemical shift difference
(∆δ ) 0.05) between the δ 34.6 resonance in PAN and the
[Cy3PCH2CH2CN]Cl resonance maybe due to the presence of
PAN in the polymer solutions or to ion-pairing effects. These
Figure 7. 13C{1H} NMR spectra (nitrile region) in DMSO-d6 at 60 °C of
PAN obtained with different initiators: (a) 1 (entry 4), (b) PCy3 (entry
11), (c) Me2CuLi‚LiI (entry 14), and (d) commercial PAN prepared by
radical polymerization. Entry numbers refer to experiments in Table 4. The
vertical scales are adjusted so that the intensity of the -CH2- resonance is
normalized. Assignments are based on the labeling scheme shown in Figure
8.
Figure 8. 13C NMR assignments for branches in PAN. Letters and asterisks
correspond to labeled peaks in Figures 6, 7, and 11. Assignments are based
on ref 7c.
ethylene or hexene polymerizations, and most importantly, the
presence of significant branching show that polymerization of
AN by 1 proceeds by an anionic mechanism.
Identification of Initiating Species in AN Polymerizations
by 1. Three possible scenarios for initiation of anionic poly-
merization of AN by 1 are (i) initiation by PCy3 released from
1 or the minor species 3,36 (ii) initiation by the Cu-Me group
-
of 1, or (iii) initiation by the CuMe2 anion of 3. Initiation by
PCy3 would yield phosphonium cations, which are either
covalently bound to the polymer chain or present as Cy3PCH2-
CH2CN+ cations (Scheme 2).36d-f The evidence discussed in
this section strongly supports a significant role for PCy3
initiation.
(36) (a) Triaryl- and trialkylphosphines have been reported to initiate AN
polymerization. (b) Horner, L.; Jurgeleit, W.; Klupfel, K. Liebigs Ann.
Chem. 1955, 591, 108. (c) Morita, K.; Suzuki, Z.; Hirose, H. Bull. Chem.
Soc. Jpn. 1968, 41, 2815. (d) Jaacks, V.; Eisenbach, C. D.; Kern, W.
Makromol. Chem. 1972, 161, 139. (e) Eisenbach, C. D.; Jaacks, V.;
Schnecko, H.; Kern, W. Makromol. Chem. 1974, 175, 1329. (f) Eisenbach,
C. D.; Franzmann, G.; Jaacks, V.; Schnecko, H.; Kern, W. Makromol.
Chem. 1974, 175, 1789. (g) Markevich, M. A.; Kochetov, E. V.; Ranogajec,
F.; Enikolopyan, N. S. J. Macromol. Sci., Chem. 1974, 8, 265. (h) Kuran,
W.; Borycki, J. J. Polym. Sci., Polym. Chem. Ed. 1981, 19, 1251 (i)
Heisenberg, E.; Jurgeleit, W. (Vereinigte Glanzstoff-Fabriken) US 2,921,-
055, 1960.
(35) The presence of branching is also indicated by the 1H NMR spectra of the
PAN samples. In addition to backbone CH2 and CH(CN) signals at δ 2.06
and 3.15, a broad singlet at δ 2.7 is observed, which was assigned to the
-CH2CN units at branch or chain ends (ref 7c,d). We also observe a broad
singlet at δ 2.4 with approximately the same intensity as the δ 2.7 resonance.
The position and intensity suggest that this signal is also associated with
branch or chain ends and probably corresponds to the penultimate methylene
group (-CH2CH(CN)CH2CN).
(37) Bestmann, H. J.; Do¨tzer, R. Synthesis 1989, 3, 204.
(38) (a) In contrast, the 31P resonance for Cy3PdO appears at δ 50. Davies, J.
A.; Dutremez, S.; Pinkerton, A. A. Inorg. Chem. 1991, 30, 2380. (b) Control
experiments show that the δ 34.1-34.4 resonances are not derived from
[Cy3PH]Cl.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2119

A R T I C L E S
Schaper et al.
Figure 10. 1H NMR spectra (alkyl region) in DMSO-d6 at 60 °C of PAN
obtained with different initiators: (a) Me2CuLi‚LiI (entry 14), (b) 1 (entry
4), and (c) PCy3 (entry 11). Entry numbers refer to experiments in Table 4.
The vertical scales are adjusted so that the intensity of the -CH2- resonance
is normalized.
counterion. However, addition of “Cu(PCy3)+” (in form of an
equimolar [Cu(NCMe)4][PF6]/PCy3 mixture) to an active PCy3-
initiated AN polymerization (which was otherwise free of Cu)
1
after /3 of the polymerization time reduced the polymer yield
to 1/3 of that obtained in an otherwise identical polymerization.
This result implies that “Cu(PCy3)+” strongly coordinates the
active RCH2CH(CN)- anions and quenches the polymerization
(cf. the inhibition of AN polymerization by Li+, vide infra).
Thus, while Me- transfer from 1 to AN clearly occurs, this
process is not expected to lead to efficient PAN formation.
It is more difficult to assess the importance of initiation by
Figure 9. Cy3PR+ region of 31P{1H} NMR spectra (60 °C in DMSO-d6):
(a) PAN obtained with 1, (b) PAN obtained with 1 and reprecipitated from
DMF/MeOH, (c) PAN obtained with PCy3, and (d) [Cy3PCH2CH2CN]Cl.
results show that PCy3 is a significant initiator in AN polymer-
izations by 1. The initiating PCy3 is most likely derived from 1
rather than 3, since 5 is inactive for AN polymerization.
If 1 initiates AN polymerization by Me- transfer to AN, CH3-
CH2CH(CN)- end groups should be present in the PAN. The
1H NMR spectra of PAN produced by 1 contain a signal at δ
1.03, which is not present in the spectra of PAN obtained with
PCy3 (Figure 10b,c). This signal appears as two overlapping
triplets with a separation of ∆δ ) 0.004 at 120 °C. The same
triplet resonances are observed in much higher intensity if AN
polymerization is initiated by CuMe2Li‚LiI (Table 4, entry 14;
Figure 10a). We assign these resonances to CH3CH2CH(CN)-
CH2CH(CN)- chain ends with different tacticity, generated by
initiation by Me- addition to AN. For comparison, the CH3-
CH2- resonances of d,l- and meso-3,5-heptanediol were observed
at δ 0.93 and 0.94, and those for d,l- and meso-3,5-diiodoheptane
appear at δ 0.76 and 0.78.39 Several new, small peaks are present
in the 13C NMR spectrum of PAN produced by CuMe2Li‚LiI
(arrows in Figure 6c), which probably also arise from the CH3-
Me- transfer from the CuMe2 unit in 3, since this process
-
would yield a more coordinatively saturated Cu(PCy3)2+ cation.
As mentioned above, CuMe2Li‚LiI does polymerize AN, but it
is less active than 1 (Table 4, entry 14). Additionally, the PAN
produced with CuMe2Li‚LiI is less branched than that from 1
or PCy3 (Figures 6c and 7c, Table 4). The low activity and
branching in CuMe2Li‚LiI-initiated AN polymerization result
from strong ion pairing involving the Li+ cation.7d,32a,41 Although
it is clear from these experiments that CuMe2- anions do initiate
AN polymerization, the facts that (i) 3 is a very tight ion pair
and (ii) 1 rapidly polymerizes AN under conditions where the
concentration of 3 is vanishingly small (e.g., Table 4, entry 3)42
are inconsistent with a significant role of 3 in AN polymeriza-
tions by 1.
(40) Alternatively, the CH3CH2CH(CN)- chain ends may arise from the reaction
of 1 with the Cy3P+CH2CH(CN)- chain ends or the Cy3P+CH2CH(CN)-
zwitterion formed by PCy3 initiation. Phosphonium cations with â-cyano
alkyl substituents are known to react with nucleophiles, such as RO-, by
nucleophilic substitution (see ref 40b). In our hands, however, the reaction
of [Cy3PCH2CH2CN]Br with 1 in CD3CN produced CH4 and AN, probably
by deprotonation of the â-CH2 group of [Cy3PCH2CH2CN]Br and
subsequent dissociation of the regenerated zwitterion to AN and PCy3. (b)
Grayson, M.; Keough, P. T.; Johnson, G. A. J. Am. Chem. Soc. 1959, 81,
4803.
1
CH2CH(CN)CH2CH(CN)- chain ends. The intensity of the H
δ 1.03 signal in PAN produced by 1, compared to the CH2
resonance at δ 2.1, indicates that ca. 20% of the methyl groups
in 1 are transferred to the polymer chain (Table 4).40 In contrast,
when CuMe2Li‚LiI is used as the initiator, one methyl group
per Me2CuLi unit reacts with the monomer (Table 4, entry 14).
Initiation of AN polymerization by Me- transfer from 1
would generate an active propagating anion and a “Cu(PCy3)+”
(41) Me- transfer from Me2CuLi‚LiI yields CuMe (the fate of which is unknown)
and a Li+ cation. Retardation of polymerization rates by 1-2 orders of
magnitude in the presence of Li+ was reported for PPh3-initiated AN
polymerizations (see ref 36f). In our hands, the presence of 2 equiv of LiI
completely quenches AN polymerization by 1 (under conditions otherwise
identical to those of entry 4 in Table 4). LiI does not polymerize AN under
the conditions used for Me2CuLi‚LiI-initiated polymerization (entry 14,
Table 4).
(39) (a) Yang, G. K.; Bergman, R. G. J. Am. Chem. Soc. 1983, 105, 6045. (b)
NMR data for 3,5-dicyanoheptane or 2,4-dicyanohexane have not been
reported.
9
2120 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

Acrylonitrile Polymerization
A R T I C L E S
Scheme 3
The presence of both polymer-bound phosphonium species
and CH3CH2CH(CN)- chain ends in PAN produced by 1
indicates that both initiation mechanisms, PCy3 addition to AN
and Me- transfer, occur in AN polymerization by 1. However,
the identical microstructures of polymers obtained with 1 and
PCy3 and the efficient quenching of active AN polymerizations
by “Cu(PCy3)+” imply that chain growth occurs mainly at
polymer anions that have a phosphonium counterion. Therefore,
we conclude that anionic AN polymerization by 1 primarily
results from initiation by PCy3, which is probably liberated from
1 by substitution by AN.43
(Bipy)2FeEt2 (2). The iron(II) complex (bipy)2FeEt2 (2) was
synthesized by Yamamoto et al.44 by reaction of Fe(acac)3 with
Al(OEt)Et2 in the presence of bipy. Complex 2 slowly decom-
poses in solution at room temperature with evolution of ethane
and ethylene (ca. 1/1).10,45 The decomposition of 2 is inhibited
by excess bipy but promoted by coordinating solvents (MeCN
> DMF > acetone > THF) and was proposed to occur by partial
or complete dissociation of a bipy ligand, â-hydride elimination
to yield ethylene and a transient Fe-H species, and reductive
elimination to yield ethane and (bipy)2Fe (Scheme 3).10a,45
Compound 2 was reported to initiate AN polymerization.10
The polymerization was proposed to proceed via π-coordination
of the AN vinyl group to a vacant site created by partial
dissociation of the bipy ligands, followed by insertion into the
Fe-Et bond. Compound 2 also polymerizes methyl methacrylate
and cyclodimerizes butadiene but is inactive for ethylene
polymerization.10
Figure 11. 13C{1H} NMR spectrum in DMSO-d6 at 60 °C of PAN (top,
nitrile region; bottom, alkyl region) produced with 2 (entry 16, Table 4).
Assignments are based on the labeling scheme in Figure 8.
values for PAN obtained with 1 or PCy3. These results provide
strong evidence that 2 polymerizes AN by an anionic mecha-
nism.
To confirm that AN polymerization by 2 does not occur by
a radical mechanism, AN/vinyl chloride copolymerization by
2 was investigated. Vinyl chloride is readily incorporated in
radical AN polymerizations.47 However, the reaction of 2 with
an AN/vinyl chloride mixture produced PAN with no vinyl
chloride incorporation (Table 4, entry 17).48 The branch patterns
and tacticity of this material are very similar to those of PAN
produced by 2 in the absence of vinyl chloride. Additionally, 2
does not homopolymerize vinyl chloride. These observations
rule out a radical mechanism for AN polymerization by 2.
Possible Initiators in AN Polymerization by 2. Free bipy
does not react with AN over 24 h at ambient temperatures and
therefore is not the initiator in AN polymerization by 2.
AN Polymerization by 2. Dissolution of 2 in neat AN
produces a deep blue solution. After a 15 min induction period,
gas evolution followed by fast exothermic polymerization of
AN to high molecular weight PAN is observed (Table 4, entry
16).46 The evolved gas is mainly ethylene (ethylene/ethane )
10/1).
The 13C NMR spectra of PAN produced by 2 display
resonances characteristic of branch units that are identical to
those observed in PAN produced by 1 or PCy3 (Figure 11).
The branch number and tacticity are similar to the corresponding
1
The H NMR spectrum of PAN generated by 2 contains a
signal at δ 1.31 that splits into two doublets at δ 1.33 and 1.35
at 120 °C (Figure 12). These resonances are not observed in
the spectra of PANs produced by either 1 or PCy3. We assign
these resonances to CH3CH(CN)CH2CH(CN)- chain ends
generated by initiation by hydride addition to AN. For com-
parison, the 1H NMR spectrum of PAN obtained by Li[Et3BH]
initiation contains the same signal (see Supporting Information),
and the CH3CH(CN)- resonances of 2,4-dicyanopentanes in
DMSO-d6 at 140 °C occur at δ 1.31 (meso) and 1.33 (d,l).32b,49
The most likely source of the initiating hydride species in
AN polymerization by 2 is an Fe-H intermediate formed by
decomposition of 2. As noted above, coordinating solvents such
as MeCN accelerate the decomposition of 2, probably by
stabilizing unsaturated intermediates formed by bipy dissocia-
tion.10a,45 Analogously, substitution of a bipy ligand by AN
(42) By use of the ∆H° and ∆S° values for the 1 ) 3 equilibrium in Table 2,
the concentrations of 1 and 3 under the conditions of entry 3 in Table 4
are calculated to be 1.5 × 10-3 M and 5.7 × 10-7 M, respectively. However,
the actual concentration of 3 will be significantly lower than the latter value,
since the formation of 3 is strongly inhibited by the presence of a large
excess of AN (see text).
(43) To investigate whether the release of PCy3 is due to impurities in the solvent
or AN, a typical polymerization mixture (i.e., 2 mL of AN and 5.5 mL of
toluene) was first stirred with 10 mg of 1 at 23 °C until polymerization
started, as indicated by development of a yellow color and precipitation of
PAN. The volatiles (toluene and AN) were then vacuum-transferred to
another flask containing 10 mg of 1 at -196 °C. The flask was sealed and
placed in an ice/water bath and polymerization was allowed to proceed.
As shown in entry 5 of Table 4, the percent conversion, tacticity, branching,
and end groups are similar to those for AN polymerization by 1 under
standard conditions at 0 °C (entry 4). The differences are ascribed to the
slow warming from -196 to 0 °C during polymerization. We thus conclude
that PCy3 is not liberated from 1 due to the presence of impurities in the
solvent or monomer.
(47) (a) Koenig, J.; Sueling, C. (Bayer AG) US 4118556, 1978. (b) Koenig, J.;
Sueling, C.; Boehmke, G. (Bayer AG), DE 76-2604630, 1977. (c) Guillot,
J.; Vialle, J.; Guyot, A. J. Macromol. Sci., Chem. 1971, 5, 735. (d) Koenig,
J.; Wendisch, D. Angew. Makromol. Chem. 1980, 92, 191.
(48) For a radical batch copolymerization with AN and vinyl chloride present
in equal starting concentrations (Table 4, entry 17), at least 20% vinyl
chloride incorporation into the polymer is expected on the basis of the
monomer reactivity ratios. See Odian, G. Principles of Polymerization, 3rd
ed.; Wiley-Interscience: New York, 1991; p 479.
(49) Matsuzaki, K.; Uryu, T.; Okada, M.; Ishigure, K.; Oki, T.; Takeuchi, M.
J. Polym. Sci., Part B 1966, 4, 487.
(44) Yamamoto, A.; Morifuji, K.; Ikeda, S.; Saito, T.; Uchida, Y.; Misono, A.
J. Am. Chem. Soc. 1968, 90, 1878.
(45) Lau, W.; Huffman, J. C.; Kochi, J. K. Organometallics 1982, 1, 155.
(46) The resulting PAN is pale red, presumably due to residual Fe species.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2121

A R T I C L E S
Schaper et al.
Scheme 5
Figure 12. 1H NMR spectrum (DMSO-d6 at 120 °C) of PAN produced
by 2 (entry 16, Table 4).
Scheme 4
there is no evidence in the literature for olefin insertion into
tertiary alkyl-metal bonds in polymerization reactions. For
palladium and nickel catalysts, which typically yield branched
polymers, ethylene insertion does not occur into tertiary alkyl-
metal bonds.50b,52 We thus exclude the possibility that an
insertion mechanism yields the branched PANs produced by 1
and 2.
should promote â-H elimination of 2 and formation of an Fe-H
species. We propose that the Fe-H intermediate undergoes net
hydride transfer to AN to initiate anionic polymerization, as
shown in Scheme 4. The details of the hydride transfer step are
unknown at this time. The evolution of ethylene, but only traces
of ethane, during the polymerization is consistent with efficient
trapping of the Fe-H intermediate by AN. In contrast, under
identical conditions, the decomposition of 2 in MeCN in the
absence of AN yields a 1/1 mixture of ethylene and ethane in
accordance with literature reports.45
Branching in Insertion Polymerization of AN. The conclu-
sion that AN polymerizations by 1 and 2 occur by an anionic
mechanism rests primarily on the observation of branch
structures, i.e., -CH2CN-ended branches on tertiary -CR2(CN)
carbon atoms, in the PANs. Therefore it is important to consider
whether insertion mechanisms can yield PANs with such
branches. Branching in insertion polymerization of ethylene with
late transition metal catalysts arises by chain-walking (repeated
â-H elimination and reinsertion), followed by monomer insertion
into a secondary alkyl-metal bond.50 As illustrated in Scheme
5, branches with -CH2CN ends could form in an AN insertion
polymerization by chain-walking of a MCH(CN)CH2{CH(CN)-
CH2}2R species (A) formed by 2,1-insertion of AN into an
alkyl-metal bond.51 However, to form the branch on a tertiary
carbon (see B), a 2,1-insertion of AN into a tertiary alkyl-
metal bond would be required. To the best of our knowledge,
Conclusions
This study provides new insights to the mechanism of AN
polymerization by metal-alkyl complexes. Cy3PCuMe (1) exists
in a temperature-dependent equilibrium with the tight ion pair
[Cu(PCy3)2][CuMe2] (3) in solution. Compound 1 polymerizes
acrylonitrile by an anionic mechanism. The major initiator in
this reaction is PCy3, which is released from 1, probably by
substitution by AN. Initiation by Cu-Me species also occurs,
but chain growth occurs mainly at polymer anions that have a
phosphonium counterion. (Bipy)2FeEt2 (2) also initiates anionic
polymerization of AN. In this case, the initiating species is
proposed to be a transient iron hydride produced by â-hydride
elimination of 2.
Experimental Section
All manipulations were performed with drybox or Schlenk techniques
under a nitrogen atmosphere or on a high-vacuum line. Nitrogen was
purified by passage through columns containing activated molecular
sieves and Q-5 oxygen scavenger. Et2O, THF, THF-d8, and toluene-d8
were distilled from sodium benzophenone; CH2Cl2 and CD2Cl2 were
distilled from CaH2. Acrylonitrile (AN) was distilled from CaCl2, which
was previously dried at 120 °C for 12 h on a high-vacuum line. NMR
solvents and AN were deoxygenated by at least three freeze-pump-
thaw cycles. Et2O, THF, and CH2Cl2 were saturated with N2 by repeated
(>20 times) degassing/N2 flush cycles. Pentane and toluene were
purified by passage through columns of activated alumina and BASF
R3-11 oxygen removal catalyst. (Bipy)2FeEt2 (2) was synthesized
according to the literature.10a All other chemicals were purchased from
Strem or Aldrich and used as received.
(50) (a) Johnson, L. K.; Killian, C. M.; Brookhart, M. J. Am. Chem. Soc. 1995,
117, 6414. (b) Ittel, S. D.; Johnson, L. K., Brookhart, M. Chem. ReV. 2000,
100, 1169.
(51) Chain-walking after 1,2-insertion of AN would yield MeCH(CN)- ended
branches. The methyl end groups observed in the PAN produced by 2 or
Li[Et3BH] are present at much lower levels than the -CH2CN ends and are
ascribed to chain initiation by hydride transfer.
(52) Shultz, L. H.; Tempel, D. J.; Brookhart, M. J. Am. Chem. Soc. 2001, 123,
11539.
9
2122 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

Acrylonitrile Polymerization
A R T I C L E S
Elemental analyses were performed by Midwest Microlab. GPC
measurements were performed on a Waters 150C high-temperature
GPC. Samples were dissolved in DMF at 40 °C. Molecular weights
were determined versus polystyrene standards and corrected by use of
experimentally determined Mark-Houwink parameters. Details are
given in the Supporting Information.
NMR (290 K, CD2Cl2): δ 2.03 (m, 6H), 1.80-1.90 (m, 24H), 1.75-
1.80 (m, 6H), 1.30-1.45 (m, 24H), 1.24 (m, 6H). 13C{1H} NMR (290
K, CD2Cl2): δ 31.8 (t, J ) 11 Hz, C1), 31.8 (s, C3), 27.3 (t, J ) 5 Hz,
C2), 26.2 (s, C4). 31P{1H} NMR (290 K, CD2Cl2) δ: 24.9 (s, PCy3),
-144.2 (septet, J ) 710 Hz, PF6-). 19F NMR (290 K, CD2Cl2): δ
1
73.7 (d, J ) 710 Hz). H NMR (190 K, CD2Cl2): δ 1.93 (m, 6H),
1.78 (m, 24H), 1.67 (m, 6H), 1.27 (m, 24H), 1.12 (m, 6H). 13C{1H}
NMR (190 K, CD2Cl2): δ 30.6 (s, C3), 29.8 (t, J ) 11 Hz, C1), 26.3
(t, J ) 6 Hz, C2), 25.2 (s, C4). 31P{1H} NMR (190 K, CD2Cl2): δ
25.1 (s, PCy3), -144.5 (septet, J ) 710 Hz, PF6-). Anal. Calcd for
C36H66CuF6P3: C, 56.20; H, 8.65. Found: C, 56.07; H, 8.54.
NMR spectra were recorded on Bruker DMX-500 or DRX-400
1
spectrometers. H and 13C chemical shifts are referenced to residual
solvent signals: CD2Cl2, δ1H ) 5.32 and δ13C ) 53.8; toluene-d8, δ1H
) 2.09 and δ13C ) 20.4; THF-d8, δ1H ) 3.58 and δ13C ) 67.4; DMSO,
δ1H ) 2.49 and δ13C ) 39.5. These values were used at all temperatures.
31P and 19F chemical shifts are reported relative to external H3PO4 (85%)
and external neat CFCl3, respectively, at room temperature. Coupling
constants smaller than 5 Hz in 13C spectra were determined by
apodization with a Gaussian function. Polymer 13C NMR spectra were
recorded for 10 wt % DMSO solutions in sealed NMR tubes at 60 °C
with an acquisition time of 2 s and without a relaxation delay. Using
a relaxation delay of 6 s yielded the same values for tacticity and
branching. An internal Ph3PO standard was used to quantify signal
intensities in 31P NMR spectra. Temperatures in VT NMR experiments
were calibrated by use of MeOH as a chemical thermometer.53
Electrospray mass spectra (ESI-MS) were recorded on freshly
prepared samples (ca. 1 mg/mL in CH2Cl2) by use of an Agilent 1100
LC-MSD spectrometer incorporating a quadrupole mass filter with a
m/z range of 0-3000. A 5 µL sample was injected by flow injection
by use of an autosampler. Purified nitrogen was used as both the
nebulizing and drying gas. Typical instrumental parameters were:
drying gas temperature 350 °C, nebulizer pressure 35 psi, drying gas
flow 12.0 L/min, and fragmentor voltage 0 or 70 V. In all cases where
assignments are given, observed isotope patterns closely matched
calculated isotope patterns. The listed m/z value corresponds to the most
intense peak in the isotope pattern.
Me2CuLi‚LiI.17c A suspension of CuI (500 mg, 2.63 mmol) in Et2O
(10 mL) was cooled to -60 °C. A solution of MeLi in Et2O (2.9 mL,
1.7 M, 4.9 mmol) was added by syringe. The suspension was stirred
for 4 h, while it warmed to -30 °C, yielding a slurry of yellow CuMe
in a colorless supernatant (ATTENTION: CuMe is explosiWe when
dry. See ref 16 for information concerning safe handling and disposal
of this compound.) The slurry was filtered cold and hexane (10 mL)
was added to the filtrate. The solvents were removed from the filtrate
under vacuum and the remaining solid was dried for 3 h under vacuum
to yield 512 mg of a yellow powder, which was stored at -40 °C and
used without further purification. This material was contaminated with
0.25 equiv of CuMe. 1H NMR spectra in THF-d8 contained signals for
Me2CuLi and Me3Cu2Li, formed by the reaction of CuMe with Me2-
CuLi.27b
Acrylonitrile Polymerizations with 1. Details are provided in Table
4. Method A: The desired amounts of initiator and toluene (typically
5.00 g) were placed in a 100 mL Schlenk flask equipped with a stirring
bar and a rubber septum. The solution was stirred for 5-15 min at the
desired temperature, and acrylonitrile (typically 2.0 mL) was added
by syringe with exclusion of light. The mixture was stirred for the
desired time (typically 40 min) to yield in most cases highly viscous
polymer suspensions. The polymerization was quenched by adding HCl/
EtOH (10 mL of a 1 M solution), followed by EtOH (80 mL). The
polymer was separated by filtration, washed repeatedly with acetone,
1 M HCl/H2O, boiling water, and acetone, and then dried for at least
12 h under vacuum. Method B: The desired amounts of initiator and
toluene were placed in a 50 mL reaction tube equipped with a Teflon
stopcock. The solution was cooled to -196 °C and the desired amount
of acrylonitrile was added by vacuum transfer. The reaction tube was
sealed and placed into a bath at the desired temperature. The mixture
was stirred for the desired amount of time, and the polymer was isolated
by the procedure described in method A.
Cy3PCuMe (1).9a A flask was charged with CuI (0.95 g, 5.0 mmol)
and PCy3 (1.5 g, 5.5 mmol), and Et2O (20 mL) was added by cannula.
The suspension was cooled to -35 °C and stirred for 15 min. A solution
of MeLi in Et2O (5.2 mL, 1.0 M, 5.2 mmol) was added by syringe and
the suspension was stirred for 45 min at -35 °C. The solvent was
evaporated while the temperature was kept below -10 °C to yield an
off-white solid. The solid was dried for 1 h at room temperature under
vacuum. The solid was extracted with toluene/pentane (1/1, 3 × 10
mL). The extracts were cooled to -40 °C for 1 week to yield colorless
crystals (820 mg, 46%). In solution, 1 is in equilibrium with 3 below
1
260 K. H NMR (190 K, CD2Cl2): δ 1.65-1.85 (m, 1 + 3), 1.55-
1.65 (m, 1), 1.35-1.45 (m, 3), 1.1-1.35 (m, 1 + 3), -0.73 (s, 3H, 3,
Me), -1.11 (s, 3H, 1, Me). 13C{1H} NMR (190 K, CD2Cl2): δ 31.2 (t,
J ) 7 Hz, 3, C1), 30.2 (d, J ) 3 Hz, 1, C3), 30.0 (d, J ) 19 Hz, 1,
C1), 29.6 (s, 3, C3), 27.1 (t, J ) 5 Hz, 3, C2), 26.6 (d, J ) 11 Hz, 1,
C2), 25.5 (s, 3, C4), 25.5 (s, 1, C4), -5.9 (br s, 3, Me), -8.6 (br s, 1,
Me). 31P{1H} NMR (190 K, CD2Cl2): δ 17.4 (s, 1), 11.8 (s, 3). NMR
data for other solvents and temperatures are given in the Supporting
Information.
[Cy3PCH2CH2CN]Cl.54 A flask was charged with PCy3 (100 mg,
0.356 mmol), [Me3NH]Cl (35.5 mg, 0.356 mmol), and MeCN (1 mL).
AN (0.7 mmol) was added by vacuum transfer at -196 °C. The
resulting slurry was heated to reflux for 10 min to give a clear solution.
The solution was cooled to room temperature and layered with ethyl
acetate (2 mL). After 12 h a crystalline precipitate was obtained. The
precipitate was isolated by filtration, washed with ethyl acetate, and
dried for 1 h under vacuum to yield a colorless powder (36 mg, 27%).
1H NMR (330 K, DMSO-d6): δ 3.00-3.05 (m, JPH ) 9 Hz, -CH2CN,
2H), 2.77-2.81 (m, JPH ) 13 Hz, -CH2P+Cy3 2H), 2.59-2.63 (q, 2JPH
{Cy3PCu(µ-I)}2 (4).25 A suspension of CuI (20 mg, 0.11 mmol)
and PCy3 (29 mg, 0.10 mmol) in CD2Cl2 was heated in a sealed NMR
1
3
) JHH ) 11 Hz, P+CH(CH2-)2, 3H), 1.90-1.92 (m, 6H), 1.80-1.83
tube to reflux for 10 min to yield a clear solution. H NMR (290 K,
CD2Cl2): δ 1.95 (m, 9H), 1.79 (m, 6H), 1.71 (m, 3H), 1.49 (m, 6H),
1.26 (m, 9H). 13C{1H} NMR (290 K, CD2Cl2): δ 32.4 (d, J ) 16 Hz,
C1), 30.5 (s, C3), 27.8 (d, J ) 11 Hz, C2), 26.6 (s, C4).
[Cu(PCy3)2][PF6] (5). PCy3 (250 mg, 0.89 mmol) and [Cu(NCMe)4]-
[PF6] (150 mg, 0.38 mmol) were dissolved in CH2Cl2 (5 mL) and the
clear solution was stirred for 2 h at room temperature. The solvent
was removed under vacuum, Et2O (10 mL) was added, and the resulting
slurry was stirred for 30 min at room temperature. The precipitate was
collected by filtration, washed with Et2O (2 × 10 mL), and dried under
vacuum to yield 250 mg (0.32 mmol, 81%) of a colorless powder. 1H
(m, 6H), 1.55-1.71 (m, 3H), 1.48-1.58 (m, 6H), 1.29-1.43 (m, 9H).
13C{1H} NMR (330 K, DMSO-d6): δ 118.5 (CN), 28.8 (d, J ) 40 Hz,
C1), 25.7 (d, J ) 4 Hz, C3), 25.6 (d, J ) 12 Hz, C2), 24.4 (d, J ) 2
Hz, C4), 10.8 (d, J ) 44 Hz, PCH2CH2CN), 10.4 (d, J ) 3 Hz,
PCH2CH2CN). 31P{1H} NMR (330 K, DMSO-d6): δ 34.6. Assignments
and coupling constants are derived from selective 1H and 31P decoupling
experiments. ESI-MS (MeOH) [Cy3PCH2CH2CN]+: Calcd m/z 334.3,
found 334.2. Anal. Calcd for C21H37PNCl‚(H2O)0.5: C, 66.56; H, 10.11;
N, 3.70. Found: C, 66.53; H, 10.09; N, 3.77. (1H NMR analysis of the
(54) Adapted from
a synthesis of [Ph3PCH2CH2CN]Cl: Teichmann, H.;
(53) Van Geet, A. L. Anal. Chem. 1970, 42, 679.
Thierfelder, W.; Kochmann, W. DD 100960, 1972.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2123

A R T I C L E S
Schaper et al.
analyzed [Cy3PCH2CH2CN]Cl sample in dry CD3CN showed presence
of 0.5 equiv of H2O due to handling of the compound in air.)
[Cy3PCH2CH2CN]Br. A flask was charged with PCy3 (200 mg,
0.71 mmol), BrCH2CH2CN (0.09 mL, 1 mmol), and benzene (5 mL).
After 7 days a colorless crystalline precipitate was isolated by filtration
and washed with benzene (5 × 2 mL) to give 45 mg (15% yield) of a
colorless powder. 1H NMR (290 K, CD3CN): δ 2.90-2.95 (m, -CH2-
Crystal Structure Determination of 1. Colorless crystals of 1 were
obtained by cooling a toluene/pentane solution to -40 °C for 1 week.
A colorless rhomb-shaped fragment (0.4 × 0.3 × 0.3 mm) was mounted
on a Bruker SMART APEX system at 100 K. A “full sphere” data set
was obtained by use of Mo KR radiation (λ ) 0.710 73 Å), which
samples approximately all reciprocal space to a resolution of 0.75 Å
(θ range ) 1.9-27.0°) by using 0.3° steps in ω and 30-s integration
times for each frame. Integration of intensities and refinement of cell
parameters were done with SAINT,55 yielding 4111 independent
reflections (Rint ) 1.8%). Absorption corrections were applied by
SADABS based on redundant diffractions with maximum and minimum
transmissions of 1.0 and 0.79, respectively.55 The space group was
determined as P1(bar) and final cell parameters were obtained from
the full data set: a ) 9.163(2) Å, b ) 9.955(2) Å, c ) 10.825(2) Å,
R ) 88.81(3)°, â ) 81.14(3)°, γ ) 76.49(3)°, V ) 948.4(3) Å3, Z )
2, dcalc ) 1.257 g/cm3, µ ) 1.230 mm-1, and F000 ) 388. The structure
was solved by direct methods.55 All non-hydrogen atoms were refined
anisotropically, and hydrogen atom positions were calculated and
refined by riding model techniques.55 Final agreement factors: R1 )
2.50%, wR2 ) 6.72% for observed [I > 2σ(I)] and GOF ) 1.070, R1
) 2.54%, wR2 ) 6.75% for all reflections. The maximum residual
2
3
CN, 2H), 2.63-2.67 (m, -CH2P+Cy3 2H), 2.56 (qt, JPH ) JHH ) 13
3
Hz, JHH ) 3 Hz, P+CH(CH2-)2, 3H), 1.87-1.95 (m), 1.84-1.87 (m,
6H), 1.71-1.76 (m, 3H), 1.50-1.57 (m, 6H), 1.33-1.47 (m, 9H). 13C-
{1H} NMR (290 K, CD3CN): δ 30.2 (d, J ) 40 Hz, C1), 27.2 (d, J )
4 Hz, C3), 26.9 (d, J ) 12 Hz, C2), 25.7 (s, C4), 12.8 (d, J ) 45 Hz,
PCH2CH2CN), 12.4 (br s, PCH2CH2CN). The -CN resonance was not
observed, probably due to overlap with the solvent signal. 31P{1H} NMR
(290 K, CD3CN): δ 34.5. ESI-MS (CH3CN) [Cy3PCH2CH2CN]+:
Calcd m/z 334.3, found 334.2. Anal. Calcd for C21H37PNBr: C, 60.87;
H, 9.00; N, 3.38. Found: C, 60.86; H, 8.85; N, 3.35.
Acrylonitrile Polymerizations with (bipy)2FeEt2 (2). Details are
provided in Table 4. A 100 mL Fisher-Porter bottle was charged with
2 (71 mg, 0.16 mmol). AN (2.0 g, 38 mmol) was added by vacuum
transfer at -196 °C. The bottle was warmed to room temperature to
afford a deep blue solution. After 15 min, gas evolution followed by
rapid polymerization of AN was observed. After the desired reaction
time, the bottle was vented and the resulting slurry was added to
acidified MeOH (100 mL of a 1 M HCl solution). The solid precipitate
was separated by filtration, washed with 1 M HCl, hot water, and finally
acetone, and then dried under vacuum, yielding 1.9 g (95% conversion)
of pale red PAN.
Analysis of Gases from the Reaction of 2 and AN. An NMR tube
was charged with 2 (15 mg, 34 µmol). AN (0.21 g, 4.0 mmol) was
added by vacuum transfer at -196 °C. The tube was warmed to room
temperature to yield a deep blue solution. After 10 min, gas evolution
followed by rapid polymerization of AN was observed. After 20 h, the
volatiles were vacuum-transferred to another NMR tube that was cooled
to -196 °C and then analyzed by 1H NMR in CD3CN, which showed
that ethylene (20 µmol, 0.58 equiv, determined vs ferrocene as internal
standard) and ethane were present in a ratio of 10/1.
electron density was 0.39 e Å-3
.
Acknowledgment. We thank Ian Steele for determination of
the crystal structure, Antoni Jurkiewicz for technical support
for NMR spectroscopy, and Bayer Polymers for GPC measure-
ments. Funding by the Department of Energy (DE-FG02-
00ER15036), Bayer Polymers, and Deutsche Forschungsge-
meinschaft is gratefully acknowledged.
Note Added after ASAP: In the version published on the
Web 1/23/2004, the structure in Scheme 3 and the TOC graphic
was incorrect. The version published 1/27/2004 and the print
version are correct.
Supporting Information Available: Additional NMR data,
details of the crystal structure of 1, details of polymer molecular
weight determinations, and further experimental details. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Analysis of Gases from the Decomposition of 2 in CD3CN. An
NMR tube was charged with 2 (15 mg, 34 µmol), and CD3CN (0.6
mL) was added by vacuum transfer at -196 °C. The tube was warmed
to room temperature to afford a deep blue solution. After 20 h, the
volatiles were vacuum-transferred to another NMR tube and analyzed
by 1H NMR, which showed that ethylene (25 µmol, 0.72 equiv,
determined vs ferrocene as internal standard) and ethane were present
in a ratio of 1/1.
JA039018A
(55) All software and sources of scattering factors are contained in the SHELXTL
(version 5.1) program library (G. Sheldrick, Bruker Analytical X-ray
Systems, Madison, WI).
9
2124 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004