Intramolecular hydrophosphination/cyclization of phosphinoalkenes and phosphinoalkynes catalyzed by organolanthanides: Scope, selectivity, and mechanism

Article abstract of DOI:10.1021/ja010811i

Organolanthanide complexes of the general type Cp′₂LnE(TMS)₂ (Cp′ = η^/5-Me₅C₅; Ln = La, Sm, Y, Lu; E = CH, N; TMS = SiMe₃) serve as effective precatalysts for the rapid intramolecular hydro

Full text of DOI:10.1021/ja010811i

J. Am. Chem. Soc. 2001, 123, 10221-10238
10221
Intramolecular Hydrophosphination/Cyclization of Phosphinoalkenes
and Phosphinoalkynes Catalyzed by Organolanthanides: Scope,
Selectivity, and Mechanism
Michael R. Douglass, Charlotte L. Stern, and Tobin J. Marks*
Contribution from the Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60208-3113
ReceiVed March 28, 2001
Abstract: Organolanthanide complexes of the general type Cp′2LnE(TMS)2 (Cp′ ) η5-Me5C5; Ln ) La, Sm,
Y, Lu; E ) CH, N; TMS ) SiMe3) serve as effective precatalysts for the rapid intramolecular hydrophos-
phination/cyclization of the phosphinoalkenes and phosphinoalkynes RHP(CH2)nCHdCH2 (R ) Ph, H; n )
3, 4) and H2P(CH2)nCtC-Ph (n ) 3, 4) to afford the corresponding heterocycles CH3CH(CH2)nPR and
Ph(H)CdC(CH2)nPH, respectively. Kinetic and mechanistic data for these processes exhibit parallels to, as
well as distinct differences from, organolanthanide-mediated intramolecular hydroamination/cyclizations. The
turnover-limiting step of the present catalytic cycle is insertion of the carbon-carbon unsaturation into the
Ln-P bond, followed by rapid protonolysis of the resulting Ln-C linkage. The rate law is first-order in [catalyst]
and zero-order in [substrate] over approximately one half-life, with inhibition by heterocyclic product intruding
at higher conversions. The catalyst resting state is likely a lanthanocene phosphine-phosphido complex, and
dimeric [Cp′2YP(H)Ph]2 was isolated and cystallographically characterized. Lanthanide identity and ancillary
ligand structure effects on rate and selectivity vary with substrate unsaturation: larger metal ions and more
open ligand systems lead to higher turnover frequencies for phosphinoalkynes, and intermediate-sized metal
ions with Cp′2 ligands lead to maximum turnover frequencies for phosphinoalkenes. Diastereoselectivity patterns
also vary with substrate, lanthanide ion, and ancillary ligands. Similarities and differences in hydrophosphination
vis-a`-vis analogous organolanthanide-mediated hydroamination are enumerated.
Introduction
potentially useful transformations. Organolanthanides have been
applied successfully to the chemo-, regio-, and enantioselective
catalytic hydrogenation of olefins, alkynes, and imines,6 hy-
drosilylation of olefins and alkynes,7 hydroboration of olefins,8
various types of polymerizations and copolymerizations,5,9 and
in particular to the hydroamination/cyclization/bicyclization of
aminoalkenes,10,11 aminoalkynes,12 and aminoallenes.13 The
The use of organolanthanide catalysts to effect a wide variety
of synthetically useful catalytic reactions is rapidly growing in
scope and diversity.1 Variation of the metal ionic radius,2 as
well as the wide variety of available nondisassociable ancillary
ligands compatible with lanthanide metal centers,3-5 renders this
area of catalysis uniquely applicable to a host of new and
(6) Hydrogenation: (a) Obora, Y.; Ohta, T.; Stern, C. L.; Marks, T. J. J.
Am. Chem. Soc. 1997, 119, 3745-3755. (b) Roesky, P. W.; Denninger,
U.; Stern, C. L.; Marks, T. J. Organometallics 1997, 16, 4486-4492. (c)
Haar, C. M.; Stern, C. L.; Marks, T. J. Organometallics 1996, 15, 1765-
1784. (d) Reference 5a.
(7) Hydrosilylation: (a) Molander, G. A.; Corrette, C. P. Organometallics
1998, 17, 5504-5512. (b) Schumann, H.; Keitsch, M. R.; Winterfeld, J.;
Muhle, S.; Molander, G. A. J. Organomet. Chem. 1998, 559, 181-190. (c)
Fu, P.-F.; Brard, L.; Li, Y.; Marks, T. J. J. Am. Chem. Soc. 1995, 117,
7157-7168. (d) Molander, G. A.; Julius, M. J. Org. Chem. 1992, 57, 6347-
6351. (e) Sakakura, T.; Lautenschlager, H.; Tanaka, M. J. Chem. Soc., Chem.
Commun. 1991, 40-41.
(8) Hydroboration: (a) Molander, G. A.; Pfeiffer, D. Org. Lett. 2001, 3,
361-363. (b) Bijpost, E. A.; Duchateau, R.; Teuben, J. H. J. Mol. Catal.
1995, 95, 121-128. (c) Harrison, K. N.; Marks, T. J. J. Am. Chem. Soc.
1992, 114, 9220-9221.
(9) (a) Koo, K.; Fu, P.-F.; Marks, T. J. Macromolecules 1999, 32, 981-
988. (b) Jia, L.; Yang, X.; Seyam, A. M.; Albert, I. D. L.; Fu, P.-F.; Yang,
S.; Marks, T. J. J. Am. Chem. Soc. 1996, 118, 7900-7913. (c) Watson, P.
L. J. Am. Chem. Soc. 1982, 104, 337-339.
(10) (a) Ryu, J.-S.; McDonald, F. E.; Marks, T. J. Org. Lett. 2001, in
press. (b) Molander, G. A.; Dowdy, E. D. J. Org. Chem. 1999, 64, 6515-
6517. (c) Molander, G. A.; Dowdy, E. D. J. Org. Chem. 1998, 63, 8983-
8988. (d) Kim, Y. K.; Livinghouse, T.; Bercaw, J. E. Tetrahedron Lett.
2001, 42, 2933-2935.
(11) (a) Gagne´, M. R.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc.
1992, 114, 275-294. (b) Gagne´, M. R.; Marks, T. J. J. Am. Chem. Soc.
1989, 111, 1, 4108-4109. (c) Reference 4b.
(1) For organolanthanide reviews, see: (a) Molander, G. A.
Chemtracts: Org. Chem. 1998, 18, 237-263. (b) Edelmann, F. T. Top.
Curr. Chem. 1996, 179, 247-276. (c) Edelmann, F. T. In ComprehensiVe
Organometallic Chemistry; Wilkinson, G., Stone, F. G. A., Abel, E. W.,
Eds.; Pergamon Press: Oxford, U.K., 1995; Vol. 4, Chapter 2. (d)
Schumann, H.; Meese-Marktscheffel, J. A.; Esser, L. Chem. ReV. 1995,
95, 865-986. (e) Schaverien, C. J. AdV. Organomet. Chem. 1994, 36, 283-
362. (f) Evans, W. J. AdV. Organomet. Chem. 1985, 24, 131-177. (g)
Marks, T. J.; Ernst, R. D. In ComprehensiVe Organometallic Chemistry;
Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; Pergamon Press: Oxford,
U.K., 1982, Chapter 21.
(2) Shannon, R. D. Acta Crystallogr. 1976, A32, 751-760.
(3) a) CGC complexes: (a) Tian, S.; Arredondo, V. A.; Stern, C. L.;
Marks, T. J. Organometallics 1999, 18, 2568-2570. (b) Shapiro, P. J.;
Cotter, M. D.; Schaefer, W. P.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem.
Soc. 1994, 116, 4623-4640.
(4) Chiral auxiliaries: (a) Giardello, M. A.; Conticello, V. P.; Brard, L.;
Sabat, M.; Rheingold, A. L.; Stern, C.; Marks, T. J. J. Am. Chem. Soc.
1994, 116, 10012-10040. (b) Giardello, M. A.; Conticello, V. P.; Brard,
L.; Gagne´, M. R.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10241-
10254. (c) Douglass, M. R.; Ogasawara, M.; Metz, M. V.; Marks, T. J.
Organometallics, in press.
(5) (a) Bis-Cp′ alkyl and hydride complexes: Jeske, G.; Lauke, H.;
Mauermann, H.; Swepston, P. N.; Schumann, H.; Marks, T. J. J. Am. Chem.
Soc. 1985, 107, 8091-8103. (b) Si-bridged and mixed Cp complexes: Jeske,
G.; Schock, L. E.; Mauermann, H.; Swepston, P. N.; Schumann, H.; Marks,
T. J. J. Am. Chem. Soc. 1985, 107, 8103-8110.
10.1021/ja010811i CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/26/2001

10222 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
success of the various types of hydro-elementation processes
catalyzed by organolanthanides raises the intriguing question
of whether analogous addition/cyclization processes are possible
for phosphorus-containing substrates. Olefin hydrophosphination
has proven generally difficult to accomplish with typical late
transition metal catalysts.14 However, the aforementioned ste-
reoelectronic tunability of organolanthanide coordination spheres,
coupled with high electrophilicity, absence of conventional
oxidative-addition/reductive-elimination pathways, and high
kinetic lability, offers an attractive alternative pathway for
accomplishing this transformation. Cyclic phosphines are as yet
understudied,15 and may have applications as ligands similar to
all-carbon ligands16 or display interesting pharmacological
properties.17 In addition, phosphorus heterocycles produced in
hydrophosphination/cyclization processes offer attractive build-
ing blocks for classes of electron-rich chiral ligands (e.g., A,
B) that have been utilized extensively for a variety of asym-
metric transformations.18 Thus, organolanthanide-catalyzed hy-
drophosphination could afford new synthetic routes to both
existing and new classes of ligands.
Scheme 1. Simplified Catalytic Cycle for
Organolanthanide-Mediated Hydroamination/Cyclization of
Aminoalkenes and Aminoalkynes
Scheme 2. Proposed Catalytic Cycle for
Organolanthanide-Mediated Hydrophosphination/Cyclization
of Phosphinoalkenes and Phosphinoalkynes
The catalytic cycle for organolanthanide-mediated hydroami-
nation/cyclization of amine substrates is mechanistically and
thermochemically reasonably well-defined (Scheme 1).11-13
After immediate and quantitative protonolysis of the hydrocarbyl
or amido precatalyst (Scheme 1, step i), the first step is turnover-
limiting insertion of the carbon-carbon unsaturation into the
lanthanide-nitrogen bond (Scheme 1, step ii). The second step
is facile intermolecular or intramolecular protonolysis of the
lanthanide-carbon bond by substrate, releasing the nitrogen-
containing heterocyclic product and regenerating a lanthanide-
(12) (a) Li, Y.; Marks, T. J. J. Am. Chem. Soc. 1998, 120, 1757-1771.
(b) Li, Y.; Marks, T. J. J. Am. Chem. Soc. 1996, 118, 9295-9306. (c) Li,
Y.; Marks, T. J. Organometallics 1996, 15, 3370-3372. (d) Li, Y.; Marks,
T. J. J. Am. Chem. Soc. 1996, 118, 707-708. (e) Li, Y.; Fu, P.-F.; Marks,
T. J. Organometallics 1994, 13, 439-440.
(13) (a) Arredondo, V. A.; McDonald, F. M.; Marks, T. J. Organome-
tallics 1999, 10, 1949-1960. (b) Arredondo, V. A.; Tian, S.; McDonald,
F. M.; Marks, T. J. J. Am. Chem. Soc. 1999, 121, 3633-3639. (c)
Arredondo, V. A.; McDonald, F. M.; Marks, T. J. J. Am. Chem. Soc. 1998,
120, 4871-4872.
(14) Hydrophosphination mediated by Pd, Pt, Ni complexes: (a) Kovacik,
I.; Wicht, D. K.; Grewal, N. S.; Glueck, D. S.; Incarvito, C. D.; Guzei, I.
A.; Rheingold, A. L. Organometallics 2000, 19, 950-953. (b) Wicht, D.
K.; Kourkine, I. V.; Kovacik, I.; Glueck, D. S.; Concolina, T. E.; Yap, G.
P. A.; Incarvito, C. D.; Rheingold, A. L. Organometallics 1999, 18, 5381-
5394. (c) Costa, E.; Pringle, P. G.; Worboys, K. Chem. Commun. 1998,
49-50. (d) Wicht, D. K.; Kourkine, I. V.; Lew, B. M.; Nthenge, J. M.;
Glueck, D. S. J. Am. Chem. Soc. 1997, 119, 5039-5040. (e) Han, L.-B.;
Tanaka, M. J. Am. Chem. Soc. 1996, 118, 1571-1572. (f) Using no
transition metal catalyst: Gaumont, A.-C.; Simon, A.; Denis, J.-M.
Tetrahedron Lett. 1998, 39, 985-988.
(15) a) Quin, L. D.; Hughes, A. N. In The Chemistry of Organophos-
phorus Compounds; Hartley, F. R., Ed.; Wiley: Chichester, 1990; Vol. 1,
pp 295-394. (b) Quin, L. D. The Heterocyclic Chemistry of Phosphorus:
Systems Based on the Phosphorus-Carbon Bond; Wiley: New York, 1981.
(16) Dillon, K. B.; Mathey, F.; Nixon, J. F. Phosphorus: The Carbon
Copy; Wiley: Chichester, 1998.
(17) Price, N. R.; Chambers, J. In The Chemistry of Organophosphorus
Compounds; Hartley, F. R., Ed.; Wiley: Chichester, 1990; Vol. 1, pp 643-
661.
(18) (a) Burk, M. J. Acc. Chem. Res. 2000, 33, 363-372. (b) Burk, M.
J.; Gross, M. F.; Martinez, J. P. J. Am. Chem. Soc. 1995, 117, 9375-9376.
(c) Burk, M. J.; Harper, G. P.; Kalberg, C. S. J. Am. Chem. Soc. 1995, 117,
4423-4424. (d) Burk, M. J.; Feaster, J. E.; Nugent, W. A.; Harlow, R. L.
J. Am. Chem. Soc. 1993, 115, 10125-10138.
nitrogen bond (Scheme 1, step iii). Thermodynamic estimates
support the viability of this cycle (∆H ≈ -13 kcal/mol for
alkenes, -35 kcal/mol for alkynes), and the broad scope of the
process, including construction of alkaloid natural products,13b
has been demonstrated. Analogous thermodynamic analysis19
for phosphines indicates that a similar catalytic cycle should
be possible, in that after initial protonlysis of the precatalyst
(Scheme 2, step i), insertion of a carbon-carbon unsaturation
into a lanthanide-phosphorus bond (Scheme 2, step ii) should
be exothermic for alkynes (∆H ≈ -33 kcal/mol) or ap-
proximately thermoneutral for alkenes (∆H ≈ +2 kcal/mol).
When coupled with subsequent protonolysis (Scheme 2, step
(19) Bond enthalpy data: (a) Nolan, S. P.; Stern, D.; Hedden, D.; Marks,
T. J. ACS Symp. Ser. 1990, 428, 159-174. (b) Nolan, S. P.; Stern, D.;
Marks, T. J. J. Am. Chem. Soc. 1989, 111, 7844-7853. (c) Skinner, H. A.;
Pilcher, G. In The Chemistry of the Metal-Carbon Bond; Hartley, F. R.,
Patai, S., Eds.; John Wiley & Sons: Chichester, 1982; Vol. 1, Chapter 2.
(d) Goldwhite, H. Introduction to Phosphorus Chemistry; Cambridge
University Press: Cambridge, 1981.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10223
in Teflon valve-sealed tubes (J. Young). HRMS studies were conducted
on a VG 70-250 SE instrument with 70 eV electron impact ionization
for substrates containing a phenyl group, and atmospheric pressure
chemical ionization for primary phosphines and the products of their
cyclizations, which were too volatile to be analyzed using EI. Elemental
analyses were carried out by Midwest Microlabs, Indianapolis, IN.
Synthesis of Pent-4-enylphosphine (1). Diethyl-4-pentenylphos-
phonate (1a). The reagent 5-bromo-1-pentene (9.4 mL, 79.5 mmol)
was subjected to reaction with P(OEt)3 (15.0 mL, 87.5 mmol) at 150
°C overnight, and the resulting phosphonate was purified by distillation
iii, ∆H ≈ -7 kcal/mol for alkynes, -17 kcal/mol for alkenes),
the data indicate a thermodynamically feasible process.20 Indeed,
a preliminary investigation indicated that organolanthanide-
mediated hydrophosphination is a viable process.21 Herein we
report a full discussion of reaction scope, metal and ancillary
ligand effects, regio- and diastereoselectivity, and kinetics and
mechanism for several series of phosphinoalkenes and phos-
phinoalkynes. It will be seen that organolanthanide-catalyzed
hydrophosphination/cyclization has a broad scope, that variation
of the metal and ligands has significant and understandable
influence on the catalytic process, that the chemo-, regio-, and
stereoselectivity of the process can be controlled to a significant
degree, and that the kinetics and mechanism are in some ways
similar to, and in some ways different significantly from, those
established for hydroamination.11-13
1
(55 °C/0.08 Torr) to yield a colorless oil (11.5 g, 70.5% yield). H
NMR (300 MHz, C6D6): δ 5.57 (m, 1H), 4.93 (m, 2H), 3.90 (m, 4H),
1.90 (q, 6.6 Hz, 2H), 1.61 (m, 4H), 1.04 (t, 7.2 Hz, 6H); 13C NMR (75
MHz, C6D6): δ 138.2, 115.9, 61.4 (d, 6.1 Hz), 35.0, 34.8, 26.9, 25.0,
22.6 (d, 4.6), 17.0 (d, 5.3); 31P NMR (121 MHz, C6D6): δ 32.3. HRMS
(m/z): [M+] calcd for C9H19O3P, 206.1072; found, 206.10715. Anal.
Calcd for C9H19O3P: C, 52.40; H, 9.29. Found: C, 52.24; H, 9.20.
Pent-4-enylphosphine (1b). The reducing agent AlCl2H was gener-
ated in situ from AlCl3 (6.48 g, 48.5 mmol) and LiAlH4 (0.74 g, 19.5
mmol) in 50 mL of tetraglyme (vacuum-distilled from Na twice).23,24
The tetraglyme was added to the LiAlH4 powder, and then the AlCl3
solid was added under an N2 flush at 0 °C in several portions. (Addition
of tetraglyme to AlCl3 results in an intractable black solid; therefore,
the order of reagent addition is critical in forming this reducing agent
in tetraglyme.) Next, phosphonate 1a (2.0 g, 9.7 mmol) was added to
the AlCl2H solution via syringe (an exotherm was observed), and the
reaction mixture was stirred until the reduction was complete by 31P
NMR (ca. 1 h). Deoxygenated water was then added slowly via cannula
at 0 °C until a clear upper layer separated from a lower gray aqueous
layer. The upper layer was separated via cannula and degassed, and
phosphine 1b was vacuum-transferred away from the tetraglyme
solution into a storage tube (cooled to -78 °C) containing activated 4
Å molecular sieves and C6D6. The final solution concentration was
∼1 M by NMR, indicating an approximately 50% yield. 1H NMR (300
MHz, C6D6): δ 5.6 (m, 1 H), 4.95-4.91 (m, 2H), 2.47 (d of t, 190.5
Hz, 2H), 1.84 (q, 7.0 Hz, 2H), 1.32 (qu, 7.5 Hz, 2H), 1.13 (m, 2H);
13C NMR (75 MHz, C6D6): δ 138.6, 115.5, 35.2 (d, 5.2 Hz), 32.9 (d,
3 Hz), 13.7 (d, 8.3 Hz); 31P NMR (121 MHz, C6D6): δ -137.8 (t,
1JP-H ) 190 Hz). HRMS (m/z): [MH+] calcd for C5H12P, 103.0625;
found, 103.0607.
Synthesis of 1-Methylpent-4-enylphosphine (3). Diethyl-1-meth-
ylpent-4-enylphosphonate (3a). The reagent diethyl ethylphosphonate
(8.66 g, 52.2 mmol) was treated with n-BuLi (1.6 M, 62.6 mmol, 39.1
mL) in 50 mL of THF at -45 °C for 30 min. Next, 4-bromo-1-butene
(5.56 mL, 54.8 mmol) was added dropwise, and the solution was
allowed to warm to room temperature and stirred for an additional 2.5
h, turning yellow and then orange during this time. The reaction was
then quenched with water, dried with brine and MgSO4 and filtered,
and the solvent was removed in vacuo. Distillation of the resulting
phosphonate (145 °C/12 Torr) afforded a colorless oil in approximately
60% yield. 1H NMR (500 MHz, C6D6): δ 5.67 (m, 1H), 4.97 (m, 2H),
3.93 (m, 4H), 2.12 (m, 1H), 1.97 (m, 2H), 1.75 (m, 1H), 1.46 (m, 1H),
1.12 (d of d, 18 Hz, 7.0 Hz, 3H), 1.04 (t, 7.0 Hz, 6H); 13C NMR (125
MHz, C6D6): δ 138.6, 115.6, 61.5 (t, 6.8 Hz), 32.0 (d, 12.5 Hz), 31.6,
30.4, 30.2 (d, 3.5 Hz), 17.0 (d, 5.0 Hz), 13.8 (d, 5.0 Hz); 31P NMR
(162 MHz, C6D6): δ 35.2. HRMS (m/z): [M+] calcd for C10H21O3P,
220.1228; found, 220.1227. Anal. Calcd for C10H21O3P: C, 54.53; H,
9.61. Found: C, 54.56; H, 9.80.
Experimental Section
Materials and Methods. All manipulations of air-sensitive materials
were carried out with rigorous exclusion of oxygen and moisture in
flame- or oven-dried Schlenk-type glassware on a dual-manifold
Schlenk line, or interfaced to a high-vacuum line (10-6 Torr), or in a
nitrogen-filled Vacuum Atmospheres glovebox with a high-capacity
recirculator (<1 ppm of O2). Argon (Matheson, prepurified) was
purified by passage through a MnO oxygen-removal column and a
Davison 4 Å molecular sieve column. The solvents, ether and THF,
were distilled before use from appropriate drying agents (sodium
benzophenone ketyl, metal hydrides, Na/K alloy) under nitrogen. The
solvents, pentane and toluene, were dried using activated alumina
columns according to the method described by Grubbs22 and were
additionally vacuum-transferred from Na/K alloy immediately before
use if employed for catalyst synthesis or catalytic reactions. Deuterated
solvents (Cambridge Isotope Laboratories; all 99+ atom % D) used
for NMR reactions and kinetic measurements (benzene-d6 and toluene-
d8) were stored in vacuo over Na/K alloy in resealable bulbs and were
vacuum-transferred immediately prior to use. All organic starting
materials were purchased from Aldrich Chemical Co., Farchan Labo-
ratories Inc., or Lancaster Synthesis Inc. and were used without further
purification unless otherwise stated. All substrates were dried a
minimum of two times as solutions in benzene-d6 or toluene-d8 over
freshly activated Davison 4 Å molecular sieves and were degassed by
freeze-pump-thaw methods. They were then stored in vacuum-tight
storage flasks. Solutions of alkenyl phosphines were protected from
light to minimize endocyclization reactions. Concentrations of substrate
solutions were determined by NMR integration after addition of a
known volume of solution to a precatalyst. The organolanthanide
precatalysts Cp′2LnCH(TMS)2 (Ln ) La, Sm, Y, Lu; Cp′ ) η5-Me5C5),5
Me2Si(Me4C5)(tBuN)SmN(TMS)2,3a and Me2Si(Cp′′)(CpR*)LnCH-
(TMS)2 (Cp′′ ) C5Me4, R* ) (-)-menthyl, (+)-neomenthyl)4a were
prepared by published procedures.
Physical and Analytical Measurements. NMR spectra were
1
recorded on either a Varian Gemini-300 (FT, 300 MHz, H; 75 MHz,
1
13C; 121 MHz, 31P), Unity- or Mercury-400 (FT, 400 MHz, H; 100
MHz, 13C; 162 MHz, 31P), or Inova-500 (FT, 500 MHz, 1H; 125 MHz,
13C; 202 MHz, 31P) instrument. Chemical shifts (δ) for 1H and 13C are
referenced to internal solvent resonances and reported relative to SiMe4.
Chemical shifts for 31P are reported relative to an external 85% H3PO4
standard. NMR experiments with air-sensitive samples were conducted
1-Methylpent-4-enylphosphine (3b). The phosphonate 3a was
reduced using AlCl2H in tetraglyme as described for 1b, and the
resulting phosphine 3b was vacuum-transferred away from the tetra-
glyme solution into a storage tube (cooled to -78 °C) containing
activated 4 Å molecular sieves and C6D6. The final solution concentra-
(20) AM-1 level calculations corroborate the thermodynamic estimates,
and indicate that addition of methyl phosphine to ethylene is exothermic
by ∼-15 kcal/mol, while addition to acetylene is exothermic by ∼-38
kcal/mol. We thank Dr. A. Israel for these calculations.
1
tion was ∼1 M by NMR, indicating an approximately 50% yield. H
(21) (a) Douglass, M. R.; Marks, T. J. J. Am. Chem. Soc. 2000, 122,
1824-1825. (b) Douglass, M. R.; Marks, T. J. Communicated in part at
the 217th National Meeting of the American Chemical Society, Anaheim,
CA, March 1999; Abstract INOR 375. (c) Preliminary observations:
Giardello, M. A.; King, W. A.; Nolan, S. P.; Porchia, M.; Sishta, C.; Marks,
T. J. In Energetics of Organometallic Species; Marthinho Simoes, J. A.,
Ed.; Kluwer: Dordrecht, 1992; pp 35-51.
NMR (400 MHz, C6D6): δ 5.63 (m, 1 H), 4.96 (m, 2H), 2.63 (d of m,
189.2 Hz, 2H), 1.94 (q, 7.0 Hz, 2H), 1.58 (sept, 6.8 Hz, 1H), 1.32 (m,
2H), 0.96, (d of d, 12.4 Hz, 7.2 Hz, 3H); 13C NMR (125 MHz, C6D6):
δ 138.7, 115.3, 39.1 (d, 10.4 Hz), 33.3 (d, 9.5 Hz), 23.7 (d, 10.0 Hz),
(23) Guillemin, J. C.; Savignac, P.; Denis, J.-M. Inorg. Chem. 1991, 30,
2170-2173.
(24) Ashby, E. C.; Prather, J. J. Am. Chem. Soc. 1966, 88, 729-733.
(22) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518-1520.

10224 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
1
22.6 (d, 7.0 Hz); 31P NMR (121 MHz, C6D6): δ -111.8 (t, JP-H
)
(m/z): [M+] calcd for C15H21O3P, 280.1228; found, 280.1205. Anal.
Calcd for C15H21O3P: C, 64.26; H, 7.56. Found: C, 64.18; H, 7.66.5-
phenyl-4-pentynylphosphine (9b). To a mixture of dichloroalane24
[AlCl3 (5.76 g, 40 mmol) and LiAlH4 (0.53 g, 14 mmol)] in 50 mL of
diethyl ether was added 9a (2.0 g, 7.1 mmol), and the gray solution
was stirred under nitrogen overnight. The reaction was then quenched
with deoxygenated water at 0 °C, and the upper organic layer was
separated via cannula, dried over MgSO4 and filtered, and solvent was
removed in vacuo to yield a light-yellow oil. Purification by distillation
(90 °C/0.07 Torr) afforded the primary phosphine 9b as a colorless oil
(0.81 g, 65% yield), which was stored in an airtight storage tube as a
188.5 Hz). HRMS (m/z): [MH+] calcd for C6H14P, 117.0828; found,
117.0807.
Synthesis of Hex-5-enylphosphine (5). Dimethyl-hex-5-enylphos-
phonate (5a). To a solution of n-BuLi (1.6 M, 80.5 mmol, 50.3 mL)
in THF (100 mL) at -45 °C was added dimethyl methylphosphonate
(80.5 mmol, 8.73 mL). The light-yellow solution turned colorless and
was stirred for an additional 45 min, during which time a white
precipitate was formed. Next, 5-bromo-1-pentene (67 mmol, 10.0 g)
was added, and after stirring at -45 °C for 30 min, the solution was
allowed to warm to room temperature. The precipitate dissolved as
the solution warmed, and the solution was stirred 1 additional h. Next,
the clear, colorless solution was quenched with water, washed with
brine, dried with MgSO4 and filtered, and the solvent was removed to
yield 5a (crude yield ∼70%) which was purified by distillation (55
1
solution in C6D6 over activated 4 A molecular sieves. H NMR (300
MHz, C6D6): δ 7.44 (m, 2H), 6.99 (m, 3H), 2.53 (d of t, 191 Hz, 7.5
Hz, 2H), 2.13 (t, 6.9 Hz, 2H), 1.46 (qu, 7.2 Hz, 2H), 1.26 (m, 2H); 13
C
NMR (75 MHz, C6D6): 132.3, 128.9, 128.2, 125.0, 90.1, 82.2, 32.7
(d, 3.0 Hz), 20.9 (d, 5.5 Hz), 13.7 (d, 9.2 Hz); 31P NMR (121 MHz,
1
°C/0.05 Torr) to afford a colorless oil. H NMR (400 MHz, C6D6): δ
1
C6D6): δ -138.2 (t, JP-H ) 190 Hz). HRMS (m/z): [M+] calcd for
5.63 (m, 1H), 4.93 (m, 2H), 3.38 (d, 6H, 11.2 Hz), 1.81 (q, 2H, 7.2
Hz), 1.50 (m, 4H), 1.19 (m, 2H); 13C NMR (100 MHz, C6D6): δ 138.8,
115.3, 52.2 (d, 6.1 Hz), 34.1, 30.5 (d, 15.9 Hz), 26.4, 25.0, 22.9 (d,
4.9); 31P NMR (162 MHz, C6D6): δ 34.9. HRMS (m/z): [M+] calcd
for C8H17O3P, 192.0915; found, 192.0913. Anal. Calcd for C8H17O3P:
C, 49.99; H, 8.92. Found: C, 50.34; H, 8.89.
Hex-5-enylphosphine (5b). The phosphonate 5a was reduced using
AlCl2H in tetraglyme as described for 1b, and product phosphine 5b
was vacuum-transferred away from the tetraglyme solution into a
storage tube (cooled to -196 °C) containing activated 4 Å molecular
sieves and C6D6. The final solution concentration was slightly less than
C11H13P, 176.0756; found, 176.0781. Anal. Calcd for C11H13P: C, 74.98;
H, 7.44. Found: C, 74.58; H, 7.24.
Synthesis of 6-Phenyl-5-hexynylphosphine (11). Diethyl-6-phenyl-
5-hexynylphosphonate (11a). The reagent 6-phenyl-5-hexynyl-1-
bromide was prepared as described previously.12b This bromide (10.63
g, 44.9 mmol) was reacted with neat P(OEt)3 (9.2 mL, 53.8 mmol) at
150 °C overnight, and the resulting phosphonate 11a purified by
distillation (160 °C/0.05 Torr) to afford 11a as a colorless oil (9.0 g,
1
68% yield). H NMR (300 MHz, C6D6): δ 7.49 (m, 2H), 6.99 (m,
3H). 3.93 (m, 4H), 2.13 (t, 7.0 Hz, 2H), 1.72 (m, 2H), 1.50 (m, 4H),
1.04 (t, 6.9 Hz, 6H); 13C NMR (75 MHz, C6D6): δ 132.2, 128.9, 128.1,
125.1, 90.3, 82.1, 61.4 (d, 6.5), 29.9 (d, 15.5), 27.1, 25.2, 22.6 (d, 5.0),
19.5, 16.9 (d, 5.6); 31P NMR (121 MHz, C6D6): δ 32.1. HRMS (m/z):
[M+] calcd for C16H23O3P, 294.1385; found, 294.1347.
6-Phenyl-5-hexynylphosphine (11b). The phosphonate 11a was
reduced using AlCl2H in diethyl ether as described for 9b. Purification
by bulb-to-bulb distillation (130 °C/0.05 Torr) gave the primary
phosphine 11b in 55% yield (0.71 g), which was stored in an airtight
storage tube as a solution in C6D6 over activated 4 A molecular sieves.
1H NMR (300 MHz, C6D6): δ 7.49 (m, 2H), 6.99 (m, 3H), 2.56 (d of
t, 190 Hz, 7.5 Hz, 2H), 2.12 (t, 6.6 Hz, 2H), 1.35 (m, 4H), 1.10 (m,
2H); 13C NMR (75 MHz, C6D6): 132.3, 128.9, 128.1, 125.1, 90.6, 82.0,
32.8 (d, 3.2 Hz), 30.2 (d, 5.5 Hz), 19.6, 13.8 (d, 8.6 Hz); 31P NMR
(121 MHz, C6D6): δ -137.3 (t, 1JP-H ) 190 Hz). HRMS (m/z): [M+]
calcd for C12H15P, 190.0912; found, 190.0917.
Synthesis of Pent-4-enylphenylphosphine (13). Pent-4-enyldi-
phenylphosphine (13a). A solution of 50 mL of KPPh2 (0.5 M in THF)
and 50 mL of THF was stirred under nitrogen. Next, 5-bromo-1-pentene
(3.0 mL, 35 mmol) was added dropwise, and the resulting viscous
orange solution was stirred for 90 min. The THF was then removed in
vacuo, and 20 mL of pentane was added to the orange residue. Next,
deoxygenated water was added via cannula, and the resulting clear,
colorless organic layer was decanted via cannula, dried over MgSO4
and filtered via cannula, and then the solvent was removed in vacuo to
yield a light-yellow oil. This material was distilled (80 °C/0.05 Torr)
to give a pure light-yellow oil (7.0 g) in 79% yield. NMR spectra agree
with the literature data for pent-4-enyldiphenylphosphine.25 1H NMR
(300 MHz, C6D6): δ 7.42 (m, 4H), 7.06 (m, 6H), 5.6 (m, 1 H), 4.97-
4.91 (m, 2H), 2.01 (q, 7.2 Hz, 2H), 1.91 (m, 2H), 1.51 (qu, 8.1 Hz,
2H); 13C NMR (75 MHz, C6D6): δ 138.6, 133.6, 133.3, 129.1, 129.0,
115.6, 35.7 (d, 13.4 Hz), 28.2 (d, 12.1 Hz), 26.0 (d, 17.0 Hz); 31P NMR
(121 MHz, C6D6): δ -15.6. HRMS (m/z): [M+] calcd for C17H18P,
253.1147; found, 253.1114. Anal. Calcd for C17H19P: C, 80.29; H, 7.53.
Found: C, 80.31; H, 7.48.
1
1 M by NMR, indicating an approximately 40% yield. H NMR (400
MHz, C6D6): δ 5.68 (m, 1 H), 4.95 (m, 2H), 2.58 (d of t, 190 Hz, 7.2
Hz, 2H), 1.84 (q, 7.2 Hz, 2H), 1.21 (m, 6H); 13C NMR (100 MHz,
C6D6): δ 139.1, 115.1, 34.3, 33.3, (d, 3.1 Hz), 30.7 (d, 5.3 Hz), 14.5
(d, 8.3 Hz); 31P NMR (162 MHz, C6D6): δ -137.2 (t, JP-H ) 189.5
1
Hz). HRMS (m/z): [MH+] calcd for C6H13P, 117.083; found, 117.083.
Synthesis of 1-Methylhex-5-enylphosphine (7). Diethyl-1-meth-
ylhex-5-enylphosphonate (7a). Phosphonate 7a was prepared in the
same manner as 3a, using diethyl ethylphosphonate (14.02 g, 84.4
mmol), n-BuLi (1.6 M, 84.4 mmol, 52.3 mL), and 5-bromo-1-pentene
(10.0 mL, 84.4 mmol). The resulting phosphonate was purified by
distillation (62 °C/0.05 Torr) to afford a colorless oil in approximately
65% yield. 1H NMR (400 MHz, C6D6): δ 5.57 (m, 1H), 4.93 (m, 2H),
3.90 (m, 4H), 1.90 (q, 6.6 Hz, 3H), 1.74 (m, 1H), 1.38 (m, 3H), 1.11
(d of d, 18.2 Hz, 7.0 Hz, 3H), 1.05 (t, 7.0 Hz, 6H); 13C NMR (100
MHz, C6D6): δ 139.0, 115.2, 61.4 (t, 6.0 Hz), 34.5, 32.5, 31.2, 30.7
(d, 3.4 Hz), 27.6 (d, 12.5 Hz), 17.3 (d, 5.3 Hz), 14.3 (d, 4.9 Hz); 31P
NMR (121 MHz, C6D6): δ 35.2. HRMS (m/z): [M+] calcd for
C11H23O3P, 234.1385; found, 234.1286. Anal. Calcd for C11H23O3P:
C, 56.39; H, 9.90. Found: C, 56.23; H, 9.92.
1-Methylhex-5-enylphosphine (7b). The phosphonate 7a was
reduced using AlCl2H in tetraglyme as described for 1b, and product
phosphine 7b was vacuum-transferred away from the tetraglyme
solution into a storage tube (cooled to -196 °C) containing activated
4 Å molecular sieves and C6D6. The final solution concentration was
slightly less than 1 M by NMR, indicating an approximately 40% yield.
1H NMR (400 MHz, C6D6): δ 5.7 (m, 1 H), 4.95-4.91 (m, 2H), 2.66
(d of t, 189 Hz, 7.5 Hz, 2H), 1.86 (q, 6.8 Hz, 2H), 1.56 (qu, 6.0 Hz,
1H), 1.26 (m, 4H), 0.98 (d of d, 12.4 Hz, 7.2 Hz, 3H); 13C NMR (75
MHz, C6D6): δ 138.6, 114.6, 38.9 (d, 8.3 Hz), 34.0, 27.8 (d, 7.5 Hz),
23.3 (d, 7.9 Hz), 22.5 (d, 5.7 Hz); 31P NMR (121 MHz, C6D6): δ
-111.5 (t, 1JP-H ) 188.6 Hz). HRMS (m/z): [MH+] calcd for C7H15P,
131.098; found, 131.096.
Synthesis of 5-phenyl-4-pentynylphosphine (9). Diethyl-5-phenyl-
4-pentynylphosphonate (9a). The reagent 5-phenyl-4-pentynyl-1-
bromide was prepared as described previously.12b This bromide (3.71
g, 16.6 mmol) was reacted with neat P(OEt)3 (3.13 mL, 18.3 mmol) at
150 °C overnight, and the resulting phosphonate was purified by
distillation (135 °C/0.05 Torr) to afford 9a as a colorless oil in
approximately 70% yield. 1H NMR (300 MHz, C6D6): δ 7.39 (m, 2H),
6.97 (m, 3H), 3.90 (m, 4H), 2.22 (t, 6.9 Hz, 2H), 1.80 (m, 4H), 1.02
(t, 6.9 Hz, 6H); 13C NMR (75 MHz, C6D6): δ 132.3, 128.9, 128.2,
89.7, 82.5, 61.7 (d, 6.7 Hz), 26.6, 24.7, 22.9 (d, 4.9 Hz), 20.7 (d, 17.6
Hz), 16.9 (d, 5.4 Hz); 31P NMR (121 MHz, C6D6): δ 31.7. HRMS
Pent-4-enylphenylphosphine (13b). To 40 mL of NH3(l) was added
0.27 g (11.8 mmol) of Na spheres, and the resulting blue solution was
stirred at -78 °C under nitrogen for 1.0 h. Next, 1.50 g (5.9 mmol)
13a was added, and the reaction mixture was stirred at -50 °C for ∼3
h until it became golden yellow. Solid NH4Cl (0.63 g, 54 mmol) was
next added under a nitrogen flush, and the colorless reaction mixture
was allowed to stir and warm to room temperature (with ammonia
(25) Sigl, M.; Schier, A.; Schmidbaur, H. Chem. Ber. 1997, 130, 951-
954.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10225
boiling off) overnight. Diethyl ether was then added, the resulting
solution decanted via cannula filtration, solvent removed in vacuo, and
the resulting oil purified by distillation (70 °C/0.05 Torr) to afford 0.75
g of a colorless oil (60% yield), which was stored in an airtight storage
tube as a solution in C6D6 over activated 4 A molecular sieves at low
and purified by distillation (120 °C/0.07 Torr) to afford 17a as a
1
colorless oil. H NMR (300 MHz, C6D6): δ 7.48 (m, 4H), 7.05 (m,
6H), 5.59 (m, 1H), 4.93 (m, 2H), 2.13 (m, 4H), 0.97 (s, 6H); 13C NMR
(75 MHz, C6D6): δ 141.3 (d, 13.6 Hz), 135.9, 133.9, 133.6, 129.1,
129.0, 128.8, 118.0, 48.5 (d, 8.5 Hz), 43.3 (d, 17.5 Hz), 34.8 (d, 14.6
Hz), 29.1 (d, 9.5 Hz); 31P NMR (121 MHz, C6D6): δ -23.8. HRMS
(m/z): [M+] calcd for C19H23P, 282.1536; found, 282.1537.
1
temperature with exclusion of light. H NMR (300 MHz, C6D6): δ
7.36 (m, 2H), 7.06 (m, 3H), 5.59 (m, 1H), 4.93 (m, 2H), 4.08 (d of t,
205 Hz, 7.2 Hz, 1H), 1.89 (q, 7.0 Hz, 2H), 1.53 (m, 2H), 1.42 (m,
2H); 13C NMR (75 MHz, C6D6): δ 138.6, 134.3 (d, 15.8 Hz), 129.0,
128.9, 128.6, 115.5, 35.5 (d, 8.6 Hz), 28.1 (d, 7.9 Hz), 23.4 (d, 11.5
Hz); 31P NMR (121 MHz, C6D6): δ -51.3 (d, 1JP-H ) 202 Hz). HRMS
(m/z): [M+] calcd for C11H15P, 178.0912; found, 178.0909.
2,2-Dimethylpent-4-enylphenylphosphine (17b). This product was
prepared from 17a in a manner analogous to preparation of 15b from
15a. Purification by bulb-to-bulb distillation (110 °C/0.10 Torr) yielded
a clear colorless oil (0.67 g) in 46% yield that was stored in an airtight
storage tube as a solution in C6D6 over activated 4 A molecular sieves
at low temperature with exclusion of light. 1H NMR (300 MHz,
C6D6): δ 7.43 (m, 2H), 7.06 (m, 3H), 5.71 (m, 1H), 5.02 (m, 2H),
4.08 (d of d of d, 206 Hz, 9.5 Hz, 7.8 Hz, 1H), 2.05 (m, 3H), 1.46 (m,
1H), 0.93 (s, 6H); 13C NMR (75 MHz, C6D6): 137.3, 135.8, 134.6,
129.1, 129.0, 128.6, 127.7, 117.9, 47.9 (d, 7.5 Hz), 38.0 (d, 13.5 Hz),
34.3 (d, 9 Hz), 28.5 (d, 8.6 Hz); 31P NMR (121 MHz, C6D6): δ -67.2
(d, 1JP-H ) 205 Hz). HRMS (m/z): [M+] calcd for C13H19P, 206.1224;
found, 206.1229.
Synthesis of 1-Methylpent-4-enylphenylphosphine (15). 1-Meth-
ylpent-4-enyldiphenylphosphine (15a). Using the published proce-
dure,26 5-hexen-2-one was reduced to the corresponding alcohol and
then converted to 5-hexene-2-methanesulfonate. Halogenation of the
alcohol did not provide acceptable yields for the next step of the
synthesis. Therefore, the mesylate (2.48 g, 14 mmol) was stirred under
nitrogen in 50 mL of THF, and KPPh2 (0.5 M in THF) was added (28
mL, 14 mmol) via syringe. The resulting viscous orange/red solution
was stirred for an additional 90 min, and then solvent was removed
under vacuum. Pentane was added to the orange residue, and then
deoxygenated water was added. The colorless organic layer was
separated via cannula, dried over MgSO4 and filtered, and the solvent
was removed to yield a light-yellow oil which was of sufficient purity
for subsequent syntheses (2.3 g, 61% yield) or was distilled (90 °C/0.1
Typical NMR-Scale Catalytic Reactions. In the glovebox, the
organolanthanide precatalyst (2-3 mg) was weighed into an NMR tube
equipped with a Teflon valve, and C6D6 was added (ca. 500 µL). On
the high-vacuum line, the tube was evacuated while the benzene solution
was frozen at -78 °C, and then the substrate (ca. 1.0 M in C6D6, 0.2-
0.3 mL, 30-150-fold molar excess) was added either via syringe under
an Ar flush or by vacuum-transfer (depending on substrate volatility).
The tube was evacuated and backfilled with Ar three times while frozen
at -78 °C, and then the tube was sealed, thawed, brought to the desired
1
Torr) to afford 15a as a colorless oil. H NMR (300 MHz, C6D6): δ
7.53 (m, 4H), 7.06 (m, 6H), 5.65 (m, 1H), 4.93 (m, 2H), 2.25 (m, 1H),
2.18 (m, 1H), 1.95 (m, 1H), 1.65 (m, 1H), 1.35 (m, 1H), 1.00 (d of d,
14.4 Hz, 6.9 Hz, 3H); 13C NMR (75 MHz, C6D6): δ 138.9, 134.7,
134.4, 134.3, 134.1, 129.2, 129.1, 129.0, 129.0, 128.9, 128.9, 115.4,
33.3 (d, 18.2 Hz), 32.2 (d, 12.1 Hz), 29.9 (d, 10.9 Hz), 16.5 (d, 15.8
Hz); 31P NMR (121 MHz, C6D6): δ -0.70. HRMS (m/z): [M+] calcd
for C18H21P, 268.1381; found, 268.1341 (loss of 1 H, [M - 1+] calcd
for C18H20P, 267.1301; found, 267.1301).
1
temperature, and the ensuing reaction was monitored by H and 31P
NMR.
Synthesis of 2-Methylphospholane (2). The precatalyst and sub-
strate 1 were combined as described above and sealed in an NMR tube;
the ensuing cyclization was monitored by 1H and 31P NMR. The
cyclization proceeded until all of the starting material had been
consumed as judged by NMR and the desired products composed 91%
of the final reaction mixture.28 The 1H assignment (similar to that
1-Methylpent-4-enylphenylphosphine (15b). To a stirring suspen-
sion of Li powder (0.37 g, 53.3 mmol) in THF (50 mL) under argon
was added 15a (2.8 g, 10.4 mmol). The gray solution gradually turned
orange, and stirring was continued ∼2.5 h. Under an Ar flush, ca. 3 g
of NH4Cl was then added, and the color changed back to gray. The
THF was then removed under vacuum, and the residue was dissolved
in 20 mL of pentane. The liquid was separated via cannula filtration,
and pentane was removed in vacuo to yield a light-yellow oil.
Distillation (40 °C/0.06 Torr) using a Vigreux column to ensure
separation of the byproduct Ph2PH yielded 0.40 g (20%) of 15b, a
colorless oil, which was stored in an airtight storage tube as a solution
in C6D6 over activated 4 A molecular sieves at low temperature with
exclusion of light. Both cis and trans forms of 15b are present in equal
amounts as assayed by 1H NMR, and the 13C NMR and 31P NMR
exhibit twice as many signals as would be expected for a single isomer.
1H NMR (300 MHz, C6D6): δ 7.39 (m, 2H), 7.06 (m, 3H), 5.65 (m,
1H), 4.93 (m, 2H), [4.5 (d of d, 206 Hz, 5.4 Hz) and 3.94 (d of d, 206
Hz, 6.3 Hz) together 1H], 2.05 (m, 3H), 1.51 (m, 1H), 1.30 (m, 1H),
[0.98 (q, 6.9 Hz) and 0.92 (q, 7.2 Hz) together 3H]; 13C NMR (75
MHz, C6D6): δ 138.8, 135.7, 135.5, 135.3, 135.1, 128.9, 128.9, 115.3,
35.5 (d, 10.4 Hz), 35.3 (d, 15.2 Hz), 32.7 (d, 3.6 Hz), 32.6 (d, 6.7
Hz)], 29.8 (d, 10.9 Hz), 29.3 (d, 9.8 Hz) 20.1 (d, 5.4 Hz) 19.0 (d, 15.2
1
previously reported)21a is H NMR (500 MHz, C6D6): δ 3.01 (d of t,
183.5 Hz, 8.0 Hz, 1H, minor isomer), 2.46 (d of m, 191 Hz, major
isomer), 2.24 (m, 1H, minor isomer), 1.71 (m, 6H), 1.39 (m, 4H), 1.09
(d of d, 10.7 Hz, 6.2 Hz, 3H), 1.07 (m, 2H), 1.02 (d of d, 18.5 Hz, 7.5
Hz, 3H); 13C NMR (75 MHz, C6D6): δ 40.5 (d, 6.1 Hz), 39.5 (d, 3.8
Hz), 32.7 (d, 8.4 Hz), 30.6 (d, 5.3 Hz), 29.5 (d, 4.6 Hz), 28.8 (d, 5.3
Hz), 21.2 (d, 31.0 Hz), 20.53 (d, 10.6 Hz), 20.45 (d, 9.9 Hz), 19.9 (s);
1
31P NMR (121 MHz, C6D6): δ -44.7 (d, JP-H ) 180.9 Hz, 61%),
1
-54.4 (d, JP-H ) 181.0 Hz, 30%). HRMS (m/z): [MH+] calcd for
C5H12P, 103.0671; found, 103.0683.
The major and minor isomer features in the 1H NMR were correlated
with the corresponding isomers in the 31P NMR using HMQC
1
spectroscopy. The structure of 2 was established through 2D H-1H
NOESY spectroscopy. The major isomer P-H proton (1H NMR: δ
2.46 (d of m, 1JP-H )191 Hz); 31P NMR: δ -44.7) exhibited a cross-
peak with the exo methyl group at δ 1.09 (d of d, 10.7 Hz, 6.2 Hz,
3H) showing that the proton and methyl group are proximate, therefore
assigning the major 31P NMR resonance at δ -44.7 ppm to isomer 2b.
In addition, another study29a assigns the same configuration based on
JP-C coupling constant arguments, and another29b assigns the same
configuration based on JP-C-CH coupling constants.
1
Hz); 31P NMR (121 MHz, C6D6): δ -27.4, -32.2 (d, JP-H ) 206
3
and 205 Hz, respectively). HRMS (m/z): [M+] calcd for C12H17P,
192.1068; found, 192.10718.
Synthesis of 2,2-Dimethylpent-4-enylphenylphosphine (17). 2,2-
Dimethylpent-4-enyldiphenylphosphine (17a). Using the published
procedure, 2,2-dimethyl-4-pentenal was prepared.27 This product was
subsequently reduced with LiAlH4 to 2,2-dimethyl-pent-4-en-1-ol, and
final traces of p-cymene (the solvent in preparation of the aldehyde)
were removed via flash chromatography. This alcohol was then
converted to the mesylate, and the mesylate was subjected to reaction
with KPPh2 to form 17a as described above for 15a (71% crude yield),
Synthesis of 2,5-Dimethylphospholane (4). The precatalyst and
substrate were combined as described above and sealed in an NMR
(28) For compound 2, 22 proton signals and 10 carbon signals are
assigned instead of 11 and 5, respectively, due to the presence of both
product diastereomers in an approximately 2:1 ratio.
(26) Becker, D.; Haddad, N. Tetrahedron 1993, 49, 947-964.
(27) Semmelhack, M. F., Ed. Org. Synth. 1984, 62, 125-133.

10226 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
tube; the ensuing cyclization was monitored by 1H and 31P NMR. The
cyclization proceeded until all of the starting material had been
consumed as judged by NMR. The ratio of the configurations of the
various products varied with catalyst (see text). On the basis of
comparisons between the predominantly cis and predominantly trans
spectra, assignments for the cis and trans products can be made. trans-
Isomer:30 1H NMR (300 MHz, C6D6): δ 2.78 (d of t, 180 Hz, 9.3 Hz,
1H), 2.35 (m, 1H), 1.86 (m, 3H), 1.63 (m, 1H), 1.28 (m, 1H), 1.08 (d
of d, 10.8 Hz, 6.6 Hz, 6H); 13C NMR (125 MHz, C6D6): δ 41.09 (d,
5.0 Hz), 39.01, 32.81 (7.5 Hz), 29.61, 22.01 (d, 30.5 Hz), 31P NMR
(121 MHz, C6D6): δ -28.3 (d, 173 Hz). cis-Isomer: 1H NMR (400
MHz, C6D6): δ 2.26 (d of m, 191 Hz, 1H), 1.88 (m, 2H), 1.65 (m,
2H), 1.28 (m, 2H), 0.97 (d of d, 16.8 Hz, 7.6 Hz, 6H); 13C NMR (125
MHz, C6D6): δ 38.50 (d, 5.0 Hz), 33.07 (d, 9.0 Hz), 21.18; 31P NMR
(121 MHz, C6D6): δ -13.4 (d, 1JP-H ) 179 Hz). HRMS (m/z): [MH+]
calcd for C6H14P, 117.0833; found, 117.0844.
Synthesis of 2-Methylphosphorinane (6). The precatalyst and
substrate were combined as described above and sealed in an NMR
tube; the ensuing cyclization was monitored by 1H and 31P NMR. The
cyclization proceeded until all of the starting material had been
consumed as judged by NMR and the major product composed 89%
of the final reaction mixture. 1H NMR (400 MHz, C6D6): δ 2.74 (d of
t, 185.2 Hz, 12.0 Hz, 1H), 1.80 (m, 2H), 1.61 (m, 1H), 1.48 (m, 2H),
1.11 (m, 4H), 1.00 (d of d, 16.8 Hz, 6.8 Hz, 3H); 13C NMR (100 MHz,
C6D6): δ 39.7 (d, 5.7 Hz), 29.1 (d, 3.0 Hz), 28.7, 26.8 (d, 7.2 Hz),
22.4 (d, 14.1 Hz), 20.0 (d, 11.4 Hz); 31P NMR (162 MHz, C6D6): δ
-44.6 (d, 183 Hz). (About 10% of the minor isomer is produced as
assayed by 31P NMR: δ -62.1 (d, 194 Hz.) HRMS (m/z): [M+] calcd
for C6H13P, 116.0755; found, 116.0754.
Synthesis of 2,6-Dimethylphosphorinane (8). The precatalyst and
substrate were combined as described above and sealed in an NMR
tube; the ensuing cyclization was monitored by 1H and 31P NMR. The
cyclization proceeded until all of the starting material had been
consumed as judged by NMR and the major product composed 86%
of the final reaction mixture. 1H NMR (400 MHz, C6D6): δ 2.74 (d of
t, 185.2 Hz, 12.0 Hz, 1H), 1.80 (m, 2H), 1.68 (m, 2H), 1.23 (m, 2H),
1.02 (d of d, 16.4 Hz, 7.2 Hz, 6H), 0.88 (m, 2H); 13C NMR (125 MHz,
C6D6): δ 38.6 (d, 5.5 Hz), 28.6 (s), 27.2 (d, 8.5 Hz), 21.7 (d, 13.5
Hz); 31P NMR (162 MHz, C6D6): δ -25.6 (d, 1JP-H ) 186 Hz). [There
are two minor isomers present in the 31P spectra: δ -35.6 (d, 182.5
Hz, ∼2%) and δ -47.5 (d, 180 Hz, ∼2%).] HRMS (m/z): [MH+] calcd
for C7H16P, 131.098; found, 131.095.
Hz, 3.9 Hz, 1H), 2.08 (m, 1H), 2.01 (m, 1H), 1.82 (m, 1H), 1.55 (m,
2H), 1.41 (m, 1H), 1.13 (m, 1H); 13C NMR (75 MHz, C6D6): δ 138.8
(d, 49.6 Hz), 133.3, 129.6, 128.8, 127.5, 33.5 (d, 5.0 Hz), 29.6 (d, 2.0
Hz), 29.3, 21.8 (d, 8.5 Hz); 31P NMR (121 MHz, C6D6): δ -53.1 (d,
1JP-H ) 191.5 Hz).
Synthesis of 2-Methyl-1-phenylphospholane (14). The precatalyst
and substrate were combined as described above and sealed in an NMR
tube, and the ensuing cyclization was monitored by 1H and 31P NMR.
The cyclization proceeded until all of the starting material had been
consumed as judged by NMR and the desired product composed
∼74.5% of the final reaction mixture.31 1H NMR (300 MHz, C6D6): δ
7.32 (m, 2H), 7.11 (m, 3H), 2.17 (m, 1H), 1.82 (m, 3H), 1.67 (m, 1H),
1.55 (m, 1H), 1.23 (d of d, 18.9 Hz, 7.5 Hz, 3H), 1.14 (m, 1H); 13C
NMR (75 MHz, C6D6): δ 131.3 (d, 16 Hz), 129.0, 128.9, 127.9, 38.7
(d, 11 Hz, CH), 37.7 (s, CH2), 28.2 (d, 3.7 Hz, CH2), 27.1 (d, 14.0 Hz,
CH2), 20.7 (d, 32.8 Hz, CH3); 31P NMR (121 MHz, C6D6): δ 1.7
(major), -6.3 (minor). HRMS (m/z): [M+] calcd for C11H15P, 178.0911;
found, 178.0926.
Synthesis of 2,5-Dimethyl-1-phenylphospholane (16). The pre-
catalyst and substrate were combined as described above and sealed in
1
an NMR tube, and the ensuing was cyclization monitored by H and
31P NMR. The cyclization proceeded until all of the starting material
had been consumed as judged by NMR and the desired product
composed ∼70% of the final reaction mixture. NMR showed conversion
to two products in approximately equal amounts, with twice as many
signals observed as would be expected for a single isomer.32 1H NMR
(500 MHz, C6D6): δ 7.54 (m, 1H), 7.40 (m, 1H), 7.13 (m, 3H), 2.52
(m, 1H), 2.20 (m, 1H), 2.05 (m, 2H), 1.76 (m, 2H), 1.56 (m, 2H), 1.41
(m, 1H), 1.35 (m, 1H), 1.26 (d of d, 19.0 Hz, 7.4 Hz, 6H), 1.25 (d of
d, 18.5 Hz, 7.5 Hz, 3H), 1.12 (m, 2H), 0.76 (d of d, 10.6 Hz, 7.0 Hz,
3H); 13C NMR (75 MHz, C6D6): δ 135.0 (d, 19.0 Hz), 131.7 (d, 16.0
Hz), 129.2, 128.9, 128.9, 128.8, 128.5, 128.1, 38.4 (d, 10.1 Hz), 37.5
(s), 36.5 (s), 36.0 (d, 12.5 Hz), 35.6 (d, 10.0 Hz), 21.6 (d, 35.6 Hz),
21.4 (d, 33.1 Hz), 15.7 (s); 31P NMR (121 MHz, C6D6): δ 19.5, 10.5.
HRMS (m/z): [M+] calcd for C12H17P, 192.1068; found, 192.1067.
Synthesis of 4,4-Dimethyl-2-methyl-1-phenylphospholane (18).
The precatalyst and substrate were combined as described above and
sealed in an NMR tube, and the ensuing cyclization was monitored by
1H and 31P NMR. The cyclization proceeded until all of the starting
material had been consumed as judged by NMR and the desired product
composed 70-90% of the final reaction mixture. NMR showed
conversion to two products in approximately equal amounts, with twice
as many signals observed as would be expected for a single isomer.33
1H NMR (500 MHz, C6D6): δ 7.53 (t, 3.9 Hz, 1H), 7.32 (t, 4.5 Hz,
1H), 7.13 (m, 3H), 2.41 (m, 2H), 1.98 (m, 2H), 1.90 (m, 2H), 1.78 (m,
2H), 1.55 (m, 1H), 1.47 (m, 1H), 1.37 (d of d, 17.5 Hz, 6.0 Hz, 3H),
1.12 (s, 3H), 0.98 (s, 3H), 0.92 (s, 3H), 0.74 (d of d, 6.3 Hz, 4.2 Hz,
3H), 0.66 (s, 3H); 13C NMR (75 MHz, C6D6): δ 134.8 (d, 20 Hz),
129.7, 129.6, 129.1, 128.8, 128.6, 128.5, 127.3, 52.4 (d, 4.6), 51.3 (d,
3.8), 42.9 (s), 41.7 (s), 41.0 (d, 13.6), 39.9 (d, 13.6), 37.5 (d, 9.8), 33.4
(d, 11.4), 31.3 (d, 4.6), 31.1 (s), 30.2 (s), 28.8 (s), 21.2 (d, 32.6), 15.9
(s); 31P NMR (121 MHz, C6D6): δ 3.7, 1.3. HRMS (m/z): [M+] calcd
for C13H19P, 206.1222; found, 206.1221.
Synthesis of 2-(1-Phenyl-2-ethenyl)phospholane (10). The pre-
catalyst and substrate were combined as described above and sealed in
1
an NMR tube, and the ensuing cyclization was monitored by H and
31P NMR. The cyclization proceeded until all of the starting material
had been consumed as judged by NMR. The product formed was stable
for several hours but decomposed slowly to many uncharacterized
products with mass spectrometric molecular weights corresponding
1
roughly to dimers. H NMR (300 MHz, C6D6): δ 7.17 (m, 3H), 7.04
(m, 3H), 3.78 (d of m, 180 Hz, 1H), 2.52 (m, 1H), 2.14 (m, 1H), 1.81
(m, 1H), 1.62 (m, 1H), 1.44 (m, 1H), 1.23 (m, 1H); 13C NMR (75
MHz, C6D6): δ 133.5 (d, 34.6 Hz), 132.2, 129.2, 128.9, 127.4, 34.5
(d, 1.9 Hz), 33.1 (d, 6.1 Hz), 21.3 (d, 4.9 Hz); 31P NMR (121 MHz,
Synthesis of n-Pentylphosphine (19). Diethyl n-Pentylphospho-
nate (19a). The reagent 1-bromopentane (16.4 mL, 132 mmol) was
reacted with P(OEt)3 (25.0 mL, 146 mmol) at 150 °C overnight, and
1
C6D6): δ -55.1 (d, JP-H ) 181 Hz). HRMS (m/z): [M+] calcd for
C11H13P, 176.0754; found, 176.0755.
Synthesis of 2-(2-Phenyl-1-ethenyl)phosphorinane (12). The pre-
catalyst and substrate were combined as described above and sealed in
(31) Product 14 was obtained as a single isomer, and the 1H and 13C
NMR reported here are for that single isomer. Formation of the six-
membered ring byproduct was sufficiently suppressed to allow complete
assignment of the product spectra.
1
an NMR tube, and the ensuing cyclization was monitored by H and
31P NMR. The cyclization proceeded until all of the starting material
had been consumed as judged by NMR. The product formed was stable
for several hours, but decomposed slowly to many uncharacterized
products with mass spectrometric molecular weights corresponding
(32) Final assignments of the 1H and 13C NMR spectra were made after
subtracting peaks from the endocyclic six-membered ring (∼30% of the
product mixture). For compound 16, 24 nonphenyl 1H signals and 8 13C
signals are observed rather than 12 and 6, respectively, due to the presence
of both product diastereomers in equal amounts. One set of signals
corresponds to the spectra reported for the trans isomer synthesized by an
alternative route (13C spectra differ slightly): Burk, M. J.; Feaster, J. E.;
Harlow, R. L. Tetrahedron: Asymmetry 1991, 2(7), 569-592.
1
roughly to dimers. H NMR (300 MHz, C6D6): δ 7.17 (m, 3H), 7.04
(m, 3H), 3.84 (d of d of m, 191 Hz, 12.3 Hz, 1H) 3.13 (d of quart, 12
(29) (a) Langhans, K. P.; Stelzer, O. Z. Naturforsch. 1990, 45b, 203-
211. (b) Couret, C.; Escudie, J.; Thaoubane, S. A. Phosphorus Sulfur 1984,
20, 81-86.
(30) Phospholane 4a (trans isomer) was prepared recently by another
method: Burk, M. J.; Pizzano, A.; Mart´ın, J. A.; Liable-Sands, L. M.;
Rheingold, A. L. Organometallics 2000, 19, 250-260.
(33) Final assignments of the 1H and 13C NMR spectra were made after
subtracting peaks from the endocyclic six-membered ring (∼5-30% of the
1
product mixture). For compound 18, 28 nonphenyl H signals and 14 13C
signals are observed rather than of 14 and 7, respectively, due to the presence
of both product diastereomers in equal amounts.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10227
the resulting phosphonate was purified twice by distillation (110 °C /8
Torr) to yield a colorless oil in approximately 80% yield. NMR spectra
agree with the literature data.34,35 1H NMR (400 MHz, C6D6): δ 3.94
(m, 4H), 1.57 (m, 4H), 1.13 (m, 4H), 1.06 (t, 6.8 Hz, 6H), 0.77 (t, 6.6
Hz, 3H); 13C NMR (100 MHz, C6D6): δ 61.5 (d, 6.1 Hz), 33.6 (d,
Table 1. Summary of the Crystal Structure Data for
[Cp′2YP(H)Ph]2 (20)
formula
cryst syst
C52H72P2Y2
tetragonal
space group
a, b
P42/m
11.1734(12) Å
16.0 Hz), 27.7, 26.3, 23.3 (d, 5.3 Hz), 23.1, 17.3 (d, 5.3 Hz), 14.6; 31
P
NMR (162 MHz, C6D6): δ 32.7.
c
V
Z
dcalc
19.129(3) Å
2388.3(4) Å3
n-Pentylphosphine (19b). Phosphonate 19a was reduced using
AlCl2H in tetraglyme as described for 1b, and the phosphine 19b was
vacuum-transferred away from the tetraglyme solution into a storage
tube (cooled to -78 °C) containing activated 4 A molecular sieves
and C7D8. Final solution concentration was about 1 M by NMR,
indicating an approximately 50% yield. 1H NMR (400 MHz, C7D8): δ
2.58 (d of t, 189 Hz, 7.2 Hz, 2H), 1.28 (t, 6.8 Hz, 2H), 1.18 (m, 2H),
1.14 (m, 4H), 0.82 (t, 6.8 Hz, 3H); 13C NMR (100 MHz, C7D8): δ
33.9 (d, 5.7 Hz), 33.8 (d, 3.8 Hz), 23.4, 14.9, 14.7 (d, 8.0 Hz); 31P
NMR (162 MHz, C7D8): δ -137.1 (t of m, 190 Hz).
Synthesis of Bis[(bispentamethylcyclopentadienyl)(µ-phenylphos-
phino)yttrium)] [Cp′2YP(H)Ph]2 (20). First, hydride precatalyst (Cp′2-
YH)2 was generated by stirring 165 mg of Cp′2Y(CHTMS)2 in 10 mL
of toluene under an H2 atmosphere overnight. Then an approximately
10-fold molar excess of PhPH2 in benzene was added via syringe under
an Ar flush to the red organoyttrium solution. The mixture turned orange
and was stirred for 1.0 h, and then all volatiles were removed in vacuo.
Toluene was next condensed onto the residue, the yellow mixture was
stirred for 1.0 h, and then the toluene was removed in vacuo. This
dissolution/evaporation process was repeated twice to ensure removal
of all CH2TMS2. The resulting solid material proved to be only slightly
soluble in pentane, even at reflux, and slow diffusion of pentane into
toluene at various temperatures did not lead to the formation of
crystalline material. Instead, the material was dissolved in toluene,
filtered, and concentrated under vacuum until solid material began to
precipitate from solution. At this point, the mixture was heated to 60
°C until a clear solution was obtained, and a color change from yellow
to red was observed. The mixture was allowed to cool slowly to room
temperature, and small colorless crystals were collected by decantation
4
1.27 g/cm3
crystal size, mm
color, habit
diffractometer
temp
0.15 × 0.13 × 0.10
colorless, trigonal bipyramidal
Bruker SMART-1000 CCD
-120 °C
µ
25.2 cm-1
transmission factors, range
radiation
0.7376-0.8132
Mo KR (λ ) 0.71069 Å)
exposure time
scan width
2θ range
25 s
0.30°
56.5°
22200 (3033, 0.163)
1203
124
0.067
intensities (unique, Ri)
intensities > 3σ
no. of params
R
Rw2 [Σ(w(F02 - FC2)2)/[Σ(w(F0 )2]1/2 0.138
2
min, max density in ∆F map
-0.65, 1.07 e-/Å3
with a methanol or ethylene glycol temperature calibrant. The sample
was then warmed quickly and inserted into the probe. Data were
acquired using four scans per time interval, with a long pulse delay (8
s) to avoid signal saturation. Reactions were monitored by measuring
the olefinic (for phosphinoalkenes) or P-H (for phosphinoalkynes)
peaks versus the EH(TMS)2 peak generated when catalytic turnover is
initiated, and were monitored over two or more half-lives. The turnover
frequency Nt (h -1) was calculated from the slope m of the least-squares
determined line [Nt ) (60 min h-1)/m], fit to one half-life because of
product inhibition observed (in some cases) at higher conversions.
Reported turnover frequencies are the average of at least two kinetic
runs.
1
to obtain 20 in a very low overall yield. H NMR (500 MHz, C7D8):
δ 7.42 (s, 2H), 7.04 (m, 8H), (δ 3.46, J1 ) 129 Hz, J2 ) 76 Hz, 2H),
2.12 (s, 60H); 13C NMR (125 MHz, C7D8): δ 138.6, 134.2, 126.4,
120.2, 13.5; 31P NMR (202 MHz, C7D8): δ -107 (br).
Results
X-ray Crystallographic Study of [Cp′2YP(H)Ph]2 (20). A suitable
crystal of 20 was mounted using oil (Infineum V8512) on a glass fiber
and mounted on a Bruker SMART-1000 CCD diffractometer. Crystal
data collection parameters are listed in Table 1. Lattice parameters were
determined from three sets of 20 frames and refined with all data. Cell
reduction calculations on all data showed 20 to crystallize in the
tetragonal space group P42/m.36 Intensity data were corrected for
Lorentz, polarization, and anomalous dispersion effects.36 A numerical
absorption correction was also applied with minimum and maximum
transmission factors of 0.7376 and 0.8132, respectively. The structure
was solved using direct methods.37 The final cycles of full-matrix least-
squares refinement were based on 3033 observed reflections (I > 3.00σ-
(I) and 124 variable parameters, and converged (largest parameter shifts
were 0.02 times the esd) with unweighted and weighted agreement
factors of R ) 0.067, Rw2 ) 0.138. Hydrogen atoms attached to carbon
atoms were given idealized positions. The hydrogen atoms on the
phosphorus centers could not be located. The Cp′ rings showed some
disorder and therefore were not refined anisotropically, and the hydrogen
atoms were not included. The largest peak in the final difference map
(1.07 e-/Å3) was located near yttrium.
The goal of this study was to examine in detail the viability,
selectivity, rates, and mechanism of hydrophosphination/cy-
clization processes mediated by organolanthanide catalysts. This
section first presents general synthetic routes to the organo-
phosphine substrates and the activation of the organolanthanide
precatalysts. Next, the catalytic intramolecular hydrophosphi-
nation/cyclization of these phosphines is described, including
the scope of the transformation, as well as the effects of varying
the metal center and ancillary ligands. Reaction kinetics and
mechanistic pathways are then discussed, followed by regio-
and diastereoselectivity, and finally the resting state and
molecular structure of the catalyst. The Discussion section brings
together various aspects of the catalytic cycle, and compares
and contrasts the results of this study with well-developed,
organolanthanide-mediated hydroamination systems.11-13
Substrate Synthesis. To investigate hydrophosphination/
cyclization processes with organolanthanide catalysts in detail,
suitable phosphinoalkene and phosphinoalkyne substrates were
required. Primary and secondary phosphines are air-sensitive
and difficult to handle, and the reported synthetic routes are
not general.38 Therefore, a general method for the synthesis of
primary phosphines was developed, consisting of two basic
steps. First an air-stable phosphonate was synthesized by either
Kinetic Studies of Hydrophosphination/Cyclization. In a typical
experiment, an NMR sample was prepared as described above (see
Typical NMR-Scale Catalytic Reactions) but kept frozen at -78 °C
until kinetic measurements were begun. Before the experiment, the
NMR probe was equilibrated to the desired temperature and checked
(34) Lowen, G. T.; Almond, M. R. J. Org. Chem. 1994, 59, 4548-4550.
(35) Patois, C.; Savignac, P. Bull. Soc. Chim. Fr. 1993, 130, 630-635.
(36) Cromer, D. T.; Waber, J. T. International Table for X-ray Crystal-
lography; The Kynoch Press: Birmingham, England, 1974; Vol. IV.
(37) SIR92: Altomare, A.; Cascarano, M.; Giacovazzo, C.; Guagliardi,
A. J. Appl. Crystallogr. 1993, 26, 343-350.
(38) Syntheses of primary and secondary phosphines: (a) Reference 23.
(b) Cabioch, J. L.; Denis, J. M. J. Organomet. Chem. 1989, 377, 227-233.
(c) Wolfsberger, W. Chem.-Ztg. 1988, 12, 379-381. (d) Kosolapoff, G.
M.; Maier, L. Organic Phosphorus Compounds; Wiley-Interscience: New
York, 1972; Vol. 1, pp 1-287.

10228 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
Scheme 3. Synthesis of Phosphine Substrates; Illustrated for
Substrates 1, 3, and 13
Figure 1. Plot of the reaction of precatalyst Cp′2YCH(TMS)2 with
∼30 equiv of n-pentyl phosphine (19) at 25 °C in C7D8.
of two methods (Scheme 3). Method A employs the Arbuzov
reaction of a phosphite with an alkyl halide,39 and the only
complication is synthesis of the corresponding alkenyl or alkynyl
fragment. Method B involves R-deprotonation of an alkyl
phosphonate, followed by electrophilic addition of the appropri-
ate alkyl chain and is more effective for formation of substrates
having branched hydrocarbon backbones.35 In either case, the
resulting phosphonate is then reduced with “dichloroalane”
AlCl2H (in this case a 5:2 molar mixture of AlCl3 and LiAlH4),40
followed by anaerobic workup with deoxygenated water.
Substrates containing a phenyl group can be isolated by removal
of the diethyl ether solvent; however, more volatile phosphines
cannot be easily separated from common organic solvents and
instead were isolated by vacuum-transfer away from a high-
boiling tetraglyme-based reducing mixture. Secondary phen-
ylphosphines were also synthesized (Method C) by first reacting
KPPh2 with the appropriate alkyl halide and then cleaving a
phenyl group from the resulting tertiary phosphine using an
alkali metal reagent.41 All phosphine substrates were stored as
solutions in C6D6 or C7D8, and the solutions were dried over
several portions of freshly activated molecular sieves. They were
Table 2. Scope of Organolanthanide-Catalyzed
Hydrophosphination/Cyclization of Phosphinoalkenes and
Phosphinoalkynesa
1
characterized by H, 13C, and 31P NMR, high-resolution mass
spectrometry, and elemental analysis (when possible) as detailed
in the Experimental Section.
Catalyst Generation. Cp′2LnE(TMS)2 precatalysts (E ) N,
CH) have been used extensively for intra-and intermolecular
hydroamination, and entry to the catalytic cycle is known to
proceed via rapid and quantitative protonolysis of the Ln-E σ
bond, with concomitant formation of a new Ln-N σ bond (step
i, Scheme 1).11-13 Although all evidence supports a proposed
catalytic cycle for hydrophosphination which is similar (Scheme
2), we observe that initial protonolysis of the Ln-E bond by
primary and secondary phosphine substrates is far slower than
by amine substrates (step i, Scheme 2). Thus, while mM Ln-
CH(TMS)2 protonolysis by mM primary and secondary amines
is essentially instantaneous at -78 °C, we find that half-lives
for the corresponding mM Ln-CH(TMS)2 protonolysis by mM
phosphine reagents can be minutes to hours at room temperature,
depending on substrate structure and also depending on substrate
and catalyst concentrations (Figure 1). This rate-ordering appears
to be a consequence of the weaker product Ln-P bond
compared to the Ln-N bond,19 as well as the softer nature of
phosphorus compared to nitrogen, suggesting a weaker interac-
tion with the hard Ln3+ metal center.42 Primary phosphines are
aAll turnover frequencies (Nt) measured in C6D6. Conversion is
b
>95% by 1H and 31P NMR spectroscopy. Cp′2LaCH(TMS)2 as
precatalyst. c Cp′2SmCH(TMS)2 as precatalyst. d Cp′2YCH(TMS)2 as
precatalyst. e Me2Si(Me4C5)(tBuN)SmN(TMS)2 as precatalyst.
reported to be slightly stronger Brønsted acids than primary
amines;43 however, it appears that hard-soft effects and also
perhaps the somewhat greater steric bulk of the organophos-
(39) (a) Delamarche, I.; Mosset, P. J. Org. Chem. 1994, 59, 5453-5457.
(b) Bhattacharya, A. K.; Thyagarajan, G. Chem. ReV. 1981, 81, 415-430.
(40) Ashby, E. C.; Prather, J. J. Am. Chem. Soc. 1966, 88, 729-733.
(41) (a) Na/NH3: ref 38c. (b) Li/THF: Chou, T.-S.; Yuan, J.-J.; Tsao,
C.-H. J. Chem. Research (S) 1985, 18-19. (c) Li and Na cleavage
methods: Budzelaar, P. H. M.; Doorn, J. A.; Meijboom, N. Recl. TraV.
Chim. Pays-Bas 1991, 110, 420-432.
(42) (a) Pearson, R. G. J. Chem. Educ. 1987, 64, 561-567. (b) Ho, T.-
L. Hard and Soft Acids and Base Principles in Organic Chemistry,
Academic Press: New York, 1977. (c) Pearson, R. G. Hard and Soft Acids
and Bases, Dowden, Hutchinson, and Ross, Inc: Straudsburg, 1973.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10229
phorus reagents compared to those of nitrogen are the dominant
factors. The lanthanide ionic radius is also important in this
initiation step, and -E(TMS)2 cleavage is more rapid for the
larger metal ions (La3+, Nd3+) and increasingly sluggish as the
ionic radius is decreased (Y3+, Lu3+).
quantities of the radical inhibitor hydroquinone does not
significantly inhibit this process. Light and heat both accelerate
this side reaction, and therefore, the present phosphine substrates
were protected from light both before and during catalytic
reactions. Substrates 1, 3, 13, 15, and 17 are affected in varying
degrees by this competing, apparently noncatalytic process, and
in some cases 1 and 3 afford these byproducts in significant
quantities, as will be discussed below. Alkynyl substrates are
not susceptible to this reaction, and olefinic substrates 5 and 7,
which would form less kinetically/thermodynamically favorable
seven-membered rings, do not undergo this side reaction at a
measurable rate, even upon heating. Attempts were also made
to effect catalytic cyclizations of hept-6-enylphosphine (C) and
1-methyl-hept-6-enylphosphine (D) to the corresponding seven-
membered rings; however, the reactions are sluggish and, upon
extended heating to 60 °C, these substrates undergo noncatalytic
oligomerization and decomposition reactions. Therefore, the
scope of intramolecular hydrophosphination/cyclization is cur-
rently limited to the formation of five- and six-membered
phosphorus-containing heterocycles.
The most effective way to activate bis-Cp′ hydrocarbyl
complexes completely and immediately for catalytic studies is
to generate the corresponding hydride (Cp′2LnH)2 in situ using
H25a and to then expose this catalyst to the phosphine substrate.
Using this protocol, immediate Ln-P bond formation occurs,
and the H2C(TMS)2 generated via hydrogenolysis can be used
as an internal NMR standard to monitor the course of the
reaction (via the substrate:catalyst ratio). Color changes are also
observed which indicate Ln-P bond formation; for example, a
yellow (Cp′2LaH)2 solution becomes orange when it undergoes
reaction with a phosphine substrate, while red (Cp′2YH)2
solutions turn bright yellow. Unlike the amine hydroamination/
cyclizations,11-13 the hydrophosphination reaction mixtures do
not return to the initial colors when substrate consumption is
complete, except at very low (e10:1) ratios of substrate:catalyst.
All of the present phosphine-based cyclizations proceed until
complete conversion (>95%) of the substrate has occurred
1
(assayed by H and 31P NMR). No catalyst decomposition is
observed, and as long as the reaction medium is free of air and
moisture, turnover continues.
Alkynyl phosphines also undergo clean and efficient cycliza-
tions (Table 2, entries 5, 6). In this case, five-membered ring
10 is formed more rapidly than is six-membered ring 12, in
accord with the trend observed for aminoalkynes.12 However,
products 10 and 12 are unstable and within hours undergo
dimerization. This dimerization process is not catalyst- or light-
dependent, and several ultimate products are detected by GC-
MS, all having parent ion molecular masses indicative of dimers.
Although the mechanism of the dimerization has not been
rigorously defined, a possible route could begin with isomer-
ization of the product to a phospha-alkene (not observed;
analogous to the enamine-imine tautomerization previously
observed in the cyclization of aminoalkynes12b,46), and phospha-
alkenes are known to add to PdC and CdC bonds (eq 2).47
Scope of Catalytic Hydrophosphination/Cyclization. Al-
kenyl phosphine substrates undergo efficient cyclization in the
presence of organolanthanide catalysts to afford phospholane
and phosphorinane products (Table 2). Phospholanes (fully
saturated five-membered phosphorus-containing rings; entries
1, 2, 7, 8, 9) and phosphorinanes (fully saturated six-membered
phosphorus-containing rings, entries 3, 4) are formed cleanly
and efficiently from alkenyl phosphines. Turnover frequencies
are roughly similar to those observed for analogous amino-
alkenes, although the dependence on ring size is not strictly 5
> 6, as observed for hydroamination/cyclization reactions.11-13
The reactions proceed until starting material is completely
consumed; however, formation of a noncatalytic side product
is also observed (eq 1). Varying quantities of a larger endocyclic
Terminal alkynyl substrate 4-pentynylphosphine (E) was also
investigated. This substrate undergoes concurrent cyclization
and oligomerization (eq 3), with the latter process presumably
analogous to that previously observed for organolanthanide-
mediated oligomerization of simple terminal alkynes.47 There-
fore, an accurate quantitation of the turnover frequency for the
hydrophosphination/cyclization process could not be obtained,
although qualitatively it is a very rapid process (Nt > 100/h-1).
The cyclized product also undergoes rapid dimerization (deter-
mined by GC-MS), as observed for phosphinoalkyne cycliza-
tion products 10 and 12.
ring are formed during substrate drying and storage; however,
this does not interfere with the catalytic reaction. This side
product could be formed independently by heating the substrate
and exposing it to light, and multinuclear NMR with reference
to literature data44 was used to confirm the phosphorinane
assignment. There is precedent for such endocyclization pro-
cesses which are thought to be radical chain in nature;45
however, storing the present phosphinoalkenes with small
(43) The Brønsted acidities of some primary and secondary phosphines
and amines have been compared, and phosphines are generally slightly more
acidic: Issleib, K.; Ku¨mmel, R. J. Organomet. Chem. 1965, 3, 84-91.
(44) The chemical shifts of phospholanes and phosphorinanes are known
to be different; see ref 15. In addition, the side product formed by the
endocyclic reaction of 3 is identical to the catalytic product formed in
organolanthanide-catalyzed cyclization 5 f 6.
(45) Examples of P-H addition to olefins to form phospolanes and
phosphorinanes: (a) Hackney, M. L.; Schubert, D. M.; Brandt, P. F.;
Haltiwanger, R. C.; Norman, A. D. Inorg. Chem. 1997, 36, 1867-1872.
(b) Brandt, P. F.; Schubert, D. M.; Norman, A. D. Inorg. Chem. 1997, 36,
1728-1731. (c) Field, L. D.; Thomas, I. P. Inorg. Chem. 1996, 35, 2546-
2548. (d) Davies, J. H.; Downer, J. D.; Kirby, P. J. Chem. Soc. (C) 1966,
245-247.
Metal and Ancillary Ligand Effects on Catalytic Hydro-
phosphination/Cyclization. Depending upon the substrate type,
organolanthanide-mediated hydroamination/cyclization pro-
cesses exhibit distinctive trends in turnover frequency with metal
ionic radius; increasing ionic radii correlates with greater

10230 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
Table 3. Metal Ion Size Effects on Turnover Frequencies for the
Hydrophosphination/Cyclization of Substrates 1 and 9
a Eight-coordinate ionic radii from reference 2. b All rates measured
in C6D6.
turnover frequencies for aminoalkenes,11 whereas decreasing
ionic radii correlate with greater turnover frequencies for
aminoalkynes,12 and lanthanides with intermediate ionic radii
exhibit the highest turnover frequencies for aminoallenes.13 The
phosphinoalkynes studied here exhibit behavior similar to that
of aminoalkenes, with larger metal ions affording greater
turnover frequencies (Table 3). Phosphinoalkenes, however,
exhibit a different trend in that the mid-sized Y3+ ion exhibits
the highest Nt values, followed by Lu3+ and Sm3+, with the
largest La3+ ion the most sluggish (Table 3). The effects of
ancillary ligands on reaction rates also differ markedly between
phosphinoalkynes and phosphinoalkenes, in that opening the
catalyst coordination sphere with the Me2Si(Me4C5)(tBuN)Ln-
(CGC; F) ligand leads to enhanced cyclization rates for alkynes
(Table 2, entry 5). In contrast, for phosphinoalkenes little
catalytic activity is observed with CGC complexes, and turnover
Figure 2. Kinetics of organolanthanide-mediated hydrophosphination/
cyclization. (A) Cyclization of pent-4-enylphosphine (1 f 2) mediated
by H2-activated Cp′2YCH(TMS)2 in C6D6 at 25 °C. The lines represent
the least-squares fit to the data points for the first half-life of the
reaction. (B) Cyclization of 1-methylpent-4-enylphosphine (3 f 4)
mediated by H2-activated Cp′2SmCH(TMS)2 in C6D6 at 25 °C. The
lines represent the least-squares fit to the data points for the first half-
life of the reaction. The deviation from zero-order kinetics is ascribed
to competitive inhibition by product (see text).
was monitored over time and normalized versus the CH2(TMS)2
internal standard (δ ∼0.05 ppm) produced by catalyst activation.
The data show that the rate of substrate consumption is initially
zero-order in substrate concentration; however, after approxi-
mately one half-life, deviations suggestive of competitive
inhibition by product11 are observed. These kinetic inhibition
effects are less evident at higher temperatures, more marked at
lower temperatures, and also depend on the steric bulk of the
product. As an example, the 1 f 2 cyclization with H2-activated
precatalyst Cp′2YCH(TMS)2 was monitored until complete
consumption of 1, then an additional aliquot of 1 was added to
the same reaction mixture, and monitoring of the reaction
continued (Figure 2A). Likewise, the cyclization of 3 f 4 with
H2-activated precatalyst Cp′2SmCH(TMS)2 was carried out in
the same way (Figure 2B). These measurements show inhibition
(deviation from zero-order kinetics) in the 1 f 2 cyclization
(Figure 2A) after one half-life in the first data set and more
severe inhibition in the second data set, where substantial
concentrations of product 2 are now present. In contrast, Figure
2B evidences substantially less inhibition for the 3 f 4
cyclization in either the first or second data set, suggesting that
the increased steric bulk of phospholane 4 (relative to 2) renders
it a less effective inhibitor. There is no decrease in initial
catalytic activity in the second data set for the 3 f 4 cyclization;
therefore, no catalyst decomposition is evident. 1H and 31P NMR
spectra also reveal that both substrate and product resonances
are broadened in cyclizations mediated by paramagnetic cata-
lysts, in contrast to the analogous hydroamination results where
only substrate resonances are severely broadened.11-13 Because
no detectable catalyst decomposition occurs in these reactions,
the observed deviation from linearity in the reaction kinetics is
most reasonably assigned to product inhibition. However, for
all cyclizations studied, data for at least the first half-life indicate
rates are too sluggish to measure. Qualitatively, the same is true
using the sterically open Me2SiCp′′2NdCH(TMS)2 catalyst, in
that it leads to enhanced rates for phosphinoalkyne cyclization,
but depressed rates are observed with phosphinoalkenes. Thus,
it appears that phosphinoalkynes are activated and undergo
cyclization via a pathway similar to that for aminoalkenes,11
while phosphinoalkene cyclizations are more sensitive to catalyst
structural variations, and large changes in turnover frequencies
occur with small changes in metal ion size and ancillary ligation.
Kinetic and Mechanistic Studies of Hydrophosphination/
Cyclization. Quantitative kinetic studies of the cyclizations of
the various phosphine substrates were carried out and monitored
by 1H and 31P NMR spectroscopy. Both Y3+ and Sm3+ catalysts
were used, with each offering distinctive advantages. In the case
of diamagnetic Cp′2YCH(TMS)2, integration of the resonances
is straightforward, while for paramagnetic Cp′2SmCH(TMS)2,
broadening of some resonances provides information about
which portions of the substrate and product structures interact
closely with the metal center. The cyclization reactions were
carried out with a 30-150-fold molar excess of substrate, and
in all cases substrate was completely consumed. The decrease
of olefinic (δ ∼5.8 ppm) or P-H (δ ∼3.5 ppm) resonances
(46) Tautomerization of enamines to imines: (a) Lammertsma, K.;
Prasad, B. V. J. Am. Chem. Soc. 1994, 116, 642-650. (b) Shainyan, B. A.;
Mirskova, A. N. Russ. Chem. ReV. 1979, 48, 107-117.
(47) a) Mackewitz, T. W.; Regitz, M. Synthesis 1998, 125-133. (b)
Mackewitz, T. W.; Peters, C.; Bergstra¨sser, U.; Leininger, S.; Regitz, M.
J. Org. Chem. 1997, 62, 7605-7613. (c) Mackewitz, T. W.; Regitz, M.
Liebigs Ann. 1996, 327-333. (d) Slany, M.; Regitz, M. Synthesis 1994,
1262-1266.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10231
Figure 3. Determination of reaction order in lanthanide concentration
for the hydrophosphination/cyclization of pent-4-enylphosphine (1 f
2) mediated by H2-activated Cp′2YCH(TMS)2 in C6D6. The line
represents the least-squares fit to the data points.
the reaction to be approximately zero-order in substrate con-
centration. Attempts to fit the kinetic data to other simple rate
laws (e.g., first-order in substrate concentration) yielded results
which were less convincing.
A plot of reaction rate versus precatalyst concentration for
the 1 f 2 cyclization mediated by H2-activated Cp′2YCH-
(TMS)2, in which substrate concentration is held constant and
precatalyst concentration varied over a more than 10-fold range,
shows the reaction to be first-order in catalyst concentration
(Figure 3). Overall, the empirical rate law (eq 4) is essentially
the same
Figure 4. (A) Normalized ratio of substrate to lanthanide concentration
as a function of time and temperature for the hydrophosphination/
cyclization of 1-methyl-pent-4-enylphosphine (3 f 4) using H2-
activated precatalyst Cp′2SmCH(TMS)2 in C6D6. (B) Eyring plot for
the hydrophosphination/cyclization of 1-methyl-pent-4-enylphosphine
(3 f 4) using H2-activated precatalyst Cp′2SmCH(TMS)2 in C6D6. The
lines are least-squares fits to the data points.
ν ) k[substrate]0[catalyst]1
(4)
as was observed for hydroamination,11-13 although competing
kinetic inhibition by the phosphorus heterocyclic product is more
pronounced. Variable-temperature kinetic measurements were
also carried out for the 3 f 4 cyclization mediated by
H2-activated Cp′2SmCH(TMS)2 and indicate the reaction to be
approximately zero-order in substrate concentration over a
greater than 30 °C temperature range (Figure 4A). Arrhenius
and Eyring (Figure 4B) analyses afford the activation parameters
Ea ) 13.0 (1.4) kcal/mol, ∆Hq ) 12.3 (1.6) kcal/mol, and ∆Sq
) -25.9 (5.2) eu.
Table 4. Effect of Catalyst on cis/trans Selectivity of
Hydrophosphination at 25 °C
Reaction Selectivity. The catalytic cyclization of a phosphi-
noalkene generates a new stereocenter at carbon, with an
additional stereocenter generated because of the slow inversion
at phosphorus,49 meaning that, in principle, two diastereomers
(two pairs of enantiomers) can be formed in the 1 f 2
cyclization (eq 5). The selectivity for phosphinoalkene ring
a Added ∼10 equiv n-propylamine.
not change after completion of the reaction (in the presence of
catalyst or after separation from the catalyst). The 31P chemical
shifts of the various products are substantially dispersed,
allowing accurate NMR assay of the product distributions.50
Phospholane 2 is consistently produced in an approximate 2:1
ratio of 2b : 2a (Table 4, entry 1), regardless of the catalyst.
The product phospholane configuration is assigned by 2D-
NOESY spectroscopy and coupling constant analysis.29,51 In
addition, the cis/trans configurational disposition of 2 is implied
closure is highly catalyst-dependent only for substrate 3 while
for substrates 1, 5, and 7 the ratios of product diastereomers
are essentially invariant with catalyst (Table 4). All ratios are
invariant with conversion as assayed by 31P NMR and also do
(48) Heeres, H. J.; Teuben, J. H. Organometallics 1991, 10, 1980-1986.
(49) The barrier to inversion at phosphorus is typically about 35 kcal/
mol: (a) Reference 15a. (b) Baechler, R. D.; Mislow, K. J. Am. Chem.
Soc. 1970, 92, 3090-3093. (c) Rauk, A.; Allen, L. C.; Mislow, K. Angew.
Chem., Int. Ed. Engl. 1970, 9, 400-414.
(50) The present phospholanes with different configurations have chemi-
cal shift differences of 10-15 ppm, and phosphorinanes are generally
downfield of phospholanes in 31P NMR: see ref 15.
(51) For details on the assignment of the configuration of 2, please see
the Experimental Section.

10232 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
Table 5. Chemical Shiftsa and Coupling Constants for the
by 31P NMR spectroscopy in the presence of a paramagnetic
catalyst, since the phosphorus lone pair should be more
accessible for the cis configuration than the trans and therefore
should interact more strongly with the catalyst. Indeed the cis-
2b resonance is appreciably broader.
Reactions of Phosphines with (Cp′2YH)2 at -80 °C
species
δ, ppm
J, 31P{1H}, Hz
J, 31P-1H, Hz
K
L
M
N
-17.4
-90.1
-110.2
-37.5
-81.6
-33.7
-39.4
-20.3
-24.1
d, 126
d, 214
d, 189
multiplet
d, 302
none
d, 206
d, 272
d, 267
d, 205
d of d, 87, 58
multiplet
d of d, 384, 51
d of d, 384, 45
d, 110
d, 68
d, 70
The distribution of phospholane product 4 stereoisomers
(Table 4, entry 2) exhibits a large variation with metal ion sizes
as high as 70% 4a for Cp′2Sm-, a marginal selectivity for
Cp′2Y-, and greater than 60% 4b for Cp′2Lu-.30 A trace
quantity of a third phospholane, presumably 4c, is observed in
all cases. Decreasing the reaction temperature to 0 °C does not
effect a significant change in product ratios for the 1 f 2 and
3 f 4 cyclizations mediated by Cp′2Y- and Cp′2Sm-. Addition
of ∼10 equiv of n-propylamine to the 3 f 4 cyclization also
results in a large change in selectivity, with a nearly 1:1 mix of
4a and 4b with Cp′2Sm-, and high selectivities for 4a with
both Cp′2Y- and Cp′2Lu-.52 Six-membered phosphorinanes 6
and 8 are both obtained in greater than 80% diastereoselectivity
(6a and 8b, entries 3, 4, Table 4) regardless of catalyst.
Secondary phenylphosphines 14, 16, and 18 undergo sluggish
cyclization and yield catalyst-invariant diastereomeric ratios of
∼10:1, 1:1, and 1:1, respectively.
O
P
d, 124
a Chemical shifts reported relative to an external 85% H3PO4
standard.
been reported, chemical shifts are not diagnostic, and phosphido
(M-PR2) ligand 31P chemical shifts can range over more than
100 ppm,53,54 while phosphine (M-PR3) ligands exhibit 31P
NMR resonance positions similar to those of the analogous free
phosphine. However, the multiplicities observed in the NMR
give useful information about the nature of species in solution,
1
31P-89
Y
and diagnostic coupling constants are in the range J
≈
50-70 Hz (e.g., J53a), J
≈ 180-200 Hz,15 and J H-
1
2
1
31P-1H
89Y
≈ 2-3 Hz.55
As noted above (eq 1), noncatalytic endocyclic ring formation
is also observed for substrates 1 and 3, and accounts for the
remainder of the product mixture (not tabulated) in Table 4. In
most cases, this product represents 10-15% of the total and is
largely formed during the drying of the substrate, and not during
the catalytic reaction. However, in a few lanthanocene cases (1
f 2 mediated by Cp′2Lu-; 3 f 4 mediated by Cp′2La-) and
also when more open ancillary ligand systems such as the CGC
ligand3a and the chiral Me2Si(Cp′′)(CpR*) ligand4 are employed,
minor formation of the exocyclic product and predominant
formation of the endocyclic product are observed. Possible
reasons and a possible alternate mechanistic pathway are
considered in the Discussion section.
Catalyst Resting State. Initial observations regarding the
catalyst resting state were made using paramagnetic metal
centers Nd3+ and Sm3+. The substrate P-H resonances are
broadened throughout the course of the reaction, indicating rapid
exchange of bound and free P-H species. At an initial level of
analysis, a phosphine-phosphido resting state for the catalyst
(G), analogous to the amine-amido resting state established
for hydroamination (H),11 seems most reasonable. Broadening
of the heterocyclic product P-H signals indicates that these
phosphines undergo rapid coordination to/dissociation from, the
lanthanide metal center, also paralleling the amine-amido
system.
Phosphinoalkene 1 was investigated initially; however, the
data proved difficult to interpret since the presence of the
endocyclic ring side product, as well as the onset of catalytic
cyclization at higher temperatures (g-20 °C), severely com-
plicate the spectra. Therefore, the noncyclizable phosphines
PhPH2 and H2PCH2(CH2)3CH3 (19) were investigated, as models
for the actual substrates. (Cp′2YH)2 was generated in situ and
treated with the phosphine in a ∼10:1 molar ratio at -78 °C.
Spectra were recorded initially at -80 °C and then at higher
temperatures as the mixture was gradually warmed to 25 °C.
1H NMR spectra were complex and not particularly informative
over the temperature range, while 31P data were more informa-
tive and are tabulated with coupling constants in Table 5.
The low-temperature 31P{1H} spectrum of the (Cp′2YH)2 +
19 reaction mixture (Figure 5A) reveals a doublet (J ) 126
Hz) at δ -17.4 ppm (K), while the 1H-coupled spectrum (Figure
5B) reveals K as a doublet of doublets (J1 ) 214 Hz, J2 ) 126
Hz). These data indicate a Y-P bond, with the proton-coupled
spectrum indicating that a single P-H linkage is also present.
By analogy to the lanthanocene alkylamido hydroamination
chemistry11 and as evidenced by phosphine broadening in the
presence of paramagnetic lanthanocene catalysts, there are likely
additional phosphorus ligands (presumably phosphines) coor-
dinated to the metal center; however, they must be exchanging
To examine the resting state of the present hydrophosphina-
tion catalysts in greater detail, low-temperature 1H and 31P NMR
studies were undertaken. The diamagnetic catalyst (Cp′2YH)2
was chosen because 89Y has I ) 1/2 (100% natural abundance)
and because several examples of Y-P bonds are known.53 Only
a few transition metal phosphine-phosphido complexes have
(54) (a) Cole, M. L.; Hibbs, D. E.; Jones, C.; Smithies, N. A. J. Chem.
Soc., Dalton 2000, 545-550. (b) Zwick, B. D.; Dewey, M. A.; Knight, D.
A.; Buhro, W. B.; Arif, A. M.; Gladysz, J. A. Organometallics 1992, 11,
2673-2685.
(52) Not surprisingly, addition of n-propylamine slows the rate of
reaction; see ref 11a.
(55) 2J
couplings are typically very small and are frequently not
89Y-1H
(53) For examples, see: (a) Fryzuk, M. D.; Haddad, T. S.; Rettig, S. J.
Organometallics 1991, 10, 2026-2036. (b) Fryzuk, M. D.; Haddad, T. S.
J. Am. Chem. Soc. 1988, 110, 8263-8265. (c) Westerhausen, M.; Hartmann,
M.; Schwarz, W. Inorg. Chim. Acta 1998, 269, 91-100.
observed. See: a) Bambirra, S.; Brandsma, M. J. R.; Brussee, E. A. C.;
Meetsma, A.; Hessen, B.; Teuben, J. H. Organometallics 2000, 19, 3197-
3204. (b) den Haan, K. H.; Boer, J. L.; Teuben, J. H.; Spek, A. L.; Jojiæ-
Prodiæ, B.; Hays, G. R.; Huis, R. Organometallics 1986, 5, 1726-1733.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10233
Figure 5. (A) Low-temperature (-80 °C) 31P{1H} spectrum of an n-pentylphosphine + (Cp′2YH)2 reaction mixture in C7D8. (B) Low-temperature
(-80 °C) 31P-1H-coupled spectrum of an n-pentylphosphine + (Cp′2YH)2 reaction mixture in C7D8. R ) n-pentyl in the Figure.
with free phosphine at a rate which is rapid on the 31P NMR
time scale, even at -80 °C. These results suggest a phosphine-
phosphido catalyst resting state such as K, analogous to that
identified in organolanthanide-mediated hydroamination,10a H.
intensity relative to signal K. In addition, the 1 f 2 cyclization
helps to clarify this assignment. After the 1 f 2 cyclization is
complete, removal of all solvent and excess phosphine, followed
by redissolution of the residue, affords a room-temperature
spectrum with a similar chemical shift. Although no measurable
coupling can be observed from the multiplet shown in Figure
5, the residue from the cyclization of 1 f 2 exhibits a triplet (J
) 56 Hz) in the 31P{1H} NMR spectrum, and this resonance
expands to a triplet of doublets (J1 ) 123 Hz, J2 ) 56 Hz) in
1
the H-coupled spectrum. This confirms the presence of two
The next higher field 31P{1H} resonance (Figure 5A) is a
doublet of doublets (J1 ) 87 Hz, J2 ) 58 Hz) at δ -90.1 ppm
(L), which expands to a doublet of doublet of doublets (J1 )
equivalent Y3+ centers and two phosphido ligands. We therefore
assign this resonance to structure M. The crystal structure of
the R ) phenyl analogue is described in detail below.
1
189 Hz, J2 ) 87 Hz, J3 ) 58 Hz) in the H-coupled spectrum
(Figure 5B). This resonance decreases in relative intensity as
the solution is warmed and is not detectable by ca. -10 °C.
Importantly, this peak is not observed upon recooling the
solution to -80 °C, whereas signals K and M reversibly increase
and decrease (with broadening) in intensity with cooling and
warming, respectively. The solution gradually changes color
upon warming, from the red color of (Cp′2YH)2 solutions to an
orange color and gradually to the final yellow color indicative
of Y-P bond formation. Furthermore, the yellow color does
not revert to red upon recooling. These results suggest that this
resonance arises from initial Y-P bond formation, a sluggish
process at low temperatures. The 31P doublet of doublets is
consistent with two inequivalent Y3+ metal centers, and the
A mixture of (Cp′2YH)2 and PhPH2 was examined in the same
manner and at -80 °C afforded spectra very similar to those
shown in Figure 5. As the temperature is increased to room
temperature, resonances analogous to K, L, and M decrease in
intensity, and two new doublets grow in at δ -37.5 (dd, J1 )
384 Hz, J2 ) 51 Hz) and δ -81.6 (dd, J1 ) 384 Hz, J2 ) 45
1
increased multiplicity in the H-coupled spectrum indicates a
P-H linkage is also present, suggesting a structure such as L.
1
Hz), with the δ -37.5 ppm resonance exhibiting H coupling
1
1
in the H-coupled spectra (J ) 302 Hz), however, but no H
coupling is observed to the δ -81.6 resonance. The signals
increase in intensity relative to the free phosphine signal as the
temperature is raised further (up to 60 °C), and do not diminish
in intensity upon cooling, indicating irreversible formation of a
1
new species. A new doublet of doublets in the H spectrum at
The final 31P{H} resonance observed, aside from free n-pentyl
phosphine at δ -132 ppm, is a multiplet at δ -110.2 ppm (M;
Figure 5A), which expands to a more complex multiplet in the
1H-coupled spectrum (Figure 5B). This resonance is phosphine
concentration-dependentsadding more phosphine lessens the
δ 5.56 (J1 ) 302 Hz, J2 ) 11 Hz) is observed, and a singlet
indicating formation of H2 is also seen. A similar species is
formed in the reaction of (Cp′2YH)2 and n-pentylphosphine, but
is formed much more slowly at comparable temperatures. The
very large coupling constant argues for a P-P bond,56 and the

10234 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
observation of two distinct phosphorus nuclei, one coupled to
a proton, implies a structure such as N.57
The low-temperature reaction of (Cp′2YH)2 + phosphino-
alkene 1 presents a more complex picture, since two doublets
with similar intensities are observed in the initial 31P{1H} NMR
spectrum at -80 °C (δ -33.7, J ) 110 Hz; δ -39.4, J ) 68
Hz); both doublets also exhibit observable P-H couplings in
the 1H-coupled spectrum of J ) 206 and 272 Hz, respectively.
These resonances indicate likely coordination of a secondary
phosphine (the side product phosphorinane) to the Y center at
-80 °C. Although any phosphido resonance immediately
analogous to K in Figure 5 is not detected, a phosphine-
phosphido structure such as O seems most reasonable.58
Figure 6. Molecular structure of [Cp′2YP(H)Ph]2 (20). Thermal
ellipsoids are drawn at the 50% probability level. Carbon atoms on the
Cp′ rings were refined isotropically.
Table 6. Selected Bond Distances (Å) and Angles (deg) for
[Cp′2YP(H)Ph]2 (20)
Bond Distances
Y1
C1
C3
C5
Y1
Y1
Y1
C8
C10
P1
C2
C4
C6
C7
C9
C11
C17
C19
3.021(3)
1.35(2)
1.36(3)
1.40(2)
2.68(2)
2.66(3)
2.66(3)
1.56(4)
1.58(4)
P1
C1
C3
C5
C6
1.86(2)
1.37(3)
1.31(3)
1.40(2)
2.63(2)
2.64(3)
1.66(3)
1.57(4)
1.55(4)
C2
C4
C1
Y1
Y1
C7
C9
C11
After allowing the cyclization of 1 f 2 to proceed to partial
completion and then recooling the sample to -80 °C, two
additional doublets (δ -20.3, J ) 70 Hz; δ -24.1, J ) 124
Hz) are observed; both doublets also exhibit P-H couplings in
C8
C10
C16
C18
C20
1
the H-coupled spectrum, J ) 267 and 205 Hz, respectively,
and all four doublets have similar intensities. These new features
indicate binding of the product phospholane to the catalyst as
well. The large number of peaks indicates that several structures
are present, including type O and also a structure such as P. As
the temperature is raised, all of these resonances broaden and
decrease in intensity, suggesting rapid interchange of free
phosphine, free cyclized phosphorinane, and free cyclized
phospholane in the resting state of the catalyst at temperatures
above -20 °C.
P1-Y1-P1’
Y1-P1-C1
119.6(1)
130.6(7)
a Cg ) ring centroid.
Molecular Structure of [Cp′2YP(H)Ph]2 (20). To understand
1
more fully the various solution-phase species detected by H
1
and 31P NMR, and to obtain further information on the catalyst
resting state, attempts were made to crystallize a Ln-phosphido
species. The molecular structure of the product of the reaction
of (Cp′2YH)2 + PhPH2 20 (see Experimental Section for
isolation and characterization procedures) was determined by
low-temperature single-crystal X-ray techniques. The result is
(56) In LnMP2 complexes, P-P coupling constants are typically greater
than 300 Hz; for examples see: (a) Etkin, N.; Benson, M. T.; Courtenay,
S.; McGlinchey, M. J.; Bain, A. D.; Stephan, D. W. Organometallics 1997,
16, 3504-3510. (b) Ho, J.; Breen, T. L.; Ozarowski, A.; Stephan, D. W.
Inorg. Chem. 1994, 33, 865-870.
(60) (C5H5)2Lu(µ-PPh2)2Li(tmed): Schumann, H.; Palamidis, E.; Schmid,
G.; Boese, R. Angew. Chem., Int. Ed. Engl. 1986, 25, 718-719.
(61) a) [Cp2Ti(µ-P(H)Cy)]2: Xin, S.; Woo, H. G.; Harrod, J. F.; Samuel,
E.; Lebuis, A.-M. J. Am. Chem. Soc. 1997, 119, 5307-5313. (b) Ho, J.;
Drake, R. J.; Stephan, D. W. Organometallics 1993, 12, 3145-3157. (c)
[Cp2Zr(µ2-PCyH)]2: Ho, J.; Stephan, D. W. Organometallics 1991, 10,
3001-3003.
(62) For other crystal structures containing Y-P bonds, see: (a) Karsch,
H. H.; Graf, V.; Reisky, M.; Witt, E. Eur. J. Inorg. Chem. 1998, 1403-
1406. (b) Fryzuk, M. D.; Love, J. B.; Rettig, S. J. J. Am. Chem. Soc. 1997,
119, 9071-9072. (c) Fryzuk, M. D.; Love, J. B.; Rettig, S. J. Organome-
tallics 1992, 11, 2967-2969.
(57) This type of structure has some precedent in titanocene chemistry;
see ref 61a.
(58) The difference in observed 31P chemical shifts for the phosphido
resonance of O vs that of K may reflect cyclic phosphine coordination to
the metal in O. This could donate more electron density than a linear
phosphine and therefore affect the resonance position of the linear phosphido
ligand.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10235
resolved (δ 3.46, J1 ) 129 Hz; J2 ) 76 Hz), indicating both
Attempts to study cyclization selectivity with these more
encumbered substrates (Table 2, entry 8) were thwarted by the
low reaction rates, and furthermore, cyclization rates are
surprisingly insensitive to geminal dimethyl substitution (the
Thorpe-Ingold effect,67 entry 9). However, the cyclizations of
primary phosphinoalkenes are significantly more rapid and more
informative. With regard to rates, the turnover frequency for
the 3 f 4 transformation is significantly greater than that for 1
f 2 (Table 2) and indicates that in the proposed seven-
membered pseudo-chair transition state (vide infra), the methyl
substituent of 3 actually enhances the rate of cyclization. The
greater 5 f 6 cyclization rate versus 1 f 2 is surprising, given
that increased ring size normally correlates with depressed
aminoalkene, aminoalkyne, and aminoallene hydroamination/
cyclization rates11-13 and also considering that process 7 f 8
is slower than 3 f 4, and process 11 f 12 is slower than 9 f
10 (Table 2). The difference between ring-size effects for
hydroamination and hydrophosphination must lie in the transi-
tion state and will be discussed in a following mechanistic
section.
3
1JP-H coupling and JP-H coupling.64 Although the 31P NMR
resonance is too broad to resolve P-H coupling, the location
at δ -107 agrees with the assigned dimeric structure type M
(vide supra, R ) phenyl).65 The location of the H atoms in the
structure is not clear, however since the two phenyl groups in
the highly symmetric structure reside in only one plane, the
two most likely positions are as drawn in Q and R.66 This
dimeric structure is somewhat different than the diffraction-
characterized amine-amido complex, Cp′2La(NHCH3)(NH2-
CH3) (S),11a which is related by Lewis base ring-opening of an
analogous [Cp′2La(NHCH3)]2 dimer. In the present case, the
isolation of 20 apparently reflects, under the reaction conditions,
greater stability and lower solubility of a dimer versus a
phosphine-phosphido complex.
Metal and Ancillary Ligand Effects on Catalyst Activation
and Activity. The ability to control the coordination sphere by
varying metal ionic radius and ancillary ligation and conse-
quently, ground-state and transition-state ligand-ligand and
substrate-ligand nonbonded interactions, is one intriguing
feature of homogeneous organolanthanide catalysis.1 Accord-
ingly, Cp′2LnCH(TMS)2 catalysts with the following six-
coordinate ionic radii: Ln ) La3+ (1.160 Å), Sm3+ (1.079 Å),
Y3+ (1.019 Å), and Lu3+ (0.977 Å),2 were examined in the
present study. As the metal-ion size is decreased, the rate of
protonolytic activation of the Cp′2LnCH(TMS)2 precatalyst
(Scheme 2, step i) slows dramatically as judged by coordinated
-CH(TMS)2:dissociated CH2(TMS)2 ratios in the 1H NMR. For
example, a ∼30:1 ratio of 11:Cp′2LaCH(TMS)2 at mM con-
centrations undergoes complete La-CH(TMS)2 protonolysis in
under 60 min at 25 °C. However, a 30:1 ratio of mM 19:Cp′2-
YCH(TMS)2 does not undergo complete protonolysis even after
12 h at 25 °C. It is difficult to compare these results to the
analogous precatalyst activation in hydroamination, because at
-78 °C, protonolysis of the Ln-C bond is essentially instan-
taneous for solutions of comparable concentrations.11 If Ln-C
protonolysis of Cp′2YCH(TMS)2 with a primary amine is
assumed to require ∼1 s, as compared to ∼12 h with the
analogous phosphine substrate, then Ln-C cleavage is at least
5 × 104 times faster for amines than for phosphines. Because
of this slow initiation step, most hydrophosphination reactions
were carried out by first activating the catalyst with H2 in situ5a
and then adding the phosphine substrate to the resulting
organolanthanide hydride.
The two different classes of phosphine substrates studied
exhibit different reactivity trends with respect to variation of
metal ionic radius and ancillary ligation and therefore will be
discussed separately. For phosphinoalkynes, the hydrophosphi-
nation/cyclization reaction rate increases with increasing Ln3+
ionic radius, analogous to the trend observed for aminoalkenes11
but opposite to that observed for aminoalkynes (Table 3).12 For
aminoalkynes, the very large exothermicity estimated for the
insertion step (Scheme 1, step ii) was postulated to correlate
with a more reactant-like transition state, and therefore less
sensitivity to metal-ionic radius and structure of ancillary
ligands.12 For phosphinoalkynes, other factors are likely opera-
tive and are based on differences in the heteroatom, in that the
softer and larger P forms significantly longer68 and weaker19
Discussion
Scope of Organolanthanide-Catalyzed Hydrophosphina-
tion/Cyclization. The goal of this research was to explore the
viability and scope of organolanthanide-mediated intramolecular
hydrophosphination/cyclization. The results in Table 2 indicate
that Cp′2LnCH(TMS)2 complexes are effective precatalysts for
the formation of five-membered (phospholane) and six-
membered (phosphorinane) heterocycles with a variety of alkyl
and aryl substituents. The reaction is efficient for both phos-
phinoalkenes and phosphinoalkynes, paralleling previous work
for organolanthanide-catalyzed hydroamination of amino-
alkenes11 and aminoalkynes.12 Although the cyclic products
derived from phosphinoalkyne precursors undergo secondary
reactions, products derived from phosphinoalkene cyclizations
are stable and have potential uses as building blocks for known
and new phosphine ligands.18 The organolanthanide catalysts
remain active, and turnover continues, as long as the reaction
medium remains free of air and moisture. Further addition of
substrate to a completed reaction mixture results in re-com-
mencement of turnover and formation of additional cyclized
product.
Both primary and secondary phosphines undergo cyclization;
however, rates are modest for secondary phenylphosphines.
(63) The broadening of the peaks suggests the dimer may be in
equilibrium with a monomeric species at room temperature.
(64) The most likely cause of the second splitting observed in the 1H
NMR is 3JP-H or 2JP-P coupling, because Y-P-H couplings are expected
to be very small (see ref 55). The observed coupling of 76 Hz is significantly
greater than typical 3JP-H values. However, for coupling across a lanthanide
metal center, two-bond P-Ln-P couplings are reported to be as large as
50-80 Hz (e.g., in structure Ja), so that a large three-bond H-P-Y-P
coupling constant is reasonable. For examples of 3JP-H couplings across a
metal center, see: (a) Grobe, J.; Grosspietsch, T.; Le Van, D.; Krebs, B.;
La¨ge, M. Z. Naturforsch. B 1993, 48, 1203-1211. (b) Fryzuk, M. D.; Joshi,
K.; Chadha, R. K.; Rettig, S. J. J. Am. Chem. Soc. 1991, 113, 8724-8736.
For examples of 2JP-P couplings across a metal center, see: (c) Bleeke, J.
R.; Rohde, A. M.; Robinson, K. D. Organometallics 1995, 14, 1674-1680.
(d) Bleeke, J. R.; Rohde, A. M.; Robinson, K. D. Organometallics 1994,
13, 401-403.
(65) In addition, the structure of 20 confirms the absence of a P-P bond,
because the P-P distance in the crystal is 3.040(3) Å, whereas typical P-P
bond distances are ∼2.2 Å. See Cowley, A. H. Compounds Containing
Phosphorus-Phosphorus Bonds; Dowden, Hutchinson, and Ross, Inc.:
Straudsburg, 1973; pp 1-9.
(66) A Pd-H-P interaction is suggested in the complex [Pd2(µ-PtBu2)(µ-
PHtBu2)(PHtBu2)2]+, see: Leoni, P.; Pasquali, M.; Sommovigo, M.; Laschi,
F.; Zanello, P.; Albinati, A.; Lianza, F.; Pregosin, P. S.; Ru¨egger, H.
Organometallics 1993, 12, 1702-1713.
(67) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183-278.

10236 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
Table 7. Activation Parameter Comparison for
Hydrophosphination and Hydroamination
Scheme 4. Possible Pathways for Insertion and Cyclization
of Phosphinoalkenes
an endocyclic insertive pathway such as that set forth in Scheme
4 cannot be rigorously ruled out. Although the proposed 2,1-
insertion does not appear to be conformationally facile, it might
be facilitated by the greater Ln-P and C-P bond lengths versus
Ln-N and N-C bond lengths, as well as the more constricted
C-P-C versus C-N-C bond angles.69 Although 1,2-additions
are observed in most lanthanide-mediated catalytic processes,
there are examples in which competing 2,1-additions are also
observed (e.g., eq 6).7c The present alternative insertion regio-
chemistry is only operative for substrates 1 and 3 and only in
cases where the “fit” is not optimal70 and is a plausible
explanation for the observed formation of the endocyclic six-
membered ring.
bonds to the metal center than N. That the rate of phosphi-
noalkyne cyclizations increases in the presence of more open
supporting ligand systems (Table 2, entry 5, Cp′2Sm-: Nt )
2.0 (40°) vs CGCSm-: Nt ) 13.5 (22 °C)) is consistent with
the aforementioned ionic radius effects and is again analogous
to aminoalkenes.3a Similar metal-ionic radius and ancillary
ligand effects are also qualitatively observed with other phos-
phinoalkynes, so that the effects observed are self-consistent
within this class of substrates.
Phosphinoalkene cyclizations exhibit different reactivity
trends with metal and ancillary ligands than do those of
phosphinoalkynes. Variation of Ln3+ for Cp′2Ln- catalysts
reveals that the intermediate-sized Y3+ effects the highest
turnover frequencies, with an overall progression Y3+ > Lu3+
> Sm3+ > La3+ (Table 3). This appears to reflect an interplay
of the weaker Ln-P bond vis-a`-vis the Ln-N bond, and also
the lower steric accessibility of the CdC bond versus the CtC
bond. It appears that turnover is most rapid when the “fit”
between metal, substrate, and ligands is optimal. Very low rates
for phosphinoalkene cyclization are also qualitatively observed
with more open ligation, such as CGC ligands and chiral Me2-
Si(Cp′′)(CpR*) ligands. The lower rates with larger metal ions
and more open coordination spheres may also be correlated with
a larger number of coordinated phosphine substrate and product
molecules in the catalyst coordination sphere (i.e., Cp′2Ln(PHR)-
(PH2R)n, where n > 1) while for smaller ions, other phosphine
ligands are less likely to access the catalyst and interfere with
the cyclization. The Cp′2Y- catalyst consistently exhibits the
highest turnover frequencies for all phosphinoalkenes cyclized,
so that once again catalyst effects are constant within a class of
similar substrates.
Another unusual aspect of the 1 f 2 and 3 f 4 transforma-
tions is the occasional but in some cases substantial formation
of an endocyclic six-membered ring (eq 1). This product
becomes kinetically competitive when the substrate-catalyst “fit”
is not optimal, such as for the large La3+ ion and the more open
CGC ligand system, where the hydrophosphination/cyclization
rate is slow. In these specific cases, the competing six-membered
ring formation is significantly more rapid than if there were no
catalyst present, suggesting that an alternate cyclization mech-
anism may be operative. Since all data indicate an irreversible
insertion of the double bond, any reinsertion/skeletal isomer-
ization pathway of 2 to a six-membered ring seems unlikely,
Kinetics and Mechanism of Hydrophosphination/Cycliza-
tion. The cyclizations reported here exhibit initial zero-order
kinetics in substrate concentration (Figure 2) and first-order
kinetics in catalyst concentration (Figure 3). These data imply
that the same basic pathway is operative as was observed for
catalytic hydroamination/cyclization,11-13 with turnover-limiting
insertion of the C-C unsaturation into the Ln-heteroatom bond,
followed by rapid Ln-C protonolysis to afford the product
heterocycle and regenerate the metal-heteroatom bond (Schemes
1, 2). Although, as noted above, the protonolysis of the initial
precatalyst by phosphine substrates is substantially slower than
that observed for amines (step i), this lower protonolytic
reactivity does not affect the observed catalytic cycle and derived
rate law in a major way. The protonolysis of the cyclized
heterocycle Ln-C linkage by substrate (step iii) should be
sterically more facile than initial Ln-CH(TMS)2 protonolysis,
and, judging from the findings for hydroamination, may well
be intramolecular.11a Thus, the insertion of the carbon-carbon
unsaturation into the metal-heteroatom bond (step ii) remains
the turnover-limiting step in the catalytic cycle under typical
reaction conditions.
For structurally similar substrates, the hydrophosphination/
cyclization activation parameters are identical within experi-
mental error to those for aminoalkene hydroamination (Table
7).11a,12b Considering the disparities in the energetics19 of bonds
being broken (D(Sm-PR2) ) 36 kcal/mol vs D(Sm-NR2) )
52 kcal/mol; D(R2P-H) ) 77 kcal/mol vs D(R2N-H) ) 100
(69) sp3-sp3 C-N bonds average 1.47 Å, while P-C bonds average
1.84 Å; the C-P-C bond angles of PMe3 are 98.3°, while the C-N-C
bond angles in NMe3 are 106°; see Gilheany, D. G. The Chemistry of
Organophosphorus Compounds; Hartley, F. R., Ed.; Wiley: Chichester,
1990; Vol. 1, pp 9-49.
(70) It is interesting that this side reaction is not observed for hexenyl
substrates 5 and 7, but not completely surprising since they would form
less kinetically favored seven-membered rings.
(68) In Nd[N(TMS)2]3, for example, the Nd-N bond length is 2.29(2)
Å, see Andersen, R. A.; Templeton, D. H.; Zalkin, A. Inorg. Chem. 1978,
17, 2317-2319. In the similar compound Nd[P(TMS)2]3(THF)2, the Nd-P
bond length is 2.82(3) Å, see Rabe, G. W.; Ziller, J. W. Inorg. Chem. 1995,
34, 5378-5379.

Organolanthanide-Catalyzed Hydrophosphination
J. Am. Chem. Soc., Vol. 123, No. 42, 2001 10237
kcal/mol) and formed (D(R2P-CH3 ) 63 kcal/mol vs D(R2N-
CH3 ) 80 kcal/mol) as the respective hydroamination and
hydrophosphination reaction coordinates are traversed, there are
remarkable overall energetic similarities for the two processes.
For both processes, there is a relatively modest enthalpic barrier,
suggesting a concerted transition state with significant bond
formation compensating for bond-breaking. The large negative
hydrophosphination entropy of activation indicates a highly
ordered transition state, similar to that postulated for hydroami-
nation (T, X ) N), for example, a seven-membered chairlike
conformation in the case of pentenyl phosphines (T, X ) P).
Scheme 5. Effect of Metal Ionic Radius on Phospholane
Ring-Closure Diastereoselectivity
Table 8. Comparison of Organolanthanide-Mediated
Hydroamination and Hydrophosphination
In the present study, cyclization rates for closure of five- and
six-membered phosphorus-containing rings are found to be
similar, instead of observing a clear 5 > 6 trend as in the case
of amines (eqs 7, 8).11,12 Ring-strain differences for five- and
six-membered rings are expected to be small,71 and low-level
AM1 and PM3 calculations show that ∆Hf for homologous
phosphorus- and nitrogen-containing heterocycles are approxi-
mately equal. Therefore, complex and subtle factors in the
transition state for hydrophosphination must play a role in
determining the rates of ring closure versus the size of the ring.
Factors causing the deviation from hydroamination trends are
likely to include differences in N versus P hardness and
softness,42 differences in Ln-N, N-C versus Ln-P, P-C
ground-state bond lengths and angles,69 and differences in
conformational mobility of N versus P heterocycles (phosphorus-
containing heterocycles are generally less mobile).72
Cp′2Sm- catalyst center, the coordination of other phosphine
bases is likely more probable, leading to a more crowded
transition state and forcing the proximate methyl substituent into
an equatorial disposition, thus leading to a trans-rich product
(Scheme 5). The more congested Cp′2Y- and Cp′2Lu- catalysts
centers would discourage participation by other bases, leading
to a less congested transition state, which allows the more
sterically demanding axial orientation of the methyl group and
favors formation of the cis-rich product. A similar steric/
stereochemical effect was also observed in aminoalkene cy-
clizations, and in these studies, increased congestion caused by
addition of other bases (e.g., n-propylamine) led to an increased
trans:cis product ratio.11a Adding n-propylamine to the present
phosphine cyclizations is found to alter the catalyst/substrate
interplay, in that large Ln3+ ions now exhibit less selectivity,
while smaller Ln3+ ions afford increased trans:cis ratios (Table
4). In these cases, added amine likely binds more strongly to
the metal center than the phosphine, perhaps biasing the
cyclizing substrate toward the more favorable equatorial methyl
group orientation. This trans selectivity effect increases as the
metal ionic radius decreases, again indicating the high sensitivity
of the phospholane ring closure stereochemistry to the sur-
rounding steric environment.
Diastereoselectivity of Hydrophosphination/Cyclization.
As noted above, the cyclization 3 f 4 exhibits increasing
product trans-diastereomer selection with increasing Ln3+ ionic
radius (Table 4, entry 2, excluding Cp′2La-mediated reactions
which proceed principally via the Scheme 4 endocyclic path-
way). The increase in the trans ring closure selectivity can be
explained using the same argument used to explain trans:cis
selection processes in formation of trans-rich 2,5-dimethyl-
pyrollidine from 1-methyl-pent-4-enylamine (analogous to the
present 3 f 4 cyclization).10a With the (comparatively) larger
Diastereoselectivity in the cyclization of other phosphine
substrates is much less sensitive to the nature of the organo-
lanthanide catalyst. Selectivity in the 1 f 2 cyclization is
virtually catalyst-independent, indicating (as proposed above)
that the additional methyl group of 3 must exert considerable
influence on stereoselection in the transition state. Closure of
larger rings (5 f 6 and 7 f 8) is also relatively insensitive to
catalyst identity (Table 4), and the same is true for formation
(71) The difference in ring strain between five- and six-membered
saturated hydrocarbon rings is ∼6 kcal/mol, see: (a) Dudev, T.; Lim, C. J.
Am. Chem. Soc. 1998, 120, 4450-4458. (b) Isaacs, N. S. Physical Organic
Chemistry; John Wiley & Sons: New York, 1987, Chapter 8.
(72) For conformational mobilities of heterocyclic rings, see: (a) Gupta,
R. R.; Kumar, M.; Gupta, V. Heterocyclic Chemistry; Springer: Berlin,
1998; Vol. 1, pp 129-146. (b) Lambert, J. B. Top. Stereochem. 1971, 6,
19-105.

10238 J. Am. Chem. Soc., Vol. 123, No. 42, 2001
Douglass et al.
of rings that are sterically more encumbered (13 f 14, 15 f
16, and 17 f 18).
These differences appear to reside in interplay of Lewis (initial
binding to the Ln3+ center) and Brønsted (R2PH species are
more acidic than R2NH)43 acid-base interactions. Both alkenyl
and alkynyl phosphines undergo cyclization, and the mechanism
of the transformations is broadly similar to that of the analogous
hydroaminations. Another similarity is in the maximum turnover
frequencies of the cyclizations, with the rates of phosphine
cyclizations generally of the same order of magnitude as the
analogous hydroamination/cyclizations. Interestingly, phosphi-
noalkene maximum rates are greater than phosphinoalkyne
maximum ratessthe opposite of what is observed for hydroami-
nation.11,12 Thus, the heteroatom identity has a greater overall
effect on the rates of the cyclization process than does the steric
accessibility of the unsaturation. Turnover frequencies vary with
metal-ionic radius and ancillary ligation in understandable ways
for both hydroamination and hydrophosphination. Although
competitive inhibition by cyclized product is more evident for
phosphines than for amines, the experimental rate laws for the
two processes are the same. Despite all of these differences
between phosphorus and nitrogen, the cyclization activation
parameters are surprisingly similar (Table 7). The resting state
of the hydrophosphination catalyst is complicated, and at low
ratios of substrate:catalyst a Ln2P2 dimer is evident, but at typical
catalytic concentrations, the resting state appears to be similar
to that postulated for hydroamination. Desirable trans-selectivity
is at present not as high for the 2,5-dimethyl phospholane (4a)
as for the analogous pyrollidine;11 however, cyclizations with
promising chiral organolanthanide catalysts are far more selec-
tive and will be reported elsewhere.4c
Catalyst Resting State. The catalyst resting state, as surveyed
using paramagnetic catalysts and variable temperature NMR
spectroscopy, is rather complex. The labile Ln-P bonding
affords more rapid exchange of free and bound substrate than
was observed with amines, and this interchange is rapid on the
NMR time scale down to -80 °C. With noncyclizable phos-
phines such as n-pentyl phosphine (19) or phenylphosphine, a
Y-PHR phosphido bond is detectable; however, exchange of
free and bound phosphines remains rapid on the NMR time
scale. In addition, there is significant coordination of the product
heterocycles to the metal center. This coordination is qualita-
tively assessed through broadening of the product P-H reso-
nances in the presence of paramagnetic catalysts, and complex-
ation is directly observed at low temperature during the Cp′2Y-
mediated cyclization of 1 f 2. It is therefore clear that the
cyclized phosphines participate in the catalyst resting state and
perhaps are bound more strongly to the catalyst than linear
substrate at low temperature. At room temperature, free linear
phosphine, bound linear phosphine, and cyclized phosphine all
interchange at rates far in excess of the catalytic turnover
frequencies.
On the basis of the 31P NMR data and the crystal structure
of a species isolated from solution, a µ-PHR dimer (M; Figure
6) can be formed but is only present in significant concentrations
at very low substrate:catalyst ratios, and at most catalytic
concentrations does not make a significant contribution to the
catalyst resting state. As the concentration of phosphine is
increased, phosphine-phosphido complexes predominate, and
the dimer is not present in significant concentrations. From the
present results and those for the analogous organolanthanide-
amine chemistry, the probable resting state of the catalyst under
true catalytic conditions is a mixture of species such as U and
U’.
Conclusions
The present results show that intramolecular hydrophosphi-
nation/cyclization catalyzed by organolanthanides is a general
reaction that can be used to construct a variety of phosphorus-
containing heterocycles. The turnover frequencies and selectivi-
ties are highly dependent on lanthanide ion, ancillary ligands,
and substrate, and trends are in some cases similar to, and in
some cases very different from, those observed for analogous
hydroamination/cyclization processes. A mechanism qualita-
tively similar to that previously defined for amine hydroami-
nation/cyclization appears operative in the present catalytic
systems.
Hydroamination versus Hydrophosphination. Table 8
presents a comparison of organolanthanide-mediated hydroami-
nation11,12 and hydrophosphination processes. Slow generation
of the active catalyst for phosphines versus amines represents
a substantial difference between the two processes and is
surprising when considering that this protonolytic reactions is
more exothermic for phosphines than for amines (eqs 9a, 9b).19
Acknowledgment. Financial support by the NSF (CHE-
0078998) is gratefully acknowledged. M.R.D. thanks Dr. Y. Wu
for assistance with 2D and low temperature NMR experiments.
Supporting Information Available: An X-ray crystal-
lographic file in CIF format for 20. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA010811I