Constrained geometry organoactinides as versatile catalysts for the intramolecular hydroamination/cyclization of primary and secondary amines having diverse tethered C-C unsaturation

Published on Web 03/20/2007

Constrained Geometry Organoactinides as Versatile Catalysts

for the Intramolecular Hydroamination/Cyclization of Primary

and Secondary Amines Having Diverse Tethered

C-C Unsaturation

Bryan D. Stubbert and Tobin J. Marks*

Contribution from the Department of Chemistry, Northwestern UniVersity,

EVanston, Illinois 60208-3113

Received September 14, 2006; E-mail: t-marks@northwestern.edu

Abstract: A series of “constrained geometry” organoactinide complexes, (CGC)An(NMe)₂(CGC ) Me₂-

Si(η⁵-Me₄C₅)(^tBuN); An ) Th, 1; U, 2), has been prepared via efficient in situ, two-step protodeamination

routes in good yields and high purity. Both 1 and 2 are quantitatively converted to the neutrally charged,

solvent-free dichlorides (1-Cl₂, 2-Cl₂) and slightly more soluble diiodides (1-I₂, 2-I₂) with excess Me₃Si-X

(X ) Cl, I) in non-coordinating solvents. The new complexes were characterized by NMR spectroscopy,

elemental analysis, and (for 1 and 2) single-crystal X-ray diffraction, revealing substantially increased metal

coordinative unsaturation vs the corresponding Me₂SiCp′′₂AnR₂(Cp′′ ) η⁵-Me₄C₅; An ) Th, R ) CH₂-

(SiMe₃), 3; An ) U, R ) CH₂Ph, 4) and Cp′₂AnR₂(Cp′ ) η⁵-Me₅C₅; An ) Th, R ) CH₂(SiMe₃), 5; An )

U, R ) CH₂(SiMe₃), 6) complexes. Complexes 1-6 exhibit broad applicability for the intramolecular

hydroamination of diverse C-C unsaturations, including terminal and internal aminoalkenes (primary and

secondary amines), aminoalkynes (primary and secondary amines), aminoallenes, and aminodienes. Large

turnover frequencies (N_tup to 3000 h^-1) and high regioselectivities (g95%) are observed throughout, along

with moderate to high diastereoselectivities (up to 90% trans ring closures). With several noteworthy

exceptions, reactivity trends track relative 5f ionic radii and ancillary ligand coordinative unsaturation.

Reactivity patterns and activation parameters are consistent with a reaction pathway proceeding via turnover-

limiting CdC/CtC insertion into the An-N σ-bond.

thanides,^3,4current intramolecular HA/cyclization research

activity is wide-ranging²and spans the entire Periodic Table,^5,6

Introduction

The regioselective formation of C-N bonds is an important

transformation in chemistry and biology, attracting considerable

interest in both academic and industrial research.¹Hydroami-

nation (HA), defined as the formal addition of an N-H bond

across a unit of C-C unsaturation,²is an atom-economical route

to constructing both simple and complex organonitrogen

skeletons and is capable of doing so efficiently and with high

selectivity. Initially dominated by studies involving lan-

from Ca⁷to the coinage metals.⁸While late transition metal

catalysts generally enjoy greater functional group tolerance, they

often require acidic conditions and N-protected amines and may

(3) (a) Seyam, A. M.; Stubbert, B. D.; Jensen, T. R.; O’Donnell, J. J., III;

Stern, C. L.; Marks, T. J. Inorg. Chim. Acta 2004, 357, 4029-4035. (b)

Ryu, J.-S.; Marks, T. J.; McDonald, F. E. J. Org. Chem. 2004, 69, 1038-

1052. (c) Hong, S.; Kawaoka, A. M.; Marks, T. J. J. Am. Chem. Soc. 2003,

125, 15878-15892. (d) Hong, S.; Tian, S.; Metz, M. V.; Marks, T. J.

J. Am. Chem. Soc. 2003, 125, 14768-14783. (e) Ryu, J.-S.; Li, G. Y.;

Marks, T. J. J. Am. Chem. Soc. 2003, 125, 12584-12605. (f) Hong, S.;

Marks, T. J. J. Am. Chem. Soc. 2002, 124, 7886-7887. (g) Ryu, J.-S.;

Marks, T. J.; McDonald, F. E. Org. Lett. 2001, 3, 3091-3094. (h)

Arredondo, V. M.; Tian, S.; McDonald, F. E.; Marks, T. J. J. Am. Chem.

Soc. 1999, 121, 3633-3639. (i) Arredondo, V. M.; McDonald, F. E.; Marks,

T. J. Organometallics 1999, 18, 1949-1960. (j) Tian, S.; Arredondo,

V. M.; Stern, C. L.; Marks, T. J. Organometallics 1999, 18, 2568-2570.

(k) Arredondo, V. M.; McDonald, F. E.; Marks, T. J. J. Am. Chem. Soc.

1998, 120, 4871-4872. (l) Li, Y.; Marks, T. J. J. Am. Chem. Soc. 1998,

120, 1757-1771. (m) Roesky, P. W.; Stern, C. L.; Marks, T. J. Organo-

metallics 1997, 16, 4705-4711. (n) Li, Y.; Marks, T. J. J. Am. Chem. Soc.

1996, 118, 9295-9306. (o) Li, Y.; Marks, T. J. Organometallics 1996,

15, 3770-3772. (p) Li, Y.; Marks, T. J. J. Am. Chem. Soc. 1996, 118,

707-708. (q) Giardello, M. A.; Conticello, V. P.; Brard, L.; Gagne, M.

R.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10241-10254. (r) Giardello,

M. A.; Conticello, V. P.; Brard, L.; Sabat, M.; Rheingold, A. L.; Stern,

C. L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10212-10240. (s) Li,

Y.; Fu, P.-F.; Marks, T. J. Organometallics 1994, 13, 439-440. (t) Gagne,

M. R.; Brard, L.; Conticello, V. P.; Giardello, M. A.; Stern, C. L.; Marks,

T. J. J. Am. Chem. Soc. 1992, 114, 2003-2005. (u) Gagne, M. R.; Stern,

C. L.; Marks, T. J. J. Am. Chem. Soc. 1992, 114, 275-294. (v) Gagne,

M. R.; Marks, T. J. J. Am. Chem. Soc. 1989, 111, 4108-4109.

(1) For general C-N bond-formation references, see: (a) Beller, M.; Seayad,

J.; Tillack, A.; Jiao, H. Angew. Chem., Int. Ed. 2004, 43, 3368-3398. (b)

Hartwig, J. F. Science 2002, 297, 1653-1654. (c) Beller, M.; Riermeier,

T. H. In Transition Metals for Organic Synthesis, 1998, 1, 184-194. (d)

Hegedus, L. S. Angew. Chem., Int. Ed. 1988, 27, 1113-1126.

(2) For general hydroamination reviews, see: (a) Odom, A. L. Dalton Trans.

2005, 225-233. (b) Hultzsch, K. C. AdV. Synth. Catal. 2005, 347, 367-

391. (c) Hultzsch, K. C.; Gribkov, D. V.; Hampel, F. J. Organomet. Chem.

2005, 690, 4441-4452. (d) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004,

37, 673-686. (e) Doye, S. Synlett 2004, 1653-1672. (f) Beller, M.; Tillack,

A.; Seayad, J. In Transition Metals for Organic Synthesis, 2nd ed.; Beller,

M., Bohm, C., Eds.; Wiley-VCH: Weinheim, 2004; pp 91-94. (g) Roesky,

P. W.; Mueller, T. E. Angew. Chem., Int. Ed. 2003, 42, 2708-2710. (h)

Pohlki, F.; Doye, S. Chem. Soc. ReV. 2003, 32, 104-114. (i) Bytschkov,

I.; Doye, S. Eur. J. Org. Chem. 2003, 2003, 935-946. (j) Seayad, J.; Tillack,

A.; Hartung, C. G.; Beller, M. AdV. Synth. Catal. 2002, 344, 795-813. (k)

Togni, A. In Catalytic Heterofunctionalization, 1st ed.; Togni, A., Gru-

etzmacher, H., Eds.; Wiley-VCH: New York, 2001; pp 91-141. (l) Nobis,

M.; Driessen-Holscher, B. Angew. Chem., Int. Ed. 2001, 40, 3983-3985.

(m) Eisen, M. S.; Straub, T.; Haskel, A. J. Alloys Compd. 1998, 271-273,

116-122. (n) Mueller, T. E.; Beller, M. Chem. ReV. 1998, 98, 675-703.

9

J. AM. CHEM. SOC. 2007, 129, 4253-4271

4253

A R T I C L E S

Stubbert and Marks

Scheme 1. Scope of Intramolecular HA/Cyclization Mediated by Organo-4f-Element Catalysts (See Refs 2d, 3, and 4)

be further plagued by low efficiency and short catalyst lifetimes.²

Early transition metal catalysts typically exhibit enhanced

activity; however, with one noteworthy exception,⁵ⁱthese

electrophilic catalysts normally have limited functional group

tolerance and sluggish reaction rates vs larger, more electrophilic

organolanthanide catalysts.^2-5

Beginning with terminal aminoalkenes,^3u,vintramolecular

organolanthanide-catalyzed HA/cyclization processes have been

extensively investigated (Scheme 1),^2d,3,4typically displaying

near-quantitative yields in addition to high regio- and diaste-

reoselectivities (>95%) in addition to moderate/good enanti-

oselectivities (up to 95%).^{2b-d,3b-d,q,r,4b,d,f,i-o,9}The highly exo-

thermic intramolecular HA/cyclization of simple aminoalkynes

proceeds with surprising ease for both terminal and 1,2-

disubstituted alkynes when catalyzed by organolanthanide com-

plexes.^3n,sFurthermore, single amines tethered to CdC and CtC

linkages undergo cascade reactions involving sequential C-N/

C-C bond fusions with high regioselection to afford indolizi-

dine, pyrrole, pyrrolizidine, and pyrazine scaffolds, while

(4) For additional examples of group 3- and lanthanide-catalyzed HA, see:

(a) Bambirra, S.; Tsurugi, H.; van Leusen, D.; Hessen, B. Dalton Trans.

2006, 1157-1161. (b) Gribkov, D. V.; Hultzsch, K. C.; Hampel, F. J. Am.

Chem. Soc. 2006, 128, 3748-3759. (c) Panda, T. K.; Zulys, A.; Gamer,

M. T.; Roesky, P. W. Organometallics 2005, 24, 2197-2202. (d) Collin,

J.; Daran, J.-C.; Jacquet, O.; Schulz, E.; Trifonov, A. Chem. Eur. J. 2005,

11, 3455-3462. (e) Kim, J. Y.; Livinghouse, T. Org. Lett. 2005, 7, 4391-

4393. (f) Kim, J. Y.; Livinghouse, T. Org. Lett. 2005, 7, 1737-1739. (g)

Molander, G. A.; Hasegawa, H. Heterocycles 2004, 64, 467-474. (h)

Lauterwasser, F.; Hayes, P. G.; Brase, S.; Piers, W. E.; Schafer, L. L.

Organometallics 2004, 23, 2234-2237. (i) Hultzsch, K. C.; Hampel, F.;

Wagner, T. Organometallics 2004, 23, 2601-2612. (j) O’Shaughnessy,

P. N.; Gillespie, K. M.; Knight, P. D.; Munslow, I. J.; Scott, P. Dalton

Trans. 2004, 2251-2256. (k) O’Shaughnessy, P. N.; Scott, P. Tetrahe-

dron: Asymmetry 2003, 14, 1979-1983. (l) O’Shaughnessy, P. N.; Knight,

P. D.; Morton, C.; Gillespie, K. M.; Scott, P. Chem. Commun. 2003 1770-

1771. (m) Kim, Y. K.; Livinghouse, T.; Horino, Y. J. Am. Chem. Soc.

2003, 125, 9560-9561. (n) Molander, G. A.; Pack, S. K. Tetrahedron 2003,

59, 10581-10591. (o) Molander, G. A.; Pack, S. K. J. Org. Chem. 2003,

68, 9214-9220. (p) Kim, Y. K.; Livinghouse, T. Angew. Chem., Int. Ed.

2002, 41, 3645-3647. (q) Bu¨rgstein, M. R.; Berberich, H.; Roesky, P. W.

Chem. Eur. J. 2001, 7, 3078-3085. (r) Kim, Y. K.; Livinghouse, T.;

Bercaw, J. E. Tetrahedron Lett. 2001, 42, 2933-2935. (s) Molander,

G. A.; Dowdy, E. D. J. Org. Chem. 1999, 64, 6515-6517. (t) Molander,

G. A.; Dowdy, E. D. J. Org. Chem. 1998, 63, 8983-8988.

(5) For group 4 catalysts see: (a) Thomson, R. K.; Bexrud, J. A.; Schafer,

L. L. Organometallics 2006, 25, 4069-4071. (b) Lee, A. V.; Schafer,

L. L. Organometallics 2006, 25, 5249-5254. (c) Esteruelas, M. A.; Lopez,

A. M.; Mateo, A. C.; Onate, E. Organometallics 2006, 25, 1448-1460.

(d) Bexrud, J. A.; Beard, J. D.; Leitch, D. C.; Schafer, L. L. Org. Lett.

2005, 7, 1959-1962. (e) Esteruelas, M. A.; Lopez, A. M.; Mateo, A. C.;

Onate, E. Organometallics 2005, 24, 5084-5094. (f) Kim, H.; Lee, P. H.;

Livinghouse, T. Chem. Commun. 2005, 5205-5207. (g) Hoover, J. M.;

Petersen, J. R.; Pikul, J. H.; Johnson, A. R. Organometallics 2004, 23,

4614-4620. (h) Gribkov, D. V.; Hultzsch, K. C. Angew. Chem., Int. Ed.

2004, 43, 5542-5546. (i) Knight, P. D.; Munslow, I.; O’Shaughnessy,

P. N.; Scott, P. Chem. Commun. 2004, 894-895. (j) Li, C.; Thompson,

R. K.; Gillon, B.; Patrick, B. O.; Schafer, L. L. Chem. Commun. 2003,

2462-2463. (k) Ackermann, L.; Bergman, R. G. Org. Lett. 2002, 4, 1475-

1478. (l) Ackermann, L.; Bergman, R. G.; Loy, R. N. J. Am. Chem. Soc.

2003, 125, 11956-11963.

selectively retaining unsaturations amenable to subsequent

functionalization.^{3j,l,p,n,4o,p}Initial difficulties in directly adding

N-H bonds across 1,2-disubstituted CdC linkages^3i,u,4s,twere

overcome via implementation of more open, thermally robust

organolanthanide structures at high temperatures.^3b,h,4oEn route

to such advances, heterocyclic skeletons of greater structural

complexity were accessed from allenic^3h,i,kand dienic^3c,d,f

substrates. In all cases, the tunable ionic radii of the lanthanide

series (Table 1)¹⁰proved vital to understanding those kinetic

(6) See for example: (a) Krogstad, D. A.; Cho, J.; DeBoer, A. J.; Klitzke,

J. A.; Sanow, W. R.; Williams, H. A.; Halfen, J. A. Inorg. Chim. Acta

2006, 359, 136-148. (b) Zhang, J.; Yang, C.-G.; He, C. J. Am. Chem.

Soc. 2006, 128, 1798-1799. (c) Kadzimirsz, D.; Hildebrandt, D.; Merz,

K.; Dyker, G. Chem. Commun. 2006, 661-662. (d) Bajracharya, G. B.;

Huo, Z.; Yamamoto, Y. J. Org. Chem. 2005, 70, 4883-4886. (e) Field,

L. D.; Messerle, B. A.; Vuong, K. Q.; Turner, P. Organometallics 2005,

24, 4241-4250. (f) Bender, C. F.; Widenhoefer, R. A. J. Am. Chem. Soc.

2005, 127, 1070-1071. (g) Zulys, A.; Dochnahl, M.; Hollmann, D.;

Lo¨hnwitz, K.; Herrmann, J.-S.; Roesky, P. W.; Blechert, S. Angew. Chem.

Int. Ed. 2005, 44, 7794-7798. (h) Lutete, L. M.; Kadota, I.; Yamamoto,

Y. J. Am. Chem. Soc. 2004, 126, 1622-1623. (i) Schlummer, B.; Hartwig,

J. F. Org. Lett. 2002, 4, 1471-1474.

(7) Crimmin, M. R.; Casely, I. J.; Hill, M. S. J. Am. Chem. Soc. 2005, 127,

2042-2043.

(8) (a) Kadzimirsz, D.; Hildebrandt, D.; Merz, K.; Dyker, G. Chem. Commun.

2006, 661-662. (b) Zhang, J.; Yang, C.-G.; He, C. J. Am. Chem. Soc.

2006, 128, 1798-1799. (c) Muller, T. E.; Grosche, M.; Herdtweck, E.;

Pleier, A.-K.; Walter, E.; Yan, Y.-K. Organometallics 2000, 19, 170-

183.

(9) For reviews of enantioselective catalysis with lanthanides, see: (a) Aspinall,

H. C. Chem. ReV. 2002, 102, 1807-1850. (b) Inanaga, J.; Furuno, H.;

Hayano, T. Chem. ReV. 2002, 102, 2211-2225. (c) Molander, G. A. Pure

Appl. Chem. 2000, 72, 1757-1761.

(10) Shannon, R. D. Acta Crystallogr. 1976, A32, 751-67.

9

4254 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

Table 1. Comparison of Ionic Radii for Relevant Lanthanide and

Actinide Ions (See Ref 10)

In light of the broad scope of substrates amenable to intramo-

lecular HA/cyclization catalyzed by organolanthanides^2d,3,4along

with the intriguing similarities and differences between 4f- and

5f-element organometallic chemistries,¹²the investigation of

organoactinide HA/cyclization catalysts having increased co-

ordinative unsaturation, along with an additional covalently

bonded ligand (vs Ln), offers an ideal opportunity to examine

the roles of bond covalency, ionic radius, metal oxidation state,

and ancillary ligation, as well as the mechanistic significance

of these factors.

ion

coordination no.

ionic radius (Å)

Th⁴⁺

U⁴⁺

9

8

1.09

1.05

1.16

1.079

1.019

0.977

La³⁺

Sm³⁺

Y³⁺

Lu³⁺

and mechanistic factors governing HA ring closure. Furthermore,

the possibility of broadening the scope of this powerful C-N

bond-forming methodology by tuning ancillary ligand archi-

tecture motivated the exploratory synthesis of additional catalyst

classes. With open coordination spheres having documented

applications throughout the Periodic Table,¹¹the “constrained

geometry” ligand (“CGC”) was applied to organolanthanides.

The resulting catalysts are particularly efficacious for activating

sterically encumbered and otherwise less-reactive amine sub-

strates,especiallywhencombinedwithlargerradiuslanthanides.^3a-j

We previously communicated the synthesis and characteriza-

tion of several members of a new class of CGC organoactinide

complexes, along with initial investigations of their catalytic

competence for intramolecular HA/cyclization.¹⁵From these

preliminary results, we envisioned a mechanistic pathway as

depicted in Scheme 2, with an insertive transition state such as

A. In this contribution, we report improved syntheses of

Actinide (An) and lanthanide (Ln) chemistries differ primarily

in the degree of covalency and extent of f-orbital participation

in bonding and reactivity.^12,13Unlike the poorly shielded and

primarily nonbonding Ln 4f orbitals, limited involvement of

An 6d and 5f orbitals has been invoked to explain observed

molecular geometries and stabilities, as well as unique chemical

reactivity patterns,¹⁴and has been ascribed to the increased

importance of relativistic effects^12a,h-land improved shielding

of the 5f orbitals. Although 6d and 5f orbital participation is

still relatiVely modest, the degree of bond covalency in actinides

is significantly enhanced over that in lanthanides,¹²presenting

unique possibilities for new and interesting reactivity modalities.

(CGC)An complexes and a detailed investigation of their

catalytic intramolecular HA/cyclization activity with respect to

a broad series of representative substrates. We compare/contrast

with less-open organoactinide complexes and discuss these

results relative to electrophilic organolanthanide and group 4

metal catalysts. The scope of organoactinide-catalyzed HA/

cyclization established includes 1,2-disubstituted alkenes, al-

lenes, and diene C-C unsaturations tethered to primary (1°)

amines. Furthermore, we show that secondary (2°) aminoalkenes

and aminoalkynes are rapidly converted to tertiary (3°) pyrro-

lidine and enamine heterocycles, requiring the agency of an

M-N σ-bonded intermediate in the insertive transition state (A,

Scheme 2). In a companion contribution, we present a detailed

mechanistic investigation of (CGC)An -mediated HA.¹⁶

(11) For reviews of constrained geometry catalysts, see: (a) Gromada, J.;

Carpentier, J.-F.; Mortreux, A. Coord. Chem. ReV. 2004, 248, 397-410.

(b) Okuda, J. Dalton Trans. 2003, 2367-2378. (c) Arndt, S.; Okuda, J.

Chem. ReV. 2002, 102, 1953-1976. (d) Britovsek, G. J. P.; Gibson, V. C.;

Wass, D. F. Angew. Chem., Int. Ed. 1999, 38, 428-447.

(12) For reviews of 4f- and 5f-element chemistry, see: (a) Burns, C. J.; Neu,

M. P.; Boukhalfa, H.; Gutowski, K. E.; Bridges, N. J.; Rogers, R. D. In

ComprehensiVe Coordination Chemistry II, 1st ed.; McCleverty, J. A.,

Meyer, T. J., Eds.; Elsevier Pergamon: Amsterdam, 2004; Vol. 3, pp 189-

345. (b) Edelmann, F. T.; Freckmann, D. M. M.; Schumann, H. Chem.

ReV. 2002, 102, 1851-1896. (c) Molander, G. A.; Dowdy, E. D. Top.

Organomet. Chem. 1999, 2, 119-154. (d) Edelmann, F. T. Angew. Chem.,

Int. Ed. 1995, 34, 2466-88. (e) Edelmann, F. T. In ComprehensiVe

Organometallic Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G.,

Eds.; Pergamon: New York, 1995; Vol. 4, pp 11-213. (f) Schaverien,

C. J. AdV. Organomet. Chem. 1994, 36, 283-362. (g) Molander, G. A.

Chem. ReV. 1992, 92, 29-68. (h) The Chemistry of the Actinide Elements;

Katz, J. J., Seaborg, G. T., Morss, L. R., Eds.; Chapman and Hall: New

York, 1986. (i) Fundamental and Technological Aspects of Organo-f-

Element Chemistry; Marks, T. J., Fragala`, I. L., Eds.; Reidel: Dordrecht,

1985. (j) Marks, T. J.; Ernst, R. D. In ComprehensiVe Organometallic

Chemistry; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon:

New York, 1982; Vol. 3, pp 173-270. (k) Marks, T. J. Chemistry and

Spectroscopy of f Element Organometallics, Part II. The Actinides. Progress

in Inorganic Chemistry 1979, 25, 224. (l) Marks, T. J. Chemistry and

Spectroscopy of f Element Organometallics, Part I. The Lanthanides.

Progress in Inorganic Chemistry 1978, 24, 51.

Experimental Section

General Considerations. 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, interfaced to a high-vacuum manifold (10^-6Torr), or in

(14) See, for example: (a) Barnea, E.; Eisen, M. S. Coord. Chem. ReV. 2006,

250, 855-899. (b) Burns, C. J. Science 2005, 309, 1823-1824. (c) Evans,

W. J.; Kozimor, S. A.; Ziller, J. W. Science 2005, 309, 1835-1838. (d)

Castro-Rodriguez, I.; Meyer, K. J. Am. Chem. Soc. 2005, 127, 11242-

11243. (e) Evans, W. J.; Kozimor, S. A.; Ziller, J. W.; Kaltsoyannis, N.

J. Am. Chem. Soc. 2004, 126, 14533-14547. (f) Castro-Rodriguez, I.;

Nakai, H.; Zakharov, L. N.; Rheingold, A. L.; Meyer, K. Science 2004,

305, 1757-1760. (g) Castro-Rodriguez, I.; Nakai, H.; Gantzel, P.; Zakharov,

L. N.; Rheingold, A. L.; Meyer, K. J. Am. Chem. Soc. 2003, 125, 15734-

15735. (h) Diaconescu, P. L.; Arnold, P. L.; Baker, T. A.; Mindiola, D. J.;

Cummins, C. C. J. Am. Chem. Soc. 2000, 122, 6108-6109.

(13) (a) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. AdVanced

Inorganic Chemistry, 6th ed.; John Wiley & Sons, Inc.: New York, 1999.

(b) Kettle, S. F. A. Physical Inorganic Chemistry: A Coordination

Chemistry Approach; Oxford University Press: New York, 1998. (c)

Huheey, J. E.; Keiter, E. A.; Keiter, R. L. Inorganic Chemistry: Principles

of Structure and ReactiVity, 4th ed.; HarperCollins College Publishers: New

York, 1993.

(15) Stubbert, B. D.; Stern, C. L.; Marks, T. J. Organometallics 2003, 22, 4836-

4838.

(16) Stubbert, B. D.; Marks, T. J. J. Am. Chem. Soc. 2007, 129, in press.

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4255

A R T I C L E S

Stubbert and Marks

Scheme 2. Proposed Mechanism for the Intramolecular HA/Cyclization of C-C Unsaturations Tethered to 1° and 2° Amines, Including

a

Alkene, Alkyne, Allene, and Diene Substrates under Mild Conditions in C₆D₆or C₇D₈

^aThis mechanistic scenario is consistent with the observed rate law, ν ∼ [(CGC)An]¹[substrate]⁰, and is similar to the σ-bond insertive pathway proceeding

through an L₂Ln-N(H)R intermediate proposed for organolanthanide-catalyzed HA/cyclization processes (see refs 2-4).

a nitrogen-filled Vacuum Atmospheres glovebox with a high-capacity

recirculator (<2 ppm of O₂). Argon (Matheson, prepurified) was

purified by passage through a MnO oxygen-removal column and a

Davison 4A molecular sieve column. All solvents were dried and

degassed over Na/K alloy and transferred in vacuo immediately prior

to use. All substrates for catalytic experiments were dried at least three

times as solutions in benzene-d₆or toluene-d₈over freshly activated

Davison 4A molecular sieves and were degassed by repeated freeze-

pump-thaw cycles. Aminoalkynes were predried by stirring over BaO

overnight as C₆D₆solutions before rigorous drying over freshly activated

molecular sieves. Substrates were then stored under Ar in vacuum-

tight storage flasks. Full experimental details for improved syntheses

of substrates 21, 25, and 41, as well as 1-X₂and 2-X₂, are provided in

the Supporting Information.

off-white solid that was further dried under high vacuum to afford 1.895

g of a roughly 1:1 mixture of Th(NMe₂)₄and Li[Th(NMe₂)₅](THF)_n,

as determined by ¹H NMR spectroscopy (ca. 4.64 mmol of

“Th(NMe₂)₄”, 35% yield).¹⁷

A 100 mL round-bottom flask in the glovebox containing the isolated

mixture (homoleptic amide + ate complex), a magnetic stir bar, and a

high-vacuum adapter was charged with pentane (15 mL) and H₂CGC

(1.18 g, 4.67 mmol) at 20 °C. The flask was moved to the high-vacuum

manifold, where volatile byproduct HNMe₂was periodically removed

via Ar flush and/or partial evacuation over 3 d. Complete evacuation

of volatiles, followed by condensation of fresh solvent (either pentane

or toluene) into the reaction flask at -196 °C, was carried out once

every 12 h. Reaction progress can be monitored by ¹H NMR

spectroscopy. The product was next isolated by extraction with pentane

(2 × 25 mL) and filtration. The solution was concentrated to a volume

of ca. 10 mL, and 1 was isolated by slow cooling to -78 °C. Collection

of the colorless powder on a glass frit afforded 1.94 g of 1 (74% yield

based on “Th(NMe₂)₄”). ¹H NMR (C₆D₆, 500 MHz, 25 °C): δ 2.81 (s,

12 H, N(CH₃)₂), 2.20 (s, 6 H, C₅(CH₃)₄), 2.02 (s, 6 H, C₅(CH₃)₄), 1.30

(s, 9 H, NC(CH₃)₃), 0.67 (s, 6 H, Si(CH₃)₂). Anal. Calcd: C, 40.06; H,

6.90; N, 7.38. Found: C, 39.84; H, 6.90; N, 7.35.

(CGC)U(NMe₂)₂(2). The one-pot preparative route to (CGC)An(NR₂)₂

complexes is detailed here for complex 2. A 250 mL sidearm flask

with a magnetic stir bar was charged in the glovebox with UCl₄(3.00

g, 7.90 mmol) and LiNMe₂(1.57 g, 30.8 mmol, 3.9 equiv). The flask

was evacuated under high vacuum before pentane (20 mL) and THF

(80 mL) were condensed onto the solids at -78 °C. The dark red

mixture was protected from light and allowed to warm slowly with

stirring to 20 °C under Ar.¹⁷After 20 h, the volatiles were removed in

vacuo and replaced with pentane (80 mL) at -78 °C. The mixture was

warmed to 20 °C, and a solution of H₂CGC (2.35 g, 9.34 mmol, 1.18

equiv) in pentane (5 mL) was added by syringe through the sidearm

Physical and Analytical Measurements. NMR spectra were

recorded on a ^UNITYInova-500 (FT, 500 MHz, ¹H; 125 MHz, ¹³C)

1

instrument. Chemical shifts (δ) for H and ¹³C spectra are referenced

to internal solvent resonances and reported relative to SiMe₄. NMR

experiments on air-sensitive samples were conducted in Teflon valve-

sealed tubes (J. Young). EI-HRMS data were obtained using a Thermo

Finnigan MAT-XL900 spectrometer at 70 eV. Elemental analyses on

air-sensitive samples were performed at the Micro-Mass Facility at the

University of California, Berkeley; air-stable samples were analyzed

at Midwest Microlabs (Indianapolis, IN).

(CGC)Th(NMe₂)₂(1). A small flip-frit apparatus with a magnetic

stir bar was charged in the glovebox with ThCl₄(5.00 g, 13.4 mmol)

and LiNMe₂(2.65 g, 52.0 mmol, 3.9 equiv). After the system was

evacuated under high vacuum, pentane (10 mL) and THF (60 mL) were

condensed onto the solids at -196 °C. The stirring mixture was allowed

to slowly warm to 20 °C under Ar and allowed to react for an additional

24 h before all volatiles were removed in vacuo. Pentane (30 mL) was

condensed onto the resulting oily residue at -78 °C and then allowed

to warm to 20 °C under Ar. After stirring overnight, the pale-yellow

pentane solution was filtered and concentrated in vacuo to afford an

(17) Jamerson, J. D. Ph.D. Thesis, University of Alberta, 1974.

9

4256 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

the tube was evacuated and backfilled with Ar three times, the valve

was closed. The sample remained frozen in the -78 °C bath until it

was thawed and placed directly into the Varian ^UNITYInova 500 MHz

NMR spectrometer, pre-equilibrated at 25.0 °C ((0.2 °C), calibrated

with a neat ethylene glycol standard. Data were obtained with only a

single transient to avoid signal saturation, and substrate and/or product

concentrations were determined relative to the intensity of an internal

standard resonance, typically downfield of C₆D₅H in the aromatic region

of Si(p-tolyl)₄(7.74 ppm), over three or more half-lives. All data were

fit by least-squares analysis according to eq 1, where C₀is the initial

concentration of substrate and t is time. The turnover frequency (N_t)

was calculated according to eq 2, where E is the normalized substrate:

catalyst ratio determined from the relative intensities of non-overlapping

resonances in each ¹H NMR spectrum, typically with (5% precision.

C ) mt + C₀

(1)

(2)

m

N_t)

E

(

)

- C₀

Results

In this section we begin with an account of (CGC)AnX₂

synthesis, molecular structures, and reaction chemistry. We then

discuss the intramolecular HA/cyclization of representative

terminal and internal aminoalkene, aminoalkyne, aminoallene,

and aminodiene substrates mediated by constrained geometry

organoactinide complexes, as well as by other organoactinides

of progressively less coordinative unsaturation. Catalytic reac-

tions with terminal and 1,2-disubstituted 1° and 2° aminoalkenes

are discussed in detail. Intramolecular terminal and 1,2-

disubstituted alkyne HA/cyclization with both 1° and 2° amines

is then described, followed by extension of the reaction scope

to an illustrative selection of aminoallenes and aminodienes.

The effects of both metal ionic radius and ancillary ligation are

analyzed in each instance, along with relevant kinetic and

activation parameters for ring closure involving each C-C

unsaturation.

Figure 1. Solid-state structures of (CGC)An(NMe₂)₂complexes and

selected bond distances (Å) and angles (°). (a) Complex 1, distances: Cp′′-

(centroid)-Th, 2.498; Th-N1, 2.315(4); Th-N2, 2.266(5); Th-N3, 2.273-

(5). Angles: N2-Th-N3, 107.6(2); N1-Th-N2, 114.0(2); N1-Th-N3,

112.8(2). (b) Complex 2, distances: Cp′′(centroid)-U, 2.425; U-N1, 2.278-

(4); U-N2, 2.207(4); U-N3, 2.212(4). Angles: N2-U-N3, 109.6(2); N1-

U-N2, 111.0(2); N1-U-N3, 114.2(2). Ellipsoids are drawn at 50%

probability (see ref 15).

(CGC)AnX₂Synthesis and Characterization. Constrained

geometry organoactinide complexes 1 and 2 are prepared in

good yields (ca. 70%) and high purity from AnCl₄reagents via

an improved, general procedure using the corresponding in situ

generated homoleptic tetrakis(dialkylamido)actinide complexes,

followed by protodeaminative substitution with the diprotic

H₂CGC proligand (eq 3). NMR-scale syntheses proceed in

quantitative yield and in >95% purity (¹H NMR spectroscopy),

indicating that the yield-limiting step in (CGC)An(NR₂)₂

preparation is precipitation and crystallization from alkane or

aromatic solvents. Yields are optimized by incorporating a slight

stoichiometric excess of H₂CGC, counteracting the formation

of ate complex contaminants (i.e., Li[An(NMe₂)₅(THF)_n],

An ) Th, U) and driving the product-forming equilibrium to

the right (eq 3). Pentane-insoluble impurities (primarily residual

LiCl) are removed by filtration prior to -78 °C recrystallization

to afford analytically pure (CGC)An(NMe₂)₂complexes 1 and

2. The absence of coordinated Lewis bases in the final products

allows direct application as precatalysts without further

purification.^3a,u,18The solid-state structures of 1 and 2, deter-

mined by single-crystal X-ray diffraction (Figure 1), are

under a rapid Ar purge. Byproduct HNMe₂was periodically removed

via Ar flow and/or partial evacuation over 2 d. Complete evacuation

of volatiles, followed by condensation of fresh pentane into the flask

at -196 °C, was carried out once every 6-12 h. Reaction progress

1

can be monitored by H NMR spectroscopy. Product 2 was isolated

and purified by filtration, followed by concentration to a volume of

<10 mL and slow cooling to -78 °C. Decantation of the mother liquor

afforded 3.20 g of brown microcrystalline 2 in 71% yield (from UCl₄).

¹H NMR (C₆D₆, 500 MHz, 25 °C): δ 22.24 (s, 12 H, N(CH₃)₂), 15.28

(s, 6 H, C₅(CH₃)₄), 1.31 (s, 6 H, C₅(CH₃)₄), -18.93 (s, 9 H, NC(CH₃)₃),

-22.08 (s, 6 H, Si(CH₃)₂). Anal. Calcd: C, 39.64; H, 6.83; N, 7.30.

Found: C, 38.37; H, 6.75; N, 7.27.

Kinetic Studies of Hydroamination Reactions. In a typical

experiment, the precatalyst (ca. 1-6 mg, 2-11 µmol) and the internal

standard Si(p-tolyl)₄(ca. 2-5 mg, 5-12 µmol) were dissolved in C₆D₆

(0.80 mL) and added to a Teflon-valved NMR tube in the glovebox. A

preliminary ¹H NMR spectrum was recorded to verify the relative

concentrations of precatalyst and standard. The NMR tube was then

attached to a high-vacuum manifold via a glass adapter with a secondary

Teflon valve and was placed in a -78 °C bath before being evacuated

to ca. 5 × 10^-6Torr and resealed. The upper portion of the valve was

then backfilled with Ar, and a 1 M solution of substrate in C₆D₆(0.20

mL) was added to the adapter under a flush of Ar. The primary valve

was next opened and the substrate solution pulled into the NMR tube,

freezing the substrate solution just above the precatalyst solution. After

(18) Gillespie, R. D.; Burwell, R. L., Jr.; Marks, T. J. Langmuir 1990, 6, 1465-

1477.

(19) EIMS (M^+•) data show fragments of m/z > MW_monomer, suggesting that

(CGC)AnX₂are dimeric (or trimeric) species. See Supporting Information

for experimental details.

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4257

A R T I C L E S

Stubbert and Marks

Table 2. Aminoalkene HA/Cyclization Processes Catalyzed by

(CGC)An Complexes 1 and 2 in C₆D₆Showing the Effects of

Ionic Radius on N_t

discussed below in the context of their substantially more open

ligand frameworks and the impact on catalytic properties (vide

infra).

Neutrally charged Lewis base-free halide complexes 1-Cl₂,

2-Cl₂, 1-I₂, and 2-I₂are cleanly prepared by reaction of 1 or 2

with a 6-10-fold excess of Me₃Si-X (X ) Cl, I; eq 4) in non-

coordinating solvents. The dihalides are isolated cleanly and

quantitatively from pentane as fine powders. Although the

solubility of diiodides 1-I₂and 2-I₂is greater than that of

dichlorides 1-Cl₂and 2-Cl₂in Et₂O and THF, insolubility in

suitable non-coordinating solvents precludes cryoscopic mo-

lecular weight determination.¹⁹

^aYields g95% by ¹H NMR spectroscopy. ^bDetermined by ¹H NMR

spectroscopy.

the Th-based complex mediates the transformation with mark-

edly greater efficiency, presumably owing to the larger ionic

radius (Table 1, vide infra).

Larger piperidine rings are also accessible using the present

organoactinide precatalysts. Again, the marked influence

of geminal dialkyl substitution is evident, effecting a ca. 10×

increase in N_ton going from substrate 13 to 15 to 17 when

mediated by 1 for R ) H < Me < Ph (entries 7, 9, and 11).

The formation of five-membered pyrrolidine vs six-membered

piperidine rings is significantly more efficient in most cases

for either Th or U, with the notable exception of substrate 15,

where HA/cyclization proceeds faster for 2 than for 1 (at 25

°C for An ) U vs 60 °C for An ) Th). Compared with substrate

9 (N_t) 2.5 h^-1at 25 °C; Table 2, entry 4), this pronounced

difference in reactivity is intriguing and may reflect Th vs U

electronic properties or unique steric factors introduced by the

CGC ancillary ligand (vide infra). This effect also obtains for

R ) Ph substitution in substrate 17, where N_t) 4.8 h^-1at 25

°C (Table 2, entry 12). The results for 13 and 17 cyclization to

piperidines 14 and 18, respectively (Table 2, entries 7, 8, 11,

and 12), parallel those for pyrrolidine formation and suggest

that similar Th and U losses in cyclization efficiency are

operative upon increasing the product ring size from five to six

members.

Hydroamination/Cyclization. Substrate and Catalyst

Trends. In this section, we present the results of catalytic

intramolecular HA studies for a range of C-C unsaturations,

revealing an impressive scope rivaled only by organolanthanide

catalysts. Substrates surveyed with precatalysts 1-6 include

primary (7-18) and secondary (19-24) terminal aminoalkenes,

primary internal (27-32) aminoalkenes, primary (33-42) and

secondary (25, 26) aminoalkynes, primary aminoallenes (43-

48), and primary aminodienes (49-54).

(a) Aminoalkenes. Intramolecular HA/cyclization of terminal

as well as 1,2-disubstituted (internal) amino-olefins is effectively

mediated by (CGC)An complexes. Results for simple 1° amine

substrates are presented first, followed by 2° amines and internal

olefins. The scope of precatalyst 1- and 2-mediated terminal

aminoalkene HA is summarized in Table 2. Thus, unsubstituted

substrate 7 is converted to exocyclic product 8 by either 1 or 2

at roughly comparable N_tvalues, although rather sluggishly at

60 °C (Table 2, entries 1 and 2). Introduction of geminal

dimethyl substituents in 9f10 cyclization (Table 2, entries 3

and 4) effects a reactivity increase of ca. 10× for 1 (N_t) 15

h^-1, An ) Th) vs 2 (N_t) 2.5 h^-1, An ) U) at 25 °C. A similar,

though more pronounced effect is observed for 11f12, where

geminal disubstitution effects are enhanced by phenyl substitu-

tion (Table 2, entries 5 and 6). Here, precatalyst 1 effects HA/

cyclization with remarkable efficiency at 25 °C - N_t) 1460

h^-1as compared to the impressive N_t) 430 h^-1for 2. Again,

The consequences of the steric openness of (CGC)An relative

to more encumbered actinocene complexes are particularly

evident in Table 3. Here, transformations 9f10 (entries 7-10)

and 11f12 (entries 1-6) are compared for precatalysts 1-6.

Note that Cp′₂An complexes 5 and 6 display markedly

diminished efficiencies and require heating at 60 °C for useful

9

4258 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

Table 3. Aminoalkene HA/Cyclization of Substrates 9 and 11

Catalyzed by Organoactinide Complexes 1-6. Comparison of N_t

as a Function of Metal Ionic Radius and Ancillary Ligand

Environment in C₆D₆

^aYields g95% by ¹H NMR spectroscopy. ^bDetermined by ¹H NMR

spectroscopy.

HA/cyclization of the two most reactive aminoalkene substrates,

9 (N_t) 0.4 h^-1, An ) Th; N_t) 0.2 h^-1, An ) U; entries 9 and

10) and 11 (N_t) 2.8 h^-1, An ) Th; N_t) 1.0 h^-1, An ) U;

entries 5 and 6). This reactivity-coordinative unsaturation trend

is evident in comparing these results with those for precatalysts

3 and 4, having intermediate Me₂SiCp′′₂An ancillary ligation.

With the extremely reactive substrate 11 (N_t) 120 h^-1, An )

Th; N_t) 29 h^-1, An ) U at 25 °C; entries 3 and 4), the greater

than 10× reactivity difference between complexes 1 vs 3 and

2 vs 4 is illustrative of the importance of the ancillary ligation

and substrate accessibility to the catalytic center.

Figure 2. Kinetic plots for representative HA/cyclization of aminoalkenes.

(a) Plot of the normalized ([aminoalkene 9]/[(CGC)Th ]) ratio × 100%

([(CGC)Th ] from precatalyst 1 relative to Si(p-tolyl)₄internal standard) at

25 °C in C₆D₆. The line depicts the least-squares fit, with N_tdetermined

from the slope of the line according to eqs 1 and 2. (b) Plot of ([aminoalkene

9])/[(CGC)Th ]) × 100% (normalized to Si(p-tolyl)₄internal standard) over

two consecutive catalytic runs under identical conditions. Following

completion of the first HA/cyclization reaction, all volatiles were removed

in vacuo (including product 10; 10^-6Torr) and replaced with fresh C₆D₆

solvent and fresh substrate 9 (see Experimental Section for complete details).

A representative plot of normalized [substrate]:[An] ratio vs

1

time (Figure 2a) was determined from the H NMR spectral

intensities of precatalyst + Si(p-tolyl)₄internal standard solu-

tions. The slope of the plot is constant in all cases, indicating

zero-order dependence on [substrate]. Near complete conversion

(low [substrate]), some departure from linearity is observed.

However, plots of multiple catalytic runs with a single loading

of precatalyst in a single NMR tube (Figure 2b) indicate that

this deviation is caused not by catalyst deactivation, but rather

by competitive inhibitory binding of product, interfering with

turnover-limiting CdC insertion into the An-N bond.^2d,3u

The cyclization of more complex 2° amine substrates to form

products containing 3° amine skeletons was also explored for

(CGC)An complexes 1 and 2 (Table 4). Aminodiene 19

undergoes initial HA and cyclization to form the intermediate

2° amine 20a with an efficiency analogous to that for the

structurally similar substrate 9 (N_t) 12 h^-1, An ) Th and

N_t) 1.5 h^-1, An ) U; Table 4 entries 1 and 2). Interestingly,

subsequent reaction at 100 °C cleanly converts 20a to bicyclic

20b in quantitative yield (eq 5). The 2° amine analogue of

4 and proceeds smoothly at 60 °C (N_t) 0.6 h^-1, An ) Th;

N_t) 0.8 h^-1, An ) U; entries 5 and 6). Sterically more

encumbered 23 bears an N-benzyl substituent and requires 100

°C reaction temperatures to achieve significant rates (N_t) 0.1

h^-1; Table 4, entries 7 and 8). The steric and electronic obstacles

presented by 23 are reflected in the sluggish turnover frequency

and its sluggish protonolytic activation of precatalysts 1 and 2

at 60 °C (typically instantaneous at -78 °C in C₇D₈). Also note

that the 25f26 conversion proceeds at turnover frequencies

(N_t) 80 h^-1, An ) Th; N_t) 120 h^-1, An ) U; Table 4, entries

9 and 10) far greater than those for analogous 1° aminoalkyne

39 at 25 °C (vide infra).

The scope of organoactinide-mediated 1,2-disubstituted (in-

ternal) aminoalkene HA/cyclization was also investigated (Table

5). These transformations are most efficient for “activated”

styrene conversion 27f28, where (CGC)An precatalysts 1 and

2 effect cyclization under mild conditions with reasonable

efficiency (N_t) 1.3 h^-1, An ) Th; N_t) 0.9 h^-1, An ) U;

entries 1 and 2). Ancillary ligation effects on the HA/cyclization

were also explored, with significantly lower turnover frequencies

observed when more coordinatively saturated Me₂SiCp′′₂An

and Cp′₂An precatalysts 3-6 were used (Table 5, entries 1-6).

At 100 °C, ansa-metallocenes 3 and 4 afford 28 at reasonable

rates (N_t) 5.8 and 7.0 h^-1for 3 and 4, respectively; entries 3

substrate 11, N-methyl substrate 21, displays the greatest

turnover frequency of all aminoalkene transformations in Table

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4259

A R T I C L E S

Stubbert and Marks

Table 6. HA/Cyclization Results with Aminoalkyne Substrates

Table 4. Organoactinide-Catalyzed HA/Cyclization of Secondary

Amine Substrates Tethered to Alkene and Alkyne C-C

Unsaturations. Comparison of N_tas a Function of Metal Ionic

Radius in C₆D₆

Mediated by Organoactinide Precatalysts 1-6 in C₆D₆

^aYields g95% by ¹H NMR spectroscopy and GC/MS. ^bDetermined by

¹H NMR spectroscopy.

^aYields g95% by ¹H NMR spectroscopy. ^bDetermined by ¹H NMR

spectroscopy.

than 29f30, in accord with (CGC)U catalyst-substrate

specificity effects observed with geminal dimethylpiperidine

formation (Table 2, entries 9-12; vide infra).

Table 5. Organoactinide-Mediated HA/Cyclization of

1,2-Disubstituted Aminoalkene Substrates in C₆D₆Catalyzed by

1-6. Comparison of N_tas a Function of Metal Ionic Radius and

Ancillary Ligand Environment in C₆D₆

Representative terminal aminoalkene conversion 17f18,

mediated by Th catalyst 1, was examined over a broad

temperature range, and from Eyring and Arrhenius analyses,²⁰

∆H^q) 12.6(5) kcal/mol, ∆S^q) -30(1) eu, and E_a) 13.3(7)

kcal/mol are obtained from 25 to 82 °C. For comparison with

internal aminoalkene substrates, the 27f28 HA/cyclization

promoted by precatalyst 2 was similarly investigated²⁰between

44 and 84 °C, affording ∆H^q) 16(2) kcal/mol, ∆S^q) -28(7)

eu, and E_a) 17(2) kcal/mol (Table 9).

(b) Aminoalkynes. Terminal and 1,2-disubstituted (internal)

aminoalkyne HA/cyclization is effectively mediated by organo-

actinide complexes 1-6. As for organo-4f-element catalysts,^2d,3n,s

this transformation is efficient and highly regioselective at room

temperature. The scope of the present aminoalkyne HA/

cyclizations includes five- and six-membered ring formation.

Conversion 33f34 was examined with precatalysts 1-6 to

explore the effects of ligand environment (Table 6, entries 1-6).

The (CGC)An framework is by far the most efficient in

mediating this transformation, proceeding rapidly at 25 °C

(N_t) 82 h^-1, An ) Th; N_t) 3000 h^-1, An ) U; Table 6,

entries 1 and 2). Interestingly, in the case of more coordinatively

saturated precatalysts 3-6, N_tat 60 °C increases with decreasing

steric openness (N_t) 149 h^-1, 3; N_t) 190 h^-1, 5; N_t) 9.8

h^-1, 4; N_t) 15.8 h^-1, 6; Table 6, entries 3-6), although not

for (CGC)An catalysts 1 and 2. Additionally, the superiority

of the (CGC)An catalyst framework over Cp′₂An is evident

for substrates 35 (An ) U; Table 6, entries 8 and 9) and 39

^aYields g90% by ¹H NMR spectroscopy. ^bDetermined by ¹H NMR

spectroscopy. ^cPolymeric byproduct precipitate observed. Single product

1

detected in solution by H NMR spectroscopy and GC/MS.

and 4), whereas unbridged actinocenes are sluggish (N_t) 0.2

and 0.5 h^-1for 5 and 6, respectively; entries 5 and 6). With

unactivated substrates, transformations 29f30 and 31f32

proceed only at elevated temperatures (100 °C). Pyrrolidine

product 30 is obtained with similar N_tvalues for both 1 and 2

(N_t) 0.6 h^-1, An ) Th; N_t) 0.2 h^-1, An ) U; Table 5, entries

7 and 8), while piperidine 32 formation proceeds with similar

efficiency for 1 (N_t) 0.2 h^-1, An ) Th; Table 5, entry 9). As

for transformation 15f16, precatalyst 2 mediates 31f32 (N_t

) 2.2 h^-1, An ) U; Table 5, entry 10) with far greater efficiency

(20) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2nd ed.;

McGraw-Hill, Inc.: New York, 1995.

9

4260 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

Table 7. HA/Cyclization Results with Aminoallene Substrates in C₆D₆Mediated by Organoactinide Precatalysts 1-6

^aYields g95% by ¹H NMR spectroscopy and GC/MS. ^bDetermined by ¹H NMR spectroscopy. ^cDetermined in situ by ¹H NMR spectroscopy. ^dByproduct

1

precipitate observed. Single set of isomers detected in solution by H NMR spectroscopy.

(An ) Th; Table 6, entries 12 and 14), where (CGC)An N_t

values are far greater.

The scope of organoactinide-mediated aminoalkyne HA/

cyclization and An⁴⁺ionic radius effects is summarized in Table

6. As noted above, conversion of Me₃Si-substituted 33 is most

efficient for both An ) Th and U, likely owing to -SiMe₃group

transition-state electronic stabilization (vide infra). Unsubstituted

terminal alkyne 35 is rapidly cyclized/tautomerized to 36,

especially by 2, where N_tis ca. 10³× greater than that for 1

(N_t) 7.8 h^-1, An ) Th; N_t) 1210 h^-1, An ) U at 25 °C;

Table 6, entries 7 and 8). More sterically demanding 37f38

conversion (Table 6, entries 10 and 11) proceeds rapidly,

although with considerably less efficiency with 1 than with 2

(nearly 100×), clearly indicating finely controlled steric and

electronic requirements. Electronically assisted transformation

39f40 (vide infra) exhibits decreased N_tvs that of sterically

less-encumbered substrates 35 and 37 for 2 but, for precatalyst

1, only vs 35 (N_t) 4.3 h^-1, An ) Th; N_t) 51 h^-1, An ) U

at 25 °C; Table 6, entries 12 and 13). The relative rates of

aminoalkyne HA/cyclization vary differently for 1 and 2, with

-SiMe₃> -H > -Ph > -Me for An ) Th, and -SiMe₃>

-H . -Me > -Ph for An ) U. The HA/cyclization of 41,

yielding benzyl-substituted tetrahydropyridine 42, proceeds less

efficiently, requiring higher temperatures (N_t) 1.4 h^-1, An )

Th; N_t) 5.0 h^-1, An ) U at 60 °C; Table 6, entries 15 and

16). As observed with aminoalkenes, metal ionic radius effects

diminish considerably in closures to larger-ring products. Plots

of [substrate]:[An] vs time (e.g., Figure 3) again exhibit

deviations from linearity at low [substrate], which is attributed

to product inhibition^2dat high conversion. Additionally, repre-

sentative transformation 37f38, mediated by Th complex 1,

was examined over a 50 °C temperature window to yield²⁰

∆H^q) 14(2) kcal/mol, ∆S^q) -27(5) eu, and E_a) 15(2) kcal/

mol (Table 9).

Figure 3. Plot of the normalized ([aminoalkyne 37]/[(CGC)U ]) ratio ×

100% ([(CGC)U ] from precatalyst 2 relative to Si(p-tolyl)₄internal

standard) at 25 °C in C₆D₆. The line depicts the least-squares fit, and N_tis

determined from the slope of the line according to eqs 1 and 2.

element catalysts,^2d,3h,i,ksuch transformations are efficient, highly

regioselective, and moderately to highly diastereoselective. The

scope of aminoallene HA/cyclization was explored for both five-

and six-membered ring products (Table 7). Cyclization of

internal aminoallenes 43, 45, and 47 proceeds with >90%

regioselectivity in forming smaller-ring pyrrolidine or piperidine

(kinetically favored) products 44, 46, and 48, with 60-90%

trans-selective product diastereoselectivities (Table 7, entries

1-8). For conversion 43f44, the effect of ancillary ligand

openness on both N_tand diastereoselectivity was explored for

precatalysts 1-6. Selectivity for trans product 44 is comparable

for either An ) Th or U, regardless of ancillary ligation,

although selectivity falls noticeably from (CGC)An to more

congested Me₂SiCp′′₂An and Cp′₂An frameworks (90:10 trans:

cis, 1; 80:20, 2; 60:40, 3 and 4; 70:30, 5; 65:35, 6; Table 7,

entries 1-6). The rate of this transformation varies far more

upon opening or closing the actinide coordination geometry,

with precatalyst 2 exhibiting superior activity vs 4 and 6, which

(c) Aminoallenes. Intramolecular HA/cyclization of 1,3-

disubstituted (internal) aminoallenes is mediated with high

regioselectivity by (CGC)An complexes. As for organo-4f-

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4261

A R T I C L E S

Stubbert and Marks

Table 8. HA/Cyclization Results with Aminodiene Substrates in C₆D₆Promoted by Organoactinide Precatalysts 1-6

1

^aYields g95% by H NMR spectroscopy and GC/MS. ^bDetermined by H NMR spectroscopy. ^cDetermined in situ by H NMR spectroscopy.

Table 9. Summary of Organolanthanide- and Organoactinide-Mediated HA/Cyclization Activation Parameters Calculated from Standard

Eyring and Arrhenius Analyses^a

^aData were obtained using Cp′₂Ln[CH(SiMe₃)₂] (Ln ) La, Sm, Lu) and 1 or 2 as noted over a 40-55 °C temperature range in C₆D₆for each class of

C-C unsaturation. The columns listed as step (i), C-C insertion into M-N bond, and step (ii), protonolysis of cyclized substrate, refer to the proposed

mechanism depicted in Scheme 2.

require 100 °C reaction temperatures for moderate rates (N_t)

29 h^-1, 2, 25 °C; N_t) 0.3 h^-1, 4, 25 °C; N_t) 0.8 h^-1, 6,

100 °C; Table 7, entries 2, 4, and 6). A similar trend is evident

for An ) Th, where precatalyst 1 is far more efficient than either

3 or 5 (N_t) 2.7 h^-1, 1, 25 °C; N_t) 0.03 h^-1, 4, 25 °C; N_t)

0.3 h^-1, 5, 100 °C; Table 7, entries 1, 3, and 5).

46, diastereoselectivities are again similar, but 2 exhibits

somewhat greater selectivity at 60 °C (75:25 trans:cis vs 60:

40; Table 8, entries 7 and 8). Furthermore, ca. 10× rate

enhancement is again observed for An ) U vs An ) Th, while

for 47f48, precatalysts 1 and 2 effect sluggish conversion for

this aminoallene substrate at 100 °C (N_t) 0.01 h^-1, An ) Th;

N_t) 0.1 h^-1, An ) U; Table 7, entries 9 and 10). Activation

parameters for the present aminoallene HA/cyclizations were

quantified for representative transformation 43f44 mediated

by Th complex 1 between 25 and 68 °C. From Eyring and

Arrhenius analyses,²⁰∆H^q) 13(2) kcal/mol, ∆S^q) -29(6)

eu, and E_a) 14(2) kcal/mol (Table 9).

The effect of An⁴⁺ionic radius on rate and selectivity was

also investigated for (CGC)An -promoted aminoallene HA/

cyclization reactions (Table 7, entries 1, 2, 7-10). The

aforementioned 43f44 conversion proceeds smoothly at 25 °C

when mediated by either 1 or 2, although ca. 10× more

efficiently for An ) U (N_t) 2.7 h^-1, An ) Th; N_t) 29 h^-1

,

An ) U; Table 7, entries 1 and 2). Diastereoselectivities here

are comparable, with 1 displaying slightly greater selectivity

(90:10 trans:cis vs 80:20). In the case of cyclization to piperidine

(d) Aminodienes. Intramolecular HA/cyclization of terminal

and 1,4-disubstituted aminodienes is effectively mediated by

(CGC)An complexes. As for organo-4f-element catalysts,^2d,3c,f

9

4262 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

such transformations are efficient and exhibit moderate/high

regioselectivity at moderate temperatures. The scope spans

pyrrolidine and piperidine skeletons, with products evidencing

mixed regioselectivities (Table 8), where rate and selectivity

vary with actinide identity. Complexes 1 and 2 mediate

conversion of aminodiene 49 to allylpyrrolidine 50a, along with

small quantities of (1-propenyl)pyrrolidine byproduct 50b (N_t

) 3.0 h^-1, 98:2 a:b, An ) Th; N_t) 5.5 h^-1, 93:7 a:b, An )

U; Table 8, entries 1 and 2). Cyclization 51f52a/b is highly

regioselective when promoted by 1 and proceeds more rapidly

than 49f50a/b at 60 °C, owing to the geminal dimethyl

substitution (Table 8, entry 7). As with substrates having similar

structural characteristics (15 and 31, vide infra), this transforma-

tion is considerably more rapid than that with An ) Th or with

2 in the 49f50a/b conversion (N_t) 19.9 h^-1, 81:19 a:b, An

) U at 60 °C; Table 8, entry 8). Conversion of styrenic substrate

53 to pyrrolidines 54a and 54b proceeds with considerably lower

N_tvalues vs that of substrates 49 and 51, as well as with

diminished regioselectivity (N_t) 0.02 h^-1, 69:31 a:b, An )

Th; N_t) 0.3 h^-1, 79:21 a:b, An ) U; Table 8, entries 9

and 10).

Cp₂An(NR₂)₂actinocenes have been prepared in this manner

from the corresponding homoleptic An(NR₂)₄reagents and

excess HCp (eq 6).^17,22,23However, activation of more encum-

bered rings (e.g., HCp′) has been less successful, even at

elevated temperatures, presumably reflecting unfavorable steric

interactions and diminished Brønsted acidity (eq 7).²⁴For the

The influence of actinide ancillary ligation was also inves-

tigated for aminodiene HA/cyclization (Table 8, entries 1-6).

Both ansa-metallocene Me₂SiCp′′₂Th (3) and unbridged Cp′₂Th

(5) promote the transformation 49f50a/b with moderate

regioselectivity (71:29 a:b, 3; 76:24 a:b, 5; Table 8, entries 3

and 5), whereas the An ) U congeners exhibit moderate but

opposing regioselectivity (15:85 a:b, 4; 35:65 a:b, 6; Table 8,

entries 4 and 6). Regarding conversion rates, note that the

Cp′₂An catalysts require 80 °C to achieve sluggish activity, in

contrast to the more open organoactinide catalysts (N_t) 0.2

h^-1, An ) Th and U; Table 8, entries 5 and 6). While similar

steric concerns may be an issue for Me₂SiCp′′₂U -mediated

formation of 50, N_tvalues are comparable for both 3 and 4 (N_t

) 1.9 h^-1, An ) Th; N_t) 0.8 h^-1, An ) U; Table 8, entries 3

and 4). Activation parameters for representative aminodiene HA/

cyclization 49f50a/b promoted by complex 2 are ∆H^q) 19-

(3) kcal/mol, ∆S^q) -16(9) eu, and E_a) 19(3) kcal/mol

between 40 and 82 °C.²⁰Activation parameters for all organo-

actinide-catalyzed intramolecular HA/cyclization reactions are

summarized in Table 9.

present H₂CGC proligand, HCp′′ steric bulk is partially offset

by chelation in the final complex.¹³Interestingly, in the synthesis

of (CGC)Ln-N(SiMe₃)₂organolanthanides (Ln ) Nd, Sm, Yb,

Lu), initial HN(SiMe₃)₂displacement and concomitant ligand

binding are found to proceed via the HN(^tBu) moiety,^3jlikely

reflecting unfavorable steric interactions (eq 8). The HCp′′ group

subsequently coordinates via a second protodeamination, both

requiring prolonged reflux and HN(SiMe₃)₂byproduct removal

in vacuo.

Discussion

We first discuss synthetic and characterization aspects of

(CGC)An(NMe₂)₂chemistry relating to catalytic intramolecular

HA/cyclization reactivity and selectivity. The significance of

metal ionic radius and ancillary ligation for catalysis is

discussed. Reactivity trends for intramolecular C-N bond-

forming processes are analyzed, along with basic mechanistic

insights and comparisons with transformations mediated by

organo-4f-element and organo-transition metal catalysts. A

detailed mechanistic analysis of (CGC)An-mediated intramo-

lecular HA reactions will be presented in a companion contribu-

tion.¹⁶

Constrained geometry organoactinide complexes 1 and 2 are

conveniently prepared in good yield and high purity from AnCl₄

via one-pot in situ generation of the corresponding tetrakis(di-

alkylamide) reagents, followed by addition of the H₂CGC

proligand (eq 3). In earlier (CGC)An(NRR′)₂work,¹⁵similar

results were obtained using sublimed, isolated homoleptic

amidoactinides. Likewise, LiCl byproduct removal and isolation

(22) (a) Bagnall, K. W.; Yanir, E. J. Inorg. Nucl. Chem. 1974, 36, 777-779.

(b) Watt, G. W.; Gadd, K. F. Inorg. Nucl. Chem. Lett. 1973, 9, 203-205.

(23) (a) Ossola, F.; Rossetto, G.; Zanella, P.; Arudini, A.; Jamerson, J. D.; Takats,

J.; Dormond, A. Inorg. Synth. 1992, 29, 234-238. (b) Arduini, A. L.;

Takats, J. Inorg. Chem. 1981, 20, 2480-2485. (c) Arduini, A. L.; Jamerson,

J. D.; Takats, J. Inorg. Chem. 1981, 20, 2474-2479. (d) Arduini, A. L.;

Edelstein, M.; Jamerson, J. D.; Reynolds, J. G.; Schimd, K.; Takats, J.

Inorg. Chem. 1981, 20, 2470-2474. (e) Jamerson, J. D.; Takats, J.

J. Organomet. Chem. 1974, 78, C23-C25. (f) Jones, R. G.; Karmas, G.;

Martin, G. A., Jr.; Gilman, H. J. Am. Chem. Soc. 1956, 78, 4285-4286.

(24) (a) Taft, R. W.; Bordwell, F. G. Acc. Chem. Res. 1988, 21, 463-469. (b)

Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463. (c) Bordwell, F. G.;

Bausch, M. J. J. Am. Chem. Soc. 1983, 105, 6188-6189.

Precatalyst Synthesis and Characterization. Protodeami-

nation routes to d- and f-block metal complexes avoid coordi-

nating solvents and undesired “ate” byproducts.²¹For example,

(21) (a) Berthet, J. C.; Ephritikhine, M. Coord. Chem. ReV. 1998, 178-180,

83-116. (b) Kuhlman, R. Coord. Chem. ReV. 1997, 167, 205-232. (c)

Bradley, D. C.; Chisholm, M. H. Acc. Chem. Res. 1976, 9, 273-280.

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4263

A R T I C L E S

Chart 1

Stubbert and Marks

of the crude “An(NR₂)₄” prior to reaction with H₂CGC are

equally successful (see Experimental Section for 1). NMR-scale

syntheses, monitored in situ, indicate that crude or pure An-

(NR₂)₄protonolysis with H₂CGC is essentially quantitative in

benzene-d₆and that appreciable (CGC)An(NR₂)₂solubility in

hydrocarbon solvents limits the isolated yields of analytically

pure 1 and 2. The highest yield was obtained for (CGC)U-

(NEtMe)₂, which presumably combines -NMe₂compactness

with -NEt₂steric screening (preventing formation of

U(NEtMe)₄“ate” contaminants) to promote crystallization in

analytical purity and ca. 80% yield from UCl₄.¹⁵In all cases,

optimized yields are obtained by incorporating a slight H₂CGC

excess to counteract ate contaminant formation and to drive the

protodeamination equilibrium to the right (eq 9). Contrary to

(CGC)Ln-N(SiMe₃)₂chemistry,^3jthe order of ligand attachment

Single crystals of 1 and 2 suitable for diffraction studies were

obtained by slow cooling of saturated pentane solutions. ORTEP

depictions of the solid-state structures of 1 and 2 are shown in

Figure 1. These complexes display a pseudo-tetrahedral geom-

etry about the An⁴⁺center, along with a considerably more open

coordination environment than that of previously characterized

organoactinide complexes (Chart 1). The two An-NMe₂

distances within each complex are nearly identical (within 1σ)

and similar to other An-N σ-bonded distances,²⁵with average

An-N bond lengths of 2.27(1) Å (1) and 2.21(1) Å (2). In

comparison, the average Th-N distance of 2.25(3) Å in (η⁵-

Me₇C₉)₂Th(NMe₂)₂^25aand the U-NEt₂distance of 2.220(5) Å

in (η⁴-[(SiMe₃)NCH₂CH₂]₃N)U(NEt₂)^25bdiffer only slightly

from those in 1 and 2, attributable to differences in interligand

steric interactions (Chart 1). However, owing to the silyl moiety

electron-withdrawing and R-anion stabilization properties,²⁶in

addition to geometric constraints imposed by the ligand

structure, the An-N(^tBu) σ-bond lengths are considerably longer

(>3σ) than their An-NR₂counterparts in both 1 and 2. This

trend is also observed in a series of tetravalent (triamidoamine)-

uranium complexes, namely (η⁴-[(SiMe₃)NCH₂CH₂]₃N)U-X

(X ) Cl, Br, I, NEt₂, or Cp′),^25b,daryl-substituted Cp′₂U[N(H)-

Ar]₂(Ar ) 2,6-dimethylphenyl),^25cand bis[8]annulene (η⁸-

COT)Th[N(SiMe₃)₂]₂^25ecomplexes. Considering the difference

in ionic radii between the two metals (Table 1),¹⁰note that this

effect appears to be less pronounced in the Th-N(^tBu) bond

length in 1 vs the U-N(^tBu) distance in 2, perhaps a

consequence of geometric constraints imposed by the chelating

CGC^2-ligand. The Cp′′ rings in 1 and 2 are coordinated to the

respective actinide centers in an η⁵manner, as indicated by the

C-C bond lengths and Cp′′(c)-An distances, without evidence

for significant η⁵fη³ring-slippage²⁷(Figure 1).

1

(determined in situ by H NMR spectroscopy) more closely

follows the expected thermodynamic pathway of protonolysis,

whereby the more acidic HCp′′ moiety (pK_a∼24) binds to the

An center first, followed by the HN(^tBu) moiety (pK_a∼33),

likely owing to the smaller HNR₂leaving groups employed for

the tetravalent actinides.^21,23Also note that the rate at which

equilibrium is reached parallels the byproduct amine volatility

(rate: HNMe₂, bp 7 °C . HNEtMe, bp 36 °C > HNEt₂, bp

55 °C).

In the case of complexes 1 and 2 (An ) Th, U; R ) Me),

the “An(NR₂)₄” species most likely contain some “ate” con-

tamination, e.g., Li[An(NMe₂)₅],^17,21-23resulting from insuf-

ficient -NMe₂group steric screening. This leads to small

amounts of byproduct tentatively formulated as Li₂[Me₂Si(η⁵-

Me₄C₅)(^tBuN)], which, along with any other pentane-insoluble

impurities, is removed by 20 °C filtration prior to -78 °C

recrystallization, affording analytically pure (CGC)An(NMe₂)₂

products. The absence of coordinated Lewis bases here allows

direct application of 1 and 2 as precatalysts in C-N bond-

forming reactions without further purification or undesirable

catalytic inhibition.^3a,u,18

Compared with previously characterized organoactinide com-

plexes bearing similar ancillary ligation,²⁸the solid-state

structures of 1 and 2 indicate considerably greater metal center

steric openness/coordinative unsaturation (Chart 2). The Cp-

(25) (a) Trnka, T. M.; Bonanno, J. B.; Bridgewater, B. M.; Parkin, G.

Organometallics 2001, 20, 3255-3264. (b) Roussel, P.; Alcock, N. W.;

Boaretto, R.; Kingsley, A. J.; Munslow, I. J.; Sanders, C. J.; Scott, P. Inorg.

Chem. 1999, 38, 3651-3656. (c) Straub, T.; Frank, W.; Reiss, G. J.; Eisen,

M. S. Dalton Trans. 1996, 2541-2546. (d) Scott, P.; Hitchcock, P. B.

Dalton Trans. 1995, 603-609. (e) Gilbert, T. M.; Ryan, R. R.; Sattelberger,

A. P. Organometallics 1988, 7, 2514-2518.

(26) (a) Brinkman, E. A.; Berger, S.; Brauman, J. I. J. Am. Chem. Soc. 1994,

116, 8304-8310. (b) Hopkinson, A. C.; Lien, M. H. J. Org. Chem. 1981,

46, 998-1003.

9

4264 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

Chart 2

(c)-An-L angles (L ) N(^tBu), Cp′′, Cp′) of 91.3° and 93.3°

determined for 1 and 2, respectively, evidence considerably

greater openness vs 3 (118.3°),²⁹closely related “ate” complex

Me₂SiCp′′₂U(µ-Cl)₄[Li(TMEDA)]₂(114.1°),³⁰(HCp′′)₂UMe₂

(134.3°),^28aCp′₂ThMe₂(133.9°),^28bor Cp′₂UMe₂(140.5°),^28b

indicating enhanced An center accessibility. The less constrained

Me₂SiCp′′₂An (3 and 4) and Cp′₂An (5 and 6) ligand

environments are quantifiably less accessible to incoming

substrates by ca. 25° and 40°, respectively, vs (CGC)An 1

and 2.

halide complexes^12,13,19,32suggests that these complexes may

exist as dimers or trimers (vide infra). ¹H NMR studies in THF-

d₈solution are in agreement with a single C_s-symmetric

instantaneous species at 20 °C, consistent with formation of a

monomeric, fluxional THF adduct or symmetric [(CGC)An(µ-

X)₂]₂(B) or [(CGC)AnX(µ-X)]₂(C) structures. Interestingly,

while the solubility of these dihalides in THF is typical of

organoactinide halides, the solubility of 1-Cl₂and 2-Cl₂in

diethyl ether or hydrocarbons is far less, prompting the synthesis

of analogous diiodides 1-I₂and 2-I₂. These display improved

solubility in diethyl ether, but not in hydrocarbons.

Compared to charge-neutral (CGC)Ln-E(SiMe₃)₂(Ln ) Sm,

E ) N; Ln ) Lu, E ) CH),^3jcomplexes, the present

organoactinide coordinative openness is greater, with larger

lanthanide ions displaying Cp(c)-Ln-N(^tBu) angles ranging

from 95.3° (Ln ) Sm) to 100.9° (Ln ) Lu) (Chart 2). This

trend can be rationalized in terms of metal ionic radius, where

eight- and nine-coordinate ionic radii¹⁰for selected Ln³⁺and

An⁴⁺ions are listed in Table 1. Overall, trends in the Cp′′(c)-

An-N1 angle indicate that the steric openness in f-element CGC

complexes increases in the order Lu , Sm < U < Th. Versus

related tetravalent group 4 complexes, the considerably less open

Cp(c)-M-N(^tBu) of 100.2° in (CGC)Zr(NMe₂)₂,³¹among

other steric and electronic factors, is in accord with the

diminished catalytic activity in mediating aminoalkene HA/

cyclization.^2,16

Closely related [(CGC)ThCl(µ-Cl)₂Li(OEt₂)₂]₂was prepared

in 43% yield via a traditional salt elimination route from ThCl₄

with Li₂CGC in diethyl ether³³and exhibits solid-state structure

D, with each Th center retaining an additional chloride ligand.

While it is unlikely that catalytic activity would be observed

for any of these species without the addition of a cocatalytic

reagent (e.g., MAO and/or AlR₃),³⁴the absence of residual

Lewis bases in the coordination spheres of 1-Cl₂, 2-Cl₂, 1-I₂,

and 2-I₂reduces the possibility of catalyst inhibition. Further-

more, base-free dihalide complexes 1-X₂and 2-X₂can be

obtained from crude, AnCl₄-derived (CGC)An(NR₂)₂complexes

in high overall yield.

Analytically pure (CGC)AnX₂(An ) Th, 1-Cl₂and 1-I₂;

An ) U, 2-Cl₂and 2-I₂) dihalide complexes are prepared in

quantitative yield by reaction of 1 or 2 with excess Me₃SiX

reagents (eq 4). These complexes are insoluble in non-

coordinating solvents and precipitate immediately from solution

as fine powders, allowing removal of the byproduct amine in

vacuo or by suitable washing. This synthetic strategy is

particularly useful for one-pot reactions, proceeding through the

dihalides 1-X₂and 2-X₂. Although poor solubility precludes

cryoscopic molecular weight determination, the open coordina-

tion spheres and frequent observation of d⁰- and f-block bridging

Catalytic C-N Bond Formation. The following discussion

focuses on (CGC)An catalytic properties in mediating HA/

cyclization of diverse primary and secondary amines, tethered

(27) See, for example: Basolo, F. New J. Chem. 1994, 18, 19-24.

(28) (a) Evans, W. J.; Kozimor, S. A.; Hillman, W. R.; Ziller, J. W.

Organometallics 2005, 24, 4676-4683. (b) Jantunen, K. C.; Burns, C. J.;

Castro-Rodriguez, I.; Da Re, R. E.; Golden, J. T.; Morris, D. E.; Scott,

B. L.; Taw, F. L.; Kiplinger, J. L. Organometallics 2004, 23, 4682-4692.

(29) (a) Fendrick, C. M.; Schertz, L. D.; Day, V. W.; Marks, T. J. Organome-

tallics 1988, 7, 1828-1838. (b) Fendrick, C. M.; Mintz, E. A.; Schertz,

L. D.; Marks, T. J. Organometallics 1984, 3, 819-821.

(32) See, for example: Fagan, P. J.; Manriquez, J. M.; Marks, T. J.; Day, C. S.;

Vollmer, S. H.; Day, V. W. Organometallics 1982, 1, 170-180 and

references therein.

(33) Golden, J. T.; Kazul’kin, D. N.; Scott, B. L.; Voskoboynikov, A. Z.; Burns,

C. J. Organometallics 2003, 22, 3971-3973.

(34) For general reviews, see: (a) Barron, A. R. Metallocene-Based Polyolefins

2000, 1, 33-67. (b) Chen, E. Y.-X.; Marks, T. J. Chem. ReV. 2000, 100,

1391-1434 and references therein.

(30) Schnabel, R. C.; Scott, B. L.; Smith, W. H.; Burns, C. J. J. Organomet.

Chem. 1999, 591, 14-23.

(31) Carpenetti, D. W.; Kloppenburg, L.; Kupec, J. T.; Petersen, J. L.

Organometallics 1996, 15, 1572-1581.

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4265

A R T I C L E S

Stubbert and Marks

to diverse C-C unsaturations. Beginning with simple terminal

aminoalkenes and progressing through additional substrate

classes, the expansive scope and powerful C-N bond-forming

ability of (CGC)An catalysts are placed in context with relevant

literature and likely mechanistic scenarios.

(a) Aminoalkenes. (CGC)An complexes rapidly and regi-

oselectively mediate intramolecular HA/cyclization of terminal

and disubstituted aminoalkenes. Compared to other catalyst

systems, the efficacy of precatalysts 1 and 2 is impressive,

displaying scope and activity comparable to those of complexes

of intermediate-sized Ln³⁺ions stabilized by similar ligands,^2d

and the aminoalkene HA/cyclization rates far exceed those of

systems mediated by group 4 catalysts, all of which require

heating to g100 °C and/or Lewis acid activators.^5d,f,h,iRates of

organolanthanide-catalyzed aminoalkene conversion to the cor-

responding pyrrolidines and piperidines are highly dependent

on ancillary ligation and ionic radius, with complexes that offer

greater access to the Ln center being the most active catalysts.

For example, the 9f10 conversion, promoted by trivalent

Cp′₂Ln- catalysts, proceeds more rapidly with larger radius

metals, i.e., La (N_t) 95 h^-1) > Sm (4.8 h^-1) > Lu (0.03 h^-1

)

at 25 °C. Expanding the degree of coordinative unsaturation

yields a similar trend: (CGC)Lu- (90 h^-1) . Me₂Si(Cp′′)-

(Cp)Lu- (4.2 h^-1) > Me₂SiCp′′₂Lu- (1.6 h^-1) > Cp′₂Lu-

(0.03 h^-1) at25 °C.^3j,u,35Examples of 1° aminoalkene HA/

cyclizations mediated by early transition metal complexes are

limited, proceeding similarly to those with the aforementioned

5d

Ln³⁺, but with far less efficiency (e.g., with Ti(NMe₂)₄and

chelated diamide E).^5fNotably, cationic â-diketiminatoscandium

complex F displays significantly increased catalytic activity vs

the neutral LScMe₂analogue, most likely a consequence of the

more accessible, more electrophilic metal center.^4hSimilar

cationic group 4 species mediate HA/cyclization of 2° ami-

noalkenes, with apparent deactivation by 1° amine substrates

(vide infra).^5h,i

Figure 4. Plots of N_tvs increasing steric openness (360° - θ; Chart 2)

for intramolecular HA/cyclizations of (a) aminoalkene 11f12, (b) ami-

noalkyne 33f34 (ln(N_t) vs increasing openness), (c) aminoallene 43f44,

and (d) aminodiene 49f50 mediated by the indicated organoactinide

complexes 1-6 (L₂An corresponds to 3 and 4, L^2-) [Me₂SiCp′′₂]^2-) in

C₆D₆and extrapolated to 25 °C. Curves are drawn as a guide to the eye.

Reactivity patterns for organoactinide-mediated HA/cycliza-

tion of 1° aminoalkenes largely parallel those of other highly

electrophilic catalysts, with considerably greater activities than

for group 4 catalysts.^5d,f,h,iTerminal aminoalkene conversions

are efficient and highly regioselective at room temperature,

without an induction period,³⁶exclusively providing exocyclic

products. In general, larger, more unsaturated An complexes

mediate this transformation most rapidly. Thus, (CGC)An

complexes 1 and 2 are the most active organoactinide catalysts

yet discovered, with N_tincreasing ca. 10× upon proceeding from

Cp′₂An f Me₂SiCp′′₂An f (CGC)An (Chart 2). Interest-

ingly, the most abrupt activity increase occurs from Cp′₂An to

Me₂SiCp′′₂An , although the Cp(c)-An-L angles discussed

above transition in roughly equal degrees, reflecting a remark-

able sensitivity of this transformation to ancillary ligand

openness. Nevertheless, the N_tincrease on proceeding to more

open precatalysts 1 and 2 is indeed significant, with greater

thermal stability^29,30,37than for 3 and 4 and greater access to

more demanding transformations (vide infra). This trend in

increasing activity with increasing coordinative unsaturation,

as indexed by Cp(c)-M-L (Chart 2), is summarized graphi-

cally in Figure 4a for 11f12, perhaps approaching an asymp-

totic limit.

(35) Where necessary, N_tvalues at 25 °C estimated on the basis of activation

parameters determined in the lead reference.

9

4266 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

The scope of 1° aminoalkene HA/cyclization was further

investigated with 1 and 2 to assess the importance of actinide

ionic radius (Tables 1 and 3). Overall, the data for the wide

range of substrate structures and substitutions generally parallels

4f trends,^2dnamely that larger metal centers more rapidly

mediate HA/cyclization, geminal dialkyl substitution accelerates

cyclizations, and five-membered pyrrolidines form more rapidly

than six-membered piperidines. The Thorpe-Ingold effect³⁸in

stabilizing the likely four-centered, insertive transition state (A)

is pervasive here,^39,40typically resulting in 100-fold rate

enhancements upon geminal dialkyl substitution, exemplified

by substrates 7 vs 9, 13 vs 15, and again for 9 vs 11, 15 vs 17.

Interestingly, this is not the case for the 15f16 conversion

promoted by 2, where N_tis greater than that for 17f18.

Remarkably, 15f16 is also considerably more efficient than

9f10, in spite of the five- to six-membered ring size increase,

contrary to the Baldwin ring-closure rules.⁴¹This interesting

result is likely a consequence of the finely tuned steric

requirements for intramolecular HA/cyclization and is also

observed for similar substrates in reactions catalyzed by

(CGC)U . Internal alkene 31f32 (Table 5) and conjugated 1,3-

diene 51f52a/b (Table 8) conversions, where each substrate

possesses the same H₂NCH₂C(Me)₂CH₂CH₂- structure adjacent

to the CdC fragment, appear to create an almost enzyme-like⁴²

“sweet spot” for cyclization (e.g., eq 10).

demanding cyclizations. As expected on the basis of its structural

similarity to 9, simple 1° amine substrate 19 is rapidly converted

to 20a at 25 °C, with strained bicyclic 20b requiring heating to

100 °C. An identical N_tis observed for 23f24 under similar

conditions (Table 4), where more hindered N-benzyl substitution

is combined with reduced N nucleophilicity. Less forcing

conditions are required for less encumbered N-methyl substrates

21 and 25 (Table 4), indicating that steric factors are important,

although the An⁴⁺size dependence suggests that ancillary

ligation openness is also important. Note also the rate enhance-

ment for 25f26 vs analogous N-unprotected aminoalkyne

39f40 (Tables 5 and 7). In that 2° aminoalkenes present greater

steric impediment to cyclization than their N-unprotected

congeners (see structure A), this result is consistent with

previously reported rate increases with decreasing Ln³⁺ionic

radius in aminoalkyne HA/cyclization (vide infra).^3n,sA similar

enhancement in activity is seen in sequential C-N/C-C bond-

forming processes mediated by Cp′₂Sm- catalysts (eq 11).^3l

Interestingly, in the case of Cp′₂Sm-, alkyne terminus phenyl

substitution leads to a decrease in cyclization rate, indicative

of unfavorable steric interactions during C-C insertion into the

M-N bond in transition state A.

The competence of 1 and 2 in mediating facile 2° aminoalk-

ene transformations argues against an easily accessible AndNR

intermediate/transition state in C-N bond formation (cf. Scheme

3).⁴³Less plausible high-energy species such as H or I must be

invoked in place of more plausible J. Ion pair H, presumably

The catalytic addition of 2° N-H bonds across CdC

functionality was also explored, with 1 and 2 typifying the ability

of more open (CGC)An complexes to mediate more sterically

generated via unimolecular R-H⁺abstraction, features An-

coordinated hypervalent N, while zwitterion I (by way of

N-based electron pair transfer to the An-N bond) places a

formal negative charge on electrophilic Th⁴⁺or U⁴⁺(no change

in ¹H NMR paramagnetic shifts is observed vs similar 1° amine

HA/cyclization processes). In regard to possible radical path-

ways, the high regioselectivity observed for An-mediated

cyclizations of 2° aminoalkene and aminoalkyne substrates is

inconsistent with radical chain processes.⁴⁴Additionally, the

required but energetically unfavorable An⁴⁺fAn³⁺electron

transfer/reduction for An-N(H)R homolysis (particularly for

An ) Th),^12,13the rate dependence on metal ionic radius, the

zero-order dependence on [substrate],²⁰and pronounced effects

(36) Induction periods have been reported in support of a [2+2] cycloaddition

pathway proceeding via a rapid pre-equilibrium for intermolecular HA

mediated by group 4 and An metals. The proposed active species, L₂Md

NR, is postulated to form via R-elimination from L₂M[N(H)R]₂species

(see refs 2a,h-n and 16 for details).

(37) Stubbert, B. D. Ph.D. Thesis, Northwestern University, 2006. Prolonged

reaction times and forcing conditions (T > 120 °C) in alkane and aromatic

solvents do not induce catalyst decomposition with similar 1° and 2° amine

substrates.

(38) Sammes, P. G.; Weller, D. J. Synthesis 1995, 1205-1222.

(39) (a) Motta, A.; Fragala, I. L.; Marks, T. J. Organometallics 2006, 25, 5533-

5539. (b) Motta, A.; Lanza, G.; Fragala, I. L.; Marks, T. J. Organometallics

2004, 23, 4097-4104.

(40) (a) Tobisch, S. Dalton Trans. 2006, No. 35, 4277-4285. (b) Tobisch, S.

Chem. Eur. J. 2006, 12, 2520-2531. (c) Tobisch, S. J. Am. Chem. Soc.

2005, 127, 11979-11988.

(41) March, J. AdVanced Organic Chemistry: Reactions, Mechanisms, and

Structure, 4th ed.; John Wiley & Sons: New York, 1992.

(42) See, for example: (a) Shakhnovich, E. I. Nat. Struct. Biol. 1999, 6, 99-

102. (b) Stryer, L. Biochemistry, 4th ed.; W. H. Freeman and Co.: New

York, 1995; pp 181-206.

(43) Hazari, N.; Mountford, P. Acc. Chem. Res. 2005, 38, 839-849.

(44) (a) Walton, J. C.; Studer, A. Acc. Chem. Res. 2005, 38, 794-802. (b)

Kemper, J.; Studer, A. Angew. Chem., Int. Ed. 2005, 44, 4914-4917. (c)

Newcomb, M.; Burchill, M. T.; Deeb, T. M. J. Am. Chem. Soc. 1988, 110,

6528-6535.

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4267

A R T I C L E S

Stubbert and Marks

Scheme 3. Proposed [2+2] Cycloaddition Mechanistic Pathway

Proceeding through an L₂MdNAr Intermediate (M ) Ti, Zr)^a

regioselectivity and X effects on rate (eq 13).^3e“Unactivated”

substrates 29 and 31 are converted to 30 and 32, respectively,

at elevated temperatures (100 °C; Table 5) but otherwise closely

follow patterns observed for 9f10 and 15f16.

(b) Aminoalkynes. The present (CGC)An complexes rapidly

and regioselectively catalyze aminoalkyne HA/cyclization under

mild conditions. Previous observations with organolanthanide

catalysts indicate extremely high activity for Ln complexes of

smaller ionic radius.^3n,sFor example, 35f36 conversion is

efficiently mediated at 21 °C by Cp′₂Ln-R precatalysts, with

the observed relative rates Lu (5.3) > Sm (4.3) > Nd (1.5) >

La (1). This trend is paralleled by the ca. 40× fall in activity

upon opening the catalyst coordination sphere of Cp′₂Sm- (77

h^-1) to Me₂SiCp′′₂Sm- (2 h^-1) in 39f40 HA/cyclization,

exactly contrary to aminoalkene trends.^2dGroup 4 catalysts also

mediate the intramolecular HA/cyclization of highly reactive

aminoalkynes, displaying greater reactivity with M ) Ti than

with M ) Zr, although electronic considerations and ancillary

ligand steric requirements are less clearly defined.^2a,h-nRates

with Ti⁴⁺complexes (proposed to proceed through a [2+2]

cycloaddition pathway, Scheme 3) are, in some cases, compa-

rable to those obtained with Ln³⁺complexes.²

Consistent with these observations, intramolecular ami-

noalkyne HA/cyclization mediated by 1 and 2 proceeds rapidly

and regioselectively (Table 6). N_tvalues are comparable to or

greater than those of L₂Ln- catalysts.³ⁿThus, conversions

35f36 and 37f38, mediated by 2, proceed at ca. twice the

rate of reactions with Cp′₂Sm- catalysts, establishing 2 as the

most efficient catalyst reported for this transformation. While

the 2-mediated conversion 33f34 is extremely rapid (N_t≈ 3000

h^-1, 25 °C), it is even more so when catalyzed by similarly

sized L₂Ln- species.^3n,sRegarding transition state A and the

tendency of Me₃Si- functionalities to stabilize R-carbanions

and â-carbocations²⁶(outweighing steric impediments to CtC

insertion), this difference in reactivity may reflect a reduced

tendency for An⁴⁺(vs Ln³⁺) ions to stabilize γ-agostic interac-

tions with proximate Si-C bonds.^12fOr, considering the

presence of a second, covalently bonded substrate moiety, this

difference may reflect a balance in steric vs electronic factors

in the highly polarized transition state (vide infra). Notably, more

electrophilic Th⁴⁺appears to favor arene coordination/transition-

state stabilization (as discussed above for 27f28) more ef-

fectively than U⁴⁺, perhaps reflecting differences in 5f-element

bond covalency.^12,13Moreover, the dispersion in rates with

alkyne substitution for 1 (-SiMe₃> -H > -Ph > -Me) vs

the more Cp′₂Sm-like³ⁿbehavior of 2 (-SiMe₃. -H . -Me

> -Ph) suggests a complex interplay of sterics and electronics.

^aThis mechanistic scenario is consistent with the observed rate law, ν

∼ [L₂M⁴⁺]¹[alkyne]¹[amine]^-1, in the intermolecular HA of alkynes with

bulky (aromatic) amines. A similar pathway is proposed for Cp′₂An⁴⁺

-

mediated processes displaying zero-order dependence on [alkyne] (see refs

2 and 47-50).

of ancillary ligation on N_tand cyclization diastereoselectivity

all render radical pathways unlikely, especially in view of the

poor regioselectivity and overall yield in radical-mediated

aminoalkene cyclizations.^44b,cResults for cationic group 4 metal-

catalyzed aminoalkene HA/cyclizations, where 1° aminoalkenes

deactivate L₂Ti-⁺and L₂Zr-⁺catalysts,^5h,isuggest that, in these

cases, neutral L₂MdNR imido species are unreactive with

respect to CdC insertion (eq 12), unlike the highly reactive

intermediate proposed in group 4 intermolecular HA.^2aAd-

ditional experiments probing the detailed mechanism of in-

tramolecular 1° and 2° amine HA/cyclization mediated by

(CGC)An and (CGC)Zr catalysts will be reported elsewhere.¹⁶

Regarding internal (1,2-disubstituted) aminoalkene HA/cy-

clizations, trends are similar to those observed for terminal

olefinic substrates and those mediated by 4f catalysts.^2dSpecif-

ically, Th-mediated transformations are generally more rapid

than those mediated by U⁴⁺, and more open (CGC)An and Me₂-

SiCp′′₂An complexes are more active than Cp′₂An catalysts

(Table 5). Specifically, (CGC)An catalysts are ∼10× more

active than the corresponding Me₂SiCp′′₂An catalysts. In the

conversion 27f28, it is plausible that the electron-withdrawing

phenyl ring stabilizes the δ^-charge in the proposed transition

state K (R ) Ph), thereby assisting turnover-limiting CdC

insertion, despite unfavorable steric interactions (Table 4).^3e,12

Additional stabilization may be provided by arene ring π-co-

ordination to the electrophilic L₂An center (K). Studies of Ln-

mediated intermolecular substituted styrene HA invoked a

similar pathway to explain the impressive anti-Markovnikov

9

4268 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

Ancillary ligand effects on aminoalkyne HA/cyclization were

also investigated with 1-6. Reactivity patterns largely parallel

those of the Cp′₂Ln- and Me₂SiCp′′₂Ln- systems,³ⁿwith rates

following the general trend Cp′₂An > Me₂SiCp′′₂An , and ca.

10× greater activity for the larger Th⁴⁺center with each ligand

set. However, a sharp deviation from this trend is observed with

the markedly more open, coordinatively unsaturated (CGC)An

catalysts, with relative rates for 33f34 of roughly 82 (1):12

(3):16 (5) h^-1for An ) Th and 3000 (2):1 (4):2 (6) h^-1for An

) U. This reveals a remarkable increase in catalytic activity

(>10³-fold; Figure 4b) with (CGC)An catalysts and provides

a dramatic demonstration of successfully combining the elec-

tronic and coordinative unsaturation characteristics of CGC

ligation in organometallic catalysis.¹¹

(c) Aminoallenes. Intramolecular (CGC)An -mediated HA/

cyclization of 1,3-disubstituted allenes covalently linked to 1°

amines proceeds under mild conditions, displaying moderate

trans-2,5-pyrrolidine and trans-2,6-piperidine diastereoselec-

tivity along with >90% regioselectivity. Compared to similar

transformations effected by group 4^5k,land Ln complexes,^3h,i,k

(CGC)An complexes display comparable activities and activity

trends. The scope of such conversions previously investigated

with organo-group 3 and 4f catalysts demonstrated impressive,

regioselective routes to natural products.^3hMechanistic analysis

revealed an unusual dependence on Ln³⁺ionic radius, with

maximum N_tvalues for intermediate ionic radii, and an “alkyne-

like” dependence on catalyst ancillary ligation, with (CGC)Y-

catalyzed reactions exhibiting marked diminution of cyclization

rates vs Cp′₂Y-.³ⁱFurthermore, group 4 catalysts with chelating

bis(sulfonamide) ligands (likely providing greater steric open-

ness and electrophilicity) exhibit striking enhancements in

activity as well as scope over traditional group 4 metallocenes.^5k,l

Interestingly, regiospecificities for Ln- and Ti-based processes

are reversed, with Ln³⁺and An⁴⁺catalysts selectively generating

exocyclic rings (pathway a, eq 14) vs formation of endocyclic

regioisomers with Ti⁴⁺catalysts (pathway b, eq 14);^3i,5kboth

pathways are favorable within the Baldwin ring-closure rules.⁴¹

Zr-mediated transformations also operate via pathway a and

cyclize 1,3,3-trisubstituted aminoallenes in moderate yields and

with >90% regioselectivity, although more sluggishly than those

with similarly ligated Ti catalysts.^5l

A (Scheme 2) and rate-limiting CdC insertion into the An-N

bond (cf. G, G^q, and J). Likewise, the additional bond covalency

of An (vs Ln) ions may help stabilize A. A qualitative

comparison of coordinative unsaturation effects on N_t(according

to the Cp(c)-M-L angle θ in Chart 2) is presented in Figure

4c for 43f44 conversion mediated by precatalysts 1-6.

More striking are Me₂SiCp′′₂An f (CGC)An ancillary

ligand effects, where cyclization rates are comparable to those

of rapid, Cp′₂Ln-promoted reactions.³ⁱWith the exception of

the “preorganized” aminoallene 43, which exhibits a 10× rate

enhancement for cyclization mediated by Cp′₂Sm- vs (CGC)U ,

there is virtually no loss in activity from Cp′₂Ln- to (CGC)U

catalysts. However, the ca. 10× decrease in activity observed

for 1 vs 2 reflects a Ln-like dependence on ionic radius in the

sterically sensitive intramolecular 1,3-disubstituted allene HA/

cyclization. Although substrates bearing geminal dialkyl sub-

stitutions were not investigated, the substantial increase in

activity of both 1 and 2 for transformations 45f46 vs 47f48

(Table 7) is consistent with the more advantageous orientation

of the reactive substrate moiety in the transition state owing to

R-alkyl substitution.³⁸

Although diastereoselectivity was not reported in detail for

group 4 catalysts,^5k,lorganolanthanides consistently afford 2,5-

trans-pyrrolidine and 2,6-cis-piperidine heterocyclic products.³ⁱ

In all instances, organoactinide catalysts (Table 7, entries 1-8)

consistently favor 2,5-trans-pyrrolidine (43f44, precatalysts

1-6) and 2,6-trans-piperidine (45f46, precatalysts 1 and 2)

diastereoselection. The former trend tracks that observed with

organolanthanide catalysts and is explicable in terms of preferred

orientation and minimization of unfavorable substrate-ancillary

ligand steric repulsion. However, the latter result differs from

that obtained with L₂Ln- catalysts, likely reflecting the second,

covalently bonded substrate interfering with CdCdC insertion

at the An-N bond. Added steric obstruction in the already

congested transition state (cf. A and G^q) is consistent with

reduced diastereoselectivity and reduction in catalytic activity

upon moving from coordinatively unsaturated (CGC)An com-

plexes, where diastereoselectivities up to 90% are achieved, to

their less open 3-6 analogues, where diastereoselectivies are

uniformly modest and reaction rates are markedly more slug-

gish.

(d) Aminodienes. While ample precedent exists for inter-

molecular 1,3-diene HA across the Periodic Table,^2kthere have

been few investigations of the intramolecular variant of this

useful transformation. Notably, scope and mechanism have been

thoroughly examined for Ln catalysts with diverse ancillary

ligands, from conventional Cp′₂Ln- to chiral bis(oxazolinato)

complexes.^3c,d,fThese studies reveal a dependence on Ln³⁺ionic

radius and coordinative unsaturation that is far more pronounced

than in aminoalkene intramolecular HA/cyclization, where

larger, more accessible metal centers with more open coordina-

tion spheres are optimal for rate and regioselectivity. Although

recent computational studies⁴⁰suggest that Ln-C σ-bond

protonolysis following aminodiene cyclization may be turnover-

limiting,⁴⁵experimental data argue for turnover-limiting C-C

unsaturation insertion into the Ln-N bond (Scheme 1),^3cas do

Overall, (CGC)An catalyst selectivity and activity (Table 7)

more closely resemble organolanthanide patterns than those of

Ti/Zr- catalysts. However, comparing Me₂SiCp′′₂An vs Cp′₂An

coordinative unsaturation, precatalysts 3-6 exhibit sluggish

efficiencies and moderate diastereoselectivities, unlike the

general trend observed with lanthanocenes (increasing reactivity

with decreasing coordinative unsaturation).³ⁱRather the An trend

resembles smaller, tetravalent Ti/Zr patterns, suggesting that

increased encumbrance introduced by an additional covalent

σ-bond may impose greater steric constraints on transition state

(45) Note that the marked dependence of the cyclization rate on ring size and

significant C-C substituent electronic effects for alkene, alkyne, allene,

and diene unsaturations seem to preclude turnover-limiting heterocycle

protonolysis after C-C insertion into M-N (i.e., k_protonolysis> k_insertion

appears operative here).

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4269

A R T I C L E S

Stubbert and Marks

computational results for aminoalkene and aminoalkyne HA/

cyclization.³⁹

alkenes), the rate dependence on metal ionic radius is negligible.

In qualitative terms, the softer, more electron-rich substrates

(i.e., alkynes > allenes > dienes) more rapidly undergo insertion

at the slightly softer¹³U⁴⁺metal center, whereas the less soft

alkene substrates more rapidly undergo insertion at the slightly

harder¹³Th⁴⁺metal center. This ordering does not hold in

organo-4f-element HA/cyclization,^2dsuggesting that increased

bond covalency^12,13may be operative in organoactinide vs

organolanthanide catalysis.

Intramolecular 1,3-aminodiene HA/cyclization, involving both

terminal and 1,4-disubstituted (internal) diene units, is effectively

catalyzed by (CGC)An complexes (Table 8). As for organo-

4f-element catalysts, such transformations are efficient and

highly regioselective. With the noteworthy exception of the

2-promoted 51f52a/b conversion (vide supra), relative ring-

closure rates follow expected trends where smaller pyrrolidines

form more rapidly from unsubstituted 1,3-dienes than do

piperidines, and 1,4-disubstituted diene conversions are slow

vs those of terminal dienes.^3cThis trend is typical of ring-closing

reactions in general,^38,41,46and (CGC)An -catalyzed aminodiene

HA is clearly governed by similar steric and electronic demands.

Furthermore, product distributions typically favor thermody-

namically less favored (∼3 kcal/mol) terminal, monosubstituted

olefinic products (50a and 52a, Table 8). That product distribu-

tions are time-independent and favor the kinetic cyclization

product (a in eq 15) suggests that subsequent isomerization via

Mechanistic Implications. A priori, the present tetravalent

organoactinide complexes could mediate catalytic intramolecular

HA/cyclization via either of two distinct, limiting mechanistic

scenarios: (1) a polar, highly organized four-centered, insertive

transition state involving L₂An species possessing two An-N

σ-bonds (“Ln-like”; Scheme 2), or (2) a [2+2] cycloaddition

pathway proceeding via reactive L₂AndNR species, analogous

to that proposed for intermolecular alkyne HA with bulky

amines, mediated by Cp₂MR₂, Cp₂MdNR(THF), and Cp′₂-

AnMe₂catalysts (M ) Ti, Zr; An ) Th, U) at elevated

temperatures (Scheme 3).^{2a,e,f,h-n,47,48}In group 4 cases, scenario

(2) has been shown to involve a rapid pre-equilibrium between

Cp₂M[N(H)Ar]₂and Cp₂MdNAr(H₂NAr) (B) species, followed

by turnover-limiting [2+2] RCtCR′ cycloaddition to the

Cp₂MdNAr species (step (ii), Scheme 3), in agreement with

the observed rate law (eq 16a).⁴⁷Slightly different kinetics were

reported for Cp′₂AnMe₂-catalyzed intermolecular additions of

bulky and unhindered aliphatic 1° amines to terminal alkynes,

although with zero-order dependence on [alkyne] (rates were

independent of alkyne identity), suggesting rapid CtC addition

following turnover-limiting amine elimination, and with com-

peting alkyne oligomerization processes.^48,49Although incon-

sistent with noncatalytic L₂AndNR chemistry,^12-14,28,30an

alternative scenario can also be envisioned where an MdNR

species forms irreversibly, followed by rapid CtC addition

(eq 16b).

reinsertion at the An center (b in eq 15) is unimportant, in

contrast to most Ln-catalyzed transformations.^3dThis trend for

Ln catalysts is explicable in terms of well-precedented η³-allyl/

benzyl intermediates (L)^12,13and suggests that such intermediates

are more stable for large Ln³⁺ions than for similarly sized An⁴⁺

ions, having two An-N σ-bonded substrates in the resting state

(M).

-

In contrast, more active cationic U(NEt₂)₃⁺BPh₄displays

behavior similar to that of Cp′₂An - and Cp₂Zr -catalyzed

intermolecular terminal alkyne HA with 1° amines, but also

mediates HNEt₂addition to Me₃SiCtCH,⁵⁰precluding an

obvious AndNR intermediate. Moreover, cationic organozir-

(47) (a) Baranger, A. M.; Walsh, P. J.; Bergman, R. G. J. Am. Chem. Soc. 1993,

115, 2753-2763. (b) Walsh, P. J.; Baranger, A. M.; Bergman, R. G.

J. Am. Chem. Soc. 1992, 114, 1708-1719.

The effects of ancillary ligand openness, as assayed by

Cp(c)-M-L (Chart 2) on the present An-catalyzed process

(Figure 4d) parallel those in organolanthanide catalysis and those

generally seen in Ln- or An-catalyzed aminoalkene HA/

cyclization.^2d,15In contrast, the rate dependence on An⁴⁺ionic

radius (rate: U > Th) resembles more thermodynamically

favored N-H bond additions to aminoalkyne CtC and ami-

noallene CdCdC units. From a Hammond postulate stand-

point,^20,41such behavior suggests a diene HA transition state

more closely resembling reactants than products in each

transformation class.

(48) (a) Straub, T.; Haskel, A.; Neyroud, T. G.; Kapon, M.; Botoshansky, M.;

Eisen, M. S. Organometallics 2001, 20, 5017-5035. (b) Haskel, A.; Straub,

T.; Eisen, M. S. Organometallics 1996, 15, 3773-3775.

(49) For the following pathway ([A] ) [MdNR]),

d[A]

) k₁[M(NHR)₂] - k₂[A][RNH₂] - k₃[A][CtC] ) 0

dt

w k₂[A][RNH₂] + k₃[A][CtC] ) k₁[M(NHR)₂]

To rationalize organoactinide reactivity patterns for all HA

reactions, it is beneficial to consider the estimated enthalpies

associated with each class of cyclization (Table 9). For

transformations where Scheme 2, step (i) is more exothermic

vs step (ii), the U⁴⁺-catalyzed processes are more rapid; when

steps (i) and (ii) are close in exothermicity (i.e., 1,2-disubstituted

k₁[M(NHR)₂]

[A] )

k₂[RNH₂] + k₃[CtC]

If the rate of product formation is ν ) k₃[A][CtC], then

k₁k₃[M(NHR)₂][CtC]

ν )

k₂[RNH₂] + k₃[CtC]

Therefore, (i) for k₃very large, ν ∼ [M(NHR)₂]¹[RNH₂]⁰[CtC]⁰; (ii) for

k₂very large, ν ∼ [M(NHR)₂]¹[RNH₂]^-1[CtC]¹; and (iii) for k₃very small,

ν ∼ [M(NHR)₂]¹[RNH₂]^-1[CtC]¹.

(46) (a) Illuminati, G.; Mandolini, L. Acc. Chem. Res. 1981, 14, 95-102. (b)

Mandolini, L. J. Am. Chem. Soc. 1978, 100, 550-554.

9

4270 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

Constrained Geometry Organoactinides as Catalysts

A R T I C L E S

Activation parameters derived for each C-C unsaturation

(Table 9) also suggest a mechanistic pathway closely resembling

that proposed for organolanthanide-mediated HA processes,

where deviation from the generally preferred trivalent Ln³⁺state

is unlikely.¹³Activation parameters for the present (CGC)An -

mediated HA/cyclizations are similar to those for the Cp′₂AnMe₂-

mediated intermolecular terminal alkyne HA (∆H^q) 11.7(3)

kcal/mol, ∆S^q) -44.5(8) eu).⁴⁸The more negative ∆S^qis not

unexpected in comparison to intramolecular aminoalkene (∆H^q

) 12.6(2) kcal/mol, ∆S^q) -29.6(3) eu) and aminoalkyne HA/

cyclization (∆H^q) 13.9(2) kcal/mol, ∆S^q) -27(5) eu; Table

9) processes and suggests that (CGC)An -catalyzed intramo-

lecular HA/cyclizations proceed through a transition state

energetically similar to that postulated for Ln-catalyzed HA/

cyclization (A). Activation parameters for (CGC)An -catalyzed

HA/cyclization are very similar to (within 3σ) those for

Ln-catalyzed reactions (Table 9),^2d,3in accord with similar,

highly organized insertive transition states. Additional mecha-

nistic data and arguments will be published separately.¹⁶

conium and organotitanium complexes effect intramolecular

aminoalkene HA/cyclizations, with scope limited to 2° amine

substrates, again suggesting that a neutral L₂MdNR species

cannot be an intermediate (eq 12).^5h,iNote that monomeric

tetravalent actinide imido complexes are normally observed with

sterically encumbered (e.g., Cp′₂An ) ancillary ligands and result

frombis(amido)actinideorrelatedprecursors(e.g.,Cp′₂An[N(H)Ar]₂

or Cp′₂An(Me)[N(H)Ar]) at elevated temperatures via R-H

elimination.^{12-14,28,30,43,47,48}These considerations argue against

the pathway of eq 16b, where the predicted rate law, although

in agreement with the presently observed (intramolecular HA/

cyclization) rate law, assumes a stable, monomeric L₂AndNR

intermediate. Under the mild reaction conditions typical of HA/

cyclizations reported here, involvement of L₂AndNR intermedi-

ates seems unlikely, considering the coordinative unsaturation

of the (CGC)An species involved, the reactivity of sterically

unencumbered substrates, the facile 2° amine cyclizations, the

absence of induction periods prior to catalytic turnover,³⁶and

the substantial rate differences observed for different C-C

unsaturations. Furthermore, the rate acceleration at low [sub-

strate] (L₂MdNR formation favored) and rate deceleration at

high [substrate] (L₂M[N(H)R]₂favored) seen in [2+2] cycload-

dition pathways⁴⁷are not observed here. On the contrary,

(CGC)An -mediated intramolecular 1° or 2° amine HA/cycliza-

tions display opposing behavior (namely, high [product] mod-

erately inhibits catalytic turnover via competitive binding;

Figures 2 and 3)^2d,3v,15and exhibit the rate law ν ∼ [An]¹-

[amine]⁰without an induction period, inconsistent with scenario

(2) and signifying that turnover is not preceded by a rapid pre-

equilibrium. Aminoalkene and aminoalkyne HA/cyclizations

mediated by monosubstituted L₂M(NR₂)X complexes, incapable

of MdNR formation with 1° or 2° amine substrates,¹⁶further

support mechanistic scenario (1), characterized by a highly

organized transition state with turnover-limiting⁴⁵C-C insertion

into a σ-bonded M-N species (cf. A). This is also in agreement

with DFT-level computational studies of organolanthanide-

mediated HA/cyclization.³⁹

Conclusions

The catalytic competence of readily accessible constrained

geometry (CGC)An(NR₂)₂complexes in mediating intramo-

lecular hydroamination/cyclization reactions of a large variety

of primary and secondary amines tethered to multiple classes

of terminal and disubstituted C-C unsaturations has been

investigated in detail. The reaction scope includes aminoalkenes,

aminoalkynes, aminoallenes, and aminodienes, with catalytic

efficiencies comparable to those of the most active organolan-

thanide catalysts and far exceeding activities reported for most

tetravalent transition metal complexes. Reactivity trends reveal

a marked dependence on An⁴⁺ionic radius and ancillary

ligation, with pronounced reactivity enhancement afforded by

the CGC framework, regardless of C-C unsaturation. Experi-

mentally determined ∆H^q, ∆S^q, and E_aparameters are similar

to those for organolanthanide-mediated processes, consistent

with highly organized transition states exhibiting considerable

bond-forming character. These facts, along with the rate law

and other similarities to organolanthanide-catalyzed HA/cy-

clization processes, suggest a “Ln-like” σ-bond insertive

mechanistic pathway proceeding through an An-N(H)R inter-

mediate (Scheme 2).

Acknowledgment. We gratefully acknowledge the National

Science Foundation (CHE-0415407) for funding this research.

We also thank Ms. Charlotte L. Stern for collection and

refinement of single-crystal X-ray diffraction data (ref 15).

Supporting Information Available: Full experimental details

for substrate preparation, including improved synthetic proce-

dures for 21, 25, and 41, and (CGC)An complexes. This

material is available free of charge via the Internet at

http://pubs.acs.org.

(50) Wang, J.; Dash, A. K.; Kapon, M.; Berthet, J.-C.; Ephritikhine, M.; Eisen,

M. S. Chem. Eur. J. 2002, 8, 5384-5396.

JA0665444

9

J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4271

Article Doi

Constrained geometry organoactinides as versatile catalysts for the intramolecular hydroamination/cyclization of primary and secondary amines having diverse tethered C-C unsaturation

DOI: 10.1021/ja0665444

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja0665444

Full text of DOI:10.1021/ja0665444

Products guided by the article

R&D Labs maybe for 5913-89-3

Relevant to this article

Hot Product