Mechanistic studies of copper(I)-catalyzed allylic amination

Article abstract of DOI:10.1021/ja0751072

The reactions of nitrosobenzene and N,N′-diethyl-4-nitrosoaniline with [Cu(CH₃CN)₄]PF₆ provide novel Cu(I) complexes, [Cu(PhNO)₃]PF₆ (1) and [Cu(Et ₂NPhNO)₃]PF₆ (2); in 2 the copper atom is N-coordinated to the nitrosoarenes in a distorted trigonal planar geometry. Complex 1 is strongly implicated as a reactive intermediate in the Cu(I)-catalyzed allylic amination of olefins based on (i) its isolation from the catalytic reaction, (ii) its stoichiometric regioselective allylic amination of α-methyl styrene (AMS), (iii) the non-involvement of free PhNO in its amination of AMS, and (iv) its function as a catalyst for the amination of alkenes from phenylhydroxylamine. The reaction between AMS and 1 (80°C, dioxane) is first order in both alkene and 1. Relative rate studies of the reaction of 1 with para substituted AMS derivatives gives a Hammett ρ value of -0.035. Alkene adducts isolated from the reaction of 1 with styrene and α-methylstyrene are formulated as [(PhNO)₃Cu(η²- alkene)]PF₆ (7,8) on the basis of spectroscopic characterization and thermolysis. PM3 and DFT MO calculations support the role of [(alkene)Cu(RNO)₃]⁺ and (η¹- or η³-allyl)Cu(RNO)₂(RNHOH)⁺ complexes as probable catalytic intermediates and address the origin of the distinctive reaction regioselectivity. A mechanistic scheme is proposed which is consistent with the accumulated experimental and computational results.

Full text of DOI:10.1021/ja0751072

Published on Web 11/16/2007
Mechanistic Studies of Copper(I)-Catalyzed Allylic Amination
Radhey S. Srivastava,*,† Nathan R. Tarver,† and Kenneth M. Nicholas‡
Contribution from the Department of Chemistry, UniVersity of Louisiana at Lafayette,
Louisiana 70504, and Department of Chemistry and Biochemistry, UniVersity of Oklahoma,
Norman, Oklahoma 73109
Received July 10, 2007; E-mail: rss1805@louisiana.edu
Abstract: The reactions of nitrosobenzene and N,N′-diethyl-4-nitrosoaniline with [Cu(CH3CN)4]PF6 provide
novel Cu(I) complexes, [Cu(PhNO)3]PF6 (1) and [Cu(Et2NPhNO)3]PF6 (2); in 2 the copper atom is
N-coordinated to the nitrosoarenes in a distorted trigonal planar geometry. Complex 1 is strongly implicated
as a reactive intermediate in the Cu(I)-catalyzed allylic amination of olefins based on (i) its isolation from
the catalytic reaction, (ii) its stoichiometric regioselective allylic amination of R-methyl styrene (AMS), (iii)
the non-involvement of free PhNO in its amination of AMS, and (iv) its function as a catalyst for the amination
of alkenes from phenylhydroxylamine. The reaction between AMS and 1 (80 °C, dioxane) is first order in
both alkene and 1. Relative rate studies of the reaction of 1 with para substituted AMS derivatives gives
a Hammett F value of -0.035. Alkene adducts isolated from the reaction of 1 with styrene and
R-methylstyrene are formulated as [(PhNO)3Cu(η2-alkene)]PF6 (7,8) on the basis of spectroscopic
characterization and thermolysis. PM3 and DFT MO calculations support the role of [(alkene)Cu(RNO)3]+
and (η1- or η3-allyl)Cu(RNO)2(RNHOH)+ complexes as probable catalytic intermediates and address the
origin of the distinctive reaction regioselectivity. A mechanistic scheme is proposed which is consistent
with the accumulated experimental and computational results.
lamines (Scheme 1). Nitroarene-16 and aminoarene-based,17
metal-catalyzed allylic aminations have also been developed by
Introduction
Catalytic reactions that transfer an -NR group with C-H
insertion can convert hydrocarbons into valuable amines. The
development of new catalytic reactions for the efficient synthesis
of allyl amines from inexpensive feedstocks, remains an enticing
goal in both industry and in the laboratory, owing to the
importance of these amines as intermediates in organic synthesis
and as useful end-products.1-5 Convenient methods for direct
allylic amination have been sought, including alkene reactions
with S/Se-imido reagents6,7 and ene-reactions with azo-,8
N-carboalkoxynitroso-,9 and N-sulfinylcarbamate derivatives.10
Transition-metal-promoted allylic amination of unsaturated
hydrocarbons presents an attractive alternative means to func-
tionalize olefins via C-N bond formation.11,12 We and others
contributed to the early development of Mo(VI)13 and Fe(II,-
III)14,15-catalyzed allylic amination of alkenes by aryl hydroxy-
us and the Cenini/Ragaini group. More recently, we18 and a
Chinese group19 independently found that the Cu-catalyzed
reactions of olefins with aryl hydroxylamines produce moderate
(6) (a) Sharpless, K. B.; Hori, T. J. Org. Chem. 1976, 41, 176. (b) Kresze, G.;
Braxmeier, H.; Munsterer, H. Org. Synth. 1987, 65, 159. (c) Katz, T. J.;
Shi, S. J. Org. Chem. 1994, 59, 8297.
(7) (a) Sharpless, K. B.; Hori, T.; Truesdale, L. K.; Dietrich, C. O. J. Am.
Chem. Soc. 1976, 98, 269. (b) Bruncko, M.; Khuong, T. A. V.; Sharpless,
K. B. Angew. Chem., Int. Ed. Engl. 1996, 35, 454.
(8) Brimble, M. A.; Heathcock, C. H. J. Org. Chem. 1993, 58, 5261. Leblanc,
Y.; Zamboni, R.; Bernstein, M. A. J. Org. Chem. 1991, 56, 1971.
(9) Keck, G. E.; Webb, R. R.; Yates, J .B. Tetrahedron 1981, 37, 4007. Knight,
G. T. Chem. Commun. 1970, 1016. Bank, R. E.; Hazeldine R. N.; Miller,
P. J. Tetrahedron Lett. 1970, 4417.
(10) Whitesell, J. K.; Yaser, H. K. J. Am. Chem. Soc. 1991, 113, 3526.
(11) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004, 37, 673. Pohlki, F.; Doye, S.
Chem. Soc. ReV. 2003, 32, 104. Kolb, H. C.; Sharpless, K. B. Transition
Met. Org. Synth. 1998, 2, 243. Mueller, P.; Fruit, C. Chem. ReV. 2003,
103, 2905. Katsuki, T. Comp. Coord. Chem. 2 2004, 9, 207.
(12) Jorgensen, K. A. In Modern Amination Methods; Ricci, A., Ed.; Wiley-
VCH: Weinheim, Germany, 2000; Chapter 1, p 1-35.
† University of Louisiana at Lafayette.
(13) (a) Srivastava, A.; Ma, Y.; Pankayatselvan, R.; Dinges, W.; Nicholas, K.
M. J. Chem. Soc., Chem. Commun. 1992, 853. (b) Srivastava, R. S.;
Nicholas, K. M. J. Org. Chem. 1994, 59, 5365.
‡ University of Oklahoma.
(1) Jacobson, A. E; May, E. L.; Sargent, L. J. In Medicinal Chemistry; Burger,
A., Ed.; Wiley: New York, 1970; Vol. 2, Chapter 49, pp 1330-1350.
Lednicer, D.; Mitscher, L. A. In Medicinal Chemistry; Burger, A., Ed.;
Wiley: New York, 1984; Vol. 3, pp 116, 190.
(14) (a) Srivastava, R. S.; Nicholas, K. M. Tetrahedon Lett. 1994, 35, 8739. (b)
Srivastava, R. S.; Khan, M. A.; Nicholas, K. M. J. Am. Chem. Soc. 1996,
118, 3311. (c) Srivastava, R. S.; Nicholas, K. M. J. Am. Chem. Soc. 1997,
119, 3302. (d) Singh, S.; Nicholas, K. M. Synth. Commun. 2001, 31, 3087.
(15) Johannsen, M.; Jorgensen, K. A. J. Org. Chem. 1994, 59, 214.
(16) Srivastava, R. S.; Nicholas, K. M. Chem. Commun. 1998, 2705. Kolel-
Veetil, M.; Khan, M. A.; Nicholas, K. M. Organometallics, 2000, 19, 3754.
Srivastava, R. S.; Nicholas, K. M. Tetrahedron Lett. 2002, 43, 931-934.
Cenini, S.; Ragaini, F.; Tollari, S.; Paone, D. J. Am. Chem. Soc. 1996,
118, 11964. Ragaini, F.; Cenini, S.; Tollari, S.; Tummolillo, G.; Beltrami,
R. Organometallics 1999, 18, 928.
(2) Comer, W. T.; Opomoll, A. W. In Medicinal Chemistry; Burger, A., Ed.;
Wiley: New York, 1970; Vol. 2, pp 1019-1064.
(3) Review: Ryder, N. S.; Stuetz, A.; Nussbaumer, P.ACS Symp. Ser. 1992,
6497, 192-204.
(4) Berdy, J.; Aszalos, A.; Bostian, M.; McNitt, K. L. Handbook of Antibiotic
Compounds; CRC Press, Boca Raton, FL, 1928; Vol. 1. Kennedy, J. F.;
White, C. A. BioactiVe Carbohydrates; Halstead Press: Chichester, England,
1983; pp 264-287.
(5) Myers, D. In Surfactant Science and Technology; VCH: Weinheim,
Germany, 1988; pp 20, 62-67. Kirk-Othmer Concise Encyclopedia of
Chemical Technology; Grayson, M., Ed.; Wiley and Sons: New York, 1985;
pp 82-87.
(17) Srivastava, R. S.; Nicholas, K. M. Chem. Commun. 1996, 2335.
(18) Hogan, G. A.; Gallo, A. A.; Nicholas, K. M.; Srivastava, R. S. Tetrahedron
Lett. 2002, 43, 9505.
(19) Ho, C.-M.; Lau, T.-C. New J. Chem. 2000, 24, 859.
9
15250
J. AM. CHEM. SOC. 2007, 129, 15250-15258
10.1021/ja0751072 CCC: $37.00 © 2007 American Chemical Society

Copper(I)-Catalyzed Allylic Amination
A R T I C L E S
dark-red solid was triturated with diethyl ether (2 × 10 mL), and then
dried under vacuum. 1H NMR (CD2Cl2, δ): 8.0 (d, J ) 8 Hz, 2H), 7.8
(t, J ) 8 Hz, 1H), 7.6 (t, J ) 8 Hz, 2H). IR (KBr, cm-1): 1600, 1437,
1398, 1326. Visible spectrum (CD2Cl2): 420 nm (ꢀ 19542 L‚mol-1‚cm-1).
Attempts to obtain a mass spectrum of 1 (EI, ESI, FAB) failed to
provide interpretable fragments as did efforts to obtain suitable crystals
of labile 1 for X-ray diffraction. ICP analysis of 1 digested in 1 mL of
concentrated HCl had a Cu content of 11.7%; Cu calcd for [Cu(PhNO)3]-
PF6 ) 12.0%. The C,H,N-microanalyses of samples of 1 by commercial
providers were unsatisfactory. When the above synthetic procedure was
employed with [Cu(CH3CN)4]PF6 (0.24 g, 0.64 mmol) and PhNO (0.30
g, 2.8 mmol), trituration of the reaction residue with Et2O recovered
0.083 g (0.78 mmol) of unreacted PhNO (containing a small amount
of azoxybenzene detected by GC). Assuming complete conversion of
[Cu(CH3CN)4]PF6 the Cu:PhNO ratio in the product complex is
calculated to be 1:3.1.
yields of allyl amines regioselectively, with N-functionalization
at the less substituted of the original olefinic carbons. Addition-
ally, Cu(II) salts catalyze the R-amination of saturated ketones.19
Our initial mechanistic probes of allylic amination catalyzed
by Cu(I) suggested the involvement of coordinated organoni-
trogen species as the ArN-transfer agent.18
In our preliminary report on the copper species involved in
allylic amination we described the first homoleptic nitrosoarene-
Cu(I) complexes, [Cu(PhNO)3]PF6 (1) and [Cu(Et2NPhNO)3]-
PF6 (2), and implicated 1 as an intermediate in amination on
the basis of its ability to stoichiometrically aminate alkenes.20
Other group 10 metal complexes of nitrosoarenes have also been
proposed as intermediates in enantioselective hetero-Diels-
Alder21 and O-nitroso aldol reactions.22 Further synthetic
development of the allylic amination and related reactions would
be aided by a better understanding of the reaction mechanism.
We report here the full details of our preliminary synthetic and
mechanistic studies along with new kinetic and reactivity
investigations of 1 and 2, both experimental and computational,
including the isolation of an alkene adduct 7 of a type that
appears to be on the catalytic pathway as well.
Synthesis of [Cu(Et2NPhNO)3]PF6 (2). [Cu(CH3CN)4]PF6 (0.22 g,
0.59 mmol) was dissolved in dichloromethane (50 mL) and then solid
N,N′-diethyl-4-nitrosoaniline (0.45 g, 2.5 mmol) was added. The initially
colorless solution changed to dark red-green immediately. After stirring
24 h the dark red-green solution was filtered, and the solvent was
removed from the filtrate in vacuum. The solid residue was triturated
with diethyl ether (30 mL × 3) and vacuum-dried to obtain 2 as a dark
red-green solid (70% based on [Cu(Et2NPhNO)3]PF6). Recrystallization
from CH2Cl2/diethyl ether at -20 °C provided greenish crystals suitable
for X-ray diffraction. IR (KBr, cm-1): 1600, 1420, 1375, 1330. MS
(FAB): 419 [(Et2NPhNO)263Cu]+, 421[(Et2NPhNO)2 65Cu]+. UV-vis
Experimental Section
General Methods and Materials. All reactions were performed in
an atmosphere of argon using standard Schlenk tube or drybox
techniques. Reagent grade solvents were dried, distilled, and stored
over activated (250 °C) 4 Å molecular sieves in a Schlenk flask under
argon. Phenylhydroxylamine23 and R-methylstyrene-d324 were prepared
following literature methods. FT-IR spectra were obtained as KBr
1
(nm, CH2Cl2): 421 (ꢀ 22426 L‚mol-1‚cm-1). H NMR (CD2Cl2) 8.6
(bs, 1H), 7.2 (bs, 1H), 6.8 (bs, 2H), 3.55 (q, 4H), 1.28 (t, 6H).
X-ray Structure Determination of 2. The data were collected at
120(2) K on a Bruker Apex (CCD) diffractometer using Mo KR (λ )
0.71073 Å) radiation. Intensity data, which approximately covered the
full sphere of the reciprocal space, were measured as a series of ω
oscillation frames each 0.3 for 25 s/frame. The detector was operated
in 512 × 512 mode and was positioned 6.12 cm from the crystal.
Coverage of unique data was 97.1% complete to 55.4° (2θ). Cell
parameters were determined from a nonlinear least-squares fit of 3322
reflections in the range of 2.6 < θ < 26.7°. A total of 31514 reflections
were measured. The structure was solved by the direct method using
SHELXTL system 22, and refined by full-matrix least-squares on F2
using all reflections. All the non-hydrogen atoms were refined aniso-
tropically. All the hydrogen atoms were included with idealized
parameters. Crystals were twined and required SHELXTL twin refine-
ment with HKLF 5 type data containing 2 components of 70% and
30% ratio. Crystal data and structure refinement information for 2 are
summarized in Table 2. Tables of atomic coordinates and complete
X-ray data tables were provided as Supporting Information for the
preliminary report (ref 20).
Reaction of 1 with r-Methylstyrene (AMS). AMS (0.10 mL, 0.77
mmol), 6.2 mg (0.048 mmol) of naphthalene (GC internal reference),
and 27 mg (0.051 mmol) of 1 were dissolved in 4 mL of 1,4-dioxane
and heated at 90-100 °C for 20 h. Analysis of the mixture by GC and
GC-MS indicated that 40% (0.010 mmol) of the allyl amine (4) had
formed; the calculated yield is based on a stoichiometry of 2Cu(I) +
PhNO + AMS f allyl amine.
Reaction of 2 with 2-Methyl-2-pentene. A Schlenk flask was
charged with complex 2 (0.010 g, 0.013 mmol) in dry 1,4-dioxane (5
mL), and 2-methyl-2-pentene (0.10 mL, 0.82 mmol) was added. The
reaction mixture was heated for 24 h at 100 °C. GC analysis detected
only free 4-diethylamino-nitrosobenzene, but no product derived from
reaction of 2 and the olefin.
1
pellets. H NMR spectra (FT) were acquired in CDCl3 and CD2Cl2
solutions at 270, 300, or 400 MHz; mass spectra were obtained by
direct insertion (70 eV) or FAB; UV-vis spectra were recorded on a
diode array spectrophotometer. Analytical GC was carried out using a
3-foot column packed with 3% OV 101 or a 30-meter SE-30 capillary
column with naphthalene as internal standard and flame ionization
detection. Yields of volatile products were obtained by GC analysis
with sensitivity factors determined from authentic samples relative to
napthalene. GC-MS (70 eV) analyses were obtained using packed
SPB-5 (30 m) and SE-30 (30 m) capillary columns. Copper was
analyzed by ICP (Perkin-Elmer emission spectrometer).
Typical Procedure for Catalytic Amination of Olefins Catalyzed
by [Cu(CH3CN)4]PF6 or 1. A 1,4-dioxane (2 mL) solution of olefin
(7.0 mmol) and [Cu(CH3CN)4]PF6 or 1 (0.08 g, 0.15 mmol) was heated
to 80-100 °C. A phenyl hydroxylamine (0.164 g, 1.5 mmol) solution
in 1,4-dioxane (8 mL) was delivered with a syringe pump over a period
of 5-6 h to the heated solution. After cooling, the volatiles were
removed in vacuum, and the residue was dissolved in diethyl ether.
Column chromatography was carried out on silica gel using dichlo-
romethane/petroleum ether as eluant to afford the allylamine (Table
1). The IR, MS, and NMR spectra of the allyl amines were compared
with the pure samples prepared as described previously for qualitative
analysis.14,15
Synthesis of [Cu(PhNO)3]PF6 (1). To a solution of [Cu(CH3CN)4]-
PF6 (0.22 g, 0.58 mmol) in chloroform (15 mL), solid nitrosobenzene
(0.26 g, 2.5 mmol) was added, producing a red color immediately. After
being stirred at room temperature for 3-4 days, the dark-red solution
was filtered, and the solvent was removed under vacuum. The resulting
(20) Srivastava, R. S.; Khan, M. A.; Nicholas, K. M. J. Am. Chem. Soc. 2005,
127, 7278.
(21) Yamamoto, Y.; Yamamoto, H. J. Am. Chem. Soc. 2004, 126, 4128.
(22) Momiyama, N.; Yamamoto, H. Angew. Chem. Int. Ed. 2002, 41, 2986.
Momiyama, N.; Yamamoto, H. Org. Lett. 2002, 4, 3579. Momiyama, N.;
Yamamoto, H. J. Am. Chem. Soc. 2003, 125, 6038.
Reaction of 1 with a Mixture of 2-Methyl-2-pentene and 2,3-
Dimethyl-1,3-butadiene. A dioxane solution (5 mL) containing
2-methyl-2-pentene (10 µL, 0.081 mmol), 2,3-dimethyl-1,3-butadiene
(8.5 µL, 0.0.081 mmol), and 1 (0.032 g, 0.06 mmol) was heated at
80-90 °C for 24 h. GC and GC-MS analysis of an aliquot was found
(23) Kamm, O. Org. Synth. 1925, 4, 57
(24) Volger, U. Rec. TraV. Chim. Pays-Bas 1967, 86, 677.
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15251

A R T I C L E S
Srivastava et al.
to contain allyl amines from the alkene (4, m/e 175; 0.005 mmol, 8%)
and diene (5, m/e 173; 0.011 mmol, 18%), respectively. None of the
hetero-Diels-Alder adduct 6 was detected. (Scheme 2)
Stability of Diels-Alder Adduct 6. A mixture of 0.10 g (0.54
mmol) of 6, 2-methyl-2-pentene, (70 µL, 0.57 mmol), and 0.20 g (0.55
mmol) of [Cu(CH3CN)4]PF6 in 5 mL of dioxane was heated to 100 °C
for 5 h. GC analysis detected only the starting adduct 6; no allyl amine
derived from 2-methyl-2-pentene was detected.
10-5 M in CH2Cl2, nm): 230 (ꢀ 27182 L‚mol-1‚cm-1), 282 (ꢀ 10831
L‚mol-1‚cm-1), 310 (ꢀ 6281 L‚mol-1‚cm-1) and 350 nm (ꢀ 1735
L‚mol-1‚cm-1). Separate solutions of 1 and AMS were prepared in
dry CH2Cl2; these solutions were mixed in varied proportion to
systematically change the concentration of AMS while keeping the
concentration of 1 constant (2.32 × 10-6 M). The UV-vis spectra of
these mixtures (ratios of 1 and AMS: 1:0, 1:1, 1:2, 1:5, 1:10, 1:20,
1:40) were measured at room temp and were largely unchanged from
that of 1.
On thermolysis in refluxing dioxane, the corresponding N-phenyl
allylamine, 5% and azoxybenzene, 9% were detected by GC.
Kinetic Studies. Kinetic data were obtained from reactions con-
ducted in a three-necked round-bottom flask fitted with a condenser,
argon inlet, and a silicone septum stopper. Complex 1 (0.010 g, 0.019
mmol), AMS (25 µL, 0.23 mmol), and naphthalene (0.0050 g, 0.039
mmol) were dissolved in dioxane (5 mL). The flask was placed in a
thermostated oil bath at 80 °C. Aliquots were withdrawn periodically
and cooled in ice immediately. GC was used to analyze the samples
for the allyl amine. Initial rates were determined from a plot of
concentration of allyl amine vs time during the first 10-15% conver-
sion. The concentration of each reacting component was systematically
varied, while holding the others constant. The rate constants for the
reaction of 1 with AMS at 70, 80, 100, and 120 °C were determined
and a plot of ln k vs 1/T according to the Erying equation was used to
calculate the activation parameters.
Catalytic Amination of AMS by Phenyl Hydroxylamine and 1.
A mixture of 1 (0.058 g, 0.11 mmol), naphthalene (0.030 g, 0.23 mmol)
as GC internal standard, and AMS (0.6 mL, 4.6 mmol) was heated in
1,4-dioxane (2 mL) at 90-95 °C. Phenyl hydroxylamine (1.37 mmol)
dissolved in 8 mL of dioxane was added over a period of 5-6 h with
a syringe pump. Aniline (0.07 mmol, 5%), azobenzene (0.04 mmol,
6%), azoxybenzene (0.15 mmol, 22%), and the N-phenyl allylamine 3
(0.49 mmol, 36%) were detected by GC and GC-MS after 24 h. A
second run with a lower loading of 1 was also carried out. To a dioxane
solution (2 mL) of 1 (0.023 g, 0.043 mmol), naphthalene (0.019 g,
0.15 mmol), and AMS (0.4 mL, 3.1 mmol) at 100 °C, phenyl
hydroxylamine (0.82 g, 0.75 mmol in 8 mL of dioxane) was slowly
added over a period of 5 h with a syringe pump. The amine 3 (0.26
mmol, 34%) was detected by GC and GC-MS after 24 h.
Catalytic Amination of AMS by Phenyl Hydroxylamine and 2.
To a 1,4-dioxane solution (2 mL) of 2 (0.025 g, 0.03 mmol),
naphthalene (0.019 g, 0.15 mmol) as GC internal standard, and AMS
(0.40 mL, 3.1 mmol) at 100 °C was added phenyl hydroxylamine (0.30
mmol in 5 mL of dioxane) over a period of 4-5 h with a syringe pump.
The N-phenyl allylamine 3 (0.042 mmol, 15%) and N,N-diethyl-4-
nitrosoaniline were detected by GC and GC-MS after 24 h.
Detection of 1 in the Reaction of [Cu(CH3CN)4]PF6, AMS and
Phenyl Hydroxylamine. PhNHOH (1.6 mmol) in dioxane (4 mL) was
added slowly over a period of 2.5 h to a solution of [Cu(CH3CN)4]PF6
(0.15 g, 0.40 mmol) and AMS (0.53 mL, 3.8 mmol) in dioxane (10
mL) at 90 °C and allowed to stir for an additional 0.5 h at the same
temperature. The solution was cooled, and the solvent was removed
under reduced pressure on a warm water bath. The resulting dark-red
residual mass (0.10 g, 0.19 mmol, 47%) was recrystallized from CH2-
Cl2/diethyl ether. The IR spectrum (KBr) of the product was indistin-
guishable from that of 1.
Styrene Adduct 7. A solution of 1 (0.126 g, 0.24 mmol) in dry
dichloromethane (20 mL) was treated with styrene (0.03 mL, 0.26
mmol), and the mixture was stirred at room temperature for 25 h. The
solvent was removed at reduced pressure, and the residue was triturated
with hexane (3 × 5 mL). GC analysis of the extract showed the presence
of PhNO. The resulting dark-red solid was dried under vacuum for 2
days. The visible spectrum of the product (CH2Cl2) exhibited peaks at
348 (ꢀ 7174 L‚mol-1‚cm-1) and 455 nm (ꢀ 2210 L‚mol-1‚cm-1). The
1H NMR spectrum of 7 (CH2Cl2) had peaks at 7.20-8.30 (m, 20H),
6.45 (dd, J ) 15, 10 Hz, 0.8H), 5.35 (d, J ) 15 Hz, 0.8H, overlapped
w/CHDCl2), 4.85 (d, J ) 10 Hz, 0.8H). Solid 7 was thermolyzed in an
evacuated, sealed flask for 1 h at >250 °C. After cooling, the residue
was dissolved in CH2Cl2; GC-MS analysis of the solution showed
the presence of styrene, azobenzene, and azoxybenzene.
The determination of rates of reaction of 1 with para-substituted
R-methylstyrenes was carried out analogously at 80 °C.
Kinetic D-Isotope Effect. Complex 1 (0.19 g, 0.36 mmol) was
dissolved in 10 mL of 1,4-dioxane. To this solution R-methylstyrene
(150 µL, 0.115 mmol) and R-(trideuteriomethyl) styrene (150 µL, 0.115
mmol, 90% D3) were added. The solution was heated to 80 °C for 20
h. After vacuum removal of the volatiles, the allyl amine was isolated
by chromatography on silica gel using petroleum ether and ethyl acetate
(95:5) as eluant. The ratio of protio to deuterio products was determined
by 1H NMR integration of the allylic CH2 and dCH2 groups. Correcting
for the % D in the R-(tridueteriomethyl)styrene a 2.2 ratio of protio/
deuterio allylamines ()kH/kD) was calculated.
Computational Studies. PM3 and DFT computations were carried
out using the Spartan 04 software suite (Wavefunction, inc.). The
energetically minimized structures (shown) were determined in the
semiempirical PM3 mode, followed by single energy point calculations
with the DFT method (B3LYP with 6-31G basis set) to provide final
calculated energies for comparison. A summary of the Cartesian
coordinates, calculated energies, and electrostatic charges for structures
A-E is provided in the Supporting information.
Results and Discussion
Synthetic and General Catalytic Reactivity Studies. Slow
addition of phenyl hydroxylamine to a heated solution of the
alkene (in excess) and [Cu(CH3CN)4]PF6 (10 mol %) in 1,4-
dioxane provides the corresponding N-aryl-N-allylamines in fair
to good yield after 8 h (Table 1, Scheme 1). Aniline, azo-, and
r-Methylstyrene Adduct 8. A solution of 1 (0.12 g, 0.23 mmol)
and R-methystyrene (0.040 mL, 0.37 mmol) in dry dichloromethane
(15 mL) was stirred for 23 h at room temperature; the solvent was
then removed under vacuum to leave a red-brown residue. Alternatively,
to a solution of [Cu(CH3CN)4]PF6 (0.11 g, 0.28 mmol) in CH2Cl2 (20
mL), solid PhNO (0.13 g, 1.2 mmol) and AMS (0.50 mL, 3.8 mmol)
were added. The red solution was stirred at room temp for 24 h; the
solvent and excess AMS were removed at reduced pressure. The residue
from either reaction was washed with hexane (3 × 5 mL) and vacuum-
dried. A red solid (0.136 g, 74%) was obtained. 1H NMR (CDCl3, δ):
6.7-8.29 (br m, 20H), 4.99 (broad s, 1H), 5.25(broad s, 1H), 2.14 (s,
3.4 H); the shift of the -CH2 and -CH3 resonances from free AMS is
0.15-0.25 ppm. IR (KBr, cm-1): 1626 (w, CdC). UV-vis (2.9 ×
Scheme 1
azoxybenzenes are the major byproducts, derived from compet-
ing phenyl hydroxylamine side reactions. The amination reac-
tions, like the aryl hydroxylamine reactions catalyzed by Mo-
13 and Fe- salts and complexes,14,15 and those involving metal-
catalyzed nitroarene reduction by CO,16 proceed highly
regioselectively, with nitrogenation occurring at the less sub-
stituted carbon of the original unsaturated unit. Other important
9
15252 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Copper(I)-Catalyzed Allylic Amination
A R T I C L E S
features of the reactions include (1) trisubstituted and 1,1-
disubstituted alkenes give the best yield and (2) no other alkene-
derived products are observed. The distinctive regioselectivity
and alkene reactivity behavior are typical of ene-type reactions.25
This selectivity complements and is superior to that observed
in stoichiometric aminations by (RN)2X (X ) S, Se),6,7 which
favor preservation of the double bond position but with variable
selectivity. Furthermore, the catalysts used in the present system
are commercially available and inexpensive.
Table 1. Cu(I)-Catalyzed Allylic Amination of Alkenes
Figure 1. X-ray structure of the cation of 2.
exclusively, with none of the PhNO-derived Diels-Alder adduct
6 being detected (Scheme 2). A control experiment established
that 6 is stable under the amination conditions. These results
rule out the intermediacy of free nitrosobenzene in the Cu(I)-
catalyzed reactions and is consistent with (but does not prove)
the involvement of a Cu(I)-coordinated species in the transfer
of an RN- or RNO-group to the alkene.
Preparation of Catalytically Relevant Copper Complexes.
Seeking to prepare potential intermediate complexes in the Cu-
mediated allylic aminations, [Cu(CH3CN)4]PF6 was treated with
4.2 equiv of PhNO in CHCl3 at room temperature (Scheme 3).
A dark-red (λ 420 nm), labile complex 1 was isolated that was
formulated as [Cu(PhNO)3]PF6 on the basis of spectroscopic
1
evidence and its metal content. The H NMR spectrum of 1
showed a 2:1:2 (d,t,t) set of aromatic resonances downfield from
PhNO itself. Unfortunately, our efforts to determine the detailed
structure of 1 were frustrated by its sensitivity and lability. A
more stable Cu-ArNO derivative was prepared from the
reaction of [Cu(CH3CN)4]PF6 with excess N,N′-diethyl-4-
nitrosoaniline in CH2Cl2 at room temperature. The greenish-
red product 2 (70% yield) was identified with the aid of IR,
NMR, MS (FAB), and X-ray crystallography (Figure 1); bond
lengths and angles for 2 are given in Table 3. Complex 2 consists
+
of distorted trigonal planar, 16 electron Cu(ArNO)3 and
a GC yield. b Isolated yield.
Scheme 2
Trapping Studies: The Non-involvement of PhNO. Two
general catalytic pathways for metal-promoted allylic amination
have been postulated. A coordinated organonitrogen species as
the ArN-transfer agent has been implicated in the FeCl2/3-
catalyzed reactions,14 whereas those catalyzed by LMoO213 and
(phthal)Fe15 involve the intermediacy of free PhNO, a proven
enophile,26 as the aminating agent. We have investigated whether
ArNO is the active aminating agent in Cu(I)-catalyzed amination
reactions. The hetero-Diels-Alder reaction of PhNO with 2,3-
dimethyl-1,3-butadiene (DMB) was used as a trapping reaction
for PhNO.27 Initially, the effective Diels-Alder trapping of
PhNO by DMB at 80 °C (dioxane) in the presence of
R-methylstyrene was established. When the Cu(I)-catalyzed
reaction (5 mol % [Cu(CH3CN)4]PF6) of PhNHOH with
2-methyl-2-pentene was carried out in the presence of DMB
(1:1, 80 °C, dioxane), the allyl amines derived from the alkene
4 (m/e 175; 8%) and the diene 5 (m/e 173; 18%) were formed
Scheme 3
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15253

A R T I C L E S
Srivastava et al.
Table 2. Crystal Data and Structure Refinement for 2
empirical formula
molecular weight
temperature
C31H44Cl2Cu F6N6O3P
828.13
120(2) K
wavelength
0.71073 Å
crystal system
space group
unit cell dimensions
triclinic
P1h
a ) 8.396(7) Å, R ) 68.477(7)°
b )14.281(11)Å, â ) 84.072(7)°
c ) 16.138(13)Å, γ ) 86.267(7)°
1790(2)Å3, 2
vol, Z
density (calculated)
absorption coefficient
F(000)
crystal size
θ range for data collection
index ranges
1.537 Mg/m3
0.878 mm-1
856
0.28 × 0.12 × 0.10 mm3
2.39 to 27.69°
-10 < h < 10,
-17 < k < 18,
-20 < 1 < 20
31514
Figure 2. Effect of [1] on kobs for the reaction of 1 with R-methylstyrene
at 80 °C.
reflection collected
independent reflections
completeness of θ ) 25.00°
absorption correction
max. and min. transmission
refinement method
31518 [R(int) ) 0.0000]
Scheme 4
99.4%
semiempirical from equivalents
0.9174 and 0.7912
full-matrix least-squares on F2
31518/0/460
data/ restraints/parameters
GOF on f2
1.027
Final R indices [I > 2σ(I)]
R indices (all data)
Largest diff. Peak and hole
R1 ) 0.0531, wR2 ) 0.1369
R1 ) 0.0701, wR2 ) 0.1445
0.906 and -0.978 e. Å-3
forming catalytic pathway or is it a dead end? Initially, treatment
of 1 with R-methylstyrene (1:15) in dioxane (100 °C) resulted
in its conversion to the corresponding allyl amine 3 (40%,
Scheme 4). Likewise, as found in the Cu(I)-catalyzed amination
reaction, 1 reacts with alkenes without the intervention of free
PhNO. Thus, when 1 was heated with an equimolar mixture of
R-methylstyrene and 2,3-dimethyl-1,3-butadiene (DMB), the
allylic amination products 4, 5 were produced exclusively (1:
13). None of the Diels-Alder adduct 627 was detected (Scheme
2). Finally, 1 also was found to catalyze the allylic amination
of R-methylstyrene by PhNHOH (100 °C, 24 h) at an initial
rate comparable to (or somewhat faster than) that of the Cu-
catalyzed reaction (Scheme 4). These observations, coupled with
the isolation of 1 in Cu-catalyzed reactions, strongly indicate
that complex 1 is on the amination catalytic pathway.
octahedral PF6-. The copper atom is coordinated to the
nitrosoarene ligands through the N-atom with the NO-units
directed out of the CuN3-plane. The Cu-N bond lengths (av.
1.933 Å) are markedly varied, ranging from 1.898 to 1.970 Å,
as are the N-Cu-N bond angles, which range from 105 to
134°. These distortions from ideal trigonal planarity are among
the most severe found in d10-ML3 complexes28 and are probably
derived from steric and/or crystal packing effects.29 The N-O
bond lengths, however, are relatively uniform (av. 1.258 Å) and
are comparable to those in free N,N′-diethyl-4-nitrosoaniline
(1.252 Å),30 suggesting the absence of appreciable backbonding
from Cu(I) to the nitrosoarene ligand. Compound 2 constitutes
the first crystallographically characterized, homoleptic ni-
trosoarene-metal complex and the first copper complex bearing
a simple C-nitroso ligand.31
Kinetics of the Reactions of [(PhNO)3Cu]PF6 with Alk-
enes. To gain further insight into the mechanism of the
amination reaction we performed kinetic studies on the reaction
of R-methylstyrene (excess) with 1. Kinetic information gained
from the direct reaction of 1 was deemed likely to be more
informative than from the catalytic reaction because (1) the
former system would probe steps closer to the critical C-N,
C-C, and C-H transforming steps (rather than the formation
of 1 from ArNHOH) and (2) acquisition of meaningful kinetic
data for the catalytic reaction would be complicated by the
inefficiency of the batch process (all reactants together) and the
need for slow addition of the hydroxylamine. The reactions were
conducted in dioxane at 80 °C and initial rates were determined
by GC monitoring the allylamine formation at low conversion
(0-15%). On varying the concentration of 1 (9.44 × 10-6 to
3.77 × 10-5 M) at constant R-methylstyrene concentration, the
appearance of allylamine is first order in the copper complex
(Figure 2). Likewise, at a constant concentration of 1 (1.88 ×
10-5 M) the rate of reaction is also first order in alkene over
the concentration range 9.6 × 10-5 to 3.84 × 10-4 M (Figure
3). Overall, these results reveal the involvement of the alkene
Reactivity Studies of 1: Role in the Catalytic Amination
of Alkenes. Once the three-coordinate copper complex 1 was
identified, we sought to determine its relevance to the copper-
catalyzed allylic amination reactions; that is, is 1 on the product
(25) Oppolzer, W.; Snieckus, V. Angew. Chem., Int. Ed. Engl. 1978, 17, 476.
Hoffman, H. M. R. Angew. Chem., Int. Ed. Engl. 1969, 8, 556. Snider, B.
Acc. Chem. Res. 1980, 13, 426.
(26) Adam, W.; Bottke, N.; Krebs, O. Org. Lett. 2000, 2, 3293. Ensley, H. E.;
Mahadevan, S. Tetrahedron Lett. 1998, 3255. Dang, H.-S.; Keck, G. E.;
Webb, R. R.; Yates, J. B. Tetrahedron 1981, 37, 4007. Banks, R. E.;
Hazeldine, R. N.; Miller, P. J. Tetrahedron Lett. 1970, 4417. Motherwell,
W.; Roberts, J. S. Chem. Commun. 1972, 329.
(27) Taylor, E. C.; Tseng, C.-P.; Rampal, J. B. J. Org. Chem. 1982, 47, 552.
Oikawa, E.; Tsubaki, S. Bull. Chem. Soc. Jpn. 1973, 46, 1819.
(28) Most Cu(I)L3 complexes display nearly ideal trigonal planar geometries:
Kappenstein, C.; Hugel, R. P. Inorg. Chem. 1978, 17, 1945. Eller, P. G.;
Corfield, P. W. R. Chem. Commun. 1971, 105. Gruf, E. S.; Koch, S. A. J.
Am. Chem. Soc. 1990, 112, 1245. A few examples with significant
distortions towards Y- or T-shapes have been reported: Coucouvanis, D.;
Murphy, C. N.; Kanodia, S. K. Inorg. Chem. 1980, 19, 2993. Gruf, E. S.;
Koch, S. A. J. Am. Chem. Soc. 1989, 111, 8762. Lewin, A. H.; Michl, R.
J. Chem. Commun. 1972, 661.
(29) Riehl, J. F.; El-Idrissi, R.; Jean, Y.; Pelissier, M. New J. Chem. 1991, 15,
239.
(30) Talberg, H. J. Acta Chem. Scand. A 1977, 31, 743.
(31) Salomon, R. G.; Kochi, J. K. J. Am. Chem. Soc. 1973, 95, 1889. Doyle,
G.; Eriksen, K. A.; Engen Van, D. Organometallics, 1985, 4, 830.
9
15254 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Copper(I)-Catalyzed Allylic Amination
A R T I C L E S
Figure 5. Hammett plot for the reaction of 1 with para substituted
R-methylstyrenes.
Figure 3. Effect of [AMS] on kobs for the reaction of 1 with R-methylstyrene
at 80 °C.
Figure 6. Calculated structure of [Cu(HNO)3]+ (A).
Figure 4. Temperature dependence of kobs for the reaction of 1 with
R-methylstyrene.
Scheme 5
before or in the rate-determining step and suggest that alkene
coordination or at least association is involved.
Temperature-dependent rate studies of the reaction of 1 with
R-methylstyrene between 70 and 120 °C afforded rate constants
which were subjected to Eyring analysis (Figure 4), yielding
the activation parameters ∆H(act) ) 9.9 kcal/mol and ∆S(act)
)
Scheme 6
-44 eu. These results, especially the large negative entropy of
activation, point to an associative, highly ordered rate-determin-
ing step.
The electronic character of the transition state was also
probed. The rate constants were determined for the reaction of
1 with a small set of para substituted R-methylstyrenes at
80 °C (Figure 5). A small Hammett F value of -0.035 indicates
a negligible effect of either electron withdrawing or electron
donating groups on the allylic amination of 1 with R-methyl-
styrene and suggests that there is little charge development in
the transition state for the reaction.
Since the amination reaction requires transfer of an allylic
hydrogen from the alkene substrate, it was of interest to
determine if allylic C-H bond-breaking is involved in the rate-
limiting step. This issue was addressed by determination of the
kinetic isotope effect (kH/kD) from a competitive reaction of
AMS and R-CD3-AMS with 1 (80 °C, dioxane, Scheme 5).
NMR analysis of the isolated allylic amine revealed a 2.2 ((0.1):
Olefin Adducts 7, 8. The above kinetics and trapping results
suggest that the alkene substrate may be coordinated during the
ArN-group transfer step. To test this hypothesis the room-
temperature reactions of 1 with excess styrene (which lacks
allylic hydrogens) and AMS were examined (Scheme 6) to
determine if adducts could be detected or isolated. After solvent
evaporation and trituration to remove residual alkene, spectro-
scopic analysis indicated the formation of new compounds 7, 8
whose 1H NMR spectra exhibited resonances in both the
aromatic and vinylic regions, shifted ca. 0.2 ppm from those of
the free alkene, but with the same patterns. Such modest
1.0 ratio of protio- to deuterio-allyl amines, hence a kH/kD
)
2.2 (ignoring secondary effects). This moderate sized primary
isotope effect indicates a significant degree of C-H bond-
breaking in the transition state for the RLS.
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15255

A R T I C L E S
Srivastava et al.
Table 3. Bond Lengths [Å] and Angles [deg] for 2
Cu(1)-N(4)
Cu(1)-N(6)
Cu(1)-N(2)
O(1)-N(2)
O(2)-N(4)
O(3)-N(6)
N(1)-C(5)
N(1)-C(2)
N(1)-C(4)
N(2)-C(8)
N(3)-C(15)
N(3)-C(12)
C(5)-C(6)
C(5)-C(10)
C(6)-C(7)
C(7)-C(8)
C(8)-C(9)
C(1)-C(2)
C(3)-C(4)
N(6)-C(28)
N(5)-C(24)
C(9)-C(10)
C(11)-C(12)
C(13)-C(14)
1.8983(15)
1.9296(19)
1.9701(18)
1.2646(16)
1.2517(16)
1.2582(16)
1.3297(19)
1.471(2)
1.4730(18)
1.358(2)
1.3455(18)
1.455(2)
1.4236(19)
1.434(2)
1.351(2)
1.404(2)
1.411(2)
C(15)-C(20)
C(15)-C(16)
C(16)-C(17)
C(17)-C(18)
C(18)-C(19)
C(19)-C(20)
C(21)-C(22)
C(23)-C(24)
C(25)-C(30)
C(25)-C(26)
C(26)-C(27)
C(27)-C(28)
C(28)-C(29)
C(29)-C(30)
Cl(2)-C(31)
1.422(2)
1.424(2)
1.349(2)
1.4081(19)
1.388(2)
1.359(2)
1.511(2)
1.509(2)
1.422(2)
1.430(2)
1.342(2)
1.414(2)
1.400(2)
1.359(2)
1.756(2)
1.734(2)
1.4703(19)
1.3784(18)
1.3347(18)
1.465(2)
133.81(6)
120.96(6)
104.96(4)
122.30(11)
121.13(12)
116.39(11)
O(1)-N(2)-C(8)
O(1)-N(2)-Cu(1)
C(8)-N(2)-Cu(1)
C(15)-N(3)-C(12)
C(15)-N(3)-C(14)
C(12)-N(3)-C(14)
O(2)-N(4)-C(18)
O(2)-N(4)-Cu(1)
C(18)-N(4)-Cu(1)
C(25)-N(5)-C(22)
C(25)-N(5)-C(24)
C(22)-N(5)-C(24)
O(3)-N(6)-C(28)
O(3)-N(6)-Cu(1)
C(28)-N(6)-Cu(1)
N(1)-C(2)-C(1)
N(1)-C(4)-C(3)
N(1)-C(5)-C(6)
N(1)-C(5)-C(10)
C(6)-C(5)-C(10)
C(7)-C(6)-C(5)
C(6)-C(7)-C(8)
N(2)-C(8)-C(7)
N(2)-C(8)-C(9)
C(7)-C(8)-C(9)
C(10)-C(9)-C(8)
118.11(11)
113.50(9)
C(9)-C(10)-C(5)
N(3)-C(12)-C(11)
N(3)-C(14)-C(13)
N(3)-C(15)-C(20)
N(3)-C(15)-C(16)
C(20)-C(15)-C(16)
C(17)-C(16)-C(15)
C(16)-C(17)-C(18)
N(4)-C(18)-C(19)
N(4)-C(18)-C(17)
C(19)-C(18)-C(17)
C(20)-C(19)-C(18)
C(19)-C(20)-C(15)
N(5)-C(22)-C(21)
N(5)-C(24)-C(23)
N(5)-C(25)-C(30)
N(5)-C(25)-C(26)
C(30)-C(25)-C(26)
C(27)-C(26)-C(25)
C(26)-C(27)-C(28)
N(6)-C(28)-C(29)
N(6)-C(28)-C(27)
C(29)-C(28)-C(27)
C(30)-C(29)-C(28)
C(29)-C(30)-C(25)
Cl(1)-C(31)-Cl(2)
121.27(12)
111.60(13)
112.29(12)
121.00(13)
121.40(12)
117.61(12)
121.33(13)
120.10(13)
118.15(12)
122.57(13)
119.25(12)
121.48(13)
119.98(13)
112.56(13)
112.74(12)
121.94(12)
120.36(13)
117.70(12)
121.38(13)
120.54(13)
118.66(12)
122.59(12)
118.75(12)
121.56(13)
120.06(13)
112.11(10)
127.30(10)
122.28(12)
121.68(12)
116.02(11)
117.80(12)
117.11(9)
124.31(10)
123.07(12)
122.04(12)
114.88(11)
117.92(12)
116.40(9)
125.43(9)
Cl(1)-C(31)
112.82(11)
113.56(11)
122.26(13)
120.61(12)
117.13(12)
120.65(13)
121.65(12)
118.40(12)
123.31(13)
118.29(13)
120.96(13)
N(3)-C(14)
1.513(2)
1.520(2)
1.3671(18)
1.468(2)
1.343(2)
N(4)-C(18)
N(5)-C(25)
N(5)-C(22)
N(4)-Cu(1)-N(6)
N(4)-Cu(1)-N(2)
N(6)-Cu(1)-N(2)
C(5)-N(1)-C(2)
C(5)-N(1)-C(4)
C(2)-N(1)-C(4)
1.509(2)
1.518(2)
32
Scheme 7
the fact that Cu(I) is effective for reducing PhNHOH to PhNH2
and suggests that Cu(I) complexes present in the catalytic
reactions would be capable of reducing an intermediate allyl
hydroxylamine.
Computational Studies. Semiempirical (PM333) and density
functional calculations were conducted on several potential
reaction intermediates to evaluate their viability, structures, and
reactivity. To economize computational time the RNO ligand
and alkene were modeled by HNO and 1-butene, respectively.
coordination shifts are common for Cu(I)-olefin complexes31
in which there is rapid exchange on the NMR time scale between
free and coordinated olefin. The electronic spectra of these
adducts 7, 8 (5.7 × 10-4 to 2.94 × 10-5 M in CH2Cl2) exhibit
peaks at 348 and 455 nm for 7 (and 230, 282, and 310 nm for
8), which are quite different from 1. Although the adducts 7, 8
were isolable from very concentrated solutions, in dilute
solutions (ca. 10-6 M) their electronic spectra were similar to
that of [Cu(PhNO)3]+ (1) and the spectrum of 1 in dilute solution
was little changed by the addition of excess (2-20 equiv) of
these alkenes. These observations indicate that alkene coordina-
tion is reversible and at lower concentrations the equilibrium is
shifted toward 1 and the free alkene.
+
The structure of Cu(HNO)3 (A) at the PM3 level minimized
to trigonal planar as defined by the copper and the nitrogen
atoms of the HNO ligands (Figure 6). The oxygen atoms are
out of the CuN3-plane as observed in solid state (X-ray), but
the distorted Y-geometry seen in the X-ray structure was not
reproduced. As noted earlier (op cit) this distortion is most likely
not electronic in origin.
The association of 1-butene with A was probed by computa-
tions on [(Cu(HNO)3(1-butene)]+ (B, Figure 7). Complex B is
an 18 e- species and minimizes in an expected pyramidal
geometry. The DFT-calculated enthalpy of B (-2188.76292
hartrees; 1 H ) 627.5 kcal) is lower (-8.0 kcal/mol) than the
sum of those of A (-2031.53252 H) and free 1-butene
(-157.21775 H). While this comparison ignores the entropy
penalty accompanying alkene association with 1, it does suggest
that formation of a [Cu(RNO)3(alkene)]+ species is an energeti-
cally viable step. This is also in line with the isolability of
[(styrene)Cu(PhNO)3]PF6, the first-order kinetic dependence on
the alkene concentration, and the large ∆Sact value that would
accompany olefin association in or before the rate-limiting step.
For energy comparisons with other prospective intermediate
complexes the relative energy of B was set to zero kcal/mol.
We envision two alternative pathways for the allylic-H
The thermolysis of 7 produced styrene, azobenzene, and
azoxybenzene while heating 8 gave the N-phenyl allyl amine
and the azo and azoxy compounds in moderate yield. Overall,
the spectroscopic features of 7 and 8 and the thermolysis results
indicate the formation of 1:1 adducts that we formulate as
[(PhNO)3Cu(η2-alkene)]PF6. The isolation of these adducts lends
additional support to the hypothesis that a complex having the
alkene and ArNO simultaneously coordinated is an intermediate
in the amination process.
RN or RNO Group Transfer? An important mechanistic
issue is whether the product allyl amine is formed via an
arylimido (RN)-group insertion into the alkene directly from 1
or whether an arylnitroso (RNO) unit is initially transferred,
giving an intermediate allyl hydroxylamine that is subsequently
reduced to the amine (Scheme 7). Evidence supporting the
viability of the RNO-transfer route was provided indirectly by
(32) Wathen, S. P.; Czarnik, A. W. J. Org. Chem. 1992, 57, 6129. Matsuguma,
H. J.; Audrieth, L. F. J. Inorg. Nucl. Chem. 1959, 12, 186. Srivastava, R.
S. Tetrahedron Lett. 2003, 44, 3271.
(33) Cundari, T. R.; Deng, J. J. Chem. Inform. Comp. Sci. 1999, 39, 376. Cundari,
T. R.; Deng, J.; Fu, W. Int. J. Quantum Chem. 2000, 77, 421. Ball, D. M.;
Buda, C.; Gillespie, A. M.; White, D. P.; Cundari, T. R. Inorg. Chem.
2002, 41, 152.
9
15256 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Copper(I)-Catalyzed Allylic Amination
A R T I C L E S
Figure 7. Computed structures and energies of potential reaction intermediates in Cu-catalyzed allylic amination.
transfer to coordinated RNO from intermediate [Cu(RNO)3-
(alkene)]+ (B). In the first of these H-transfer is accompanied
by an η2 to η3 shift in Cu bonding, giving [(η3-1-methallyl)-
Cu(HNO)2(NHOH)]+ (C). Such an 18 e- species is calculated
to have a pyramidal geometry with a ∆Hform substantially higher
than the η2-alkene adduct (+30 kcal/mol). Alternatively, allylic
H-transfer in B could occur together with an η2- to η1-Cu shift
to produce [(η1-methallyl)Cu(HNO)2(NHOH)] (D). Complex D
and its isomer D′ could also be generated from the η3-allyl
species C since facile η1/η3 equilibria are likely for Cu-allyl
complexes.34 The η1-allyl derivative D is calculated to be
substantially lower in energy (∆∆H 14.5 kcal) than the η3-
isomer C and is predicted to have nearly a square planar
geometry, as expected for a d8-four coordinate species. The
isomeric, branched η1-allyl species, D′, is calculated to be
substantially less stable than the terminal isomer (∆∆H 9.6 kcal),
probably the result of a weaker Cu-(2°)-C bond, greater steric
crowding, and the higher energy, terminal CdC. The calculated
greater stability of the η1-allyl complexes D, D′ relative to C
contrasts with recent DFT calculations on the reductive elimina-
tion of (allyl)CuMe2 complexes which found the η3-allyl
derivatives to be lower in energy.34d The origin of this difference
could lie in the differing charges, total electron count, and
auxiliary ligands in the two systems. In any event the H-transfer
conversion from alkene-complex B to either of these high energy
intermediates C, D, or D′ is likely the rate-limiting step in the
allylic amination. This would account for the experimentally
observed D-isotope effect, the olefin rate dependency, and the
negative ∆Sact. With the crude assumption that the activation
energies to form C, D, or D′ from η2-alkene complex B parallel
the energies of the intermediates, the η1-intermediate D appears
to be the most energetically accessible.
C-N bond formation to form an allyl hydroxylamine
derivative, for example, η2-complexes E, E′, could proceed from
any of the allyl-Cu complexes C, D by reductive elimination.
This step would determine the product regiochemistry. Its highly
exothermic nature (30-45 kcal calculated) would likely make
it fast and essentially irreversible. Little energy difference is
calculated between the isomeric η2-allyl hydroxylamine com-
plexes E, E′, so their relative stability (thermodynamic control)
would not appear to be the origin of the high reaction
regioselectivity. Reductive elimination from the most stable of
the allyl complexes, D to E, would produce the allyl hydroxy-
lamine (and eventually the allyl amine) derived from N-
functionalization of the original, less substituted olefinic carbon,
as observed experimentally. RE (internal NHOH-transfer) from
the η3-methallyl complex C could also produce E or E′,
depending on to which carbon (C5 vs C6) the -NHOH unit is
delivered. We have not addressed this possibility in depth
computationally but the calculated structure of C does show
significant asymmetry of the Cu-allyl unit, Cu-C1 (2.15 Å)
versus Cu-C3 (2.03 Å), C1-C2 (1.42 Å) versus C2-C3 (1.46
Å), and electrostatic charges, C1 (+0.18) versus C3 (-0.33),
which suggests a predisposition to form the allyl hydroxylamine
isomer that is observed experimentally.
(34) (a) Liepins, V.; Backvall, J.-E. Eur. J. Org. Chem. 2002, 21, 3527. (b)
Kondo, J.; Inoue, A.; Shinokubo, H.; Oshima, K. Tetrahedron Lett. 2002,
43, 2399. (c) Levisalles, J.; Rudler-Chauvin, M.; Rudler, H. J. Organomet.
Chem. 1977, 136, 103. (d) Yamanaka, M.; Kato, S.; Nakamura, E. J. Am.
Chem. Soc. 2004, 126, 6287.
Plausible Catalytic Mechanism. The above observations,
together with the distinctive regioselectivity of the Cu-catalyzed
reactions, lead us to suggest as a plausible reaction pathway
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15257

A R T I C L E S
Scheme 8
Srivastava et al.
the one shown in Scheme 8. Initially, some Cu(II) is likely
generated by Cu(I) reduction of ArNHOH (PhNH2 is detected);
the available Cu(II) then oxidizes PhNHOH to PhNO with
formation of the C-nitroso complex A. Alkene association with
A produces the (η2-alkene)Cu(RNO)3]+ species (B) which can
be converted by allyl-H transfer and Cu-haptotropic shift, to
an allyl-Cu complex, most likely the η1-allyl species D.
Reductive elimination from D could produce an allyl hydroxy-
lamine derivative, either coordinated (as in E) or free. Reduction
of the hydroxylamine by Cu(I) would give the allyl amine and
regenerate Cu(II) for return into the catalytic cycle.
nation has been further elucidated with trapping experiments,
isotope and substituent effects, and computational studies which
point to the involvement of [(allyl)Cu(RNO)2(NHOH)]+ as
important reaction intermediates. Efforts are underway to
expand the synthetic scope and generality of this reaction and
to develop enantioselective variants.
Acknowledgment. We thank Dr. Masood Khan (O.U.
Analytical Services) for determination of the X-ray structure
of 2 and Dr. R. D. Braun (UL-L, Chemistry Department) for
ICP measurements. Financial support provided by the National
Science Foundation (to K.M.N.) and the Research Corporation
(Cottrell College Science Award, CC6234 to R.S.S.) is greatly
appreciated.
Conclusions
The Cu(I)-catalyzed amination of alkenes with aryl hydroxy-
lamines constitutes a useful and regioselective method for the
direct preparation of N-aryl-N-allylamines. We have established
the first structurally verified copper complex of a C-nitroso
compound. The involvement of 1 as the aminating agent in [Cu-
(CH3CN)4]PF6-catalyzed allylic amination also has been im-
plicated on the basis of kinetics and isolation reactions with
alkenes. The mechanism of the copper-catalyzed allylic ami-
Supporting Information Available: The DFT calculational
output (Cartesian coordinates, calculated energies, and electro-
static charges for structures A-E). This material is available
free of charge via the Internet at http://pubs.acs.org.
JA0751072
9
15258 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007