Dearomatization reactions of N-heterocycles mediated by group 3 complexes

Article abstract of DOI:10.1021/ja908489p

Group 3 (Sc, Y, Lu, La) benzyl complexes supported by a ferrocene diamide ligand are reactive toward aromatic N-heterocycles by mediating their coupling and, in a few cases, the cleavage of their C-N bonds. When these complexes reacted with 2,2′-bipyridine or isoquinoline, they facilitated the alkyl migration of the benzyl ligand onto the pyridine ring, a process accompanied by the dearomatization of the N-heterocycle. The products of the alkyl-transfer reactions act as hydrogen donors in the presence of aromatic N-heterocycles, ketones, and azobenzene. Experimental and computational studies suggest that the hydrogen transfer takes place through a concerted mechanism. An interesting disproportionation reaction of the dearomatized, alkyl-substituted isoquinoline complexes is also reported.

Full text of DOI:10.1021/ja908489p

Published on Web 12/03/2009
Dearomatization Reactions of N-Heterocycles Mediated by
Group 3 Complexes
Kevin L. Miller,† Bryan N. Williams,† Diego Benitez,‡ Colin T. Carver,†
Kevin R. Ogilby,† Ekaterina Tkatchouk,‡ William A. Goddard III,‡ and
Paula L. Diaconescu*,†
Department of Chemistry & Biochemistry, UniVersity of California, Los Angeles,
California 90095, and Materials and Process Simulation Center, California Institute of
Technology, Pasadena, California 91125
Received October 13, 2009; E-mail: pld@chem.ucla.edu
Abstract: Group 3 (Sc, Y, Lu, La) benzyl complexes supported by a ferrocene diamide ligand are reactive
toward aromatic N-heterocycles by mediating their coupling and, in a few cases, the cleavage of their
C-N bonds. When these complexes reacted with 2,2′-bipyridine or isoquinoline, they facilitated the alkyl
migration of the benzyl ligand onto the pyridine ring, a process accompanied by the dearomatization of the
N-heterocycle. The products of the alkyl-transfer reactions act as hydrogen donors in the presence of
aromatic N-heterocycles, ketones, and azobenzene. Experimental and computational studies suggest that
the hydrogen transfer takes place through a concerted mechanism. An interesting disproportionation reaction
of the dearomatized, alkyl-substituted isoquinoline complexes is also reported.
through a σ-bond metathesis mechanism,16,20-22 in which the
Introduction
alkyl group on the metal engages the hydrogen atom on
Early transition metal alkyl or hydride complexes often react
with aromatic N-heterocycles to produce ortho-metalated
complexes.1-19 Such ortho-metalation reactions take place
the R-carbon of the heterocyclic ring and is eliminated as the
corresponding hydrocarbon. Less commonly, alkyl ligands of
early transition metal complexes will undergo a 1,3-alkyl
migration, when they are transferred to the R-carbon of the
coordinated N-heterocycle.23-25 Alkyl migration may occur
occasionally to other positions of the heteroaromatic ring.19 In
some instances, the alkyl migration26 is followed by the C-N
bond cleavage of the aromatic heterocycle,27-29 a reaction
relevant to the study of hydrodenitrogenation processes.30-35
† University of California, Los Angeles.
‡ California Institute of Technology.
(1) Diaconescu, P. L. Curr. Org. Chem. 2008, 12, 1388–1405.
(2) Boaretto, R.; Roussel, P.; Kingsley, A. J.; Munslow, I. J.; Sanders,
C. J.; Alcock, N. W.; Scott, P. Chem. Commun. 1999, 1701–1702.
(3) Boaretto, R.; Roussel, P.; Alcock, N. W.; Kingsley, A. J.; Munslow,
I. J.; Sanders, C. J.; Scott, P. J. Organomet. Chem. 1999, 591, 174–
184.
(4) den Haan, K. H.; Wielstra, Y.; Teuben, J. H. Organometallics 1987,
6, 2053–2060.
(19) Sugiyama, H.; Aharonian, G.; Gambarotta, S.; Yap, G. P. A.;
Budzelaar, P. H. M. J. Am. Chem. Soc. 2002, 124, 12268–12274.
(20) Bercaw, J. E. Pure Appl. Chem. 1990, 62, 1151–1154.
(21) Watson, P. L. J. Am. Chem. Soc. 1983, 105, 6491–6493.
(22) Watson, P. L.; Parshall, G. W. Acc. Chem. Res. 1985, 18, 51–56.
(23) Jantunen, K. C.; Scott, B. L.; Hay, P. J.; Gordon, J. C.; Kiplinger,
J. L. J. Am. Chem. Soc. 2006, 128, 6322–6323.
(5) Dormond, A.; El Bouadili, A. A.; Mo¨ıse, C. J. Chem. Soc., Chem.
Commun. 1985, 914–916.
(6) Duchateau, R.; Brussee, E. A. C.; Meetsma, A.; Teuben, J. H.
Organometallics 1997, 16, 5506–5516.
(7) Duchateau, R.; van Wee, C. T.; Teuben, J. H. Organometallics 1996,
15, 2291–2302.
(24) Cameron, T. M.; Gordon, J. C.; Scott, B. L.; Tumas, W. Chem.
Commun. 2004, 1398–1399.
(8) Jantunen, K. C.; Scott, B. L.; Kiplinger, J. L. J. Alloys Compd. 2007,
444-445, 363–368.
(25) Reardon, D.; Conan, F.; Gambarotta, S.; Yap, G.; Wang, Q. J. Am.
Chem. Soc. 1999, 121, 9318–9325.
(9) Jordan, R. F.; Guram, A. S. Organometallics 1990, 9, 2116–2123.
(10) Kiplinger, J. L.; Scott, B. L.; Schelter, E. J.; Pool Davis Tournear,
J. A. J. Alloys Compd. 2007, 444-445, 477–482.
(11) Klei, B.; Teuben, J. H. J. Chem. Soc., Chem. Commun. 1978, 659–
660.
(26) Scott, J.; Gambarotta, S.; Korobkov, I.; Budzelaar, P. H. M. J. Am.
Chem. Soc. 2005, 127, 13019–13029.
(27) Pool, J. A.; Scott, B. L.; Kiplinger, J. L. Chem. Commun. 2005, 2591–
2593.
(12) Klei, E.; Teuben, J. H. J. Organomet. Chem. 1981, 214, 53–64.
(13) Pool, J. A.; Scott, B. L.; Kiplinger, J. L. J. Alloys Compd. 2006,
418, 178–183.
(28) Gray, S. D.; Weller, K. J.; Bruck, M. A.; Briggs, P. M.; Wigley,
D. E. J. Am. Chem. Soc. 1995, 117, 10678–10693.
(29) Weller, K. J.; Gray, S. D.; Briggs, P. M.; Wigley, D. E. Organo-
metallics 1995, 14, 5588–5597.
(14) Radu, N. S.; Buchwald, S. L.; Scott, B.; Burns, C. J. Organometallics
1996, 15, 3913–3915.
(30) Katzer, J. R.; Sivasubramanian, R. Catal. ReV.-Sci. Eng. 1979, 20,
155–208.
(15) Scollard, J. D.; McConville, D. H.; Vittal, J. J. Organometallics 1997,
16, 4415–4420.
(31) Ho, T. C. Catal. ReV.-Sci. Eng. 1988, 30, 117–160.
(32) Perot, G. Catal. Today 1991, 10, 447–472.
(16) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan,
M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am.
Chem. Soc. 1987, 109, 203–219.
(33) Prins, R. In A Handbook of Heterogeneous Catalysis; Ertl, G.,
Kno¨zinger, H., Weitkamp, J., Eds.; VCH Verlagsgesellschaft: Wein-
heim, Germany, 1997.
(17) Watson, P. L. J. Chem. Soc., Chem. Commun. 1983, 276–277.
(18) Jantunen, K. C.; Scott, B. L.; Gordon, J. C.; Kiplinger, J. L.
Organometallics 2007, 26, 2777–2781.
(34) Kabe, T.; Ishikawa, A.; Qian, W. Hydrodesulfurization and Hy-
drodenitrogenation; Wiley-VCH: Chichester, UK, 1999.
9
342 J. AM. CHEM. SOC. 2010, 132, 342–355
10.1021/ja908489p  2010 American Chemical Society

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
The cleavage of aromatic C-N bonds is relatively rare27,36-52
and the dearomatization of N-heterocycles may be the first step
in developing a methodology for their C-N bond cleavage.1,42,53
Our group has previously reported several reactions mediated
by the scandium, yttrium, and lutetium alkyl complexes
(NNfc)M(CH2Ar)(THF), 1M-THF (NNfc ) fc(NSitBuMe2)2, fc
) 1,1′-ferrocenylene; M ) Sc, Lu: Ar ) 3,5-Me2C6H3; M )
Y: Ar ) C6H5), in which pyridines have been C-H activated
at the R-carbon (Scheme 1).51,54,55 These ortho-metalated
Scheme 1. Reactions of 2-Phenylpyridine Mediated by the Alkyl
Complexes 1M-THF (M ) Sc/Y/Lu)
complexes were supported by the silylated ferrocene diamide
51,54-58
ligand NNfc
and underwent subsequent coupling and
isomerization reactions (Scheme 1).51,55 The isomerization
reaction entailed the migration of a hydrogen atom within a
single heterocyclic ring. On the basis of DFT computational
studies, we proposed that the mechanism involved a transition
state in which the heterocyclic ring adopted a boat-like
conformation, allowing the hydrogen to migrate between the
two carbon atoms involved in the process.
In addition to the C-H activation of aromatic N-heterocycles,
we have found that an alkyl migration takes place when
substrates such as isoquinoline and 2,2′-bipyridine are employed
in reactions with group 3 (scandium, yttrium, lutetium, and
lanthanum) benzyl complexes. Herein we report the synthesis
and characterization of these products and their subsequent
reactivity with unsaturated substrates. These reactions also
involve hydrogen transfer; however, in contrast to the previous
isomerization process, the hydrogen is transferred between two
separate rings and the mechanism involves metalocycles as
intermediates. DFT computational studies were undertaken to
provide insight and support the proposed mechanism. The
present report represents the first combined experimental and
computational systematic study of alkyl migration to aromatic
N-heterocycles involving group 3 complexes and of the
subsequent reactions of these products. These results comple-
ment those reported for transition-metal olefin polymerization
catalysts supported by derivatized aromatic N-heterocyclic
ligands.25,26,59,60
(35) Furimsky, E.; Massoth, F. E. Catal. ReV.-Sci. Eng. 2005, 47, 297–
489 and references therein.
(36) Gray, S. D.; Weller, K. J.; Bruck, M. A.; Briggs, P. M.; Wigley,
D. E. J. Am. Chem. Soc. 1995, 117, 10678–10693.
(37) Gray, S. D.; Smith, D. P.; Bruck, M. A.; Wigley, D. E. J. Am. Chem.
Soc. 1992, 114, 5462–5463.
(38) Weller, K. J.; Gray, S. D.; Briggs, P. M.; Wigley, D. E. Organo-
metallics 1995, 14, 5588–5597.
(39) Weller, K. J.; Filippov, I.; Briggs, P. M.; Wigley, D. E. Organome-
tallics 1998, 17, 322–329.
(40) Weller, K. J.; Filippov, I.; Briggs, P. M.; Wigley, D. E. J. Organomet.
Chem. 1997, 528, 225–228.
(41) Allen, K. D.; Bruck, M. A.; Gray, S. D.; Kingsborough, R. P.; Smith,
D. P.; Weller, K. J.; Wigley, D. E. Polyhedron 1995, 14, 3315–
3333.
(42) Weller, K. J.; Fox, P. A.; Gray, S. D.; Wigley, D. E. Polyhedron
1997, 16, 3139–3163 and references therein.
(43) Kleckley, T. S.; Bennett, J. L.; Wolczanski, P. T.; Lobkovsky, E. B.
J. Am. Chem. Soc. 1997, 119, 247–248.
Results and Discussion
(44) Bonanno, J. B.; Veige, A. S.; Wolczanski, P. T.; Lobkovsky, E. B.
Inorg. Chim. Acta 2003, 345, 173–184.
Synthesis and Characterization of Complexes. The complexes
1M-THF had been previously characterized for scandium,
yttrium, and lutetium.51,55,56 In the present work, this chemistry
is extended to lanthanum, (NNfc)La(CH2Ph)(THF) (1La-THF),
which allows a comparison of the reactivity behavior of the
group 3 benzyl complexes across a wider range of atomic radii
than before. In addition, the complex La(CH2Ar)3(THF)3 (Ar
) 3,5-Me2C6H3) was prepared and characterized as a precursor
to 1′La-THF (Ar ) 3,5-Me2C6H3). The 3,5-Me2C6H3 group was
chosen initially because it improved the solubility of the
corresponding scandium and lutetium complexes; however,
La(CH2Ar)3(THF)3 (Ar ) 3,5-Me2C6H3) was more difficult to
prepare and store than its scandium and lutetium counterparts.
Therefore, most reactivity studies presented in this work were
carried out with 1La-THF (Ar ) Ph), for which La(CH2Ph)3-
(THF)3 was used as a precursor.
(45) Covert, K. J.; Neithamer, D. R.; Zonnevylle, M. C.; LaPointe, R. E.;
Schaller, C. P.; Wolczanski, P. T. Inorg. Chem. 1991, 30, 2494–
2508.
(46) Neithamer, D. R.; Parkanyi, L.; Mitchell, J. F.; Wolczanski, P. T.
J. Am. Chem. Soc. 1988, 110, 4421–4423.
(47) Fout, A. R.; Bailey, B. C.; Tomaszewski, J.; Mindiola, D. J. J. Am.
Chem. Soc. 2007, 129, 12640–12641.
(48) Bailey, B. C.; Fan, H.; Huffman, J. C.; Baik, M. H.; Mindiola, D. J.
J. Am. Chem. Soc. 2006, 128, 6798–6799.
(49) Huertos, M. A.; Perez, J.; Riera, L. J. Am. Chem. Soc. 2008, 130,
5662–5663.
(50) Huertos, M. A.; Pee´rez, J.; Riera, L.; Menee´ndez-Vela´zquez, A. J. Am.
Chem. Soc. 2008, 130, 13530–13531.
(51) Carver, C. T.; Diaconescu, P. L. J. Am. Chem. Soc. 2008, 130, 7558–
7559.
(52) Monreal, M. J.; Khan, S.; Diaconescu, P. L. Angew. Chem., Int. Ed.
2009, 48, 8352–8355.
(53) Sanchez-Delgado, R. A. Organometallic modelling of the hydrodes-
ulfurization and hydrodenitrogenation reactions; Kluwer Academic
Publishers: Dordrecht, The Netherlands, 2002.
(54) Carver, C. T.; Diaconescu, P. L. J. Alloys Compd. 2009, 488, 518-
523.
(55) Carver, C. T.; Benitez, D.; Miller, K. L.; Williams, B. N.; Tkatchouk,
E.; Goddard, W. A.; Diaconescu, P. L. J. Am. Chem. Soc. 2009, 131,
10269–10278.
(56) Carver, C. T.; Monreal, M. J.; Diaconescu, P. L. Organometallics
2008, 27, 363–370.
(57) Monreal, M. J.; Diaconescu, P. L. Organometallics 2008, 27, 1702–
1706.
In order to test the scope of the aromatic N-heterocycle ortho-
(58) Monreal, M. J.; Carver, C. T.; Diaconescu, P. L. Inorg. Chem. 2007,
46, 7226–7228.
metalation reaction,51,54,55 a biheterocyclic substrate, 2,2′-
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 343

A R T I C L E S
Miller et al.
Scheme 2. Reactions of Isoquinoline Mediated by the Benzyl
Complexes 1M-THF (M ) Sc/Y/Lu/La)
Figure 1. ORTEP representation of 2Sc (left) and 3Y-iqnMe (right) with
ellipsoids drawn at the 50% probability level (irrelevant hydrogen atoms
were removed for clarity).
bipyridine, was mixed with the alkyl complexes 1M-THF (M
) Sc, Lu). However, instead of the C-H activation reaction
encountered with pyridines, the product of alkyl transfer was
observed (eq 1, 2M). The concomitant dearomatization of one
of the pyridine rings was indicated in the 1H NMR spectrum of
the scandium product 2Sc by the upfield shift of the hydrogen
resonances of the 4 and 4′-positions, from 7.20 ppm in free
2,2′-bipyridine (C6D6, 25 °C) to 4.08 ppm (4-position, dearo-
matized ring) and 6.90 ppm (4′-position) in the dearomatized
complex. The protons on the 3 and 5-positions showed an
109.89(33), and 110.52(34)° are in agreement with the above
structural assignment.
Since the reaction of 1Sc-THF with 8-methylisoquinoline
led to the expected ortho-metalated product,55 a study of the
reactivity behavior of the alkyl metal complexes 1M-THF with
isoquinoline (iqn) was undertaken next. The reaction of
1M-THF with isoquinoline first led to THF displacement by
the N-heterocycle (1M-iqn, Scheme 2). Upon further reaction,
the inspection of the 1H NMR spectrum of the product revealed
neither free mesitylene nor toluene, which would result from
the C-H activation of the heterocycle. Rather, a 1,3-alkyl
migration of the benzyl group onto the R-carbon of isoquinoline
took place at room temperature (M ) Y, Lu, La) or 50 °C (M
) Sc), resulting in the formation of 3M-THF (Scheme 2). This
migration was accompanied by the appearance in the 1H NMR
(500 MHz, C6D6) spectrum of the lutetium complex 3Lu-THF
(the other complexes show analogous NMR characteristics) of
a doublet at 5.98 ppm, as well as a doublet of doublets at 5.20
ppm. The appearance of those signals was concomitant with
the disappearance of aromatic isoquinoline peaks. By using
deuterium-labeled isoquinoline, in which deuterium incorpora-
tion at the 1 and 3-positions was 36 and 54%, respectively, an
1
upfield shift as well, to the olefinic region of the H NMR
spectrum: a doublet of doublets at 5.54 ppm (3-position) and a
doublet of triplets at 4.76 ppm (5-position) were assigned to
these protons (coupling of the C3 and C5 protons occurs). The
1H NMR spectrum of the lutetium product was analogous to
1
that of 2Sc. A H-1H COSY experiment (25 °C, C6D6) was
performed on 2Lu, confirming the assignments of the 1H NMR
spectra (see the Supporting Information for details) and the
coupling between the C3 and C5 protons.
X-ray crystallography indicated that the alkyl transfer oc-
curred to the 4-position of one of the pyridine rings to give 2M
(Figure 1, M ) Sc).19 It is likely that an initial 1,3-alkyl transfer
to the 2-position takes place to give 2M-1,3, which is similar
to the product of the 1,3-alkyl transfer from Lu(CH2SiMe3)3-
(THF)2 to terpyridine,23 and then an isomerization leads to 2M,
likely to be the thermodynamically stable product. Similar
behavior has been reported for the reactions of group 3 hydride
complexes with pyridine.61-65 Metrical parameters were con-
sistent with the dearomatization of one of the pyridine rings.
For example, the C-C distances in the aromatic ring are
1.3774(41), 1.3729(44), 1.3841(45), and 1.3961(42) Å, while
in the dearomatized one they are 1.3458(44), 1.5008(50),
1.5058(49), and 1.3354(45) Å, with the two longest distances
to the sp3-carbon atom. Also, the Sc-N distance to the
dearomatized pyridine ring, 2.1397(25) Å, is shorter by 0.15 Å
than the Sc-N distance (2.2827(24) Å) to the aromatic ring.
The CCC angles around the sp3-carbon atom of 107.5(3),
2
investigation of the H NMR spectrum for the corresponding
product was possible. Its examination allowed the assignment
of the peak at 5.20 ppm to the C1 proton of the dearomatized
isoquinoline as well as the peak at 7.58 ppm to the C3 proton.
1
These assignments were corroborated by an H-1H COSY
experiment (25 °C, C6D6) that revealed that the resonances at
5.98 ppm and 7.58 ppm belong, indeed, to protons on the
adjacent C4 and C3 carbons, respectively. The methylene
protons of the benzyl group were found at 3.38 and 2.83 ppm
as two doublets of doublets that integrated to one proton each.
Furthermore, each methylene proton and the C1 proton were
1
1
found to be related by the H-1H COSY experiment. The H
NMR spectroscopy assignment was confirmed by 13C NMR
spectroscopy as well: two peaks, at 108.9 and 98.8 ppm, were
assigned to C3 and C4 of isoquinoline, respectively, since those
carbons are olefinic. Five sharp peaks were found in the aliphatic
region of the spectrum from 40.2 ppm to 20.6 ppm. These peaks
were assigned to the methylene carbon, C1 of isoquinoline, the
two methyl carbons of the benzyl moiety, and the remaining
carbons of the NNfc ligand (see the experimental section for
the full assignment).
(59) Vidyaratne, I.; Scott, J.; Gambarotta, S.; Duchateau, R. Organome-
tallics 2007, 26, 3201–3211.
(60) Knijnenburg, Q.; Gambarotta, S.; Budzelaar, P. H. M. Dalton Trans.
2006, 5442–5448.
(61) Evans, W. J.; Meadows, J. H.; Hunter, W. E.; Atwood, J. L. J. Am.
Chem. Soc. 1984, 106, 1291–1300.
(62) Deelman, B.-J.; Stevels, W. M.; Teuben, J. H.; Lakin, M. T.; Spek,
A. L. Organometallics 1994, 13, 3881–3891.
(63) Gountchev, T. I.; Tilley, T. D. Organometallics 1999, 18, 2896–
2905.
(64) Kirillov, E.; Lehmann, C. W.; Razavi, A.; Carpentier, J.-F. Eur.
J. Inorg. Chem. 2004, 2004, 943–945.
An X-ray diffraction study of 3Lu-THF confirmed the
identity of the product; unfortunately, the data collection was
of poor quality and only connectivity information could be
(65) Robert, D.; Voth, P.; Spaniol, T. P.; Okuda, J. Eur. J. Inorg. Chem.
2008, 2008, 2810–2819.
9
344 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
Figure 3. ORTEP representation of 6Lu-(iqn)2 (left) and 8Y (right) with
ellipsoids drawn at the 50% probability level (irrelevant hydrogen atoms
were removed for clarity).
Figure 2. ORTEP representation of 4Lu-(iqn)2 (left, 50% probability) and
4′Y-1,5-(py)2 (right, 35% probability); irrelevant hydrogen atoms were
removed for clarity.
is present in 4Lu-(iqn)2 because of the absence of the benzyl
group in the proximity of the metal center. Metrical parameters
are similar to those observed for 3Y-iqnMe, with long C-C
and N-C distances of 1.5110(40) and 1.4721(36) Å, respec-
tively, and an NCC angle of 113.17(24)° at the sp3-carbon atom.
Inspection of the crude reaction mixture by 1H NMR
spectroscopy revealed that there was a second product present,
5Ar (Scheme 2). The two products, 4M-(iqn)2 and 5Ar, were
separated by fractional crystallization from hexanes. The
comparison of the 1H NMR spectrum of 5Ar with spectra
reported for 1-(3,4-dimethylbenzyl)isoquinoline and 1-(4-me-
thylbenzyl)isoquinoline66 allowed its identification as 1-ben-
zylisoquinoline or 1-(3,5-dimethylbenzyl)isoquinoline. Also, the
identity of 5Ar, Ar ) Ph (1-benzylisoquinoline), was verified
by the X-ray crystal structure of another reaction product, in
which 1-benzylisoquinoline was coordinated to the metal center
(see below, Figure 3).
gathered. However, the analogous yttrium complex, 3Y-iqnMe
,
in which isoquinoline and THF were replaced by 3-methyliso-
quinoline (iqnMe), provided good quality data. X-ray crystal-
lography (Figure 1) confirmed that the alkyl transfer occurred
to the 1-position of isoquinoline. Although an argument could
be construed that for 3Y-iqnMe the 3-position bears the methyl
group and, therefore, steric hindrance would prevent transfer
of the benzyl group to that position, the NMR spectroscopy
data support the preferred transfer to the 1-position for both
isoquinoline and 3-methylisoquinoline, as does the connectivity
information obtained from the structure of 3Lu-THF. Metrical
parameters are consistent with the dearomatization of the
pyridine ring. For example, the N-C distances in the coordi-
nated isoquinoline are 1.3272(47) and 1.3830(048) Å, while in
the dearomatized one they are 1.3761(48) and 1.4693(44) Å,
with the longest distance to the sp3-carbon atom. In addition,
C-C distances also reflect the dearomatization of one isoquino-
line ligand: the values for the pyridine ring of the coordinated
isoquinoline are 1.3990(50), 1.4140(55), 1.3693(52), and
1.4191(52) Å, while in the dearomatized one they are 1.4204(53),
1.5186(51), 1.4395(52), and 1.3690(53) Å, with the longest
distance involving the sp3-carbon atom. The Y-N distance to
the dearomatized isoquinoline (2.2770(31) Å) is shorter by
almost 0.2 Å than the Y-N distance (2.4675(32) Å) to the
aromatic, coordinated one. The angles around the sp3-carbon
atom of CCC - 109.77(28)° and NCC - 110.15(27), 113.08(32)°
also support the above structural assignment.
The mixture of 4M-(iqn)2 and 5Ar can also be accessed
directly from the reaction of 1M-THF with four equivalents
of isoquinoline. For 4M-(iqn)2 to be produced, a hydrogen must
migrate from the benzyl-substituted isoquinoline to one of the
coordinated isoquinoline ligands, resulting in the dearomatization
of the latter along with the concurrent rearomatization of the
former. A hydrogen migration of this type can take place through
one of several different mechanisms, which will be discussed
below in conjunction with the results from computational
studies. After rearomatization, the coordinated 5Ar is likely
replaced by isoquinoline.
During the course of the reaction of 1M-THF with isoquino-
line, it was noticed that employing an excess of the substrate
led to the formation of a new product. Consequently, the reaction
of three equivalents of isoquinoline with 3Y-THF for six hours
or with 3Lu-THF for thirty hours, at 50 °C, allowed the
1
isolation of 4M-(iqn)2 (Scheme 2). The H NMR (300 MHz,
The reaction of 1M-THF with pyridine has always given a
mixture of products that proved intractable. In one instance
though, when the yttrium complex 1Y-THF was employed, a
product reminiscent of 4Y-(iqn)2 was isolated (eq 2). However,
instead of identifying 4′Y-1,3-(py)2, in which the dearomatized
pyridine bears the additional hydrogen atom at the 2-position,
as was the case for isoquinoline, the product of the hydrogen
migration to the 4-position, 4′Y-1,5-(py)2, was identified by
X-ray crystallography (Figure 2). Unfortunately, this reaction
suffered from poor reproducibility and attempts to characterize
4′Y-1,5-(py)2 by other methods were thwarted by its contami-
nation with other products. As will be discussed in the
C6D6) spectrum of 4Lu-(iqn)2 showed the appearance of two
doublets corresponding to one proton each, at 6.43 and 5.69
ppm, assigned to the C3 and C4 protons, respectively, of the
dearomatized isoquinoline ligand. A singlet at 4.56 ppm, which
integrated to two protons, was also identified and assigned to
the protons on the 1-position of the dearomatized isoquinoline.
Integration of the remaining peaks supported the presence of
two additional coordinated isoquinoline ligands. The 1H NMR
spectrum of the yttrium product was analogous to that of the
lutetium one.
The compound 4Lu-(iqn)2 was also characterized by single-
crystal X-ray diffraction (Figure 2), which confirmed that two
isoquinoline ligands were coordinated in 4Lu-(iqn)2 instead of
the one found in 3Y-iqnMe. It is likely that less steric crowding
(66) Uff, B. C.; Kershaw, J. R. J. Chem. Soc. (C) 1969, 666–668.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 345

A R T I C L E S
Miller et al.
Scheme 3. Hydrogen Transfer to Ketones from 3M-THF (M ) Sc/Y/La/Lu)
computational-study section, both the alkyl and hydrogen-
transfer reactions have high activation barriers that are similar
in magnitude in the case of pyridine. Furthermore, the product
of the hydrogen-transfer reaction is not as stabilized with respect
to the starting material as in the case of isoquinoline. When
these considerations are taken into account, along with the
possibility that other reactions between 1Y-THF and pyridine
could also occur,54,55 it is not surprising that the product
4′Y-1,5-(py)2 could not be obtained reproducibly.
moiety to one of the nitrogen atoms of azobenzene (eq 3). The
single-crystal X-ray structure of 8Y (Figure 3) allowed the
unambiguous assignment of 1-benzylisoquinoline, formed by
the rearomatization of the N-heterocyclic fragment, as mentioned
earlier. Furthermore, the metrical parameters of the newly
formed hydrazide fragment are consistent with the reduction
of azobenzene: the N-N distance of 1.4626(61) Å is similar to
N-N distances in other [η2-(PhNH-NPh)]- complexes,70-72
while the Y-N distances are significantly different from each
other (2.2525(48) and 2.4359(45) Å).
It was reasoned that the hydrogen migration from 3M-THF
may also occur in the presence of other substrates. Indeed, the
reaction of 3M-THF with 2-adamantanone or benzophenone
led to the formation of new products, 6Lu-iqnAr and 7M-iqnAr
(M ) Sc, Y, La), respectively (Scheme 3). Although 1H NMR
spectroscopy showed that the alkoxide formation had taken place
in the presence of the ketone substrate alone, the addition of
two equivalents of isoquinoline was necessary to facilitate the
crystallization of the products as 6Lu-(iqn)2 and 7M-(iqn)2 (M
) Y, La).
Since different types of mechanisms can be envisioned for
the hydrogen transfer, ꢀ-hydride elimination and direct transfer
from one ligand to another (see below), it was reasoned that if
hydride transfer operates then it should also occur to nonpolar
substrates, such as alkenes and alkynes.73 Therefore, reactions
of 3M-THF with 2-butyne, diphenylacetylene, and ethylene
were carried out. Initially, a transformation of the starting
material was observed, but since the same outcome was
identified in all of those reactions, it was reasoned that no
reaction took place with any of the substrates. Indeed, when
3M-THF was heated in C6D6 by itself (eq 4), the previous
reaction mixture was also obtained. Interestingly, either toluene
(M ) Y, La) or mesitylene (M ) Sc, Lu) was produced as a
byproduct in the reaction, suggesting a possible ꢀ-alkyl elimina-
tion of the benzyl moiety as a reaction pathway. ꢀ-Alkyl
elimination on group 3 metals is not without precedent and is
as common as ꢀ-hydride elimination as a decomposition
pathway for organo-f-element complexes.74,75 However, inspec-
tion of the 2H NMR spectrum of the reaction products showed
1
Analysis of the H NMR (500 MHz, C6D6) spectrum of
6Lu-(iqn)2 showed a new peak at 4.55 ppm corresponding to
the hydrogen on the newly sp3-hybridized carbon, the adamantyl
C2. This proton displayed the expected downfield shift relative
to the other adamantyl protons, which were located significantly
more upfield, in the region from 2.32 to 1.39 ppm. The
complexes 7M-(iqn)2 were characterized by 1H and 13C NMR
spectroscopy and elemental analysis. Their structural assignment
was based upon comparison of their 1H NMR spectra to that of
6Lu-(iqn)2. X-ray crystallography of 6Lu-(iqn)2 confirmed the
solution structure (Figure 3): the Lu-O distance of 2.0368(51)
Å is similar to other Lu-O distances in alkoxide complexes.67-69
Also, the C-O distance of 1.4160(75) Å and the OCC/CCC
angles of 111.07(53), 112.97 (0.55), and 108.11(52)° involving
the former ketone carbon are consistent with the presence of
an sp3-hybridized carbon atom.
Another substrate used to probe the hydrogen transfer from
3M-THF was azobenzene. The reaction between 3Y-THF and
azobenzene gave the product 8Y, in which a hydrogen atom had
been transferred from the dearomatized 1-benzylisoquinoline
(70) Yuan, F. Synth. React. Inorg. ME 2005, 35, 703–708.
(71) Evans, W. J.; Kociok-Koehn, G.; Leong, V. S.; Ziller, J. W. Inorg.
Chem. 1992, 31, 3592–3600.
(72) Zambrano, C. H.; Fanwick, P. E.; Rothwell, I. P. Organometallics
1994, 13, 1174–1177.
(67) Konkol, M.; Spaniol, T. P.; Kondracka, M.; Okuda, J. Dalton Trans.
2007, 4095–4102.
(73) Burger, B. J.; Thompson, M. E.; Cotter, W. D.; Bercaw, J. E. J. Am.
Chem. Soc. 1990, 112, 1566–1577.
(68) Anwander, R.; Munck, F. C.; Priermeier, T.; Scherer, W.; Runte,
O.; Herrmann, W. A. Inorg. Chem. 1997, 36, 3545–3552.
(69) Beer, P. D.; Drew, M. G. B.; Kan, M.; Leeson, P. B.; Ogden, M. I.;
Williams, G. Inorg. Chem. 1996, 35, 2202–2211.
(74) Watson, P. L.; Roe, D. C. J. Am. Chem. Soc. 1982, 104, 6471–6473.
(75) Jeske, G.; Lauke, H.; Mauermann, H.; Swepston, P. N.; Schumann,
H.; Marks, T. J. J. Am. Chem. Soc. 1985, 107, 8091–8103.
9
346 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
1.4687(73) Å and the N-C distance of 1.3720(65) Å are
consistent with this interpretation.
It is likely that two molecules of 3M-THF are required to
form the two reaction products 4M-(THF)2 and 9M. Although
an in-depth mechanistic study of this process is beyond the scope
of the present article, the reaction of 3La-THF-d (for the
formation of 3La-THF-d, isoquinoline deuterated at the 1-36%
and 3-54% positions was used) showed the formation of
deuterated toluene (eq 5). The integration of the methyl-
deuterons from the toluene obtained (ca. 30%) was consistent
with toluene formation from the benzyl and the deuteron on
the C1 carbon, but given that a substrate that was 100%
deuterium labeled at this position could not be obtained, other
possibilities cannot be excluded. The reaction in eq 4 is
somewhat similar to that reported by Cronin et al. for phenan-
thridinium salts:85 in the presence of amines, the phenanthri-
dinium salt was dearomatized and a product corresponding to
3M-THF was formed. In the presence of a base, the dearoma-
tized phenanthridinium transferred one proton to a molecule of
the initial phenanthridinium salt, leading to the formation of a
product corresponding to 4M-(THF)2, and underwent rearo-
matization, giving a product corresponding to 9M.
It was observed that the reaction in eq 4 was faster for the
lanthanum complex than for the yttrium one, an observation
supporting the proposal that the reaction is bimolecular in nature.
The larger lanthanum center would allow the two coordination
spheres to come into closer proximity than the yttrium one
would. Another factor that may influence the outcome of the
reaction is the fact that both the yttrium and lanthanum
complexes 3M-THF liberate toluene, while the scandium and
lutetium complexes eliminate mesitylene. It is possible that the
electron-donating properties of the two methyl groups disfavor
the formation of mesitylene. Other factors may also be
responsible, given the complex nature of the reaction outcome.
Computational Studies. Calculations at the B3LYP/
LACV3P++**(2f) level were performed to gain insight into
different pathways for the reactions studied. For consistency,
all calculations employed yttrium complexes, using the full NNfc
ligand. First, the electronic nature of the benzyl transfer from
yttrium to the 2-position of pyridine (to simplify computer effort)
was investigated (Scheme 4a) as a model for the transformation
1Y-THF to 3Y-THF (Scheme 2). By using one molecule of
pyridine as a model substrate, the barrier for alkyl transfer for
1Y-py was found to be 29.8 kcal/mol (Scheme 4a). When an
additional molecule of pyridine was coordinated to the yttrium
center, the barrier for alkyl transfer was slightly lower, at 25.4
kcal/mol (Scheme 4b), than in the previous case, and the
transformation led to 3′Y-py, which was 12.3 kcal/mol more
stable than 3′Y. An extra molecule of pyridine lowered the
Figure 4. ORTEP representation of 4Y-(THF)2 (left) and 9La (right) with
ellipsoids drawn at the 50% probability level (irrelevant hydrogen atoms
were removed for clarity).
that the toluene/mesitylene observed did not contain deuterium,
so the deuterated solvent was not involved in its formation.
The workup of the reaction mixture revealed that two
products, 4M-(THF)2 and 9M, were formed in a 1:1 ratio (eq
4). The two products were separated by fractional crystallization
giving red and orange crystals. Because of differing solubility
properties, the orange crystals were characterized by X-ray
crystallography for the yttrium complex and the red crystals
for the lanthanum one. For scandium and lutetium, analogous
reaction mixtures were formed based on their 1H NMR spectra,
but the products could not be isolated. X-ray crystallography
indicated that the orange crystals corresponded to 4Y-(THF)2,
while the red ones to 9La (Figure 4). The reaction stoichiometry
is wrong for the number of THF molecules involved, but it is
possible that an additional molecule of THF is necessary only
during the crystallization of 4M-(THF)2 and the decomposition
of 9M may provide it. The X-ray crystal structure of 4Y-(THF)2
is similar to that of 4Lu-(iqn)2, with long C-C and N-C
distances of 1.5092(47) and 1.4749(41) Å, respectively, and an
NCC angle of 112.79(28)° at the sp3-carbon atom. The X-ray
crystal structure of 9La shows the formation of a lanthanum
pseudo-η3-allyl fragment and the rearomatization of the pyridine
ring; the La-Callyl distances of 2.8832(55), 2.9489(53), and
3.0622(55) Å and La-Niqn distance of 2.5827(44) Å are
consistent with this formulation. It is proposed that the longer
La-Callyl distances than those found in other lanthanum-allyl
complexes76-81 are a consequence of the charge delocalization
over four bonds to include the N-C bond of the isoquinoline
fragment. This delocalization is supported by the shorter
La-Niqn distance than those found in other lanthanum-pyridine
complexes.82-84 The two C-C distances of 1.4032(71) and
(76) Kretschmer, W. P.; Brummelhuis, B. t.; Meetsma, A.; Teuben, J. H.
Z. Anorg. Allg. Chem. 2006, 632, 1933–1935.
(77) Karsch, H. H.; Appelt, A.; Mu¨ller, G. Angew. Chem., Int. Ed. 1986,
25, 823–824.
(78) Taube, R.; Windisch, H.; Hemling, H.; Schumann, H. J. Organomet.
Chem. 1998, 555, 201–210.
(83) Katkova, M. A.; Kurskii, Y. A.; Fukin, G. K.; Averyushkin, A. S.;
Artamonov, A. N.; Vitukhnovsky, A. G.; Bochkarev, M. N. Inorg.
Chim. Acta 2005, 358, 3625–3632.
(79) Taube, R.; Windisch, H.; Go¨rlitz, F. H.; Schumann, H. J. Organomet.
Chem. 1993, 445, 85–91.
(80) Taube, R.; Windisch, H.; Weienborn, H.; Hemling, H.; Schumann,
H. J. Organomet. Chem. 1997, 548, 229–236.
(84) Giesbrecht, G. R.; Gordon, J. C.; Clark, D. L.; Scott, B. L. Appl.
Organomet. Chem. 2005, 19, 98–99.
(81) Taube, R.; Windisch, H.; Maiwald, S.; Hemling, H.; Schumann, H.
J. Organomet. Chem. 1996, 513, 49–61.
(85) Parenty, A. D. C.; Smith, L. V.; Pickering, A. L.; Long, D.-L.; Cronin,
L. J. Org. Chem. 2004, 69, 5934–5946.
(82) Giesbrecht, G. R.; Collis, G. E.; Gordon, J. C.; Clark, D. L.; Scott,
B. L.; Hardman, N. J. J. Organomet. Chem. 2004, 689, 2177–2185.
(86) Tsuruta, H.; Yamaguchi, K.; Imamoto, T. Chem. Commun. 1999,
1703–1704.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 347

A R T I C L E S
Miller et al.
Scheme 4. Potential Free Energies (∆G) at 298 K for the Alkyl Transfer from (a) 1Y-py to 3′Y and (b) 1Y-(py)2 to 3′Y-py, Showing
Selected Bond Distances for the Transition Structure 1TS3′Y-(py)2
Scheme 5. Proposed Mechanisms for the Transformation of
Ponndorf-Verley reduction of ketones by various metal
3M-THF to 4M-(iqn)2
catalysts,90-104 a reaction similar to the transformation of
3M-iqn to 4M-(iqn)2. In addition to the two mechanisms
considered by us, a radical mechanism had also been proposed
for the Meerwein-Ponndorf-Verley reduction of ketones.105 The
first mechanism, the hydridic route, is often invoked when
explaining the behavior of transition metal catalysts and involves
the formation of a metal hydride followed by hydride transfer
from the metal to the substrate.106-108 The second mechanism,
the direct-hydrogen transfer, used mostly for aluminum-group
complexes, implies a concerted-hydrogen transfer, where both
the hydrogen donor and the hydrogen acceptor are held in close
proximity by the metal center, and involves a cyclic transition
state. Recently, a combined computational and experimental
study presented evidence supporting the direct hydrogen-transfer
mechanism for catalysis by aluminum complexes.104
energy of the product further, by 7.7 kcal/mol, consistent with
the experimental observations, since the isolated complexes were
analogous to 3′Y-(py)2.
(90) Meerwein, H.; Schmidt, R. Liebigs Ann. Chem. 1925, 444, 221–
In order to understand the details of the alkyl-transfer reaction,
238.
a natural bond orbital analysis on 1TS3′Y-(py)2 was performed.
(91) Ponndorf, W. Angew. Chem. 1926, 39, 138–143.
A negative charge buildup was observed at the migrating benzyl
carbon (-0.68) and at the N atom (-0.77) of the heterocycle
undergoing dearomatization. The natural localized molecular
orbital/natural population analysis87,88 bond orders showed a
bond order of 0.08 for Y-CBz (2.697 Å distance) and a bond
order of 0.55 for CBz-Cpy (2.126 Å distance). These values
indicate a late transition state where the benzyl carbon migrates
as a carbanion and has almost completely detached from yttrium,
while the receiving pyridine is partially dearomatized at the
potential energy saddle point. Our results suggest that the
activation barriers in the dearomatization of heterocycles will
likely correlate with the lost aromatic stabilization energy.89
(92) Verley, M. Bull. Soc. Chim. Fr. 1925, 871–874.
(93) Cha, J. S. Org. Proc. Res. DeV. 2006, 10, 1032–1053.
(94) de Graauw, C. F.; Peters, J. A.; van Bekkum, H.; Huskens, J. Synthesis
1994, 1994, 1007–1017.
(95) Yan, P.; Nie, C.; Li, G.; Hou, G.; Sun, W.; Gao, J. Appl. Organomet.
Chem. 2006, 20, 338–343.
(96) Fukuzawa, S.-i.; Nakano, N.; Saitoh, T. Eur. J. Org. Chem. 2004,
2004, 2863–2867.
(97) Klomp, D.; Maschmeyer, T.; Hanefeld, U.; Peters, J. A. Chem.sEur.
J. 2004, 10, 2088–2093.
(98) Nakano, Y.; Sakaguchi, S.; Ishii, Y. Tetrahedron Lett. 2000, 41, 1565–
1569.
(99) Nishikido, J.; Yamamoto, F.; Nakajima, H.; Mikami, Y.; Matsumoto,
Y.; Mikami, K. Synlett 1999, 1999, 1990–1992.
(100) Evans, D. A.; Nelson, S. G.; Gagne, M. R.; Muci, A. R. J. Am. Chem.
Soc. 2002, 115, 9800–9801.
Next, the transformation from 3M-THF to 4M-(iqn)2
(Scheme 2) was studied. Two types of mechanisms were
envisioned (Scheme 5): (1) ꢀ-hydride elimination and (2) direct
transfer from one heterocyclic ligand to another. Both types of
mechanisms have been proposed to explain the Meerwein-
(101) Lebrun, A.; Namy, J.-L.; Kagan, H. B. Tetrahedron Lett. 1991, 32,
2355–2358.
(102) Namy, J. L.; Souppe, J.; Collin, J.; Kagan, H. B. J. Org. Chem. 1984,
49, 2045–2049.
(103) Fernandez, I.; Sierra, M. A.; Cossio, F. P. J. Org. Chem. 2007, 72,
1488–1491.
(104) Cohen, R.; Graves, C. R.; Nguyen, S. T.; Martin, J. M. L.; Ratner,
M. A. J. Am. Chem. Soc. 2004, 126, 14796–14803.
(105) Ashby, E. C. Acc. Chem. Res. 1988, 21, 414–421.
(106) Zassinovich, G.; Mestroni, G.; Gladiali, S. Chem. ReV. 1992, 92,
1051–1069.
(87) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899–
926.
(88) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.;
Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO. Version 5.0;
Theoretical Chemistry Institute, University of Wisconsin: Madison,
WI, 2001.
(107) Hanasaka, F.; Fujita, K.-i.; Yamaguchi, R. Organometallics 2004,
23, 1490–1492.
(108) Kitamura, M.; Tsukamoto, M.; Bessho, Y.; Yoshimura, M.; Kobs,
U.; Widhalm, M.; Noyori, R. J. Am. Chem. Soc. 2002, 124, 6649–
6667.
(89) Mandado, M.; Otero, N.; Mosquera, R. A. Tetrahedron 2006, 62,
12204–12210.
9
348 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
Scheme 6. Potential Free Energies (∆G) at 298 K for the (a)
ꢀ-Hydride elimination is not common for group 3 metals,
although it can take place.109-112 Past studies have found that,
in general, ꢀ-hydride elimination mechanisms for lanthanide and
actinide metal complexes would be much more endothermic
than their counterparts in the late transition metals.113 The lack
of reactivity toward olefins and alkynes points to the direct
hydrogen transfer as being the viable route. Accordingly, when
the ꢀ-hydride elimination pathway was evaluated, no saddle
point on the pathway where the hydrogen was transferred from
the ligand to yttrium was found. All attempts led to a stepwise
process, where the acidic hydrogen detached first, leading to
rearomatization of the heterocycle, followed by the formation
of an yttrium hydride species, with an activation barrier of ∼56
kcal/mol in toluene. The large activation barrier is not surprising
since the migrating hydrogen carries a partial positive charge
(as a consequence of the nature of the rearomatization process)
while the metal is also positively charged (+2.0).
Pyridine-to-Pyridine, (b) Pyridine-to-Isoquinoline, and (c)
Isoquinoline-to-Isoquinoline H-Transfer, Displaying the Natural
Charge on the N Atoms of the Heterocycles and on the Migrating
H Atom (b)
In the case of the direct transfer pathway to the 2-position of
pyridine, a transition state was found at 25.2 kcal/mol starting from
3′Y-(py)2 and leading to 4′Y-(py)2, which was 6.2 kcal lower in
energy than 3′Y-(py)2 (Scheme 6a). Given that an analagous
product to 4′Y-1,5-(py)2 was observed experimentally, a pathway
for the direct transfer to the 4-position of pyridine was also studied
and was found to have an activation barrier 15.4 kcal/mol higher
in energy than that calculated for the transfer to the 2-position,
albeit the products were virtually isoenergetic. The value of 40.6
kcal/mol for the direct transfer to the 4-position suggests that the
transfer to the 2-position is likely to occur first, giving 4′Y-(py)2,
which possibly isomerizes to 4′Y-1,5-(py)2.
It was anticipated that the barrier for the direct transfer from
pyridine to isoquinoline would be considerably lower than for
pyridine-to-pyridine transfer because the relative magnitude of
the lost (aza-ring in isoquinoline) versus regained (pyridine)
aromatic stabilization energy is higher for pyridine than for
isoquinoline (the total aromatic stabilization energy for iso-
quinoline is higher than that for pyridine, however, the aromatic
stabilization energy of the N-containing ring of isoquinoline is
lower than that of pyridine).89 In order to evaluate this, the direct
transfer from a dearomatized benzylpyridine to isoquinoline was
studied. It was found that the activation barrier was indeed much
lower than for the pyridine case (only 13.0 kcal/mol versus the
starting material).
Natural population analyses87,88 on the N atom of both
reacting heterocycles showed an inverted charge profile for
each reaction. The charges at the N atom on the isoquinoline
ligand changed from -0.59 in 3′Y-(iqn)2, to -0.75 in the
transition state 3′TS4′Y-(iqn)2, and -0.92 in 4′Y-(iqn)2
(Scheme 6b), while the charges at the N atom on the
substituted pyridine ligand varied in the reverse order (-0.91,
-0.75, and -0.59, respectively). The charge at the migrating
H was almost constant from +0.15 in 3′Y-(iqn)2, to +0.18
in the transition structure, and +0.17 in 4′Y-(iqn)2; the
charge of the metal center underwent a negligible change
(∼0.01 e) throughout the transformation. All these data
suggest that the hydrogen atom migrates more as a proton
than as a hydride, while a significant electronic reorganization
is happening at both ligands that, however, does not affect
their bonding with the metal center. The reacting ligands were
localized at the transition structure in a displaced face-to-
face arrangement with the H atom at 1.32 Å from the
benzylpyridine carbon and 1.49 Å from the isoquinoline
carbon, resembling a chairlike, six-member metalocycle (see
the Supporting Information for details).
The direct transfer from isoquinoline to isoquinoline (Scheme
6c) was also studied: a barrier of 20.6 kcal/mol was found,
consistent with a six-hour reaction time at 50 °C. In addition, our
hypothesis, that the energetic profile of H-transfer is dependent on
the net aromatic stabilization energy, was supported by the relative
free energies (∆G) of the three reactions analyzed (Scheme 6):
transformations where both reacting heterocycles were the same
(small changes in the aromatic stabilization energy, Scheme 6a
and 6c) were less exergonic (-6.2 and -4.2 kcal/mol, respectively)
than the reaction where the participating heterocycles were different
and there was a gain in the aromatic stabilization energy (-16.5
kcal/mol in the case of H-transfer from pyridine to isoquinoline,
Scheme 6b).
(109) Ephritikhine, M. Chem. ReV. 1997, 97, 2193–2242.
(110) Watson, P. L. J. Am. Chem. Soc. 1982, 104, 337–339.
(111) Evans, W. J.; Meadows, J. H.; Wayda, A. L.; Hunter, W. E.; Atwood,
J. L. J. Am. Chem. Soc. 1982, 104, 2015–2017.
(112) Bazan, G. C.; Schaefer, W. P.; Bercaw, J. E. Organometallics 1993,
12, 2126–2130.
Conclusions
The reactivity behavior of the benzyl complexes 1M-THF
toward 2,2′-bipyridine (M ) Sc, Lu) and isoquinoline (M )
Sc, Y, Lu, La) was investigated. The observed alkyl transfer to
(113) Bruno, J. W.; Marks, T. J.; Morss, L. R. J. Am. Chem. Soc. 1983,
105, 6824–6832.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 349

A R T I C L E S
Miller et al.
Synthesis of 1′La-THF. La(CH2Ar)3(THF)3 (600 mg, 0.849
mmol) was divided into two equal portions and each was dissolved in
1: 1 THF: Et2O (8 mL) and cooled to -35 °C. Each solution was
combined with 1 equiv of fc(NHSitBuMe2)2 (190 mg each, 0.850 mmol
total), as a cooled solution in Et2O and stirred for 3 h at 0 °C. The
solvent was removed under vacuum, the resulting orange solid was
washed with hexanes (0.5 mL), extracted with a 1:1 solution of toluene:
hexanes, and filtered through Celite. The volatiles were removed under
vacuum, the resulting solid was dissolved in a minimum amount of
Et2O that was layered with n-pentane, and the solution was placed in
a -35 °C freezer. The product was collected as crystals by decanting
the mother liquor. Yield: 600 mg, 92.3%. 1H NMR (500 MHz, C6D6):
δ, ppm: 6.36 (s, 2H, o-C6H2), 6.10 (s, 1H, p-C6H), 4.01 (br s, 4H,
fc-CH), 3.71 (br s, 4H, OCH2CH2), 3.23 (s, 4H, fc-CH), 2.60 (s, 2H,
La-CH2), 2.17 (s, 6H, C6H3-(CH3)2), 1.35 (br s, 4H, OCH2CH2), 0.99
(s, 18H, SiC(CH3)3), 0.18 (s, 12H, Si(CH3)2). 13C NMR (126 MHz,
C6D6): δ, ppm: 154.1, 142.6, 118.6, 118.1, 106.5, 70.3, 66.8, 65.7,
65.0, 27.5, 25.4, 22.3, 20.3, -2.9. Anal. (%) for C35H59FeLaN2OSi2
Calcd: C, 54.40; H, 7.43; N, 3.60; Found: C, 54.29; H, 7.62; N, 3.70.
Synthesis of 1La-THF. 1La-THF was prepared analogously
to the preparation of 1′La-THF, using La(CH2C6H5)3(THF)3:
La(CH2C6H5)3(THF)3 (1.236 g, 1.97 mmol) was divided in 4 equal
portions, dissolved in a 1:1 mixture of THF/Et2O (8 mL) and cooled
to -35 °C. This solution was combined with 1 equiv of H2(NNfc)
(219 mg each, 1.97 mmol total), as a cooled solution in Et2O and
stirred for 3 h at 0 °C. The solvent was removed under vacuum,
the resulting orange solid was washed with hexanes (1.5 mL),
extracted with a 1:1 solution of toluene: hexanes, and filtered
through Celite. The volatiles were removed, the solid was dissolved
in a minimum amount of Et2O that was layered with n-pentane
and the solution was placed in a -35 °C freezer. The product was
collected as crystals by decanting the mother liquor. Yield: 950
mg, 64.8% as two crops of crystals. 1H NMR (500 MHz, C6D6): δ,
ppm: 7.14 (m, 2H, C6H5), 6.65 (d, 2H, C6H5), 6.32 (t, 1H, C6H5),
4.02 (s, 4H, fc-CH), 3.63 (br s, 4H, OCH2CH2), 3.20 (s, 4H, fc-
CH), 2.65 (s, 2H, La-CH2), 1.31 (br s, 4H, OCH2CH2), 0.99 (s,
18H, SiC(CH3)3), 0.15 (s, 12H, Si(CH3)2).
these substrates is in contrast to ortho-metalation reactions
reported for a variety of other substituted pyridines and metal
complexes. In addition, the products of the 1,3-alkyl-transfer
reaction to isoquinoline undergo subsequent transformations.
The reactions with polar (isoquinoline, ketones, azobenzene)
and nonpolar (olefins, alkynes) unsaturated substrates were
investigated and it was found that a hydrogen transfer occurred
from the dearomatized isoquinoline ring to polar unsaturated
substrates. Olefins and alkynes did not react and, instead, a
disproportionation reaction of the dearomatized, benzyl-
substituted isoquinoline complexes 3M-THF was observed.
Both types of reactions, the alkyl transfer to isoquinoline and
the hydrogen transfer from one isoquinoline to another, were
investigated computationally. It was found that the hydrogen
transfer occurred through a concerted mechanism akin to that
proposed for the Meerwein-Ponndorf-Verley reduction of
ketones by aluminum alkoxide complexes. The reactions
reported herein add to and contrast the list of reactions observed
for group 3 alkyl complexes with aromatic N-heterocycles, a
statement of the rich reactivity behavior of these metal centers
when they are supported by a ferrocene diamide ligand.
Experimental Section
All experiments were performed under a dry nitrogen atmosphere
using standard Schlenk techniques or an MBraun inert-gas glove-
box. Solvents were purified using a two-column solid-state purifica-
tion system by the method of Grubbs114 and transferred to the
glovebox without exposure to air. NMR solvents were obtained
from Cambridge Isotope Laboratories, degassed, and stored over
activated molecular sieves prior to use. Scandium, yttrium, and
lutetium oxides were purchased from Stanford Materials Corpora-
tion, 4 Meadowpoint, Aliso Viejo, CA 92656, and used as received.
LaBr3, K(CH2Xy-3,5), KCH2Ph, La(CH2C6H5)3(THF)3,115 1Sc-
THF, 1Y-THF, and 1Lu-THF were prepared following pub-
lished procedures.51,56 The aromatic heterocycles were distilled or
recrystallized before use; deuterium-labeled isoquinoline was
synthesized by modifying a literature procedure.116 All other
Synthesis of 2Sc. 1Sc-THF (200 mg, 0.295 mmol) was dissolved
in toluene (10 mL) and 1.3 equiv of 2,2′-bipyridine (60 mg, 0.383
mmol) was added as a solid. The reaction mixture turned dark
brown immediately and was allowed to stir for 1 h at room
temperature. The solvent was removed under vacuum and the
resulting brown solid was washed with hexanes, extracted with
toluene, and filtered through Celite, and the solvent was removed
under vacuum. Yield: 198.8 mg, 80.8% as a solid obtained after
toluene removal. 1H NMR (500 MHz, C6D6): δ, ppm: 8.52 (d, 1H,
bipy 6′-H), 7.40 (d, 1H, bipy 3′-H), 7.06 (s, 2H, o-C6H3), 6.90 (t,
1H, bipy 4′-H), 6.83 (s, 1H, p-C6H3), 6.61 (d, 1H, bipy 6-H), 6.46
(t, 1H, bipy 5′-H), 5.54 (dd, 1H, bipy 3-H), 4.77 (dt, 1H, bipy 5-H),
4.25-4.22 (m, 2H, fc-CH), 4.08 (m, 2H, fc-CH), 4.08 (m, 1H, bipy
4-H), 4.06 (br s, 4H, OCH2CH2), 3.53 and 3.52 (s, 1H, fc-CH),
3.14 (s, 2H, fc-CH), 3.10 (t, 2H, CH2-Ar), 2.23 (s, 6H, C6H3-
(CH3)2), 1.29 (br s, 4H, OCH2CH2), 0.93 and 0.90 (s, 9H,
SiC(CH3)3), 0.41, 0.32, -0.24, and -0.26 (s, 3H, Si(CH3)2).13C
NMR (126 MHz, C6D6): δ, ppm: 156.3, 143.7, 142.9, 138.4, 135.9,
135.7, 135.3, 133.8, 127.4, 126.4, 123.7, 118.6, 117.6, 101.9, 99.2,
99.1, 97.1, 70.8, 66.7, 66.6, 66.0, 65.9, 64.3, 64.2, 63.7, 63.6, 47.9,
36.1, 26.0, 25.9, 22.9, 19.6, 19.5, 18.6, 18.5, -4.7, -4.8, -5.1,
-5.3. Anal. (%) for C45H65FeN4OScSi2 Calcd: C, 67.36; H, 7.93;
N, 6.04; Found: C, 67.40; H, 8.31; N, 6.16.
1
materials were used as received. H NMR spectra were recorded
on Bruker300, Bruker500, or Bruker600 spectrometers (work
supported by the NSF grants CHE-9974928 and CHE-0116853) at
room temperature in C6D6 unless otherwise specified. Chemical
shifts are reported with respect to internal solvent, 7.16 ppm (C6D6).
CHN analyses were performed by UC Berkeley Micro-Mass facility,
8 Lewis Hall, College of Chemistry, University of California,
Berkeley, CA 94720 and by Midwest Microlab, LLC, 7212 N.
Shadeland Avenue, Suite 110, Indianapolis, IN 46250.
Synthesis of La(CH2-3,5-Me2C6H3)3(THF)3, La(CH2Ar)3-
(THF)3. A slurry of LaBr3(THF)3 (1.5 g, 2.52 mmol) in 1: 1 THF:
hexanes (50 mL) was cooled to -78 °C and 2.4 equiv of KCH2Ar
(0.958 g, 6.05 mmol) was added as a solid. The reaction mixture
was warmed to 0 °C and stirred for 3 h. The pale yellow slurry
was filtered through Celite and the solvent was removed under
vacuum. The resulting reddish-orange, oily solid was dissolved in
THF and layered with pentane. Yield: 0.676 g, 38% as yellow
1
crystals from THF: n-pentane. H NMR (500 MHz, C4D8O): δ,
ppm: 5.96 (s, 3H, p-C6H3), 5.74 (s, 6H, o-C6H3), 3.54 (br s, 4H,
OCH2CH2), 2.02 (s, 18H, C6H3-(CH3)2), 1.69 (br s, 4H, OCH2CH2),
1.47 (s, 6H, CH2). 13C NMR (126 MHz, C4D8O): δ, ppm: 151.8,
139.9, 129.1, 119.6, 118.1, 68.2, 26.4, 22.1. Anal. (%) for
C39H57LaO3 Calcd: C, 65.71; H, 8.06; N, 0; Found: C, 65.37; H,
8.14; N, <0.2.
Synthesis of 2Lu. 1Lu-THF (107.4 mg, 0.133 mmol) and 1 equiv
of 2,2′-bipy (20.7 mg, 0.133 mmol) were dissolved in toluene (6
mL). The dark-brown solution was heated and stirred in a Schlenk
flask for 2 d at 85 °C. The solvent was removed by vacuum. The
dark-brown solid was then taken up in hexanes and filtered through
a fiber-glass filter. The solution was concentrated under vacuum
and stored at -35 °C to afford brown crystals. A second crop of
crystals was also obtained from the mother liquor in the same
manner. Yield: 104.2 mg, 81.3%. 1H NMR (600 MHz, C6D6): 8.43
(114) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518–1520.
(115) Bambirra, S.; Meetsma, A.; Hessen, B. Organometallics 2006, 25,
3454–3462.
(116) Calf, G. E.; Garnett, J. L. J. Phys. Chem. 1964, 68, 3887–3889.
9
350 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
(d, 1H, bipy 6′-H), 7.43 (d, 1H, bipy 3′-H), 7.05 (s, 2H, o-C6H3),
6.90 (t, 1H, bipy 4′-H), 6.83 (s, 1H, p-C6H3) 6.53 (d, 1H, bipy
6-H), 6.51 (t, 1H, bipy 5′-H), 5.54 (dd, 1H, bipy 3-H), 4.80 (dt,
1H, bipy 5-H), 4.17 (br m, 2H, fc-H), 4.12 (br m, 1H, bipy 4-H),
4.03 (br s, 4H, OCH2CH2), 4.02 (br s, 2H, fc-H) 3.56 (br m, 2H,
fc-H) 3.12 (br m, 2H, fc-H), 3.12 (br m, 2H, CH2Ar), 2.23 (s, 6H,
C6H3-(CH3)2), 1.27 (br m, 4H, OCH2CH2), 0.94 and 0.91 (s, 9H
each, SiC(CH3)3), 0.31, 0.21, -0.08, -0.11 (s, 3H each, Si(CH3)2).
13C NMR (151 MHz, C6D6): δ, ppm: 159.3, 145.6, 145.1, 140.3,
137.61, 137.58, 135.7, 120.7, 120.2, 104.6, 104.3, 104.2, 99.1, 72.4,
67.8, 67.7, 67.01, 66.94, 65.7, 65.6, 65.4, 65.3, 49.9, 38.1, 35.0,
34.9, 29.4, 27.8, 27.7, 25.6, 25.1, 21.5, 20.9, 20.64, 20.58, -2.5,
-2.6, -3.3, -3.4, -4.0.
1H NMR (500 MHz, C6D6): δ, ppm: 7.62 (d, 1H, a), 7.17 (br
d, 4H, c, d, e, f), 7.08 (t, 2H, k), 7.01 (s, 2H, j), 6.96 (t, 1H, l),
5.88 (d, 1H, b), 5.20 (br s, 1H, g), 3.93 and 3.87 (br d, 4H,
fc-H), 3.59 (br s, 4H, OCH2CH2), 3.32 (s, 4H, fc-H), 3.18 (br d,
1H, h), 2.92 (br d, 1H, i), 1.09 (br s, 4H, OCH2CH2), 1.02 (s,
18H, SiC(CH3)3), 0.17 and 0.12 (d, 12H, Si(CH3)2). 13C NMR
(126 MHz, C6D6): δ, ppm: 140.3, 140.2, 134.9, 130.4, 130.2,
128.1, 127.0, 126.8, 126.4, 125.5, 125.2, 122.8, 121.4, 108.2,
97.2, 70.6, 67.0, 61.9, 40.7, 27.7, 25.2, 20.5, -2.0, -2.2. Anal.
(%) for C42H60N3OFeYSi2 Calcd: C, 61.23; H, 7.34; N, 5.10.
Found: C, 61.61; H, 7.54; N, 5.27.
Synthesis of 3Lu-THF. In a typical reaction, 1Lu-THF (177.9
mg, 0.220 mmol) was combined with 1 equiv of isoquinoline
(28.4 mg, 0.220 mmol) in a vial and dissolved in about 5 mL of
toluene. The solution became transparent red immediately. The
reaction mixture was stirred for about 24 h at room temperature,
after which it took on an orange hue. The solution was then
filtered through Celite and dried under vacuum. The resulting
orange solid was taken up in n-pentane. The n-pentane solution
was then concentrated under vacuum and stored at -35 °C to
afford yellow crystals. The mother liquor was removed, further
concentrated under vacuum, and stored at -35 °C to afford a
second crop of crystals. Total yield: 206.3 mg, 69.8%. Average
yield for this reaction, over the course of 6 separate reactions,
Synthesis of 1Sc-iqn. 1Sc-THF (300 mg, 0.442 mmol) was
dissolved in toluene (10 mL) and combined with a solution of 1 equiv
of isoquinoline (57 mg, 0.442 mmol) in toluene (1 mL). The reaction
mixture immediately turned orange and was stirred for 30 min at room
temperature. The solvent was removed by vacuum and the resulting
orange solid was washed with hexanes, extracted with toluene, filtered
through Celite, and the volatiles were removed under vacuum. Yield:
203.3 mg, 62.5% as a solid obtained after toluene removal. 1H NMR
(500 MHz, C6D6): δ, ppm: 8.92 (s, 1H, iqn-CH), 8.22 (s, 1H, iqn-
CH), 7.52 (d, 1H, iqn-CH), 7.23 (s, 2H, o-C6H3), 7.13-7.10 (m, 2H,
iqn-CH), 7.05 (t, 1H, iqn-CH), 6.62 (s, 1H, p-C6H3), 4.22 (s, 2H fc-
CH), 3.88 (s, 2H fc-CH), 3.56 (s, 4H, fc-CH), 3.01 (s, 2H, CH2-Ar),
2.37 (s, 6H, C6H3-(CH3)2), 0.93 (s, 18H, SiC(CH3)3), -0.05 and -0.13
(s, 6H, Si(CH3)2). 13C NMR (126 MHz, C6D6): δ, ppm: 153.9, 150.5,
140.3, 138.2, 136.7, 133.0, 128.9, 128.7, 126.7, 124.7, 121.6, 121.5,
106.0, 70.3, 68.5, 66.5, 55.1, 27.8, 21.9, 20.5, -2.4, -3.1. Anal. (%)
for C40H56FeN3ScSi2 Calcd: C, 65.49; H, 7.67; N, 5.71; Found: C,
65.48; H, 7.59; N, 5.65.
1
was 61.2%. H NMR (500 MHz, C6D6): δ, ppm: 7.58 (d, 1H,
iqn 3-H), 7.23, 7.21, 7.20, 7.14, 7.11, 7.10, 7.08, 7.07, 7.05,
7.02, 7.00 (m, 4H, iqn 5-H, 6-H, 7-H and 8-H), 6.87 (s, 2H,
xylyl-CH), 6.67 (s, 1H, xylyl-CH), 5.98 (d, 1H, iqn 4-H), 5.20
(dd, 1H, iqn 1-H), 3.87 (br s, 4H, fc-H), 3.64 (br s, 4H,
OCH2CH2), 3.42 (s, 2H, fc-H), 3.38 (dd, 1H, CH2), 2.83 (dd,
1H, CH2), 2.19 (s, 6H, xylyl-CH3), 1.10 (br s, 4H, OCH2CH2),
1.04 (s, 18H, SiC(CH3)3), 0.18 and 0.12 (s, 6H each, Si(CH3)2).
13C NMR (126 MHz, C6D6): δ, ppm: 141.3, 140.4, 137.1, 135.2,
131.6, 128.9, 127.5, 126.8, 125.9, 123.4, and 122.0 (aromatic C
and Cp C-N), 108.9 (iqn 3-C), 98.8 (iqn 4-C), 71.3 (Cp C-H),
67.5 (OCH2CH2), 62.6 (Cp C-H), 40.2, 27.8, 25.3, 21.5, and
20.6 (methylene CH2, isoquinoline 1-C, SiC(CH3)3, Si(CH3)2,
and C6H3(CH3)2), -2.2 (OCH2CH2 and SiC(CH3)3). Anal. (%)
for C44H64N3OSi2FeLu Calcd: C, 56.34; H, 6.88; N, 4.48. Found:
C, 56.41, H, 6.91, N, 4.57.
Synthesis of 3Sc-THF. 1Sc-THF (200 mg, 0.295 mmol) was
dissolved in C6D6 (1.5 mL) and combined with 1 equiv of
isoquinoline (38 mg, 0.295 mmol) and a drop (2.5 mg) of THF.
The reaction mixture immediately turned orange and was heated
to 50 °C for 36 h. The solvent was removed under vacuum, hexanes
was added to the resulting orange solid and filtered through Celite.
The solution was concentrated and placed in a -35 °C freezer.
The product was collected as crystals by decanting the mother
liquor. Yield: 155 mg, 72.9%. 1H NMR (300 MHz, C6D6): δ, ppm:
7.87 (d, 1H, iqn C3-H), 7.17 (m, 2H, iqn-CH), 7.01-6.93 (m, 2H,
iqn-CH), 6.80 (s, 2H, o-C6H3), 6.89 (s, 1H, p-C6H3), 5.87 (d, 1H,
iqn C4-H), 5.47 (t, 1H, iqn C1-H), 4.22 (s, 1H, fc-CH), 4.11 (s,
1H, fc-CH), 3.88-3.60 (m, 7H, fc-CH and OCH2CH2), 3.52, 3.37,
and 3.28 (s, 1H, fc-CH), 3.12 (dd, 2H, iqn-CH2), 2.18 (s, 6H,
C6H3CH3), 1.16 (br s, 4H, OCH2CH2), 1.05 and 1.02 (s, 9H,
SiC(CH3)3), 0.27, 0.23, 0.16, and 0.12 (s, 3H, SiCH3). Anal. (%)
for C44H64FeN3OScSi2 Calcd: C, 65.41; H, 7.98; N, 5.20; Found:
C, 64.58; H, 7.84; N, 5.43.
Synthesis of 3La-THF. A toluene (8 mL) solution of 1La-THF
(300 mg, 0.389 mmol) was combined with a toluene (1 mL)
solution of isoquinoline (5 mg, 0.389 mmol) and the reaction
mixture was stirred for 15 min at room temperature. The solvent
was removed under vacuum, the resulting red solid was washed
with hexanes, extracted in toluene, and filtered through Celite.
The solution was concentrated, layered with n-pentane, and
placed in a -35 °C freezer. The product was collected as crystals
1
by decanting the mother liquor. Yield: 263 mg, 75%. H NMR
Synthesis of 3Y-THF. 1Y-THF (229.0 mg, 0.335 mmol) and
isoquinoline (43.5 mg, 0.337 mmol) were combined in a capped
20-mL scintillation vial and stirred for 2 h in toluene at room
temperature. The solvent was removed under vacuum, the resulting
crude mixture was dissolved in hexanes and left at -35 °C overnight
to precipitate 3Y-THF. The yield was 74.9% (206.8 mg, 0.251
mmol) of a yellow-orange powder.
(500 MHz, C6D6): δ, ppm: 7.43 (d, 1H, iqn 3-H), 7.17 (m, 2H,
iqn-CH), 6.88 (t, 1H, iqn-CH), 6.77 (s, 2H, o-C6H5), 6.72 (s,
1H, p-C6H5), 6.70 (d, 1H, iqn-CH), 5.79 (d, 1H, iqn 4-H), 5.21
(dd, 1H, iqn 1-H) 4.06 (s, 4H, fc-CH), 3.70, (br s, 8H,
OCH2CH2), 3.51 (dd, 1H, CH2Ph), 3.14 (s, 4H, fc-CH), 2.88
(dd, 1H, CH2Ph), 2.20 (s, 6H, C6H3(CH3)2), 1.28 (br s, 8H,
OCH2CH2), 1.04 (s, 18H, SiC(CH3)3), 0.33 and 0.27 (s, 6H,
Si(CH3)2). 13C NMR (126 MHz, C6D6): δ, ppm: 140.7, 140.6,
137.1, 136.6, 128.7, 127.6, 127.0, 126.6, 121.3, 120.8, 105.4,
93.6, 70.4, 66.4, 66.1, 66.0, 63.5, 40.6, 27.7, 25.3, 21.5, 20.5,
-2.3, -2.4. Anal. (%) for C44H64FeLaN3OSi2: Calcd.: C, 58.59;
H, 7.15; N, 4.66; Found: C, 58.81; H, 6.84; N, 4.42.
Synthesis of 3Y-iqnMe. 1Y-THF (117.2 mg, 0.172 mmol) and
3 equiv of 3-methylisoquinoline (75.5 mg, 0.527 mmol) were
combined in toluene in a capped 20-mL scintillation vial and
stirred for 2 h, at room temperature. The solvent was removed
under vacuum and the resulting crude mixture was dissolved in
hexanes and left at -35 °C overnight, precipitating the desired
product. The yield was 60.5% (94.6 mg, 0.104 mmol) of a red
powder.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 351

A R T I C L E S
Miller et al.
1H NMR (500 MHz, C6D6): δ, ppm: 8.88 (s, 1H, n), 7.37 (d,
1H, s), 7.33, 7.26, 7.20, 7.10, 7.02, 6.98, 6.85 (m, 13H, c, d, e,
f, j, k, l, o, p, q, r), 6.01 (s, 1H, b), 5.57 (dd, 1H, g), 4.25 (s, 1H,
fc-CH), 4.11 (s, 1H, fc-CH), 3.85 (s, 1H, fc-CH), 3.75 (s, 1H,
fc-CH), 3.68 (s, 1H, fc-CH), 3.63 (s, 1H, fc-CH), 3.60 (s, 1H,
fc-CH), 3.54 (m, 1H, fc-CH), 2.81 (overlapped s, 2H, h, i), 2.72
(s, 3H, m), 2.71 (s, 3H, a), 0.94 (s, 9H, SiC(CH3)3), 0.92 (s, 9H,
SiC(CH3)3), 0.11 (s, 3H, SiCH3), -0.20 (s, 3H, SiCH3), -0.41
(s, 3H, SiCH3), -0.59 (s, 3H, SiCH3).13C NMR (126 MHz,
C6D6): δ, ppm: 155.0, 149.5, 148.8, 140.1, 136.8, 135.8, 132.6,
130.9, 130.4, 129.1, 126.8, 126.4, 125.6, 125.3, 125.0, 122.7,
121.0, 120.9, 110.4, 109.4, 96.5, 69.7, 68.9, 68.6, 68.2, 66.4,
66.2, 65.7, 62.9, 61.0, 40.2, 34.5, 28.9, 27.7, 27.5, 26.8, 25.2,
20.6, 20.3, 11.2, -1.2, -2.8, -3.8, -4.1. Anal. (%) for
C49H63N4FeYSi2 Calcd: C, 64.75; H, 6.99; N, 6.16. Found: C,
64.46; H, 7.02; N, 6.17.
1H NMR (300 MHz, C6D6): δ, ppm: 10.02 (s, 2H, d), 9.10 and
9.08 (d, 2H, e), 7.70 and 7.68 (d, 2H, f), 7.27 through 7.19, 7.14
through 7.01, and 6.86 through 6.81 (m, 11H, g), 6.45 and 6.42 (s,
2H, b), 5.71 and 5.68, (d, 1H, c), 4.56 (s, 2H, a), 3.93 (m, 4H,
fc-H), 3.23 (br s, 4H, fc-H), 1.03 (s, 18H, SiC(CH3)3), 0.37 (s, 12H,
Si(CH3)2). 13C NMR (126 MHz, C6D6): δ, ppm: 155.1, 148.6, 142.9,
138.5, 136.7, 132.4, 128.8, 126.90, 126.87, 126.7, 124.8, 122.2,
121.7, 120.3, 104.4, 96.3, 67.4, 66.7, 53.7, 28.2, 21.2, -1.6. Anal.
(%) for C49H60N5Si2FeLu Calcd: C, 58.50; H, 6.01; N, 6.96. Found:
C, 57.76, H, 5.51, N, 6.80.
Characterization of 5Ar.
Synthesis of 4Y-(iqn)2. 3Y-THF (95.1 mg, 0.115 mmol) and 3
equiv of isoquinoline (44.8 mg, 0.347 mmol) were combined in a
Schlenk tube and stirred in toluene for 6 h at 50 °C. The solvent
was removed under vacuum and the resulting crude mixture was
dissolved in hexanes and left at -35 °C overnight, precipitating
the desired product. The yield was 72.0% (76.5 mg, 0.0831 mmol)
of blood-red crystals and powder.
1H NMR (300 MHz, C6D6): δ, ppm: 8.58 and 8.56 (d, 1H, a),
8.04 and 8.02 (d, 1H, f), 7.37 and 7.34 (d, 1H, b), 7.08 (m, 3H, c,
d and e), 6.99 (s, 2H, h), 6.64, (s, 1H, j), 4.64 (s, 2H, g), 2.02 (s,
6H, i).
Synthesis of 6Lu-(iqn)2. 3Lu-THF (133.0 mg, 0.142 mmol) and
2 equiv of 2-adamantanone (42.6 mg, 0.284 mmol) were combined
in about 5 mL of toluene. The orange solution was stirred at room
temperature for 3.5 h, after which it was dried under vacuum. The
reaction was triturated once with hexanes, resulting in a foamy,
orange oil, which weighed 175.9 mg. All attempts to crystallize
1
the oil failed. H NMR spectroscopy showed the product to be
6Lu-iqnAr, for which a molecular weight was calculated to be
1016.093 g/mol. Based upon this observation, 2 equiv of isoquino-
line was added (44.7 mg, 0.346 mmol) and the reactants were
dissolved in ∼3 mL of toluene. The solution was stirred for 2 h at
room temperature, after which it was filtered through a fiberglass
filter. The vial was washed with hexanes and the hexanes wash
was also filtered through the fiberglass filter. The solvent was then
removed under vacuum and the solid was taken up in n-pentane
and stored at -35 °C, affording an orange powder. Total yield:
56.7 mg, 38.9%. 1H NMR (500 MHz, C6D6): δ, ppm: 10.25 (br s,
2H, iqn 1-H), 9.24 (d, 2H, iqn 3-H), 7.86 and 7.84 (d, 2H, iqn
8-H) 7.30 and 7.29 (d, 2H, iqn 5-H), 7.26 and 7.24 (d, 2H, iqn
4-H), 7.20 (t, 2H, iqn 6-H) 7.11 (t, 2H, iqn 7-H) 4.55 (s, 1H,
O-CH), 4.05 and 3.53 (s, 4H each, fc-H), 2.32 and 2.30 (d, 2H,
ad-H), 2.02 (s, 2H, ad-H), 1.85 through 1.67 (m, 8H, ad-H), 1.41
and 1.39 (d, 2H, ad-H), 1.12 (s, 18H, SiC(CH3)3), 0.22 (s, 12H,
Si(CH3)2). 13C NMR (126 MHz, C6D6): δ, ppm: 154.6, 142.7, 136.5,
132.1, 126.8, 121.6, 105.3, 80.8, 67.1, 65.6, 38.6, 38.0, 37.7, 32.1,
28.7, 28.2, 28.1, 21.1, -2.0. Anal. (%) for C50H67N4OSi2FeLu
Calcd: C, 58.47; H, 6.58; N, 5.46. Found: C, 58.42, H, 6.45, N,
5.51.
1H NMR (500 MHz, C6D6): δ, ppm: 9.90 (s, 2H, i), 8.98 (s,
2H, o), 7.65 (d, 2H, n), 7.32, 7.23 through 7.18, 7.14 through
7.01, and 6.85 through 6.82 (m, 12H, c, d, e, f, j, k, l, m), 6.50
(d, 1H, a), 5.69 (d, 1H, b), 4.66 (s, 2H, g, h), 3.97 (s, 4H, fc-H),
3.20 (s, 4H, fc-H), 1.03 (s, 18H, SiC(CH3)3), 0.37 (s, 12H,
Si(CH3)2). 13C NMR (126 MHz, C6D6): δ, ppm: 154.7, 148.2,
142.5, 138.7, 136.6, 132.2, 128.7, 126.9, 126.7, 124.7, 122.0,
121.7, 120.3, 105.2, 95.4, 67.2, 66.7, 53.4, 28.1, 20.9, -1.8.
Anal. (%) for C49H60N5FeYSi2 Calcd: C, 63.97; H, 6.57; N, 7.61.
Found: C, 63.67; H, 6.28; N, 7.46.
Synthesis of 4Lu-(iqn)2. 3Lu-THF (101.6 mg, 0.126 mmol)
and 4 equiv of isoquinoline (65.1 mg, 0.502 mmol) were
combined in a Schlenk tube with 5 mL of toluene. The solution
immediately turned red. A stir bar was added and the solution
was stirred at 50 °C for 30 h. The solution was next filtered
through a fiber-glass filter and dried under vacuum, resulting in
a red oil. The oil was triturated once with hexanes. The solid
was then taken up in hexanes and concentrated under vacuum.
The solution was stored at -35 °C, affording analytically pure
red crystals. A second crop of crystals was also obtained after
further concentration of the mother liquor. Total yield: 105.2
mg, 83.2%.
Synthesis of 7Sc-iqnAr. 3Sc-THF (280 mg, 0.380 mmol) was
combined with 1.05 equiv of benzophenone (72.8 mg, 0.400 mmol)
9
352 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
in toluene (10 mL) and stirred for 1 h at room temperature. The
solvent was removed by vacuum and the resulting yellow solid
was washed with hexanes (1 mL), extracted with toluene and filtered
through Celite. Yield: 349 mg, 99%, as a solid obtained after the
removal of the volatiles. 1H NMR (500 MHz, C6D6): δ, ppm: 8.95
(s, 1H, iqn-CH), 8.21 (d, 1H, iqn-CH), 7.96 (d, 4H, o-C6H5), 7.48
(d, 1H, iqn-CH), 7.32 (t, 4H, m-C6H5), 7.10 (t, 2H, p-C6H5), 7.05
(t, 1H, iqn-CH), 6.85 (s, 2H, o-C6H5), 6.70 (s, 1H, p-C6H5), 4.26
(br s, 2H, fc-CH), 4.18 (s, 2H, CH2-Ar), 3.82 (br s, 4H, fc-CH),
3.75 (br s, 2H, fc-CH), 2.09 (s, 6H, C6H3-(CH3)2), 0.93 (s, 18H,
SiC(CH3)3), -0.18 (s, br, 12H, Si(CH3)2). The limited solubility
of 7Sc-iqnAr prevented the collection of its 13C NMR spectrum.
Anal. (%) for C53H66FeN3OScSi2 Calcd: C, 69.34; H, 7.25; N, 4.58;
Found: C, 69.10; H, 7.18; N, 4.58.
dissolved in n-pentane and left at -35 °C overnight, precipitating
the desired product. The yield was 36.0% (73.9 mg, 0.0791 mmol)
of a brown powder.
Synthesis of 7Y-(iqn)2. 3Y-THF (51.0 mg, 0.0619 mmol) and
benzophenone (11.6 mg, 0.0637 mmol) were combined in a capped
20-mL scintillation vial and stirred in toluene (3 mL) for 5 min at
room temperature. A toluene solution (2 mL) with 2 equiv of
isoquinoline (16.0 mg, 0.124 mmol) was added and the mixture
was stirred for an additional 30 min at room temperature. The
solvent was removed under vacuum and the resulting crude mixture
was dissolved in hexanes and left at -35 °C overnight, precipitating
the desired product. The yield was 62.3% (37.5 mg, 0.0386 mmol)
of a pale-yellow powder.
1H NMR (300 MHz, C6D6, includes 1 equiv free toluene): δ,
ppm: 9.11 (br s, 1H, a), 7.72 (d, 1H, b), 7.33 - 7.37, 7.21, 7.11,
7.07, 7.03, 6.93, 6.88, 6.85 (m, 24H, Ar, toluene, c, d, e, f, j, k,
l), 6.22 (d, 1H, i), 5.04 (s, 2H, g, h), 4.15 (br s, 1H, fc-CH),
4.08 (br s, 1H, fc-CH), 3.97 (br s, 1H, fc-CH), 3.74 (br s, 1H,
fc-CH), 3.66 (br s, 1H, fc-CH), 3.40 (br s, 1H, fc-CH), 3.16 (br
s, 1H, fc-CH), 2.99 (br s, 1H, fc-CH), 2.11 (s, 3H, toluene),
0.88 (s, 18H, SiC(CH3)3), 0.13 (s, 6H, Si(CH3)2), -0.03 (s, 6H,
Si(CH3)2). Note: Because of its solubility properties, 8Y could
not be obtained pure enough for an elemental analysis study.
As a result, 8Y was transformed into 8Y-(iqn)2, which was fully
characterized (see below).
Synthesis of 8Y-(iqn)2. 8Y (73.9 mg, 0.0791 mmol) and 2 equiv
of isoquinoline (20.5 mg, 0.159 mmol) were combined in a capped
20-mL scintillation vial and stirred for 2 h in toluene at room
temperature. The solvent was removed under vacuum and the
resulting crude mixture was dissolved in n-pentane and left at -35
°C overnight to precipitate the product. The yield was 55.2% (42.5
mg, 0.0437 mmol) of a beige powder.
1H NMR (500 MHz, C6D6): δ, ppm: 9.78 (s, 2H, h), 8.92 (s,
2H, b), 7.57, 7.22, 7.08, 6.95 (m, 16H, Ar, d, e, f, g), 6.84 (br s,
2H, c), 6.64 (s, 1H, a), 4.07 (s, 4H, fc-CH), 3.42 (s, 4H, fc-CH),
1.10 (s, 18H, SiC(CH3)3), 0.33 (s, 12H, Si(CH3)2). 13C NMR (126
MHz, C6D6): δ, ppm: 154.2, 149.6, 141.7, 136.0, 131.3, 128.2,
126.9, 126.1, 125.4, 120.8, 105.7, 82.7, 66.3, 65.4, 27.5, 20.4, -2.4.
Anal. (%) for C53H63N4OFeYSi2 Calcd: C, 65.42; H, 6.53; N, 5.76.
Found: C, 65.36; H, 6.32; N, 5.74.
Synthesis of 7La-(iqn)2. A toluene (10 mL) solution of
3La-THF (263 mg, 0.291 mmol) was added to a toluene (1 mL)
solution of 1.1 equiv of benzophenone (58 mg, 0.321 mmol) and
the reaction mixture was stirred for 15 min at room temperature.
The solvent was removed by vacuum and the resulting yellow solid
was extracted with hexanes and filtered through Celite. After an
unsuccessful attempt at crystallization as the THF solvate from
hexanes, 2 equiv of isoquinoline (75 mg, 0.583 mmol) was added
and the isoquinoline solvate was crystallized from hexanes. Yield:
1H NMR (500 MHz, C6D6): δ, ppm: 9.60 (br s, 2H, g), 8.88
(br s, 2H, a), 7.56, 7.21, 7.03, 6.89, 6.66, 6.58, 6.52, 6.21 (m,
20H, Ar, b, c, d, e, f), 6.27 (s, 1H, h), 4.12 (s, 2H, fc-CH), 3.87
(s, 2H, fc-CH), 3.60 (s, 2H, fc-CH), 2.90 (s, 2H, fc-CH), 1.12
(s, 18H, SiC(CH3)3), 0.52 (s, 6H, Si(CH3)2), 0.35 (s, 6H,
Si(CH3)2). 13C NMR (126 MHz, C6D6): δ, ppm: 154.7, 148.8,
135.8, 131.1, 128.8, 126.0, 121.0, 115.9, 114.9, 113.4, 105.4,
67.3, 67.0, 66.7, 65.4, 64.2, 61.3, 27.8, 26.2, 20.6, -1.5, -2.3.
Anal. (%) for C52H63N6FeYSi2: Calcd.: C, 64.19; H, 6.53; N,
8.64. Found: C, 63.45; H, 6.15; N, 8.31.
1
168 mg, 56%. H NMR (500 MHz, C6D6): δ, ppm: 9.53 (s, 2H,
iqn-CH), 8.71, (d, 2H, iqn-CH), 7.70 (d, 4H, C6H5 or iqn-CH), 7.50
(d, 2H, C6H5 or iqn-CH), 7.22 (d, 2H, C6H5 or iqn-CH), 7.16 (t,
1H, C6H5 or iqn-CH), 7.09-7.05 (m, 8H, C6H5 or iqn-CH), 6.97
(t, 2H, C6H5 or iqn-CH), 6.71 (s, 1H, OCH), 4.13 (s, 4H, fc-CH),
3.40 (s, 4H, fc-CH), 1.03 (s, 18H, SiC(CH3)3), 0.39 (s, 12H,
Si(CH3)2).13C NMR (126 MHz, C6D6): δ, ppm: 153.8, 151.0, 141.8,
136.5, 131.7, 129.3, 128.6, 128.5, 127.5, 126.6, 126.1, 121.5, 107.4,
84.3, 65.7, 65.6, 27.6, 20.7, -2.5. Anal. (%) for C53H63FeLaN4OSi2
Calcd: C, 62.22; H, 6.21; N, 5.48; Found: C, 61.85; H, 6.25; N,
5.18.
Disproportionation Reaction of 3Y-THF. 3Y-THF (127.8
mg, 0.154 mmol) was dissolved in benzene in a Schlenk tube
and heated without stirring for 3 d at 70 °C. The solution was
then moved to a 20-mL scintillation vial, the solvent was
removed, and a minimal amount of n-pentane was added to the
crude mixture. The vial was left at -78 °C for one hour and the
resulting liquid and solid were separated, dissolved in minimal
n-pentane and left overnight at -35 °C, resulting in a orange
powder (4Y-(THF)2) and a red compound, 9Y, that could not
be fully separated before its decomposition. The yield for
4Y-(THF)2was 94.4% (58.7 mg, 0.0728 mmol).
Synthesis of 8Y. 3Y-THF (171.9 mg, 0.209 mmol) and
azobenzene (40.0 mg, 0.220 mmol) were combined in a Schlenk
tube and stirred in toluene for 24 h at 50 °C. The solvent was
removed under vacuum and the resulting crude mixture was
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 353

A R T I C L E S
Miller et al.
outside the glovebox. The X-ray data collections were carried out
on a Bruker AXS single crystal X-ray diffractometer using Mo KR
radiation and a SMART APEX CCD detector. The data was reduced
by SAINTPLUS and an empirical absorption correction was applied
using the package SADABS. The structures were solved and refined
using SHELXTL (Bruker 1998, SMART, SAINT, XPREP AND
SHELXTL, Bruker AXS Inc., Madison, Wisconsin, USA).117 All
atoms were refined anisotropically and hydrogen atoms were placed
in calculated positions unless specified otherwise. Tables with
atomic coordinates and equivalent isotropic displacement param-
eters, with all the bond lengths and angles, and with anisotropic
displacement parameters are listed in the cifs.
1H NMR (300 MHz, C6D6): δ, ppm: 7.25 (overlap m, 2H, a, c),
7.10 (overlap m, 3H, d, e, f), 5.76 (d, 1H, b), 4.75 (s, 2H, g), 4.03
(s, 4H, fc-CH), 3.86 (s, 8H, O(CH2CH2)2), 3.13 (s, 4H, fc-CH),
1.31 (s, 8H, OCH2CH2), 0.99 (s, 18H, SiC(CH3)3), 0.28 (s, 12H,
Si(CH3)2). Note: Because of its solubility properties, the complex
4Y-(THF)2 could not be obtained pure enough for an elemental
analysis study. As a result, 4Y-(THF)2 was transformed into
4Y-(iqn)2, which was fully characterized (see below).
X-Ray Crystal Structure of 1′La-THF. X-ray quality crystals
were obtained from a concentrated toluene: n-pentane solution
placed in a -35 °C freezer in the glovebox. A total of 33077
reflections (-25 e h e 25, -23 e k e 22, -15 e l e 14) were
collected at T ) 100(2) K with 2θmax ) 56.54°, of which 9094
were unique (Rint ) 0.0613). The residual peak and hole electron
density were 0.67 and -0.51 eA-3. The least-squares refinement
converged normally with residuals of R1 ) 0.0364 and GOF )
1.006. Crystal and refinement data for 1′La-THF: formula
C35H57N2Si2FeOLa, space group Pna21, a ) 19.018(3), b )
17.284(3), c ) 11.281(2), R ) ꢀ ) γ ) 90°, V ) 3708.1(11) Å3,
Z ) 4, µ ) 1.620 mm-1, F(000) ) 1600, R1 ) 0.0437 and wR2 )
0.0657 (based on all 9094 data, I > 2σ(I)).
Reaction of 4Y-(THF)2 with Isoquinoline to Form 4Y-
(iqn)2. 23.4 mg (0.0291 mmol) of 4Y-(THF)2and 2.1 equiv of
isoquinoline (7.9 mg, 0.0612 mmol) were dissolved and combined
in C6D6 in an NMR tube fitted with a Teflon stopper. The solution
was monitored and the reaction was determined to be complete
after 2 h. The solution was moved to a 20-mL scintillation vial
and the solvent was removed. A minimal amount of hexanes was
added to the crude reaction mixture and the resulting solution was
kept overnight at -35 °C, yielding a blood-red powder. Yield:
36.6% (9.8 mg, 0.0107 mmol).
X-Ray Crystal Structure of 2Sc. X-ray quality crystals were
obtained from a concentrated toluene: n-pentane solution placed
in a -35 °C freezer in the glovebox. A total of 22432 reflections
(-14 e h e 15, -19 e k e 19, -21 e l e 21) were collected at
T ) 100(2) K with 2θmax ) 56.52°, of which 12025 were unique
(Rint ) 0.0463). The residual peak and hole electron density were
1.85 and -0.69 eA-3. The least-squares refinement converged
normally with residuals of R1 ) 0.0581 and GOF ) 1.024. One
molecule of toluene solvent was found in the unit cell. Some of
the methyl groups were thermally disordered; the disorder was not
extensive and was not modeled. Crystal and refinement data for
2Sc: formula C52H73N4Si2FeScO, space group Pjı, a ) 11.4536(19),
b ) 14.490(2), c ) 16.233(3), R ) 97.870(2)°, ꢀ ) 98.337(2)°, γ
) 106.382(2)°, V ) 2512.0(7) Å3, Z ) 2, µ ) 0.511 mm-1, F(000)
) 992, R1 ) 0.0926 and wR2 ) 0.1530 (based on all 12025 data,
I > 2σ(I)).
Disproportionation Reaction of 3La-THF. 3La-THF was
prepared in situ by the reaction of 1La-THF (200 mg, 0.269 mmol)
with 1 equiv of isoquinoline (34.7 mg, 0.269 mmol) in C6D6 (1.5
mL). This reaction mixture was heated to 50 °C for 96 h. The
solvent was removed by vacuum; a red product, identified as 9La,
was extracted with hexanes, and filtered through Celite. An orange
product identified as 4La-(THF)2 was extracted with toluene and
filtered through Celite. Yield: 45 mg of 9La, 38.4%, as crystals from
n-pentane, and 30 mg of 4La-(THF)2, 26.1%, as crystals from
toluene: n-pentane. Crystals of 9La suitable for X-ray diffraction
were obtained from the analogous reaction of the 3,5-dimethyl-
benzyl complex from toluene: n-pentane. 4La-(THF)2was not
characterized because it could not be obtained in sufficient amount,
as a consequence of its solubility properties and its decomposition
to intractable products. 9La:1H NMR (500 MHz, C6D6): δ, ppm:
8.13 (d, 1H, iqn-CH), 7.94 (s, 1H, iqn-CH), 7.89 (s, 2H, iqn-CH),
7.31-7.20 (m, 4H, C6H5 or iqn-CH), 6.77 (t, 1H, C6H5 or iqn-
CH), 6.43 (s, 1H, C6H5 or iqn-CH), 6.02 (s, 1H, La-CH), 4.01 (s,
4H, fc-CH), 3.57 (s, 4H, OCH2CH2), 3.09 (s, 4H fc-CH), 1.28 (s,
4H, OCH2CH2), 0.96 (s, 18H, SiC(CH3)3), 0.01 (s, 12H, Si(CH3)2).
13C NMR (126 MHz, C6D6): δ, ppm: 154.2, 143.4, 140.6, 137.0,
132.3, 130.1, 126.5, 125.9, 125.1, 121.9, 121.2, 106.9, 79.8, 70.0,
66.5, 27.6, 25.4, 20.5, -2.8. Anal. (%) for C42H58FeN3LaOSi2
Calcd: C, 57.86; H, 6.70; N, 4.82; Found: C, 57.92; H, 6.69; N,
4.86.
X-Ray Crystal Structure of 3Y-iqnMe. X-ray quality crystals
were obtained from a concentrated hexanes solution placed in a
-35 °C freezer in the glovebox. A total of 41770 reflections (-18
e h e 19, -21 e k e 22, -26 e l e 26) were collected at T )
125(2) K with 2θmax ) 56.59°, of which 11372 were unique (Rint
) 0.0749). The residual peak and hole electron density were 0.53
and -0.43 eA-3. The least-squares refinement converged normally
with residuals of R1 ) 0.0469 and GOF ) 0.961. Crystal and
refinement data for 3Y-iqnMe: formula C49H63N4Si2FeY, space
group Pna21, a ) 14.3694(13), b ) 16.5282(14), c ) 19.6525(17),
R ) ꢀ ) γ ) 90°, V ) 4667.5(7) Å3, Z ) 4, µ ) 1.636 mm-1
,
F(000) ) 1912, R1 ) 0.0708 and wR2 ) 0.0979 (based on all 11372
data, I > 2σ(I)).
X-Ray Crystal Structure of 4Lu-(iqn)2. X-ray quality crystals
were obtained from a concentrated hexanes solution layered with
n-pentane and placed in a -35 °C freezer in the glovebox. A total
of 22108 reflections (-18 e h e 18, -18 e k e 18, -19 e l e
19) were collected at T ) 100(2) K with 2θmax ) 56.70°, of which
12008 were unique (Rint ) 0.0282). The residual peak and hole
electron density were 1.18 and -0.80 eA-3. The least-squares
refinement converged normally with residuals of R1 ) 0.0316 and
GOF ) 1.004. Some of the methyl groups were thermally
disordered; the disorder was not extensive and was not modeled.
Crystal and refinement data for 4Lu-(iqn)2: formula C52H64N5-
Si2FeLu, space group Pjı, a ) 13.908(2), b ) 13.942(5), c )
14.914(3), R ) 95.659(2), ꢀ ) 116.825(2), γ ) 98.466(2)°, V )
Disproportionation Reaction of Deuterium-Labeled 3La-
THF. 1La-THF (100 mg, 0.134 mmol) was combined with 1 equiv
of isoquinoline-d (17.5 mg, 0.134 mmol) in C6H6 (1.5 mL). A
portion of the reaction mixture (0.5 mL) was separated and the
solvent removed under vacuum. The resulting red-orange solid was
dissolved in C6D6 (0.75 mL) and placed in a J-Young NMR tube.
The remaining C6H6 solution was also placed in a J-Young tube.
1
The two samples were heated to 50 °C and monitored by H and
2H NMR spectroscopy, respectively. After 48 h a peak at 2.08 ppm
was observed by 2H NMR spectroscopy that was attributed to
deuterium incorporation in the liberated toluene.
X-Ray Crystal Structures. X-ray quality crystals were obtained
from various concentrated solutions placed in a -35 °C freezer in
the glovebox. Inside the glovebox, the crystals were coated with
oil (STP Oil Treatment) on a microscope slide, which was brought
(117) Sheldrick, G. Acta Cryst. A 2008, 64, 112–122.
9
354 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Reactions of N-Heterocycles by Group 3 Complexes
A R T I C L E S
2507.6(10) Å3, Z ) 2, µ ) 2.333 mm-1, F(000) ) 1072, R1 )
0.0408 and wR2 ) 0.0694 (based on all 12008 data, I > 2σ(I)).
X-Ray Crystal Structure of 4′Y-1,5-(py)2. X-ray quality
crystals were obtained from a concentrated n-pentane solution and
placed in a -35 °C freezer in the glovebox. A total of 23359
reflections (-36 e h e 36, -17 e k e 17, -19 e l e 19) were
collected at T ) 100(2) K with 2θmax ) 56.53°, of which 6404
were unique (Rint ) 0.0563). The residual peak and hole electron
density were 1.05 and -0.52 eA-3. The unit cell contains one
molecule of n-pentane or hexane, which was very disordered. A
good model could not be obtained for the solvent because of its
disorder and of the fact that some of the atoms sit in special
positions. After successive failed attempts to find a proper solvent
model, no carbon atoms were fitted to the electron density
corresponding to the solvent molecule. Some of the methyl groups
were thermally disordered; the disorder was not extensive and was
not modeled. The least-squares refinement converged normally with
residuals of R1 ) 0.0627 and GOF ) 1.196. Crystal and refinement
data for 4′Y-1,5-(py)2: formula C37H54N5Si2FeY, space group C2/
c, a ) 27.282(3), b ) 13.0297(15), c ) 14.8077(17), ꢀ )
99.407(1)°, V ) 5192.9(10) Å3, Z ) 4, µ ) 1.461 mm-1, F(000)
) 1616, R1 ) 0.0927 and wR2 ) 0.2318 (based on all 6404 data,
I > 2σ(I)).
(-54 e h e 53, -11 e k e 11, -23 e l e 24) were collected at
T ) 100(2) K with 2θmax ) 50.07°, of which 7562 were unique
(Rint ) 0.1188). The residual peak and hole electron density were
0.50 and -0.52 eA-3. The least-squares refinement converged
normally with residuals of R1 ) 0.0497 and GOF ) 0.991. Crystal
and refinement data for 9La: formula C44H62N3Si2FeOLa, space
group C2/c, a ) 45.634(13), b ) 9.376(3), c ) 20.379(6), ꢀ )
99.243(3)°, V ) 8606(4) Å3, Z ) 8, µ ) 1.408 mm-1, F(000) )
3728, R1 ) 0.0982 and wR2 ) 0.0907 (based on all 7562 data, I >
2σ(I)).
Computational Details. Density functional theory (DFT)
calculations were performed using the B3LYP functional (as
implemented in Jaguar 7.6, release 110). Main group atoms were
described using the 6-311++G** basis set,118 while the
LACV3P++**(2f), a small core effective potentials (angular
momentum projected pseudopotentials) including a double-ꢁ
f-type shell, was used for all metals.119-122 Structures were
optimized in the gas phase (starting with geometries obtained
from the X-ray crystal structures where available) using the
LACVP** and 6-31G** basis sets.119,123 For stationary points
and transition structures, the analytic Hessian was calculated to
obtain vibrational frequencies, which in turn were used to obtain
zero-point and vibrational thermodynamic corrections (ZPE,
H(T), and S). Solvent corrections were based on single point
self-consistent Poisson-Boltzmann continuum solvation calcula-
tions for benzene (ε ) 2.284 and R0) 2.60 Å), toluene (ε )
2.379 and R0 ) 2.76 Å) or tetrahydrofuran (ε ) 7.6 and R0 )
2.52 Å) as noted using the PBF module in Jaguar.124,125
X-Ray Crystal Structure of 6Lu-(iqn)2. X-ray quality crystals
were obtained from a concentrated hexanes solution placed in a
-35 °C freezer in the glovebox. A total of 43149 reflections (-32
e h e 32, -15 e k e 15, -22 e l e 22) were collected at T )
101(2) K with 2θmax ) 56.63°, of which 11870 were unique (Rint
) 0.0964). The residual peak and hole electron density were 0.72
and -0.88 eA-3. The least-squares refinement converged normally
with residuals of R1 ) 0.0454 and GOF ) 0.978. One of the t-butyl
groups was disordered over two sites (40 and 60% occupancy).
One of the disordered counterparts was only refined isotropically.
Crystal and refinement data for 6Lu-(iqn)2: formula C50H67N4-
Si2FeOLu, space group Pna21, a ) 24.396(10), b ) 11.830(5), c
) 16.723(7), R ) ꢀ ) γ ) 90°, V ) 4826(4) Å3, Z ) 4, µ )
2.423 mm-1, F(000) ) 2112, R1 ) 0.0668 and wR2 ) 0.0789 (based
on all 11870 data, I > 2σ(I)).
Acknowledgment. P.L.D. thanks Prof. James Mayer, University
of Washington, Seattle, for a useful suggestion, Dr. Saeed Khan,
UCLA, for help with some cyrstallography details, and Dr. Robert
Taylor, UCLA, for help with NMR spectroscopy experiments.
C.T.C. and K.R.O. thank Mr. Samuel I. Roth, Harvard-Westlake
School, for the synthesis of La(CH2C6H5)3(THF)3 and 1La-THF.
The experimental work was supported by the UCLA, Sloan
Foundation, and NSF (Grant CHE-0847735). The MSC computa-
tional facilities were funded by grants from ARO-DURIP and
ONR-DURIP. E.T. gratefully acknowledges support from UC
MEXUS-CONACyT for a postdoctoral fellowship. D.B., E.T., and
W.A.G. were supported partially by the Center for Catalytic
Hydrocarbon Functionalization, an Energy Frontier Research Center
funded by the DOE (DE-SC0001298).
X-Ray Crystal Structure of 8Y. X-ray quality crystals were
obtained from a concentrated toluene: Et2O solution placed in a
-35 °C freezer in the glovebox. A total of 41841 reflections (-49
e h e 49, -58 e k e 59, -14 e l e 14) were collected at T )
101(2) K with 2θmax ) 52.80°, of which 10820 were unique (Rint
) 0.1027). The residual peak and hole electron density were 1.94
and -0.61 eA-3. The least-squares refinement converged normally
with residuals of R1 ) 0.0579 and GOF ) 1.001. Crystal and
refinement data for 8Y ·C7H8: formula C57H70N5Si2FeY, space group
Fdd2, a ) 39.536(4), b ) 47.303(4), c ) 11.3457(10), R ) ꢀ )
γ ) 90°, V ) 21218(3) Å3, Z ) 16, µ ) 1.448 mm-1, F(000) )
8640, R1 ) 0.0807 and wR2 ) 0.1336 (based on all 10820 data, I
> 2σ(I)).
Supporting Information Available: Experimental details for
compound characterizations, structural representations indicating
a full atom-labeling scheme, DFT calculation details, and full
crystallographic descriptions. This material is available free of
charge via the Internet at http://pubs.acs.org.
X-Ray Crystal Structure of 4Y-(THF)2. X-ray quality crystals
were obtained from a concentrated n-hexane solution placed in a
-35 °C freezer in the glovebox. A total of 36317 reflections (-13
e h e 13, -38 e k e 38, -18 e l e 18) were collected at T )
JA908489P
(118) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys.
1980, 72, 650–654.
109(2) K with 2θmax ) 56.50°, of which 9909 were unique (Rint
)
(119) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299–310.
(120) Melius, C. F.; Goddard, W. A. Phys. ReV. A 1974, 10, 1541.
(121) Melius, C. F.; Olafson, B. D.; Goddard, W. A. Chem. Phys. Lett.
1974, 28, 457–462.
0.0978). The residual peak and hole electron density were 0.56 and
-0.42 eA-3. The least-squares refinement converged normally with
residuals of R1 ) 0.0521 and GOF ) 0.983. Two methyl groups
were disordered over two sites (50% occupancy each). Crystal and
refinement data for 4Y-(THF)2: formula C39H62N3Si2FeO2Lu, space
group P21/c, a ) 10.065(3), b ) 29.126(7), c ) 14.029(4), ꢀ )
100.204(3)°, V ) 4047.4(18) Å3, Z ) 4, µ ) 1.880 mm-1, F(000)
) 1704, R1 ) 0.1070 and wR2 ) 0.1082 (based on all 9909 data,
I > 2σ(I)).
(122) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114, 3408–
3420.
(123) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54,
724–728.
(124) Tannor, D. J.; Marten, B.; Murphy, R.; Friesner, R. A.; Sitkoff, D.;
Nicholls, A.; Honig, B.; Ringnalda, M.; Goddard, W. A. J. Am. Chem.
Soc. 1994, 116, 11875–11882.
X-Ray Crystal Structure of 9La. X-ray quality crystals were
obtained from a concentrated toluene: n-pentane solution placed
in a -35 °C freezer in the glovebox. A total of 30464 reflections
(125) Marten, B.; Kim, K.; Cortis, C.; Friesner, R. A.; Murphy, R. B.;
Ringnalda, M. N.; Sitkoff, D.; Honig, B. J. Phys. Chem. 1996, 100,
11775–11788.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 355