Synthesis of an Iron Parent Amido Complex and a Comparison of Its Reactivity with the Ruthenium Analog

Article abstract of DOI:10.1021/om0499660

The iron amido complex (dmpe)₂Fe(H)(NH₂) (2; dmpe = 1,2-bis(dimethylphosphino)ethane) has been prepared from the hydrido chloride complex using sodium amide in a liquid ammonia/THF solvent mixture. The complex is a mixture of trans and cis isomers in solution, although it crystallizes as the trans isomer exclusively. Complex 2 is a strong base, although somewhat weaker than the ruthenium amido complex trans-(dmpe)₂Ru(H)(NH ₂) (1). Amido complex 2 deprotonates several weak acids, and some of these deprotonations result in displacement of ammonia, while others do not. The basicity of complex 2 is further demonstrated with systems in which acidic substrates are reversibly deprotonated. Complex 2 undergoes H/D exchange with toluene-d_s, and it catalyzes H/D exchange between tolueneda and several other extremely weak acids. An investigation into the mechanism of the H/D exchange and isomerization reactions of trans-9,10-diisopropyl-9,10- dihydroanthracene (trans-4) using 1, 2, and cis-(PMe₃) ₄Ru(H)(NH₂) (5) as catalysts demonstrates that the H/D exchange reactions proceed through formation of ion pairs, although it is surprising that displacement of ammonia is never observed. Amido complexes 1 and 2, as well as the hydroxo complex trans-(dmpe)₂Ru(H)(OH), are also good nucleophiles. For example, they ring-open 1,3-di-tert-butylaziridinone in a nucleophilic manner to form trans-(dmpe)₂M(H)(XC(O)CH-(t-Bu)NH(t-Bu) ) (M = Fe, X = NH; M = Ru, X = NH, O), A second example is an unusual insertion of carbon monoxide into the N-H bond of 2 to form the carboxamide complex trans-(dmpe)₂Fe(H)(NHCHO) (6). Although CO is a poor electrophile, mechanistic experiments suggest that this reaction proceeds through direct attack of the amide nitrogen on the CO carbon.

Full text of DOI:10.1021/om0499660

1656
Organometallics 2004, 23, 1656-1670
Articles
Syn th esis of a n Ir on P a r en t Am id o Com p lex a n d a
Com p a r ison of Its Rea ctivity w ith th e Ru th en iu m An a log
Daniel J . Fox and Robert G. Bergman*
Department of Chemistry and Center for New Directions in Organic Synthesis,
University of California, Berkeley, California 94720
Received J anuary 12, 2004
The iron amido complex (dmpe)2Fe(H)(NH2) (2; dmpe ) 1,2-bis(dimethylphosphino)ethane)
has been prepared from the hydrido chloride complex using sodium amide in a liquid
ammonia/THF solvent mixture. The complex is a mixture of trans and cis isomers in solution,
although it crystallizes as the trans isomer exclusively. Complex 2 is a strong base, although
somewhat weaker than the ruthenium amido complex trans-(dmpe)2Ru(H)(NH2) (1). Amido
complex 2 deprotonates several weak acids, and some of these deprotonations result in
displacement of ammonia, while others do not. The basicity of complex 2 is further
demonstrated with systems in which acidic substrates are reversibly deprotonated. Complex
2 undergoes H/D exchange with toluene-d8, and it catalyzes H/D exchange between toluene-
d8 and several other extremely weak acids. An investigation into the mechanism of the H/D
exchange and isomerization reactions of trans-9,10-diisopropyl-9,10-dihydroanthracene
(trans-4) using 1, 2, and cis-(PMe3)4Ru(H)(NH2) (5) as catalysts demonstrates that the H/D
exchange reactions proceed through formation of ion pairs, although it is surprising that
displacement of ammonia is never observed. Amido complexes 1 and 2, as well as the hydroxo
complex trans-(dmpe)2Ru(H)(OH), are also good nucleophiles. For example, they ring-open
1,3-di-tert-butylaziridinone in a nucleophilic manner to form trans-(dmpe)2M(H)(XC(O)CH-
(t-Bu)NH(t-Bu)) (M ) Fe, X ) NH; M ) Ru, X ) NH, O). A second example is an unusual
insertion of carbon monoxide into the N-H bond of 2 to form the carboxamide complex trans-
(dmpe)2Fe(H)(NHCHO) (6). Although CO is a poor electrophile, mechanistic experiments
suggest that this reaction proceeds through direct attack of the amide nitrogen on the CO
carbon.
In tr od u ction
the synthesis and study of these species.8-17 Roundhill
and co-workers have demonstrated the nucleophilic and
insertion chemistry of CpRu(dcpe)(NH2) (dcpe ) 1,2-bis-
(dicyclohexylphosphino)ethane) with various electro-
philes.13 Recent examples that most closely relate to our
work, from the laboratories of Gunnoe and co-workers,
demonstrate the strong basicity of several TpRu-
(L)2NHR complexes15,16 and the nucleophilicity of (dt-
bpe)Cu(NHPh) (dtbpe ) 1,2-bis(di-tert-butylphosphino)-
Late-transition-metal complexes containing nondative
metal-nitrogen and -oxygen single bonds have been
shown to play important roles in many biological
systems1,2 and have been proposed as intermediates in
numerous synthetic3-7 and industrial processes.5,8 The
direct study of the metal-heteroatom complexes has
represented a particular challenge in many of these
cases, leading many groups to develop techniques for
(9) Bergman, R. G. Polyhedron 1995, 14, 3227.
(10) Bryndza, H. E.; Fultz, W. C.; Tam, W. Organometallics 1985,
4, 939.
* To whom correspondence should be addressed. E-mail:
bergman@cchem.berkeley.edu.
(1) Bertini, I.; Gray, H. B.; Lippard, S. J .; Valentine, J . S. Bioinor-
ganic Chemistry; University Science Books: Mill Valley, CA, 1994.
(2) Holm, R. H.; Kennepohl, P.; Solomon, E. I. Chem. Rev. 1996,
96, 2239.
(3) Hartwig, J . F. Angew. Chem., Int. Ed. 1998, 37, 2047.
(4) Ishiyama, T.; Hartwig, J . F. J . Am. Chem. Soc. 2000, 122, 12043.
(5) Roundhill, D. M. Chem. Rev. 1992, 92, 1.
(11) Bryndza, H. E. Organometallics 1985, 4, 1686.
(12) Fryzuk, M. D.; Montgomery, C. D. Coord. Chem. Rev. 1989, 95,
1.
(13) J oslin, F. L.; J ohnson, M. P.; Mague, J . T.; Roundhill, D. M.
Organometallics 1991, 10, 2781.
(14) Li, J . J .; Li, W.; J ames, A. J .; Holbert, T.; Sharp, T.; Sharp, P.
R. Inorg. Chem. 1999, 38, 1563.
(15) Conner, D.; J ayapraksh, K. N.; Gunnoe, T. B.; Boyle, P. D.
Organometallics 2002, 21, 5265.
(6) Schaad, D. R.; Landis, C. R. J . Am. Chem. Soc. 1990, 112, 1628.
(7) Wagaw, S.; Rennels, R. A.; Buchwald, S. L. J . Am. Chem. Soc.
1997, 119, 8451.
(16) J ayapraksh, K. N.; Conner, D.; Gunnoe, T. B. Organometallics
2001, 20, 5254.
(17) Blue, E. D.; Davis, A.; Conner, D.; Gunnoe, T. B.; Boyle, P. D.;
White, P. S. J . Am. Chem. Soc. 2003, 125, 9435.
(8) Bryndza, H. E.; Tam, W. Chem. Rev. 1988, 88, 1163.
10.1021/om0499660 CCC: $27.50 © 2004 American Chemical Society
Publication on Web 03/11/2004

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1657
ethane).17 Through the study of late-transition-metal-
heteroatom complexes, we hope to further our funda-
mental understanding of the factors that control their
overall reactivity and the mechanisms of this reactivity.
As the knowledge base of these key factors develops, a
concomitant improvement of processes that proceed
through similar transition-metal-heteroatom species
should be achieved.
Sch em e 1
We have developed techniques for the synthesis of
metal alkoxo, aryloxo, hydroxo, and amido complexes
using several different metal centers (Re, Ru, Rh, Ir,
Ni). In the process, we have observed many different
types of reactivities involving these species including,
but not limited to, insertions, deprotonations, and
nucleophilic displacements/ring openings.18
Most recently, the work in our laboratories involving
transition-metal-heteroatom complexes has involved
the study of a ruthenium complex with a parent (i.e.
unsubstituted) amido ligand, trans-(dmpe)2Ru(H)(NH2)
(1; dmpe ) 1,2-bis(dimethylphosphino)ethane).19-21 Com-
plex 1 is the first structurally characterized monomeric
late-transition-metal parent amido complex. The reac-
tivity of amido complex 1 is dominated by its exception-
ally high basicity;20,21 it was found to deprotonate weak
carbon acids such as fluorene and phenylacetylene and
even undergo H/D exchange with extremely weak acids
such as toluene and propene.20 The pKa of a representa-
tive conjugate acid of 1, [trans-(dmpe)2Ru(H)(NH3)]-
[Ph3C], was determined to be between 23 and 24.21
We pursued an expanded scope of this metal-amido
chemistry by the study of the first-row analog (dmpe)2Fe-
(H)(NH2) (2). The work presented in this paper concerns
the synthesis and reactivity of 2, which was synthesized
using the techniques previously developed for the
ruthenium amido complex 1.22 The basic and nucleo-
philic properties of 2 were explored, including an
investigation into the mechanism of H/D exchange.
Additionally, the mechanism of an unusual process
involving carbon monoxide insertion into the N-H bond
of complex 2 rather than the M-N bond was studied.
complex by lithium aluminum hydride24 and with
shorter reaction time than that required for reduction
by sodium in the presence of H2 with extraction and
25
filtration as the only necessary purification. The iron
parent amido complex (dmpe)2Fe(H)(NH2) (2) was syn-
thesized by treatment of trans-(dmpe)2Fe(H)(Cl) with
NaNH2 in a THF/NH3(l) solvent mixture22 and was
isolated in 63% yield as an orange crystalline solid
(Scheme 1). As we have noted in a previous communica-
tion,22 a significant difference between the iron amido
complex 2 and the ruthenium amido complex 1 is that,
in solution, complex 2 is a rapidly isomerizing equilib-
rium mixture of cis and trans isomers, while complex 1
is the pure trans species. We presume that the isomer-
ization of 2 proceeds through a dechelation of one of the
dmpe ligands, followed by a Berry pseudorotation26,27
and rechelation of the ligand.
Dep r oton a tion Rea ction s. The amido complex 2
was found to deprotonate several weak carbon acids
irreversibly. In some of these systems, the only observed
product was the ion pair. Reaction of 2 with fluorene in
THF resulted in an immediate color change to a brighter
orange than that of the starting amide solution, and a
small quantity of the ion pair precipitated as an orange
crystalline solid. Removal of THF under reduced pres-
sure after stirring for 3 h resulted in clean formation of
[trans-(dmpe)2Fe(H)(NH3)][Fl] (3, Fl ) fluorenide; eq 1)
Resu lts a n d Discu ssion
Syn th esis of (d m p e)2F e(H)(NH2) (2). The iron
hydrido chloride complex trans-(dmpe)2Fe(H)(Cl) was
synthesized as reported by Field and co-workers23 from
cis-(dmpe)2Fe(H)2. Although the dihydride complex was
reported previously,24 we developed a new synthesis
that entailed the reduction of trans-(dmpe)2Fe(Cl)2 by
sodium in liquid ammonia. This synthetic method was
found to produce the dihydride complex in greater yield
than that obtained by reduction of the dichloride
(18) Fulton, J . R.; Holland, A. W.; Fox, D. J .; Bergman, R. G. Acc.
Chem. Res. 2002, 35, 44.
(19) Kaplan, A. W.; Ritter, J . C. M.; Bergman, R. G. J . Am. Chem.
Soc. 1998, 120, 6828.
(20) Fulton, J . R.; Bouwkamp, M. W.; Bergman, R. G. J . Am. Chem.
Soc. 2000, 122, 8799.
(21) Fulton, J . R.; Sklenak, S.; Bouwkamp, M. W.; Bergman, R. G.
J . Am. Chem. Soc. 2002, 124, 4722.
(22) Some of the results described here were reported earlier in
preliminary form: Fox, D. J .; Bergman, R. G. J . Am. Chem. Soc. 2003,
125, 8984.
in 97% yield. The hydride resonance of 3 is shifted
upfield relative to that of amide 2 in the 1H NMR
(23) Baker, M. V.; Field, L. D.; Young, D. J . J . Chem. Soc., Chem.
Commun. 1988, 546.
(24) Whittlesey, M. K.; Mawby, R. J .; Osman, R.; Perutz, R. N.; Field,
L. D.; Wilkinson, M. P.; George, M. W. J . Am. Chem. Soc. 1993, 115,
8627.
(25) Kaplan, A. W.; Bergman, R. G. Organometallics 1997, 16, 1106.
(26) Crabtree, R. H. The Organometallic Chemistry of the Transition
Elements, 2nd ed.; Wiley: New York, 1994.
(27) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M.
Advanced Inorganic Chemistry, 6th ed.; Wiley: New York, 1999.

1658 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
spectrum (-29.61 ppm), and the selectively decoupled
31P spectrum shows a doublet at 74.0 ppm.28
Having determined that iron amido complex 2 is
strongly basic, we wished to determine its basicity
relative to ruthenium amido complex 1. Addition of 1
to ion pair 3 resulted in clean formation of 2 and [trans-
(dmpe)2Ru(H)(NH3)][Fl]. As expected from the previous
result, addition of 2 to [trans-(dmpe)2Ru(H)(NH3)][Fl]
did not result in proton transfer, demonstrating that
the ruthenium amido complex is the stronger base (eq
2). The difference in basicity between the ruthenium and
hydroxide complex from the hydroxide-bridged dimer by
thermolysis were unsuccessful.
The reactions of iron amido complex 2 with fluorene
and water showed that some anions are not capable of
displacing ammonia from the trans-(dmpe)2Fe(H)(NH3)
cation. Similar behavior was observed with the trans-
(dmpe)2Ru(H)(NH3) cation using some of the same
anions, although the behavior of the hydroxide anion
is an exception.21 In the case of hydroxide, the hydrogen
bonding present in the dimeric ion pair is probably
responsible for the lack of displacement, while the
fluorenide anion is too large to bind well to the iron
center.
As with ruthenium amido complex 1, there are cases
in which displacement of ammonia does occur. The
keteniminate complex trans-(dmpe)2Fe(H)(NCC(H)Ph)
was formed in 82% yield upon deprotonation of phenyl-
acetonitrile by amido complex 2 and displacement of
ammonia by the keteniminate anion (eq 4). This com-
iron amido complexes is most likely due to the fact that
ruthenium is more electron rich than iron, as demon-
strated by the lower frequency of the carbonyl stretch
in CpRu(CO)2CH3 relative to CpFe(CO)2CH3.29,30 We
assume that this increased electronic density is trans-
mitted to the attached nitrogen atom (and lone pair)
through the Fe-N single bond.
Our group had found that addition of water to
ruthenium amido complex 1 resulted in formation of
trans-(dmpe)2Ru(H)(OH) through loss of ammonia.31
However, when water was added to the iron amido
complex 2 in THF-d8, no loss of ammonia was observed
1
by H NMR spectroscopy. Upon addition of water to a
stirred solution of 2 in THF, the solution became yellow-
orange. A yellow crystalline solid, insoluble in pentane,
was isolated (60% yield) and recrystallized from THF
at -35 °C to give material whose chemical composition
was found by combustion analysis and NMR spectros-
copy to be consistent with [trans-(dmpe)2Fe(H)(NH3)]2-
[OH]2‚H2O. The complex displayed a hydride resonance
1
plex is characterized by a quintet hydride resonance at
-27.28 ppm in the H NMR spectrum and a ketenimi-
at -28.17 ppm in the H NMR spectrum and a doublet
1
at 70.5 ppm in the 31P{1H} NMR spectrum. An X-ray
diffraction study revealed a dimeric structure consisting
of two trans-(dmpe)2Fe(H)(NH3) moieties bridged by
hydroxide anions through hydrogen bonding of each
hydroxy oxygen atom to two ammonia hydrogen atoms;
each hydroxide anion was also hydrogen bonded to a
water molecule (eq 3, Figure 1). This dimeric structure
is similar to that of the ion pair that was formed upon
addition of p-cresol to the ruthenium amido complex 1.21
Presumably, the stability of the hydrogen-bound hy-
droxide-bridged dimer prohibits displacement of am-
monia from the iron center. Efforts to produce the iron
nate stretch at 2111 cm-1 in the IR spectrum.21 Reaction
of 2 with terminal alkynes also resulted in displacement
of ammonia and formation of the corresponding alkynyl
complexes. For example, addition of propyne to 2 in
benzene-d6 solution resulted in formation of the iron
methylacetylide complex trans-(dmpe)2Fe(H)(CCMe) in
50% yield (eq 5), in addition to smaller quantities of two
(28) The range of the proton decoupler is such that 1H resonances
upfield of -20 ppm are not decoupled.
(29) Davison, A.; McCleverty, J . A.; Wilkinson, G. J . Chem. Soc.
1963, 1133.
(30) Gansow, O. A.; Schexnayder, D. A.; Kimura, B. Y. J . Am. Chem.
Soc. 1972, 94, 3406.
(31) Fulton, J . R.; Bergman, R. G. Unpublished results.

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1659
unidentified products.32 Furthermore, complex 2 also
caused isomerization of allene, such that the alkynyl
complex trans-(dmpe)2Fe(H)(CCMe) was formed in 60%
yield, along with a smaller quantity of an unidentified
product (eq 6). These products are analogous to those
formed by similar reactions with ruthenium amido
complex 1.21
Nu cleop h ilic Rea ctivity of P a r en t Am id o Com -
p lexes. In addition to acting as bases, the amido
complexes were observed to act as nucleophiles toward
strained-ring systems. For example, addition of 2 to 1,3-
di-tert-butylaziridinone33 at 25 °C resulted in immediate
formation of the ring-opened product trans-(dmpe)2Fe-
(H)(NHC(O)CH(t-Bu)NH(t-Bu)) as a yellow crystalline
solid in 74% yield (eq 7). The product displays a hydride
F igu r e 1. ORTEP diagram of [trans-(dmpe)2Fe(H)(NH3)]2-
[OH]2‚2H2O. The hydrogen atoms on the dmpe ligands have
been removed for clarity. The hydride, ammonia, hydroxide,
and water hydrogen atoms were located from a difference
Fourier map, and their positions were refined. The struc-
ture shows the hydrogen bonds to two symmetry-related
hydroxide anions (N1-O1 ) 2.95 Å, N1-O1′ ) 2.88 Å).
Selected bond lengths (Å) and angles (deg): Fe1-N1, 2.103-
(5); Fe1-H33, 1.37(5); N1-Fe1-H33, 178(2); H-N1-
H(average), 103. Thermal ellipsoids are shown at the 50%
probability level.
νCO 1588 cm-1) were isolated, respectively (eq 7). The
product derived from the ruthenium amide displays a
hydride quintet at -18.79 ppm in the 1H NMR spectrum
and a multiplet at 43.8 ppm in the 31P{1H} spectrum,
while the hydroxide-derived product displays a hydride
1
quintet at -23.28 ppm in the H NMR spectrum and a
multiplet at 43.6 ppm in the 31P{1H} spectrum. An X-ray
diffraction study performed on a crystal of the ruthe-
nium amide derived complex confirmed the connectivity
of the complex, although the resolution of the structure
was poor due to crystal decomposition. Reaction of the
ruthenium hydroxide complex with 1,3-di-tert-butyl-
aziridinone is slower than with the amido complexes,
as complete conversion required 12 h.
H/D Exch a n ge Rea ction s. The high basicity of iron
amido complex 2 can also be demonstrated with sub-
strates that 2 deprotonates reversibly. Dissolution of 2
in toluene-d8 resulted in H/D exchange between the
toluene methyl group and the amide hydrogens; this
process also resulted in exchange into all of the bound
dmpe hydrogen positions to form (dmpe-d16)2Fe(H)(ND2)
(2-d34, eq 8). In the polydeuterated complex, only the
1
quintet at -28.41 ppm in the H NMR spectrum and a
complex multiplet at 71.2 ppm in the 31P{1H} NMR
spectrum. The complex splitting pattern in the 31P NMR
spectrum is likely the result of hindered rotation of the
large carboxamide ligand. Additionally, the CdO stretch
in the IR spectrum of the product is significantly red-
shifted relative to the CdO stretch in 1,3-di-tert-
butylaziridinone (to 1580 cm-1 from 1850 cm-1), indica-
tive of aziridinone ring opening. The observed final
product was the result of nucleophilic attack on the
carbonyl carbon, followed by ring opening by the break-
ing of the C-N bond and proton transfer to the tert-
butyl-substituted nitrogen.
Analogous reactions occurred upon addition of 1,3-
di-tert-butylaziridinone to both the ruthenium amido
complex 1 and the ruthenium hydroxo complex trans-
(dmpe)2Ru(H)(OH). With these systems, trans-(dmpe)2Ru-
(H)(NHC(O)CH(t-Bu)NH(t-Bu) (61%, νCO 1582 cm-1
)
and trans-(dmpe)2Ru(H)(OC(O)CH(t-Bu)NH(t-Bu) (77%,
hydride resonances (both trans and cis) remained inert
to exchange. The 2H{1H} NMR spectrum contained
aliphatic resonances corresponding to the dmpe ligands
(32) Ittel, S. D.; Tolman, C. A.; English, A. D.; J esson, J . P. J . Am.
Chem. Soc. 1978, 100, 7577.
(33) Sheehan, J . C.; Beeson, J . H. J . Am. Chem. Soc. 1967, 89, 362.

1660 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
Sch em e 2
Sch em e 3
and also displayed a broad N-D resonance at -4.6 ppm;
no iron-bound deuteride was observed. Although ex-
change was observed with toluene, neither the ion pair
[trans-(dmpe)2Fe(H)(NH3)][CH2Ph] nor the ammonia
displacement product trans-(dmpe)2Fe(H)(CH2Ph) was
observed (Scheme 2). However, treatment of trans-
(dmpe)2Fe(H)(CH2Ph) with excess ammonia did not
result in the formation of 2 (eq 9). Therefore, we cannot
exchange between toluene-d8 and propene.21 This ob-
servation is consistent with our conclusion derived from
proton-transfer studies that the iron amido complex is
less basic than its ruthenium analog (vide supra).
Although 2 is a catalyst for H/D exchange between
toluene-d8 and ammonia, treatment of 2 with 15N-
labeled ammonia did not result in incorporation of 15
N
into 2 (eq 11). This result indicated that deuterium
exchange into ammonia was achieved through an acid/
base mechanism rather than associative exchange of the
nitrogen ligand (Scheme 3). Although exchange of
deuterium into the complex’s dmpe ligands was ob-
served, the amido complex 2 did not catalyze H/D
exchange between toluene-d8 and free dmpe or the dmpe
ligands of trans-(dmpe)2Fe(H)(Cl), indicating that the
exchange of deuterium into the dmpe ligands of 2
probably occurred in an intramolecular fashion.
draw conclusions about the thermodynamics of the
process, but a large kinetic barrier for the transforma-
tion must exist that prevents displacement in either
direction. Treatment of trans-(dmpe)2Fe(H)(CH2Ph) with
p-cresol and p-thiocresol did result in formation of the
cresolate and thiocresolate complexes, respectively (eq
10).
Ca ta lytic Isom er iza tion Rea ction s. Displacement
of ammonia was not observed in any H/D exchange
reactions of very weak acids, consistent with two pos-
sible mechanisms: (a) exchange occurred via a concerted
mechanism and ion pairs were not formed or (b) ion
pairs were formed, but the anion that was generated
failed to displace the ammonia from the metal center
(Scheme 4). We observed that our amido complexes
catalyzed (also without displacement) the isomerization
of 1,3-cyclohexadiene to an equilibrium mixture of 1,4-
and 1,3-cyclohexadiene (eq 12) and the H/D exchange
Amido complex 2 is also a catalyst for the transfer of
deuterium from toluene-d8 to dihydrogen, ammonia,
cycloheptatriene, 9,10-dihydroanthracene, 9,10-disub-
stituted 9,10-dihydroanthracenes (exchange into 9,10-
positions), cyanocyclopropane, and the aliphatic proton
of triphenylmethane. However, unlike the ruthenium
amido complex 1, complex 2 does not catalyze H/D

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1661
Sch em e 4
Sch em e 5
Sch em e 6
between toluene-d8 and cyclohexadiene, 9,10-dihydroan-
thracene, and cycloheptatriene. Deuterium incorpora-
tion into all hydrogen positions of cycloheptatriene also
implied that an isomerization had taken place (Scheme
5; presumably the isomerization of 1,3-cyclohexadiene
occurs in a similar manner).
Ta ble 1. Con ver sion a n d H/D Exch a n ge Da ta for
Rea ction s w ith tr a n s-4 a n d cis-4
amt of H, %
k (M-1
s-1
conversn,
%
entry catalyst substrate
)
cis
trans
1
2
60% 5
60% 5
trans-4
2.18 × 10-3
trans-4-d2 7.2 × 10-4
21
62
87
62
100
47
54
56
43
37
18 (20 h)
91
6.5
7.7
10
The observation of these rearrangements suggests
that formation of an ion pair with some mobility of the
two ions is the more likely mechanistic pathway. We
reasoned that if the H/D exchange reactions occurred
via a concerted (or very tight ion pair) mechanism
(Scheme 4, path a), isomerization of 9,10-disubstituted-
9,10-dihydroanthracenes might not be expected to occur,
while ion pair formation (Scheme 4, path b) could be
expected to lead to isomerization to the thermodynami-
cally favored isomer.34 trans-9,10-Diisopropyl-9,10-di-
hydroanthracene (trans-4) was prepared according to
literature procedures,35 and the trans geometry was
confirmed by comparison to previously reported NMR
data.36 When a catalytic quantity of amido complex (1,
2, or cis-(PMe3)4Ru(H)(NH2) (5)) was added to trans-4
in benzene-d6 solution, isomerization to the thermody-
namically favored cis-4 was observed (Scheme 6).34 The
isomerization of trans-4 to the cis isomer necessitates
removal of a proton from one face of the molecule and
reprotonation on the opposite face. We surmised that
comparison of the isomerization rate with the rate of
H/D exchange could give an indication as to how closely
the mechanisms of the two processes were related.
Additionally, it could be possible to determine if H/D
exchange was faster in one isomer than in the other.
Isomerization of trans-4 to the cis isomer using 5 as
catalyst in benzene-d6 was found to be first order in both
3
4
5% 1
trans-4-d2
cis-4-d2
9.9
5% 2-d34 trans-4
20
65
90
94
90
75
87
86
100
51
79
46
6 (15 h)
5
10% t-Bu- trans-4-d2
NHLi,a cis-4-d2
45 °C
1.0
a
2 equiv of t-BuNH2 was added as a hydrogen source.
F igu r e 2. Kinetic plot showing first-order decay of trans-4
using 5 as catalyst. Initial concentrations: 25 mM trans-
4, 21 mM 5, 25 °C.
catalyst and substrate with an average rate constant
of k ) (2.18 ( 0.11) × 10-3 M-1 s-1 (Figure 2, Table 1).
When the isomerization was performed using trans-4-
d2 (deuterated in the 9- and 10-positions) as the
substrate and 5 mol % of the bis-dmpe ruthenium amido
(34) Harvey, R. G.; Cho, H. J . Am. Chem. Soc. 1974, 96, 2434.
(35) Harvey, R. G.; Davis, C. C. J . Org. Chem. 1969, 34, 3607.
(36) Zieger, H. E.; Schaeffer, D. J .; Padronaggio, R. M. Tetrahedron
Lett. 1969, 10, 5027.

1662 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
Sch em e 7
complex 1 as catalyst and hydrogen source37 in benzene-
d6, observation of the reaction solution by 1H NMR
spectroscopy after 62% conversion showed that H/D
exchange had occurred to vastly different extents in the
trans substrate and cis product (Table 1). Integration
of the benzylic resonances relative to the methine
resonances indicated that, in the cis product at this
percent conversion, the benzylic position contained 43%
hydrogen and the remaining trans starting material
contained only 9.9% hydrogen. These observations are
most easily rationalized by assuming that the ion pair
It is formed initially (Scheme 7). This species is poten-
tially capable of either transferring H or D back to the
carbanion or rearranging to the diastereomeric trans
ion pair Ic. We assume that a primary isotope effect
would encourage transfer of H preferentially to D in
both intermediates. It is notable that the product
contains 43% H, close to the 50% that would be expected
if the initial ion pair were formed and isomerized and
then the carbanion reprotonated by exclusive H+ rather
than D+ back transfer from the countercation. Since so
little H is found in the trans starting material at this
substantial extent of reaction, and the amount of H
incorporated into the cis product is much larger, the
thermodynamic preference that favors the cis over the
trans hydrocarbon in this system is felt in the rate
constants that control the relative rates of isomerization
and H/D exchange.
There are two possible ways of explaining this be-
havior. The first assumes that the initially formed ion
pair It rearranges to its stereoisomer Ic, and this
undergoes proton transfer to give the rearranged prod-
uct cis-4, relatively rapidly compared to the rate of
reconversion of It to the starting hydrocarbon trans-4
(i.e., non-Curtin-Hammett conditions). In this case,
there could be a strong driving force for rearrangement
of the trans ion pair It to the cis ion pair Ic (i.e.,
rearrangement to the opposite face of the carbanion is
more rapid than reverse proton transfer in the initial
intermediate, a situation illustrated in the free energy
diagram in Figure 3a). Once formed, the isotope effect
requires the cis ion pair to transfer H+ predominantly
over D+ to the carbanion. However, although many
proton transfers are very fast, transfers to delocalized
carbanions are an exception and often tend to be
unusually slow, due to the necessity of electron reorga-
nization in the reprotonation transition state. Therefore,
molecular tumbling in the ion pair might be faster than
reprotonation, and in this case Curtin-Hammett con-
straints might apply. In this scenario, the ratio of back
proton transfer (as monitored approximately by H/D
exchange in the trans starting material) to formation
of cis product would be determined only by the relative
free energies of the reprotonation transition states. It
is conceivable that the transition state leading to the
cis product might be the lower energy, reflecting the
greater thermodynamic stability of this diastereomer
(Figure 3b).
Whatever the situation is with respect to the relative
rates in Scheme 7, these isomerization and exchange
results are reminiscent of classical experiments carried
out by Cram and co-workers in their stereochemical
studies of carbanion rearrangements that proceed via
organic ion pairs.38,39 Therefore, they support the ion
pair mechanism for H/D exchange in weak carbon acids.
After 7 half-lives (1.5 h), 37% H/D exchange was
observed in the cis product, indicating that H/D ex-
change apparently did not progress after the isomer-
ization was complete (Table 1). Although it seemed
unlikely, there remained the possibility that H/D ex-
change simply occurred at a faster rate in cis-4 than in
trans-4. Therefore, cis-4-d2 was synthesized by addition
of trans-4 to a DMSO-d6 solution containing 5 mol %
NaH.40 When cis-4-d2 was exposed to 5 mol % 1, only
18% H/D exchange was observed after 20 h, indicating
that H/D exchange does not occur rapidly with cis-4-d2
q
(Table 1). That is, ∆Gc is large, reflecting the greater
thermodynamic stability of cis-4 over trans-4.
To monitor the reaction progress more carefully, we
used ruthenium amido complex 5 as catalyst because
it catalyzed the isomerization of trans-4 more slowly
than the bis-dmpe amido complex. Under identical
conditions, trans-4-d2 and cis-4-d2 were subjected to 60
mol % 5 in THF-d8 at 22 °C. trans-4-d2 was isomerized
by 5 with k ) (7.2 ( 0.4) × 10-4 M-1 s-1. The cis product
contained 50% hydrogen in the benzylic position, but
no H/D exchange was observed when cis-4-d2 was used
as the substrate (eq 13, Table 1), observations analogous
to those described above using 1 as the catalyst. These
results indicated that, in the isomerization reactions,
the H/D exchange that was observed was not the result
of exchange with the cis product but occurred along with
the isomerization process.
(38) Cram, D. J .; Gosser, L. J . Am. Chem. Soc. 1963, 85, 3890.
(39) Cram, D. J .; Gosser, L. J . Am. Chem. Soc. 1964, 86, 2950.
(40) Gleiter, R.; Weigl, H.; Haberhauer, G. Eur. J . Org. Chem. 1998,
1447.
(37) The catalyst contains 34 exchangeable hydrogen atoms in each
molecule.

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1663
F igu r e 3. Reaction coordinate diagrams representing the isomerization of trans-4 to cis-4 using a parent amido complex
as catalyst: (a) non-Curtin-Hammett conditions; (b) Curtin-Hammett conditions.
Sch em e 8
We also wished to determine if more traditional
anionic bases catalyzed the isomerization and H/D
exchange reactions in the same way as those catalyzed
by our amide bases. Isomerization of trans-4-d2 with
10% t-BuNHLi and 2.0 equiv of t-BuNH2 as a source of
hydrogen at 45 °C yielded results similar to those
observed with the neutral amide bases. After 51%
conversion, the benzylic position of the cis product
contained 46% H, while the starting trans substrate
contained only 1% H (Table 1). Under the same condi-
tions starting with cis-4-d2, only 6% H was observed in
the benzylic position of the substrate after the same
amount of time required for 51% conversion of the trans
substrate. These experiments indicate that the results
observed with the neutral amido bases are not unique
to those species and provide strong support for the ion
pair mechanism, which has been firmly established for
organic exchange reactions.38,39
To determine whether isotope effects had an influence
on the H/D exchange results, the isomerization of
trans-4 was carried out using 5% (dmpe-d16)2Fe(H)(ND2)
(2-d34) as the catalyst and deuterium source. The
isomerization reaction was observed over the course of
21 h (100% conversion). The amount of hydrogen in the
product varied from 91% after 10 min (20% conversion)
to 79% after 21 h (100% conversion), while it varied from
94% at 10 min to 75% at 3 h (90% conversion) in the
starting material (Table 1). This result indicated that
deuterium was less likely to be returned to the substrate
than hydrogen and agreed with the observation in the
isomerization of trans-4-d2 that hydrogen incorporation
in the cis product was approximately 50%. When a
deuterium was removed in order for isomerization to
take place, it was most often a hydrogen that was
returned to the molecule rather than a deuterium.
In ser tion of Ca r bon Mon oxid e in to th e Am id e
N-H Bon d . As reported previously,22 reaction of the
iron amido complex 2 with 1 atm of CO in toluene
resulted in net insertion of CO into the N-H bond of
the amide ligand to form trans-(dmpe)2Fe(H)(NHCHO)
(6, eq 14) in 60% isolated yield as a yellow crystalline
solid. To our knowledge, this represents the first pref-
erential insertion of CO into a late-transition-metal
amide N-H bond rather than the M-N bond. Carbon
monoxide reacts with several late-transition-metal ami-
do complexes through insertion into the M-N bond (eq
15); no instances of N-H insertion had been reported
previously.10,12,13,41,42 The M-N insertion reactions are
proposed to proceed via coordination of CO to the metal
(41) The ruthenium amido complex 1 was found to insert CO to yield
both the C-bound product, trans-(dmpe)2Ru(H)(C(O)NH2), and the
N-bound product, trans-(dmpe)2Ru(H)(NHCHO): Fulton, J . R.; Berg-
man, R. G. Unpublished results.
(42) Holland, P. L.; Andersen, R. A.; Bergman, R. G. J . Am. Chem.
Soc. 1996, 118, 1092.

1664 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
Sch em e 9
center, followed by migratory insertion of CO to yield
the observed product.
(D)(NH2) (2-FeD) was treated with CO as indicated
above for the proton-containing analog. This reaction
resulted in exclusive formation of trans-(dmpe)2Fe(D)-
(NHCHO) (6-FeD, eq 14). Path A of Scheme 9 as written
would result in scrambling of the deuterium label into
the carbonyl C-H bond of the product. Additionally,
reaction of 2 with 1 atm of CO in the presence of 1 equiv
of cis-(dmpe)2Fe(D)2 resulted in no deuterium incorpora-
tion into product 6. We therefore eliminated this mech-
anism, as illustrated, on the basis of the two labeling
experiments.46
Three possible mechanisms have been considered to
explain the unusual CO insertion. The first involves
direct attack of the amide nitrogen on the carbon of free
CO to form a zwitterionic intermediate that gives the
observed product (6) after proton transfer (Scheme 8).43
Because CO is a poor electrophile, we have also consid-
ered a second mechanism for CO insertion, illustrated
in Scheme 9, path A. This mechanism is initiated by
dissociation of one arm of a dmpe ligand from the iron
center in 2 and coordination of CO. Migratory insertion
of CO into the M-N bond would yield the intermediate
trans-(dmpe)2Fe(H)(C(O)NH2) (7).11,42,44 A similar mech-
anism is observed with lithium amide and lithium alkyl
species, which are carbonylated by CO through a
nucleophilic mechanism.43 â-Hydride elimination of
isocyanic acid, followed by insertion into the iron-
hydride bond, would give rise to complex 6. The above
mechanism is supported by experimental results ob-
tained both in our laboratory and by Field and co-
workers45 involving the insertion of several heterocu-
mulenes into cis-(dmpe)2Fe(H)2 to form products similar
to 6 (an example from our laboratory is shown in eq 16).
Although the intermediacy of the iron dihydride
complex could be ruled out, we considered the possibility
that the M-N insertion product 7 is formed during the
course of the reaction, followed by rearrangement to the
observed N-bound product 6 (Scheme 9, path B). Pho-
tolysis of cis-(dmpe)2Fe(H)2 forms the unsaturated iron-
(0) complex (dmpe)2Fe, which undergoes both C-H and
N-H bond activation with a variety of substrates.6,47,48
We reasoned that reaction of formamide with the iron-
(0) complex could lead to 7, allowing us to test whether
this complex was a likely intermediate. However, when
cis-(dmpe)2Fe(H)2 was irradiated (500 W mercury lamp,
through Pyrex) in the presence of formamide in THF-
d8, the N-H activation product 6 was observed while
C-H activation product 7 was not detected (Scheme 10).
Photolysis of cis-(dmpe)2Fe(H)2 in the presence of for-
mamide-d resulted in the formation of the complex 6-CD
1
2
(Scheme 10), as observed by H, H{1H}, and 31P{1H}
NMR spectroscopy. While this experiment did not
provide an independent route to intermediate 7, the
absence of 6-FeD excluded the possibility of C-H (or
C-D) activation in the photolysis reactions.
Finally, we considered the possible intermediacy of
(dmpe)2Fe(CO), formed by reductive elimination of NH3
from the amide starting material and coordination of
CO. The metal in this complex could serve to activate
the bound CO toward attack at its carbon by the amide
ligand of a second molecule of complex 2, leading to the
observed product. This mechanism was tested by treat-
ment of 2 with CO in the presence of 1 equiv of
However, deuterium-labeling experiments do not sup-
port the formation of isocyanic acid in the CO insertion
reaction. The iron deuterio amide complex (dmpe)2Fe-
(43) Alkali-metal salts of strongly basic nitrogen and carbon anions
are known to attack CO. See, for example: (a) Rautenstrauch, V.;
J oyeux, M. Angew. Chem., Int. Ed. Engl. 1979, 18, 85. (b) Trzupek, L.
S.; Newirth, T. L.; Kelly, E. G.; Sbarbati, N. E.; Whitesides, G. M. J .
Am. Chem. Soc. 1973, 95, 8118.
(44) Collman, J . P.; Hegedus, L. S.; Norton, J . R.; Finke, R. G.
Principles and Applications of Organotransition Metal Chemistry;
University Science Books: Sausalito, CA, 1987.
(46) It is possible that â-hydride elimination, if it occurs by deche-
lation of the dmpe ligand, would form a species with coordinated
HNCO. If this is the case, and the hydrides maintained their inequiva-
lence, scrambling of the deuterium label might not be observed.
(47) Baker, M. V.; Field, L. D. J . Am. Chem. Soc. 1986, 108, 7433.
(48) Buys, I. E.; Field, L. D.; Hambley, T. W.; McQueen, A. E. D. J .
Chem. Soc., Chem. Commun. 1994, 557.
(45) Field, L. D.; Lawrenz, E. T.; Shaw, W. J .; Turner, P. Inorg.
Chem. 2000, 39, 5632.

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1665
Sch em e 10
(dmpe)2Fe(13CO). We failed to observe incorporation of
13C into product 6, indicating that this mechanism was
not operative (eq 17). Considering the above observa-
tions as a whole, we conclude that our experimental
evidence most strongly supports the simple nucleophilic
attack mechanism illustrated in Scheme 8, although the
mechanism shown in Scheme 9, path B, cannot be ruled
out conclusively.
leads to displacement of ammonia (acetylides, keten-
iminates) while others do not lead to displacement
(fluorenide, hydroxide, benzyl). However, both the size
of the anion and the degree of hydrogen bonding in the
ion pair may play important roles in controlling this
selectivity.
The nucleophilic reactivity of amido complexes 1 and
2 (and trans-(dmpe)2Ru(H)(OH)) was also demonstrated
by the nucleophilic ring opening of 1,3-di-tert-butylaziri-
dinone. The mechanism of an unusual transformation
involving net insertion of CO into the N-H bond of
amido complex 1 was investigated; the mechanism of
this reaction appears to involve direct nucleophilic
attack of the amide group at the CO carbon.
Exp er im en ta l Section
Gen er a l Com m en ts. All manipulations were carried out
under an inert atmosphere or by using standard Schlenk or
1
vacuum-line techniques unless noted otherwise. The H, 13C-
{1H}, 2H{1H}, 31P{1H}, and 15N{1H} NMR spectra were ob-
tained on Bruker 300, 400, and 500 MHz Fourier transform
spectrometers with commercial Bruker AM or AV series
interfaces. IR spectra were obtained using a Mattson Instru-
ments Galaxy 3000 Fourier transform spectrometer. Elemental
analyses were performed at the University of California,
Berkeley Microanalytical Facility, on a Perkin-Elmer 2400
Series II CHNO/S analyzer.
Reagents were purchased from commercial vendors, checked
for purity, and used without further purification unless
otherwise noted. Liquids were degassed using three freeze-
pump-thaw cycles and dried over 4 Å activated molecular
sieves or passed through an activated alumina column. Solids
were stored in an inert-atmosphere glovebox. Toluene, ben-
zene, and pentane (UV grade, alkene free) were passed
through a column of activated alumina (A1, 12 × 32, Purifry
Co.) under nitrogen pressure and sparged with nitrogen prior
to use.49 Tetrahydrofuran and diethyl ether were distilled from
purple sodium/benzophenone ketyl under nitrogen. Benzene-
d6 and toluene-d8 were degassed by three freeze-pump-thaw
cycles, placed over 4 Å activated molecular sieves, and filtered
through glass-fiber filter paper prior to use. THF-d8 was
vacuum-transferred from purple sodium/benzophenone ketyl
and stored under an inert atmosphere. Addition of known
quantities of condensable gases was achieved using a known
Con clu sion
A technique developed for the synthesis of ruthenium
amido complex 1 was shown to be applicable to the
synthesis of its iron amido analog 2. Complex 2 was
found to be a strong base, although it is less basic than
1. The difference in basicity between 1 and 2 may be
understood in terms of the greater electron richness of
ruthenium relative to iron. Deprotonation of terminal
alkynes and phenylacetonitrile led to ammonia displace-
ment products, while the deprotonation of fluorene and
water did not lead to displacement of ammonia. The
mechanism of H/D exchange reactions between parent
amido complexes 1, 2, and 5 with very weak acids was
investigated by qualitatively comparing the relative
rates of H/D exchange and trans to cis isomerization of
trans-4 using the amido complexes as catalysts. These
experiments indicate that an ion pair is formed and that
trans-4 (due to its higher free energy of formation) is
more acidic than cis-4. Although our evidence suggests
that (presumably very tight) ion pair formation occurs
during the course of the H/D exchange, displacement
of ammonia was never observed, even in systems where
the counteranion that is presumably formed is not large.
It is not clear why ion pair formation in some systems
(49) Alaimo, P. J .; Peters, D. W.; Arnold, J .; Bergman, R. G. J . Chem.
Educ. 2001, 78, 64.

1666 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
volume bulb attached to a high-vacuum line equipped with a
digital MKS Baratron gauge.
quintets at -27.35 and -19.97 ppm, respectively) and 31P{1H}
NMR spectroscopy (doublet at 76.0 and singlet at 46.0 ppm,
respectively).
UV irradiation was performed using a 500 W medium-
pressure Hanovia mercury vapor lamp housed in a water-
cooled quartz immersion well. During irradiation, samples
were immersed in spectral grade ethylene glycol maintained
by a Neslab Endocal circulating refrigeration bath at 20 °C.
The following compounds were synthesized according to
literature procedures: trans-(dmpe)2Fe(Cl)2,50 trans-(dmpe)2Fe-
(H)(Cl),23 cis-(dmpe)2Fe(H)2,22 (dmpe)2Fe(H)(NH2) (2),22 trans-
(dmpe)2Ru(H)(NH2) (1),19 cis-(PMe3)2Ru(H)(NH2) (5),51 9,10-
dihydroanthracene-d4,40 1,3-di-tert-butylaziridinone,33 trans-
(dmpe)2Fe(H)(NHCHO) (6),22 and trans-(dmpe)2Fe(H)(N(p-
Tol)CHO).22 Details of some of the experiments discussed in
the text have also been described previously.22
(d m p e)2F e(H)(15NH2) (2-15N). To a heavy-walled glass
vessel with a fused Teflon stopcock were added a stir bar, THF
(2 mL), (dmpe)2Fe(H)(Cl) (200 mg, 0.51 mmol), and NaNH2
(33 mg, 0.82 mmol). The vessel was sealed and degassed on a
vacuum line using three freeze-pump-thaw cycles, after
which equal quantities of NH3 and 15NH3 were condensed into
the vessel (the total solution volume after warming to 25 °C
was 4 mL). The reaction vessel was warmed to 25 °C behind
a blast shield, and the mixture was stirred at this temperature
for 4 h. The ammonia was allowed to boil off under N2, and
the remaining volatile materials were removed in vacuo to give
an orange solid. The solid was extracted with pentane (40 mL),
and the extracts were filtered through Celite on a glass frit.
The resulting bright orange solution was concentrated to yield
2-15N (50% 15N label) as an orange, powdery solid (186 mg,
98%). The 1H and 31P{1H} NMR spectra of 2-15N were identical
with those obtained for 2. It is not clear why no splitting due
to the 15N label is observed. 15N{1H} NMR (C6D6; 50.5 MHz):
δ -456.3, -457.4 (mixture of trans and cis isomers; the
downfield peak is of greater intensity).
Ad d ition of 2 to [tr a n s-(d m p e)2Ru (H)(NH3)][C13H9]. The
ruthenium amido complex 1 (11 mg, 0.027 mmol), fluorene (4.5
mg, 0.027 mmol), and THF-d8 (0.3 mL) were added to an NMR
tube. After formation of [trans-(dmpe)2Ru(H)(NH3)][C13H9] was
confirmed by NMR spectroscopy (hydride at -19.78 ppm in
1
the H NMR spectrum and singlet at 40.8 ppm in the 31P{1H}
NMR spectrum),21 the iron amido complex 2 (10 mg, 0.027
mmol) was added to an NMR tube containing [trans-(dmpe)2Ru-
(H)(NH3)][C13H9] as a THF-d8 solution (0.2 mL). The 1H and
31P{1H} NMR spectra were consistent with those reported
above (hydride resonances at -19.41 and -26.65 ppm; 31P-
{1H} resonances at 73.3 and 43.1 ppm), indicating that complex
2 did not deprotonate [trans-(dmpe)2Ru(H)(NH3)][C13H9].
Rea ction of 2 w ith Wa ter . Water was degassed by
bubbling N2 through it for 1 h. A portion of the water (14.5
µL, 0.80 mmol) was then added via syringe to 1 mL of THF,
which was added to a stirred solution of 2 (100 mg, 0.27 mmol)
in THF (4 mL). The reaction mixture was stirred at 25 °C for
3 h, and the solvent was removed in vacuo to leave a yellow-
orange solid. The solid was extracted with pentane to give an
orange solution and leave behind a yellow solid (65 mg, 60%).
The pentane solution was found by NMR spectroscopy to
contain starting material and decomposition products. The
yellow solid was dissolved in THF and filtered through a glass
fiber filter and crystallized from THF at -35 °C to give a
product whose combustion and NMR analysis data were
consistent with [(dmpe)2Fe(H)(NH2)]2‚3H2O. An X-ray crystal-
lographic study revealed a dimeric structure: [trans-(dmpe)2Fe-
(H)(NH3)]2[OH]2‚2H2O (Figure 1). Note that the material used
for NMR and combustion analysis was dried in vacuo, while
the material used for X-ray diffraction analysis was not dried.
This difference in the purification method accounts for the
different molecular formulas found. 1H NMR (THF-d8; 400
MHz): δ 2.23 (br s, 10H, OH/NH), 2.15 (br m, 8H, PCH2), 1.58
(br m, 8H, PCH2), 1.52 (s, 24H, PCH3), 1.27 (s, 24H, PCH3),
-28.17 (quintet, 2H, J HP ) 47.2 Hz, FeH). 13C{1H} NMR (THF-
d8; 100.4 MHz): δ 32.5 (quintet, J CP ) 14.1 Hz, PCH2), 24.4
(m, PCH3), 15.2 (br m, PCH3). 31P{1H} NMR (THF-d8; 162
MHz): δ 70.5 (d, J PH ) 43.7 Hz). IR (Nujol): 1781 (m), 1647
(w), 1281 (w), 936 (s), 889 (m), 838 (w), 800 (w), 698 (m), 643
(m) cm-1. Anal. Calcd for C24H76N2O3P8Fe2: C, 36.02; H, 9.57;
N, 3.50. Found: C, 36.10; H, 9.65; N, 3.46.
[tr a n s-(d m p e)2F e(H)(NH3)][C13H9] (3). A THF solution (3
mL) of fluorene (45 mg, 0.27 mmol) was added dropwise to an
equal volume of a THF solution of 2 (100 mg, 0.27 mmol).
Within 5 s, precipitation of a bright orange, powdery solid was
observed. The mixture was stirred at 25 °C for 3 h. The volatile
materials were removed in vacuo to leave 140 mg (97%) of 3
as a bright orange crystalline solid. 1H NMR (THF-d8; 400
MHz): δ 7.82 (d, 2H, J ) 7.6 Hz, Ar H), 7.24 (d, 2H, J ) 8.3
Hz Ar H), 6.76 (t, 2H, J ) 7.0 Hz, Ar H), 6.39 (t, 2H, J ) 7.3
Hz, Ar H), 5.90 (s, 1H, Ar H), 1.42 (br m, 8H, PCH2), 1.13 (s,
12H, PCH3), 0.92 (s, 12H, PCH3), -1.61 (br s, 3H, NH3), -29.61
(quintet, 1H, J HP ) 48.8 Hz, FeH). 13C{1H} NMR (THF-d8;
100.4 MHz): δ 138.2 (s, C, aryl), 119.4 (s, CH, aryl), 119.2 (s,
CH, aryl), 118.7 (s, C, aryl), 117.0 (s, CH, aryl), 109.0 (s, CH,
aryl), 83.6 (s, CH, aryl), 31.8 (m, PCH2), 23.2 (m, PCH3), 13.8
(m, PCH3). 31P{1H} NMR (THF-d8; 162 MHz): δ 74.0 (d, J PH
) 45.4 Hz). IR (Nujol): 1816 (m), 1600 (w), 1568 (w), 1509
(m), 1377 (m), 1301 (s), 1220 (m), 1197 (w), 1154 (w), 1140
tr a n s-(d m p e)2F e(H)(NCC(H)P h ). Phenylacetonitrile (31
µL, 0.27 mmol) was added dropwise via syringe to a stirred
solution of 2 (100 mg, 0.27 mmol) in toluene (5 mL) in a 20
mL vial. The solution was stirred at 25 °C for 5 h, during which
time a yellow precipitate formed. The volatile materials were
removed in vacuo to leave trans-(dmpe)2Fe(H)(NCC(H)Ph) as
a canary yellow, powdery solid (100 mg, 82%). The product
was crystallized from toluene at -35 °C to give analytically
1
(w), 1063 (w), 936 (s), 838 (m), 715 (s), 636 (m), 571 (m) cm-1
.
pure material. H NMR (C6D6; 400 MHz): δ 7.31 (t, 2H, J )
7.5 Hz, Ar H), 7.08 (d, 2H, J ) 7.4 Hz, Ar H), 6.67 (t, 1H, J )
7.1 Hz, Ar H), 3.49 (s, 1H, NCC(H)Ph), 1.47 (br m, 8H, PCH2),
1.21 (s, 12H, PCH3), 0.94 (s, 12H, PCH3), -27.28 (quintet, 1H,
J HP ) 48.0 Hz, FeH). 13C{1H} NMR (C6D6; 100.4 MHz): δ 129.2
(s, C, aryl), 128.7 (s, CH, aryl), 128.5 (s, CH, aryl), 128.3 (s,
CH, aryl), 118.6 (s, NCC(H)Ph), 113.2 (s, NCC(H)Ph), 32.0
(quintet, J CP ) 13.4 Hz, PCH2), 22.4 (br m, PCH3), 15.5 (br m,
PCH3). 31P{1H} NMR (C6D6; 162 MHz): δ 72.0 (d, J PH ) 42.0
Hz). IR (Nujol): 2111 (s), 1920 (m), 1806 (m), 1749 (m), 1588
(s), 1419 (m), 1291 (m), 1226 (w), 1171 (s), 1119 (w), 1091 (w),
1062 (w), 981 (m), 929 (s), 888 (m), 731 (s), 700 (s) cm-1. Anal.
Calcd for C20H39NP4Fe: C, 50.76; H, 8.31; N, 2.96. Found: C,
50.59; H, 8.44; N, 3.34.
Anal. Calcd for C25H45NP4Fe: C, 55.67; H, 8.41; N, 2.60.
Found: C, 55.63; H, 8.37; N, 2.49.
Dep r oton a tion of 3 by 1. The iron amido complex 2 (10
mg, 0.027 mmol), fluorene (4.5 mg, 0.027 mmol), and THF-d8
(0.3 mL) were added to an NMR tube. Some orange precipitate
was observed upon mixing. After formation of 3 was confirmed
by 1H and 31P{1H} NMR spectroscopy (data same as above),
the NMR tube was opened in the glovebox and the ruthenium
amido complex 1 (11 mg, 0.027 mmol) was added as a solution
in THF-d8 (0.2 mL). The solution became a darker orange color
and the orange precipitate of 3 that had formed redissolved.
Formation of complex 2 and [trans-(dmpe)2Ru(H)(NH3)]-
[C13H9]21 was confirmed by 1H NMR spectroscopy (hydride
Rea ction of 2 w ith P r op yn e. A solution of 2 (10 mg, 0.027
mmol) and hexamethylbenzene internal standard (1.4 mg,
0.0090 mmol) in benzene-d6 (0.3 mL) was added to an NMR
tube, and the tube was fitted with a Cajon adaptor. The tube
(50) Chatt, J .; Hayter, R. G. J . Chem. Soc. 1961, 5507.
(51) Holland, A. W.; Bergman, R. G. J . Am. Chem. Soc. 2002, 124,
14684.

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1667
was degassed using three freeze-pump-thaw cycles, and
propyne (0.054 mmol) was added by vacuum transfer. The tube
was heated to 45 °C for 3 days. The NMR spectra indicated
formation of trans-(dmpe)2Fe(H)(CCCH3) as the major product
(60%, hydride at -18.71 ppm, 31P resonance at 75.5 ppm).32
An equivalent quantity of NH3 was also observed. Two other
species with hydride resonances at -20.26 (12%) and -32.00
ppm (20%) were also observed. Attempts to isolate trans-
(dmpe)2Fe(H)(CCCH3) were not successful.
Rea ction of 2 w ith Allen e. A solution of 2 (10 mg, 0.027
mmol) and hexamethylbenzene internal standard (1.4 mg,
0.0090 mmol) in benzene-d6 (0.3 mL) was added to an NMR
tube, and the tube was fitted with a Cajon adaptor. The tube
was degassed using three freeze-pump-thaw cycles, and
allene (0.054 mmol) was added by vacuum transfer. After 3
days at 45 °C, the hydride resonance of trans-(dmpe)2Fe(H)-
(CCCH3) was observed at -18.73 ppm (60% yield) with a
corresponding signal in the 31P{1H} NMR spectrum at 75.5
ppm.32 A second hydride resonance was observed at -20.26
ppm (12%). Attempts to isolate trans-(dmpe)2Fe(H)(CCCH3)
were not successful.
tr a n s-(d m p e)2F e(H)(NHC(O)C(H)(t-Bu )N(H)(t-Bu )). A
pentane solution of 2 (100 mg, 0.27 mmol) was slowly added
to a stirred solution of 1,3-di-tert-butylaziridinone (45 mg, 0.27
mmol) in pentane (total volume, 10 mL). The solution im-
mediately became cloudy, followed by formation of light orange
crystals. The solution remained a dark orange-red. The vial
was then cooled to -35 °C for 1 day. Yellow crystals of trans-
(dmpe)2Fe(H)(NHC(O)C(H)t-BuN(H)t-Bu) were isolated from
the pentane solution (107 mg, 74%). 1H NMR (C6D6; 400
MHz): δ 2.80 (d, 1H, J ) 8.0 Hz, NHC(CH3)3), 2.44 (d, 1H, J
) 7.6 Hz, CHC(CH3)3), 2.13 (br m, 2H, PCH2), 2.01 (br m, 2H,
PCH2), 1.63 (br s, 1H, NH), 1.55 (br m, 2H, PCH2), 1.43 (d,
9H, J ) 6.4 Hz, NHC(CH3)3), 1.36 (br m, 3H, PCH2, NH), 1.21
(s, 12H, PCH3), 1.20 (s, 12H, PCH3), 1.06 (d, J ) 8.0 Hz, 9H,
CHC(CH3)3), -28.41 (quintet, 1H, J HP ) 49.4 Hz, FeH). 13C-
{1H} NMR (C6D6; 100.4 MHz): δ 181.9 (s, NCO), 69.7 (s, CHC-
(CH3)3), 50.5 (s, NHC(CH3)3), 35.6 (s, CHC(CH3)3), 32.9 (m,
PCH2), 31.3 (s, NHC(CH3)3), 29.1 (s, CHC(CH3)3), 25.7 (br m,
PCH3), 24.9 (br m, PCH3), 18.0 (m, PCH3), 17.7 (m, PCH3).
31P{1H} NMR (C6D6; 162 MHz): δ 71.4 (m). IR (Nujol): 3270
(w), 1794 (s), 1580 (s), 1280 (s), 1262 (s), 1108 (m), 1079 (w),
1033 (w), 922 (s), 838 (m), 796 (m), 697 (s), 643 (s) cm-1. Anal.
Calcd for C22H54N2OP4Fe: C, 48.71; H, 10.03; N, 5.16. Found:
C, 48.71; H, 9.70; N, 5.13.
tr a n s-(d m p e)2Ru (H)(NHC(O)C(H)(t-Bu )N(H)(t-Bu )). A
pentane solution of 1 (50 mg, 0.12 mmol) was slowly added to
a stirred solution of 1,3-di-tert-butylaziridinone (20 mg, 0.12
mmol) in pentane (total volume, 10 mL). The solution im-
mediately became cloudy, followed by formation of a white
precipitate. The solution remained light yellow. The vial was
then cooled to -35 °C for 1 day. Colorless crystals of trans-
(dmpe)2Ru(H)(NHC(O)C(H)(t-Bu)N(H)(t-Bu)) were isolated from
the pentane solution (43 mg, 61%). 1H NMR (C6D6; 400 MHz):
δ 2.81 (d, 1H, J ) 8.4 Hz, NHC(CH3)3), 2.71 (br s, 1H, NH),
2.59 (d, 1H, J ) 8.4 Hz, CHC(CH3)3), 1.99 (br m, 4H, PCH2),
1.79 (br m, 4H, PCH2), 1.43 (s, 6H, PCH3), 1.35 (s, 6H, PCH3),
1.28 (s, 9H, NH(CH3)3), 1.25 (s, 9H, CH(CH3)3), 1.16 (d, 12H,
J ) 4.0 Hz, PCH3), -18.79 (quintet, 1H, J HP ) 22.0 Hz). 13C-
{1H} NMR (C6D6; 100.4 MHz): δ 172.6 (s, NCO), 69.7 (s, CHC-
(CH3)3), 50.6 (s, NHC(CH3)3), 35.5 (s, CHC(CH3)3), 32.5 (quin-
tet, J CP ) 13.6 Hz, PCH2), 32.1 (quintet, J CP ) 13.6 Hz, PCH2),
31.3 (s, NHC(CH3)3), 29.1 (s, CHC(CH3)3), 25.1 (m, PCH3), 24.2
(m, PCH3), 18.2 (m, PCH3), 17.6 (m, PCH3). 31P{1H} NMR
(C6D6; 162 MHz): δ 43.8 (m). IR (Nujol): 3339 (w), 3280 (w),
1868 (s), 1582 (s), 1286 (m), 1237 (m), 1171 (w), 1108 (w), 1032
(w), 933 (s), 887 (m), 840 (m), 797 (m), 760 (w), 723 (s), 699
(s), 645 (m) cm-1. Anal. Calcd for C22H54N2OP4Ru: C, 44.97;
H, 9.26; N, 4.77. Found: C, 45.23; H, 9.56; N, 4.56.
mmol) in THF (5 mL) was added to a glass vessel with a fused
Teflon stopcock. Degassed water (2.7 µL, 0.15 mmol) was added
to the stirred solution of 1 with a syringe. The solution was
stirred for 1 h, and the volatile materials were removed in
vacuo. The hydroxide complex was isolated as a cream-colored
powdery solid (60 mg, 95%), and crystallization from THF gave
clean material whose NMR spectra were identical with those
previously reported.52
tr a n s-(d m p e)2Ru (H)(OC(O)C(H)(t-Bu )N(H)(t-Bu )). In
an NMR tube, solutions of trans-(dmpe)2Ru(H)(OH) (6.1 mg,
0.015 mmol) and 1,3-di-tert-butylaziridinone (2.5 mg, 0.015
mmol) in C6D6 (0.3 mL total volume) were mixed. After 20 h,
complete conversion to product was observed by 1H and 31P-
{1H} NMR spectroscopy. The solution was concentrated in
vacuo to give 6.6 mg (77%) of trans-(dmpe)2Ru(H)(OC(O)C(H)-
(t-Bu)N(H)(t-Bu)) as a colorless solid. The product was crystal-
lized from THF layered with pentane at -35 °C to give
1
analytically pure material. H NMR (C6D6; 400 MHz): δ 2.83
(d, 1H, J ) 8.4 Hz, NHC(CH3)3), 2.52 (d, 1H, J ) 7.6 Hz, CHC-
(CH3)3), 2.02 (br m, 4H, PCH2), 1.85 (br m, 4H, PCH2), 1.46 (s,
6H, PCH3), 1.41 (s, 6H, PCH3), 1.28 (s, 9H, NHC(CH3)3), 1.23
(s, 9H, CC(CH3)3), 1.13 (s, 6H, PCH3), 1.09 (s, 6H, PCH3),
-23.28 (quintet, 1H, J HP ) 21.9 Hz, Ru H). 13C{1H} NMR
(C6D6; 100.4 MHz): δ 179.6 (s, NCO), 67.2 (s, CHC(CH3)3), 50.6
(s, NHC(CH3)3), 35.3 (s, CHC(CH3)3), 31.9 (m, PCH2), 31.2 (s,
NHC(CH3)3), 29.2 (s, CHC(CH3)3), 24.1 (quintet, J CP ) 13.1
Hz, PCH3), 23.1 (quintet, J CP ) 13.1 Hz, PCH3), 18.0 (quintet,
J CP ) 9.5 Hz, PCH3), 17.2 (quintet, J CP ) 9.5 Hz, PCH3). 31P-
{1H} NMR (C6D6; 162 MHz): δ 43.6 (m). IR (benzene-d6): 2967
(s), 2903 (s), 1922 (m), 1839 (w), 1588 (s), 1473 (w), 1356 (m),
1278 (w), 1231 (w), 937 (s), 888 (m), 728 (m), 644 (m) cm-1
.
Anal. Calcd for C22H53NO2P4Ru: C, 44.89; H, 9.08; N, 2.38.
Found: C, 44.41; H, 9.07; N, 2.24.
H/D Exch a n ge betw een 2 a n d Tolu en e-d 8. Complex 2
(10 mg, 0.027 mg) was dissolved in toluene-d8 (0.3 mL) in a J .
Young NMR tube. Integration relative to the residual aryl
solvent peaks revealed that the intensity of the toluene methyl
peak increased, while the NH2 and dmpe resonances of 2
decreased in intensity; the hydride resonances of 2 remained
unchanged. After 3 days at 25 °C, the NH2 and dmpe
resonances of 2 were no longer observed. The 2H{1H} spectrum
after removal of the deuterated solvent in vacuo and dissolu-
tion of the remaining orange solid in benzene-h6 revealed
deuterium signals corresponding to the dmpe and NH2 ligands
of 2; no resonance in the hydride region was observed.
tr a n s-(d m p e)2F e(H)(CH2P h ). A 1.0 M diethyl ether solu-
tion of benzylmagnesium chloride (0.61 mL, 0.61 mmol) was
added to a stirred solution of trans-(dmpe)2Fe(H)(Cl) (200 mg,
0.51 mmol) in THF (5 mL). The orange solution was stirred
for 3 h, and the solvent was removed in vacuo. The orange
solid was extracted with pentane (10 mL) and filtered through
a glass fiber filter. The pentane was removed at reduced
pressure to give an orange solid. The product was crystallized
from diethyl ether at -35 °C to give a clean product (73 mg,
1
32%). H NMR (C6D6; 400 MHz): δ 7.12-7.02 (m, 4H, Ar H),
6.88 (t, 1H, J ) 7.0 Hz, Ar H), 1.63 (br m, 4H, PCH2), 1.38 (br
m, 4H, PCH2), 1.19 (s, 12H, PCH3), 1.10 (s, 12H, PCH3), 1.07
(quintet, 2H, J HP ) 6.0 Hz, FeCH2Ph), -23.3 (quintet, 1H, J HP
) 48.00 Hz, FeH). 13C{1H} NMR (C6D6; 100.4 MHz): δ 165.4
(quintet, J CP ) 3.0 Hz, C, aryl), 128.1 (s, CH, aryl), 127.0 (s,
CH, aryl), 118.6 (s, CH, aryl), 32.4 (quintet, J CP ) 13.5 Hz,
PCH2), 28.1 (quintet, J CP ) 6.4 Hz, PCH3), 16.2 (s, quintet,
J CP ) 2.9 Hz, PCH3), 8.8 (quintet, J CP ) 10.5 Hz, FeCH2Ph).
31P{1H} NMR (C6D6; 162 MHz): δ 73.3 (s). IR (Nujol): 1738
(s), 1588 (m), 1422 (m), 1272 (m), 1206 (m), 1172 (w), 1015
(w), 982 (w), 927 (s), 884 (m), 832 (w), 788 (w), 748 (w), 691
(s), 637 (m) cm-1. Anal. Calcd for C19H40P4Fe: C, 50.91; H,
8.99. Found: C, 51.07; H, 9.24.
tr a n s-(d m p e)2Ru (H)(OH). Water was degassed by bub-
bling N2 through it for 30 min. A solution of 1 (63 mg, 0.15
(52) Burn, M. J .; Fickes, M. G.; Hartwig, J . F.; Hollander, F. J .;
Bergman, R. G. J . Am. Chem. Soc. 1993, 115, 5875.

1668 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
Th er m olysis of tr a n s-(d m p e)2F e(H)(CH 2P h ) in th e
P r esen ce of NH3. A solution of trans-(dmpe)2Fe(H)(CH2Ph)
(10 mg, 0.022 mmol) in 0.3 mL of THF-d8 was added to a J .
Young NMR tube. The tube was degassed using three freeze-
pump-thaw cycles, and 1 atm of NH3 was added. The reaction
mixture was stable indefinitely at 25 °C. After 5 h at 45 °C,
toluene was observed in the 1H NMR spectrum, but no
evidence for formation of 2 was apparent. Further heating at
45 °C resulted in further decomposition of the benzyl complex.
Dissolu tion of tr a n s-(d m p e)2F e(H)(CH2P h ) in Liqu id
NH3. A stirbar and trans-(dmpe)2Fe(H)(CH2Ph) (18 mg, 0.040
mmol) were added to a heavy-walled glass vessel with a fused
Teflon stopcock. The vessel was evacuated and frozen at -196
°C, and NH3 was condensed into it such that, when it was
warmed to ambient temperature behind a blast shield, the
total volume of the mixture was 5 mL. The mixture was stirred
for 3 h at ambient temperature. The substrate was only
slightly soluble in the NH3 solvent, resulting in a reaction
mixture with the appearance of orange juice. After 3 h, the
NH3 was allowed to boil off under N2. The yellow solid that
remained was dissolved in benzene-d6, and NMR spectra
indicated that no reaction had occurred.
Rea ction of tr a n s-(d m p e)2F e(H)(CH2P h ) w ith p-Th io-
cr esol. THF-d8 solutions of trans-(dmpe)2Fe(H)(CH2Ph) (10
mg, 0.022 mmol) and p-thiocresol (2.7 mg, 0.022 mmol) were
mixed (total volume 0.3 mL) and added to a J . Young NMR
tube. Upon mixing, the solution immediately became a darker
orange. NMR analysis indicated clean formation of a mixture
of the cis and trans isomers of (dmpe)2Fe(H)(S-p-Tol) in a 1:2
ratio with loss of toluene. The identity of the products was
confirmed by comparison of the observed FeH resonances with
those previously reported for (dmpe)2Fe(H)(SPh).53
tr a n s-(d m p e)2F e(H)(O-p-Tol). THF solutions of trans-
(dmpe)2Fe(H)(CH2Ph) (220 mg, 0.49 mmol) and p-cresol (64
mg, 0.59 mmol) were mixed (total volume, 5 mL) in a 20 mL
vial. Upon mixing, the solution immediately became a darker
orange. After 2 h, the volatile materials were removed in vacuo
to give trans-(dmpe)2Fe(H)(OTol) as a dark orange crystalline
solid (140 mg, 60%). The product was crystallized from toluene
at -35 °C to give 77 mg (33%) of analytically pure material.
1H NMR (THF-d8; 400 MHz): δ 7.04 (d, 2H, J ) 8.3 Hz, Ar
H), 6.15 (d, 2H, J ) 7.6 Hz, Ar H), 2.36 (s, 3H, ArCH3), 1.96
(br m, 4H, PCH2), 1.46 (br m, 4H, PCH2), 1.31 (s, 12H, PCH3),
1.05 (s, 12H, PCH3), -35.02 (quintet, 1H, J HP ) 49.8 Hz, FeH).
13C{1H} NMR (THF-d8; 100.4 MHz): δ 171.8 (s, CO, aryl),
129.8 (s, CH, aryl), 120.3 (s, CH, aryl), 117.2 (s, CCH3, aryl),
32.3 (quintet, J CP ) 13.2 Hz, PCH2), 23.6 (quintet, J CP ) 6.6
Hz, PCH3), 21.3 (s, ArCH3), 15.8 (br m, PCH3). 31P{1H} NMR
(THF-d8; 162 MHz): δ 70.9 (dd, J ) 32.4, 13.0 Hz). IR (Nujol;
cm-1): 1789 (s), 1601 (s), 1503 (s), 1378 (w), 1334 (s), 1291 (s),
1155 (m), 1094 (m), 1067 (w), 990 (w), 884 (m), 816 (m), 696
(s), 642 (s). Anal. Calcd for C19H40OP4Fe: C, 49.15; H, 8.68.
Found: C, 49.36; H, 8.96.
H/D Exch a n ge betw een 2 a n d H2 in Tolu en e-d 8. A
solution of 2 (10 mg, 0.027 mmol) in toluene-d8 (0.3 mL) was
added to a J . Young NMR tube. The tube was degassed using
three freeze-pump-thaw cycles, and H2 (1 atm) was added
to the tube. The resonances corresponding to the dmpe and
NH2 ligands of 2, and of H2, were observed to decrease in
intensity relative to the residual aryl resonances of the solvent.
H/D Exch a n ge betw een 2 a n d Am m on ia in Tolu en e-
d 8. A solution of 2 (10 mg, 0.027 mmol) in toluene-d8 (0.3 mL)
was added to a J . Young NMR tube. The tube was degassed
using three freeze-pump-thaw cycles, and NH3 (0.054 mmol)
was added to the tube by vacuum transfer. The resonances
corresponding to the dmpe and NH2 ligands of 2, and of
ammonia, were observed to decrease in intensity relative to
the residual aryl resonances of the solvent.
Ad d ition of 15NH3 to 2. A solution of 2 (10 mg, 0.027 mmol)
in benzene-d6 (0.3 mL) was added to a J . Young NMR tube.
The tube was degassed using three freeze-pump-thaw cycles,
and 15NH3 (0.054 mmol) was added to the tube by vacuum
transfer. Over the course of 3 days, no 14NH3 was observed by
1H NMR spectroscopy and the quantity of 15NH3 did not
decrease, indicating that 15N was not incorporated into 2.
H /D E xch a n ge b et w een 2 a n d Cycloh ep t a t r ien e in
Tolu en e-d 8. Solutions of 2 (10 mg, 0.027 mmol) and cyclo-
heptatriene (5.1 µL, 0.054 mmol) in toluene-d8 (0.3 mL) were
added to a J . Young NMR tube. The resonances corresponding
to the dmpe and NH2 ligands of 2, and all the resonances of
cycloheptatriene, were observed to decrease in intensity rela-
tive to the residual aryl resonances of the solvent.
H/D Exch a n ge betw een 2 a n d Cya n ocyclop r op a n e in
Tolu en e-d 8. Solutions of 2 (10 mg, 0.027 mmol) and cyano-
cyclopropane (2.0 µL, 0.027 mmol) in toluene-d8 (0.3 mL) were
added to a J . Young NMR tube. Over the course of 46 h, the
cyanocyclopropane resonances at 0.30 and 0.00 ppm were
observed to decrease in intensity relative to the residual
solvent aryl resonances, as did the resonances corresponding
to the dmpe and NH2 ligands of 2. At t ) 10 min, the ratio of
the cyanocyclopropane resonances to the aryl peaks was 21:1,
and after 46 h, the ratio was 10:1.
H/D Exch a n ge betw een 2 a n d P h 3CH in Tolu en e-d 8.
Solutions of 2 (10 mg, 0.027 mmol) and triphenylmethane (13
mg, 0.054 mmol) in toluene-d8 (0.3 mL) were added to a J .
Young NMR tube. The resonances corresponding to the dmpe
and NH2 ligands of 2, and those of the methine hydrogen of
triphenylmethane, were observed to decrease in intensity
relative to the residual aryl resonances of the solvent.
Attem p ted H/D Exch a n ge betw een 2 a n d P r op en e in
Tolu en e-d 8. A solution of 2 (10 mg, 0.027 mmol) in toluene-
d8 (0.3 mL) was added to a J . Young NMR tube. The tube was
degassed using three freeze-pump-thaw cycles, and propene
(0.054 mmol) was added to the tube by vacuum transfer. While
the resonances corresponding to the dmpe and NH2 ligands of
2 were observed to decrease in intensity, no decrease ((5%)
in intensity of the propene resonances was observed.
At t em p t ed H /D E xch a n ge b et w een
2 a n d tr a n s-
(d m p e)2F e(H)(Cl) in Tolu en e-d 8. Complex 2 (5.0 mg, 0.013
mmol) and trans-(dmpe)2Fe(H)(Cl) (5.0 mg, 0.013 mmol) were
dissolved in toluene-d8, and the solution was added to an NMR
tube that was flame-sealed. Over the course of 2 days,
observation of the 1H NMR spectrum showed that while the
resonances corresponding to the dmpe and NH2 ligands of 2
decreased in intensity relative to the residual aryl resonances
of the solvent, the dmpe resonances of trans-(dmpe)2Fe(H)-
(Cl) did not decrease in intensity.
Isom er iza tion of 1,3-Cycloh exa d ien e by 2. Benzene-d6
solutions of 2 (10 mg, 0.027 mmol) and 1,3-cyclohexadiene (5.1
µL, 0.076 mmol) were mixed (total volume 0.3 mL) and added
to a J . Young NMR tube. After 1 day, 1,3-cyclohexadiene and
1,4-cyclohexadiene were observed in an equilibrium ratio of
2:1.54 Complex 2 remained unreacted.
tr a n s-9,10-Diisopr op yl-9,10-d ih ydr oa n th r a cen e (tr a n s-
4). The compound was synthesized using a procedure similar
to that reported by Harvey and Davis for the synthesis of other
9,10-dialkyl-9,10-dihydroanthracenes.35 To a solution of 9,10-
dihydroanthracene (2.0 g, 11 mmol) in diethyl ether (50 mL)
and liquid ammonia (50 mL) at -35 °C was added 18 mL of a
1.5 M solution of n-butyllithium in hexane (27 mmol). The
bright red-orange solution was stirred for 30 min at -35 °C,
and 2-bromopropane (4.2 mL, 44 mmol) was added via syringe,
resulting in the solution becoming first yellow and then
colorless after 10 min. After the mixture was stirred for 30
min at -35 °C, water (15 mL) was added and the ammonia
was allowed to evaporate. Diethyl ether (75 mL) was added,
and the organic solution was washed with brine (3 × 20 mL)
(53) Boyd, S. E.; Field, L. D.; Hambley, T. W.; Young, D. J . Inorg.
Chem. 1990, 29, 1496.
(54) Taskinen, E.; Nummelin, K. J . Org. Chem. 1985, 50, 4844.

Iron Parent Amido Complex
Organometallics, Vol. 23, No. 8, 2004 1669
and dried with MgSO4. Evaporation of the diethyl ether on a
rotary evaporator led to formation of a thick yellow oil. After
2 days, clear, colorless crystals formed in the oil. These were
isolated and recrystallized from diethyl ether to give 432 mg
(15% yield) of trans-4 as white crystals. 1H NMR (C6D6; 400
MHz): δ 7.27 (m, 4H, aryl), 7.14 (m, 4H, aryl), 3.75 (d, 2H, J
) 4.8 Hz, benzylic CH), 2.25 (m, 2H, CH(CH3)2), 0.91 (d, 12H,
J ) 7.2 Hz, CH(CH3)2). Lit.36 1H NMR (no solvent indicated):
7.25 (m), 3.78 (J ) 5.0 Hz), 0.97 (J ) 6.8 Hz).
CH(CH3)2), 0.97 (d, 12H, J ) 8.8 Hz, CH(CH3)2). Lit.36 1H NMR
(no solvent indicated): 7.20 (m), 3.27 (J ) 9.5 Hz), 1.00 (J )
6.2 Hz).
H/D Exch a n ge betw een cis-4-d 2 a n d 1. A THF-d8 solution
of 1 (1.0 mg, 0.0024 mmol) was added to a THF-d8 solution of
cis-4-d2 (13 mg, 0.048 mmol), and the resulting solution (0.3
mL total volume) was added to an NMR tube. The 1H NMR
1
spectra after 10 min and 1 and 20 h revealed H incorporation
into the benzylic position of the substrate of 6.6%, 10%, and
18%, respectively.
Isom er iza tion of tr a n s-4 by 5. In a typical experiment,
benzene-d6 stock solutions of 5 and substrate were mixed such
that 0.018 mmol of 5 and 0.025 mmol of trans-4 were used.
The resulting solution was diluted to 1.0 mL and transferred
to an NMR tube that was subsequently flame-sealed. The
Isom er iza t ion of tr a n s-4-d 2 by t-Bu NHLi/t-Bu NH 2.
Solutions of trans-4-d2 (10 mg, 0.038 mmol) and t-BuNHLi (0.3
mg, 0.004 mmol) in THF-d8 (0.3 mL total volume) were mixed
and added to an NMR tube. t-BuNH2 (7.9 µL, 0.075 mmol) was
added to the tube with a syringe. The tube was then flame-
sealed. The tube was heated to 45 °C, and the reaction progress
1
progress of the reaction was monitored by H NMR spectros-
copy at regular intervals. The isomerization was repeated eight
times, and the isomerization was found to be first order in both
1
was monitored by H NMR spectroscopy at regular intervals.
After 15 h, 51% conversion was observed, with 46% H in the
benzylic position of the product and 1% H in the benzylic
position of the substrate.
catalyst and substrate with k ) (2.18 ( 0.11) × 10-3 M-1 s-1
.
t r a n s-9,10-Diisop r op yl-9,10-d ih yd r oa n t h r a ce n e -d 2
(tr a n s-4-d 2). The deuterio analog was synthesized using the
method described above with 9,10-dihydroanthracene-d440 (1.4
g, 7.6 mmol) as the substrate, 18 mmol of n-butyllithium, and
2.0 mL (30 mmol) of 2-bromopropane. The product was purified
by column chromatography (100% hexanes; silica gel) followed
by recrystallization from pentane to give 240 mg (20%) of clean
H/D Exch a n ge betw een cis-4-d 2 a n d t-Bu NH2 Ca ta -
lyzed by t-Bu NHLi. Solutions of cis-4-d2 (10 mg, 0.038 mmol)
and t-BuNHLi (0.3 mg, 0.004 mmol) in THF-d8 (0.3 mL total
volume) were mixed and added to an NMR tube. t-BuNH2 (7.9
µL, 0.075 mmol) was added to the tube with a syringe. The
tube was then flame-sealed. The tube was heated to 45 °C,
and the reaction progress was monitored by 1H NMR spec-
troscopy at regular intervals. After 15 h, 6% H was observed
in the benzylic position of the substrate.
1
trans-4-d2. The H NMR spectrum was identical with that of
the nondeuterated analog, except that the benzylic resonance
was not observed (>95% deuteration).
Isom er iza tion of tr a n s-4-d 2 by 1. A benzene-d6 solution
of 1 (1.0 mg, 0.0024 mmol) was added to a benzene-d6 solution
of trans-4-d2 (13 mg, 0.048 mmol), and the resulting solution
(0.3 mL total volume) was added to an NMR tube that was
subsequently flame-sealed. The 1H NMR spectrum after 10
min revealed that conversion to the cis product was 64%
complete. Additionally, the spectrum revealed that while the
Isom er iza tion of tr a n s-4 by 2-d 34. Solutions of 2-d34 (0.8
mg, 0.002 mmol) and trans-4 (10 mg, 0.038 mmol) in THF-d8
(0.3 mL) were mixed and added to an NMR tube. The relative
quantities of cis-4 and trans-4, and the amount of H in the
benzylic position of each, were observed at 10 min and 1, 3,
1
and 21 h by H NMR spectroscopy (Table 1).
tr a n s-(d m p e)2F e(H)(15NHCHO) (6-15N). CO (1 atm) was
added to a degassed J . Young NMR tube containing a benzene-
d6 solution (0.3 mL) of 2-15N (10 mg, 0.027 mmol). The tube
was shaken for 4 h at 25 °C. The 1H and 31P{1H} NMR spectra
of 6-15N (50% 15N label) were identical with those obtained for
6. 15N{1H} NMR (C6D6; 50.5 MHz): δ -24.1 (s).
1
remaining trans starting material contained only 10% H in
the benzylic position, the cis product contained 43% 1H in the
benzylic position. After 1.5 h, conversion to product was
complete and the product contained 40% 1H in the benzylic
position.
cis-(d m p e)2F e(D)2. trans-(dmpe)2Fe(Cl)2 (1.0 g, 2.3 mmol),
sodium (2.0 g, 87 mmol, cut finely), and THF (40 mL) were
added to a 1 L glass vessel with a fused Teflon stopcock. The
mixture was degassed using three freeze-pump-thaw cycles,
and D2 (2.5 atm) was added. The reaction mixture was stirred
at ambient temperature until the sodium became a single large
clump (1 week). The volatile materials were removed at
reduced pressure, the purple-brown residue was extracted with
benzene and toluene (70 mL each), and the extracts were
filtered through Celite on a glass frit. The solvent was removed
Isom er iza tion of tr a n s-4-d 2 by 5. A THF-d8 solution of
trans-4-d2 (7.5 mg, 0.028 mmol) was added to a 1.0 mL
volumetric flask. A THF-d8 solution of 5 (6.9 mg, 0.016 mmol)
was then added to the same flask, and the solution was diluted
to 1.0 mL and transferred to an NMR tube that was subse-
quently flame-sealed. The progress of the isomerization and
H/D exchange was observed by 1H NMR spectroscopy at
regular intervals. First-order decay of trans-4 was observed
for the isomerization with k ) (7.2 ( 0.4) × 10-4 M-1 s-1; the
cis product contained 50% hydrogen in the benzylic position.
1
in vacuo to give a yellow-orange solid (0.82 g, 97%). H NMR
cis-9,10-Diisop r op yl-9,10-d ih yd r oa n th r a cen e-d 2 (cis--
4-d 2). The procedure of Gleiter et al. for the deuteration of
9,10-dihydroanthracene was used40 with trans-4 as substrate.
A 20 mL DMSO-d6 solution of NaH (5.6 mg, 0.23 mmol) was
heated in a glass vessel with a fused Teflon stopcock at 90 °C
for 1 h. trans-4 (155 mg, 0.59 mmol) was added, and the (now
pink) solution was stirred at 80 °C for 1 h. The reaction
mixture was quenched by addition of D2O (75 µL, 4.0 mmol),
and the solution turned orange. After the mixture was stirred
for 10 min, H2O (20 mL) was added, leading to the precipita-
tion of the product as a white powdery solid. The product was
collected by filtration and washed with 5 mL of H2O. Analysis
of the crude product by 1H NMR spectroscopy revealed 95%
deuteration in the benzylic position. Purification by column
chromatography (100% hexanes; silica gel) gave a white
crystalline solid (100 mg, 64%) that was observed to be clean
cis-4-d2 by 1H NMR spectroscopy with 95% deuteration. 1H
NMR (C6D6; 400 MHz): δ 7.12 (m, 8H, aryl), 3.27 (d, 0.1H, J
) 10.0 Hz, benzylic CH), 1.76 (septet, 2H, J ) 8.8 Hz,
(C6D6; 400 MHz): δ 1.58 (br s, 10H, PCH2/PCH3), 1.35 (br m,
4H, PCH2), 1.27 (s, 12H, PCH3), 1.08 (d, 6H, J HP ) 4.4 Hz,
PCH3). 2H{1H} NMR (C6D6; 61.6 MHz): δ -13.9 (br m) 31P-
{1H} NMR (C6D6; 162 MHz): δ 77.7 (undecet, J ) 9.1 Hz),
67.8 (t, J ) 26.7 Hz).
Rea ction of 2 w ith CO in th e P r esen ce of cis-(d m p e)2-
F e(D)2. Solutions of 2 (10 mg, 0.027 mmol) and cis-(dmpe)2Fe-
(D)2 (19 mg, 0.054 mmol) in benzene-d6 and benzene-h6 (each
0.3 mL) were added to two J . Young NMR tubes. The tubes
were degassed using three freeze-pump-thaw cycles, and 1
atm of CO was added to each. The tubes were shaken for 12
h, and NMR spectroscopy (1H, 2H{1H}, and 31P{1H}) indicated
that 8 was formed with no incorporation of deuterium; cis-
(dmpe)2Fe(D)2 remained unreacted.
Rea ction of 2 w ith CO in th e P r esen ce of (d m p e)2F e-
(
13CO). A benzene-d6 (0.3 mL) solution of cis-(dmpe)2Fe(H)2
(10 mg, 0.028 mmol) was added to a J . Young NMR tube. The
tube was degassed using three freeze-pump-thaw cycles, and
1 atm of 13CO was added. The tube was irradiated for 3.5 h,

1670 Organometallics, Vol. 23, No. 8, 2004
Fox and Bergman
at which time NMR spectroscopy indicated complete conver-
sion to (dmpe)2Fe(13CO) (13C{1H} quintet at δ 222.9 ppm, 31P-
{1H} doublet at 64.8 ppm). The tube was again degassed, and
2 (10 mg, 0.027 mmol) was added to the solution. The tube
was degassed, and 1 atm of unlabeled CO was added to the
tube. After the tube was shaken for 3.5 h, the 13C label was
hydroxide, and water of crystallization were located from a
difference Fourier map and assigned on the basis of the
reasonable hydrogen-bond geometries they displayed. These
hydrogens were assigned fixed Biso values (Biso ) 2.85), and
their positions were refined (all other hydrogen atoms were
included in calculated positions but were not refined). The data
were collected using 10 s frames with an ω scan of 0.3°.
Empirical absorption corrections based on comparison of
redundant and equivalent reflections were applied using
SADABS56 (Tmax ) 0.91, Tmin ) 0.78). The maximum and
minimum peaks on the final difference Fourier map correspond
to 0.39 and -0.49 e/Å3, respectively.
1
not observed in insertion product 6 by H and 13C{1H} NMR
spectroscopy, and no loss of the label was observed in (dmpe)2Fe-
(13CO).
X-r a y Str u ctu r e Deter m in a tion for [tr a n s-(d m p e)2F e-
(H)(NH3)]2[OH]2‚H2O. The X-ray crystal structure was ob-
tained by Dr. F. J . Hollander and Dr. A. G. Oliver at the UC
Berkeley X-ray facility (CHEXRAY). The crystals were mounted
on glass fibers using Paratone N hydrocarbon oil. All measure-
ments were made on a SMART CCD area detector with
graphite-monochromated Mo KR radiation (λ ) 0.710 69 Å).
The data were collected with a detector position of 60.00 mm.
Data were integrated by the program SAINT and were
corrected for Lorentz and polarization effects. Data were
analyzed for agreement and possible absorption using XPREP.55
The structure was solved by direct methods and expanded
using Fourier techniques.
The coordinated ammonia of the iron complex forms hydro-
gen bonds to two symmetry-related hydroxide anions (N1-
O1 ) 2.95 Å; N1-O1′ ) 2.88 Å), each of which are also
hydrogen bonded to the ammonia of another symmetry-related
iron complex and to the water of crystallization (O1-O2 ) 2.55
Å). The H-bonding pattern is such that the two iron complexes
and two H-bonded hydroxides form a rhomboidal dimer about
the inversion center with two waters of crystallization attached
to the sides. The hydrogen atoms of the hydride, ammonia,
Ack n ow led gm en t. We are grateful for financial
support from the National Science Foundation (Grants
No. CHE-0094349 and CHE-0345488). We thank Dr. F.
J . Hollander and Dr. A. G. Oliver for determining the
solid-state structure of [trans-(dmpe)2Fe(H)(NH3)]2-
[OH]2‚2H2O, and Mr. D. H. Leung for performing the
X-ray diffraction study of trans-(dmpe)2RuH(NHC(O)-
CH(t-Bu)(NH)t-Bu). Drs. J . Robin Fulton, Andrew W.
Holland, and Daniela Rais are gratefully acknowledged
for helpful discussions. The Center for New Directions
in Organic Synthesis is supported by Bristol-Myers
Squibb as a sponsoring member and Novartis as a
supporting member.
Su p p or tin g In for m a tion Ava ila ble: Crystallographic
data for [trans-(dmpe)2Fe(H)(NH3)]2[OH]2‚2H2O as CIF files.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OM0499660
(55) XPREP (v. 5.03), part of the SHELXTL Crystal Structure
Determination; Siemens Industrial Automation, Inc., Madison, WI,
1995.
(56) SADABS: Siemens Area Detector Absorption Correction Pro-
gram: Sheldrick, G. M. Advance copy, private communication, 1996.