The importance of cyclopentadienyl substituent effects in group 4 metallocene dinitrogen chemistry

Article abstract of DOI:10.1139/v05-009

This article highlights some of our recent efforts and presents new data on the importance of cyclopentadienyl substituent effects on group 4 metallocene dinitrogen chemistry. Reactions such as the coordination of N₂ to an isolated titanium sandwich complex, alkali-metal reductions of zirconocene dihalide complexes, alkane reductive elimination reactions, and the hydrogenation of zirconium dinitrogen complexes are all extremely sensitive to the groups present on the cyclopentadienyl rings. These results are promising for the future of N₂ fixation, as the reactivity of a specific metallocene can be dramatically altered by subtle manipulations in ligand substituents.

Full text of DOI:10.1139/v05-009

286
The importance of cyclopentadienyl substituent
effects in group 4 metallocene dinitrogen
chemistry1
Jaime A. Pool and Paul J. Chirik
Abstract: This article highlights some of our recent efforts and presents new data on the importance of cyclopenta-
dienyl substituent effects on group 4 metallocene dinitrogen chemistry. Reactions such as the coordination of N2 to an
isolated titanium sandwich complex, alkali-metal reductions of zirconocene dihalide complexes, alkane reductive elimi-
nation reactions, and the hydrogenation of zirconium dinitrogen complexes are all extremely sensitive to the groups
present on the cyclopentadienyl rings. These results are promising for the future of N2 fixation, as the reactivity of a
specific metallocene can be dramatically altered by subtle manipulations in ligand substituents.
Key words: cyclopentadienyl, zirconium, dinitrogen, ammonia, sandwich.
Résumé : Ce travail met en relief des travaux récents de notre équipe et présente de nouvelles données sur
l’importance des effets des substituants du cyclopentadiényle sur la chimie du diazote des métallocènes du groupe 4.
Les réactions, telle la coordination du N2 sur un complexe sandwich isolé, les réductions par les métaux alcalins des
complexes dihalogénés du zirconocène, les réactions d’éliminations réductrices d’alcane et l’hydrogénation des com-
plexes de diazote du zirconium sont toutes extrêmement sensibles aux groupes présents sur les cycles cyclopentadiény-
les. Ces résultats sont remplis de promesses pour la fixation future du N2 puisque la réactivité spécifique d’un
métallocène spécifique peut être dramatiquement altérée par des manipulations subtiles des ligands substituants.
Mots clés : cyclopentadiényle, zirconium, diazote, ammoniac, sandwich.
[Traduit par la Rédaction] Pool and Chirik 295
Introduction
where ruthenium dihydrogen complexes serve as the H-atom
source (7), and molybdenum compounds that are catalyti-
cally competent (8, 9).
The isolation and characterization of the first homoge-
neous transition metal dinitrogen complex, [(H3N)5Ru(N2)]2+,
by Allen and Senoff (1) spawned a new era in nitrogen fixa-
tion. As demonstrated by Shepherd and Taube (2), synthetic
chemists were presented with the unprecedented opportunity
to rationally assemble and investigate the metal–N2 linkage
(3, 4), and, more significantly, determine its proclivity to
participate in functionalization reactions. One pervasive
theme common to all subsequent work on well-defined tran-
sition metal dinitrogen compounds is the ability to tune
metal–ligand combinations and explore structure–reactivity
relationships. This prospect paved the way for Chatt et al.’s
pioneering work (5), highlighted by the fixation of N2 to am-
monia by successive proton and electron transfers with zero-
valent group 6 dinitrogen compounds. Recently, this cycle
has been extended to include iron complexes (6), systems
In addition to studies with group 6 and 8 compounds,
there has been a long-standing interest in early transition
metal complexes for nitrogen fixation. Not long after Allen
and Senoff’s report (1), Vol’pin and Shur (10) described
ammonia production from N2 by protonation of a (η5-
C5H5)2TiCl2–alkyllithium mixture. Subsequently, van
Tamelen (11) and Brintzinger (12) independently explored
the mechanism of this reaction, leading to the pioneering
synthesis by the latter’s research group of the first well-
defined titanium dinitrogen compound, [(η5-C5Me5)2Ti]2(µ2,
η1, η1-N2) (13). This molecule, as well as its zirconium con-
gener, [(η5-C5Me5)2Zr(η1-N2)]2(µ2, η1, η1-N2), can be pro-
tonated with mineral acids to yield hydrazine (14). Since
these seminal discoveries, group 4 transition metal dinitro-
gen chemistry has experienced remarkable growth and nu-
merous successes (15). A few recent highlights include the
synthesis by Mori and co-workers (16) of N-heterocycles
and natural products from N2 using modified Vol’pin chem-
istry and Fryzuk and co-workers’ (17, 18) side-on bound
dinitrogen complex that undergoes partial hydrogenation and
reacts with both silanes and terminal alkynes.
Received 9 September 2004. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 3 March 2005.
J.A. Pool2 and P. Chirik.3 Department of Chemistry and
Chemical Biology, Baker Laboratory, Cornell University,
Ithaca, NY 14853, USA.
Inspired by these advances, we initiated a program in July
2001 aimed at incorporating atmospheric N2 into value-added
nitrogen-containing products using well-defined group 4
transition metal dinitrogen compounds. Specifically, we
hoped to gain an understanding of how dinitrogen is acti-
1This article is part of a Special Issue dedicated to Dinitrogen
Chemistry.
2Present address: Chemistry Division, Los Alamos National
Laboratory, Los Alamos, NM 87545, USA.
3Corresponding author (e-mail: pc92@cornell.edu).
Can. J. Chem. 83: 286–295 (2005)
doi: 10.1139/V05-009
© 2005 NRC Canada

Pool and Chirik
287
Fig. 1. Effect of cyclopentadienyl ligand substitution on N2 coordination to titanium sandwich complexes. Tmax refers to the maximum
temperature at which N2 coordination is observed. Temperature values are 2 °C.
vated by low-valent metal centers, elucidate the factors that
govern its hapticity (e.g., “side-on” vs. “end-on”), and trans-
late these findings onto reactions that transform and liberate
a functionalized nitrogen fragment from the metal center.
Here, we recount some of our previous results as well as
present new data that underscore the importance of cyclo-
pentadienyl substituent effects in group 4 metallocene
dinitrogen chemistry. Particularly striking is how seemingly
modest changes in substitution, such as the replacement of a
methyl group with a hydrogen atom or a tert-butyl with an
isopropyl group, can completely alter the course of N2 acti-
vation and functionalization.
Mach and co-workers (22) have since independently
prepared a family of silyl-substituted titanocenes that are re-
ported to resist N2 coordination up to pressures of 50 atm
(1 atm = 101.325 kPa).
Intrigued by these reactivity differences, we sought to un-
derstand the steric and electronic properties of cyclopenta-
dienyl ligands that govern N2 activation by the titanium
sandwiches. In an attempt to separate steric and electronic
contributions, two new titanium(II) complexes, (η5-C5Me4R)2Ti
(R = CHMe2, CMe3), were prepared. The presence of 10
alkyl substituents on the ligand framework served to simu-
late the electronic environment of N2-reactive (η5-C5Me5)2Ti
while large isopropyl or tert-butyl groups mimicked the
steric disposition of reportedly N2-inert (η5-C5Me4SiMe3)2Ti
(22). The magnetic and NMR spectroscopic data collected at
23 °C, in addition to elemental analyses, were consistent
with isolation of the sandwich complexes, (η5-C5Me4CHMe2)2Ti
and (η5-C5Me4CMe3)2Ti (23).
Results and discussion
Activation of N2 by low-valent group 4 metallocenes
Our illustration of the importance of cyclopentadienyl
substituent effects on N2 activation begins with one of the
simplest reactions imaginable for dinitrogen reduction — the
coordination of N2 to a preformed, low-valent titanium(II)
sandwich compound (Fig. 1). Previous reports from the liter-
ature hinted at subtle, yet interesting substituent effects. The
transiently generated titanium(II) sandwiches, (η5-C5Me5)2Ti
(13), (η5-C5H3-1,3-(SiMe3)2)2Ti (19), and presumably (η5-
C5Me4H)2Ti (20), react rapidly with N2 to afford end-on
bound, bridged, dimeric dinitrogen complexes. In contrast,
Lawless and co-workers (21) have isolated and crystallo-
graphically characterized (η5-C5Me4SiMe2-t-Bu)2Ti, which
is unreactive toward N2 at ambient temperature and pressure.
Attempts to grow X-ray quality crystals of (η5-
C5Me4CHMe2)2Ti by cooling a concentrated pentane solu-
tion to –35 °C in an N2 atmosphere produced a spectacular
color change from red to intense blue. Warming the solu-
tions back to room temperature resulted in vigorous evolu-
tion of gas and regeneration of the diagnostic red color of
the sandwich compound. A combination of in situ React-IR
spectroscopy, Toepler pump analysis, and X-ray diffraction
were used to identify the blue compound as the monomeric
titanocene bis-dinitrogen complex, (η5-C5Me4CHMe2)2Ti(N2)2
(23). Further investigation of other titanium sandwiches such
© 2005 NRC Canada

288
Can. J. Chem. Vol. 83, 2005
Fig. 2. Cyclopentadienyl substituent effects on the outcome of alkali-metal reduction of zirconocene dichloride complexes (1 atm =
101.325 kPa).
as (η5-C5Me4CMe3)2Ti and (η5-C5Me4SiMe3)2Ti by in situ
React-IR studies established reversible dinitrogen coordina-
tion for these molecules as well, although N2 loss is ob-
served at much lower temperatures (Fig. 1). While N2 and
CO are isoelectronic, these molecules are the first examples
of bent metallocene dinitrogen complexes with a bonding
motif frequently encountered with more familiar titanocene
dicarbonyl complexes.
Importantly, these studies have demonstrated the effect
that cyclopentadienyl substitution can have on the coordina-
afforded deep-purple, diamagnetic crystals identified as the
tion of dinitrogen to a low-valent transition metal center.
zirconium(III) monochloride dimer, [(η5-C5H4SiMe3)2Zr(µ2-
Changing two methyl groups on the ligand array to
Cl)]2 (1) (eq. [1]). Stronger reducing agents such as KC8,
isopropyl, tert-butyl, or silyl substituents blocks formation
1
CsC8, and sodium potassium alloy also yielded 1. The H
of dimeric dinitrogen compounds and thus allows reversible
NMR spectrum of 1 in benzene-d6 is consistent with a C2v
dinitrogen coordination.
symmetric compound. The cyclopentadienyl resonances, ob-
For zirconium, the situation is more complicated. Isolation
served at 5.31 and 5.44 ppm, are relatively upfield-shifted as
of a bona fide bis-cyclopentadienyl zirconium sandwich
compared to the analogous resonances of the starting di-
complex, [(CpRn)2Zr], has not been achieved, most likely
chloride complex, which appear at 6.09 and 6.69 ppm.
owing to the increased thermodynamic potential for the sec-
Cooling a concentrated pentane solution of 1 afforded
purple blocks suitable for X-ray diffraction. The solid-state
ond-row metal to attain its highest oxidation state. Typically,
access to zirconium dinitrogen compounds proceeds by re-
duction of the appropriate Zr(IV) dihalide complex with an
alkali metal or its equivalent (e.g., KC8). Again, significant
cyclopentadienyl substituent effects are operative in these re-
actions. Whereas (η5-C5Me5)2ZrCl2 undergoes smooth re-
duction with 40% sodium amalgam in the presence of N2 to
yield [(η5-C5Me5)2Zr(η1-N2)]2(µ2, η1, η1-N2) (14), related
zirconocenes such as (η5-C5H3-1,3-(EMe3)2)2ZrCl2 (E = C
(24), Si (25)) only undergo a one-electron reduction to pro-
duce the corresponding Zr(III) monochloride under similar
conditions (Fig. 2).
structure, presented in Fig. 3, establishes a dimeric molecule
possessing a planar Zr2Cl2 core where the [SiMe3] substitu-
ents are oriented away from the metallocene wedge in a
cisoid arrangement. The Zr–Zr distance of 3.512(12) Å is
shorter than the values of 3.86 and 3.94 Å reported for the
two molecules in the unit cell of paramagnetic [(η5-C5H3-
1,3-(SiMe ) ) Zr( -Cl)] (25), but is significantly longer
µ2
3 2 2
than the value of 2.96 Å2for the sum of the covalent radii for
zirconium. The presence of only one [SiMe3] on the
cyclopentadienyl ring affords a more open metallocene com-
pared to [(η5-C5H3-1,3-(SiMe3)2)2Zr(µ2-Cl)]2 and, as a result,
a much shorter Zr–Zr distance is observed.
These results prompted further investigation into the re-
duction of various substituted zirconocene dihalide com-
plexes. Reduction of (η5-C5H4SiMe3)2ZrCl2 with 0.5% sodium
amalgam in toluene under 1 atm of N2 (1 atm = 101.325 kPa)
Reduction of the corresponding zirconocene diiodide
complex, (η5-C5H4SiMe3)2ZrI2, with the range of reductants
described above again resulted in one-electron reduction to
© 2005 NRC Canada

Pool and Chirik
289
Fig. 4. Molecular structure of 2 at 30% probability ellipsoids.
Fig. 3. Molecular structure of 1 at 30% probability ellipsoids.
Hydrogen atoms are omitted for clarity.
Hydrogen atoms are omitted for clarity.
the zirconium(III) monoiodide complex, [(η5-C5H4-
SiMe3)2Zr(µ2-I)]2 (2) (eq. [1]). As with 1, 2 is diamagnetic
and has been characterized by NMR spectroscopy, elemental
analysis, and X-ray diffraction (Fig. 4). As expected with the
larger iodide ligands, the Zr–Zr distance has expanded to
3.6855(8) Å and the I(1)-Zr(1)-I(2) angle has opened to
101.870(19)° as compared to 93.24(8)° for 1. Similar chem-
istry is also observed for the reduction of (η5-C5H4SiMe2-t-
Bu)2ZrCl2 (26), where only one-electron reduction to the
diamagnetic Zr(III) monohalide dimer results (27).
duction of the dihalide complexes with activated Mg and CO
(32), the affinity of the zirconium for the incoming ligand
appears to influence the outcome of the reduction. Addition
of 1 atm of CO (1 atm = 101.325 kPa) to both 1 and 2 in-
duced disproportionation into the known zirconocene
dicarbonyl, (η5-C5H4SiMe3)2Zr(CO)2 (33), and the dihalide
complex, (η5-C5H4SiMe3)2ZrX2 (eq. [2]). Similar reactivity
has also been observed by Gambarotta and co-workers with
the unsubstituted zirconium cyclopentadienyl compounds (29).
Dimeric, diamagnetic Zr(III) monohalide complexes have
been prepared previously by Gambarotta and co-workers
(28, 29). Independent computational studies by DeKock et
al. (30) and Bénard and Rohmer (31) have established that a
partial Zr–Zr bonding interaction accounts for the observed
spin pairing. However, the metal–metal interaction can be at-
tenuated, and therefore the spin state of the molecule altered,
by both the bridging ligand and cyclopentadienyl substitu-
ents. Relatively unhindered cyclopentadienyl ligands such as
those found in 1 and 2 allow close contact (3.512–3.685 Å)
of the two zirconium centers, allowing a metal–metal inter-
action and spin pairing. Interestingly, in the case of the
larger iodide ligands present in 2, the I-Zr-I angle opens to
allow electronic communication between the two zirconium
centers. Introduction of an additional [SiMe3] substituent as
in the case of [(η5-C5H3-1,3-(SiMe3)2)2Zr(µ2-Cl)]2 limits the
contact of the metal centers, and the compound is paramag-
netic (25).
A mechanism that accounts for these observations is pre-
sented in Fig. 5. Following the initial reduction to Zr(III), an
exogenous ligand coordinates (N2 is shown in this case) and
forms the intermediate dinitrogen complex, (CpRn)2ZrX(N2),
which then disproportionates into the Zr(IV) dihalide and
the Zr(II) dinitrogen complex. The resulting zirconocene
dihalide is then recycled back to Zr(III) by the reducing
agent and the process continues. Lappert and co-workers
(34) have obtained EPR evidence for a monomeric, side-on
bound Zr(III) dinitrogen compound, Cp2Zr(CH2SiMe3)(η2-
N2), providing experimental support for such an intermedi-
ate. Thus, the ability of the Zr(III) compound to bind
dinitrogen, not the strength of the reducing agent, appears to
dictate whether a two-electron reduction occurs.
The inability to reduce the silylated zirconocenes to zirco-
nium(II) dinitrogen compounds with potent reducing agents
such as CsC8 and NaK suggested that a kinetic rather than a
thermodynamic barrier inhibited the desired two-electron re-
duction. Considering that an array of formally zirconium(II)
dicarbonyl complexes can be prepared by straightforward re-
This model, in conjunction with the known steric and
electronic properties of zirconocene complexes (32), accounts
© 2005 NRC Canada

290
Can. J. Chem. Vol. 83, 2005
Fig. 5. Proposed mechanism for the synthesis of zirconocene dinitrogen complexes by alkali-metal reduction.
for the observed reduction chemistry. Zirconocenes with
sterically demanding, electron-donating ligands such as (η5-
C5Me5) or (η5-C5Me4H) inhibit dimerization of the Zr(III)
intermediate and stabilize the Zr–N2 interaction through π-
backbonding. In contrast, zirconocenes with single silyl sub-
stituents are prone to form robust halide-bridged dimers and
have less electron density for back donation. These effects
should be considered when exploring zirconocene com-
plexes for dinitrogen activation.
tion and β-hydrogen elimination with zirconocene dihydrides
and alkyl hydrides to gain an understanding of molecular
weight control in metallocene-catalyzed olefin polymeriza-
tion. Having worked with a number of model complexes, we
became interested in extending our studies to isospecific
ansa-zirconocenes. A stock solution of rac-BpZrH2 (Bp =
Me2Si(C5H2-2-SiMe3-4-CMe3)2) in toluene was prepared for
kinetic studies and placed in the drybox freezer and essen-
tially forgotten. About a year later during one of John’s fa-
mous group cleanups, the vial (clearly labeled of course!)
was found, and we discovered that beautiful pleochroic crys-
tals had formed. X-ray diffraction established a side-on
bound, modestly activated N2 ligand (37). Subsequent NMR
experiments demonstrated that loss of H2 and coordination
of dinitrogen took place over the course of hours at ambient
temperature (eq. [3]).
The limitations of the alkali-metal reduction reactions, in
particular with cyclopentadienyl ligands containing electron-
withdrawing substituents, demonstrated the need for alterna-
tive methods for the synthesis of zirconocene dinitrogen
complexes. Alkane or dihydrogen reductive elimination from
the appropriate Zr(IV) precursor could also achieve the
desired two-electron reduction. Importantly, ground-state
considerations (vide infra) also suggest that reductive elimi-
nation will be favored for electron-poor zirconocenes and
would therefore complement reduction by alkali metals in
terms of the class of metallocenes that could be targeted.
Inspiration for this aspect of the chemistry derived from
several sources. Teuben and co-workers (20) had reported
the synthesis of an end-on bound titanocene dinitrogen com-
plex upon loss of dihydrogen from the corresponding tita-
nium(III) monohydride. Fryzuk et al. (35) also had described
a similar reaction with a dinuclear tantalum tetrahydride to
yield an unusual dinitrogen complex with a N2 ligand that
can be classified as being both side-on and end-on. This
molecule has since exhibited rich reaction chemistry (36).
Another major inspiration stems from a serendipitous dis-
covery made by one of us (PJC) while working with John
Bercaw at Caltech. We were studying rates of olefin inser-
In July 2001 at Cornell, we began a systematic investiga-
tion of the rate of alkane reductive elimination as a function
of cyclopentadienyl ligand. Zirconocene isobutyl hydrides
were chosen because these molecules could be conveniently
prepared by addition of 2 equiv. of tert-butyllithium to the
corresponding dichloride complex (38). Using this method-
ology, a series of “mixed-ring” zirconocene isobutyl hy-
drides, (η5-C5Me5)(CpRn)Zr(CH2CHMe2)H, were prepared.
Their relative rates of reductive elimination were measured
and the products formed from alkane loss identified. In each
case, loss of isobutane afforded a single diastereomer of a
cyclometalated zirconium hydride rather than the desired
dinitrogen complex. Importantly, however, it was established
that the rate of alkane reductive elimination increased with
incorporation of [SiMe3] substituents, an effect ascribed to
both steric and electronic destabilization of the Zr(IV) alkyl
© 2005 NRC Canada

Pool and Chirik
291
hydride ground state (39). Subsequent experiments with the
(η5-C5Me5)(η5-C5H3-1,3-(SiMe3)2)Zr(R)H (R = alkyl) family
of complexes demonstrated increased rates of elimination
with increasing size of the alkyl group, a phenomenon also
attributed to ground-state destabilization (40).
carbocycle is compromised to engage in bonding with the
zirconium (eq. [5]) (44). These molecules have proven to be
isolable sources of Zr(II), participating in numerous cou-
pling and ligand addition reactions, although N2 activation
has not been observed. Thus, the chemistry of low-valent
zirconium cyclopentadienyl and indenyl-substituted com-
plexes is distinctly different, again highlighting the sensitiv-
ity to “ancillary” ligand environment.
These results led to the design and synthesis of the purely
silyl-substituted zirconocene neopentyl hydride, (η5-C5H3-
1,3-(SiMe3)2)2Zr(CH2CMe3)H, which undergoes facile loss
of neopentane at ambient temperature to afford the corre-
sponding cyclometalated zirconocene hydride. Exposure of
this molecule to a blanket of N2 at –35 °C afforded crystals
of the dinitrogen complex [(η5-C5H3-1,3-(SiMe3)2)2Zr]2(µ2,
η2, η2-N2) (eq. [4]). X-ray diffraction revealed a side-on
bound N2 ligand with an elongated N—N bond of 1.47 Å.
Thus, alkane reductive elimination provides a unique syn-
thetic pathway to a dinitrogen complex that cannot be ac-
cessed by alkali-metal reduction. Unfortunately, the yield of
the dinitrogen complex is low owing to the propensity of the
cyclometalated zirconocene hydride to participate in numer-
ous side reactions. However, preliminary reactivity studies
with [(η5-C5H3-1,3-(SiMe3)2)2Zr]2(µ2, η2, η2-N2) are encour-
aging as reaction with ethyl chloroformate yields diethyl-
azodicarboxylate, a useful precursor for the synthesis of
amino acids and other organic molecules.
End-on vs. side-on coordination and the hydrogenation
of N2
The elongated N—N bond in [(η5-C5H3-1,3-(SiMe3)2)2Zr]2-
(µ2, η2, η2-N2) and the reaction of this complex with electro-
philes prompted further investigation into the origins of side-
on coordination of the dinitrogen ligand. Particularly, we
were interested in elucidating the effect of cyclopentadienyl
substituents on N2 hapticity and exploring reactivity differ-
ences between the side-on and end-on coordination modes.
Based on the observed end-on coordination in [(η5-
C5Me5)2Zr(η1-N2)]2(µ2, η1, η1-N2), we hypothesized that
side-on coordination was in fact electronically favored and
should be observed in the absence of overriding steric ef-
fects. To test this assertion, we decided to systematically in-
vestigate the hapticity of the N2 ligand as methyl groups
were removed from the pentamethylcyclopentadienyl ligand.
Ideally, the methyl substitution would maintain sufficient
electron density and steric protection to promote complete
reduction to the zirconium(II) dinitrogen complex yet allow
close contact of the metal centers for η2-coordination of the
bridging N2 ligand (Fig. 6).
Low-valent bis-indenyl zirconium chemistry
Because incorporation of [SiMe3] groups into the cyclo-
pentadienyl ligand array afforded N2-reactive cyclometalated
hydride complexes, we have also explored similar reactions
with the corresponding silylated indenyl complexes. The
indenyl anion, the benzo-substituted version of the cyclo-
pentadienyl ligand, has enjoyed considerable success as a
supporting ligand in metallocene-catalyzed olefin polymer-
ization (41), although investigations into the chemistry of
low-valent species have been limited. Moreover, replacement
of the cyclopentadienyl anion with an indenyl ligand often
results in enhanced or even unique reactivity. This behavior,
when applied to associative substitution reactions, has been
termed the “indenyl effect” (42).
The bis-tetramethyl zirconocene dichloride, (η5-
C5Me4H)2ZrCl2 (45), was first explored owing to the com-
mercial availability of the ligand and the relative ease of
synthesis of homoleptic zirconocene dichloride compounds.
Reduction of the Zr(IV) dihalide with 0.5% sodium amal-
gam under N2 afforded a green solid. Diagnostic reactions
with CO and I2 afforded the zirconocene dicarbonyl and
diiodide complexes, respectively, suggesting that the product
was indeed a dinitrogen complex. Not long after the initial
synthesis, beautiful single crystals were obtained for a dif-
fraction experiment. The solid-state structure confirmed our
formulation of the molecule’s identity, but revealed that the
N2 had changed hapticity to side-on as compared to end-on
in the case of permethylzirconocene (eq. [6]) (46). Another
important feature of the solid-state structure was the peculiar
arrangement of the cyclopentadienyl ligands in the dimer.
The two metallocene wedges are canted with respect to each
Treatment of the silylated bis-indenyl zirconocene
dichloride, (η5-C9H5-1,3-(SiMe3)2)2ZrCl2, with 2 equiv. of
isobutyllithium allows observation of the transient bis-
indenyl zirconocene isobutyl hydride, which undergoes rapid
reductive elimination of isobutane to yield a unique bis-
indenyl zirconium sandwich (43). Recent crystallographic
studies have established an unprecedented η9 bonding inter-
action of one of the indenyl rings, where the planarity of the
© 2005 NRC Canada

292
Can. J. Chem. Vol. 83, 2005
Fig. 6. Exploration of cyclopentadienyl substituent effects on N2 hapticity.
Fig. 7. Side-on vs. end-on coordination of dinitrogen and comparison of hydrogenation reactivity (1 atm = 101.325 kPa).
pected dihydride complex but rather the zirconocene
hydrazido compound, [(η5-C5Me4H)2ZrH]2(µ2, η2, η2-N2H2),
arising from hydrogenation of the N2 ligand (Fig. 7). IR and
NMR experiments in conjunction with isotopic labeling
studies quickly established the molecule’s identity, but a
crystal structure was essential. The zirconocene hydrazido
other, forming a dihedral angle of 65.3°. This angle is signif-
complex was sparingly soluble in both aromatic and
icantly distorted from the electronically preferred coplanar
aliphatic hydrocarbons, making crystallization difficult. For-
arrangement observed in less-substituted dinitrogen com-
tunately, slow diffusion of hydrogen into a concentrated
benzene-d6 solution of the dinitrogen complex afforded a
very large single crystal that yielded high-quality diffraction
data.
plexes (39).
Having isolated and crystallographically characterized a
new, side-on bound N2 complex, we were anxious to explore
its reactivity. Given the long-standing interest in the synthe-
sis of ammonia from its elements, we first studied hydroge-
nation. Based on literature precedent, we anticipated that
this nitrogen complex would behave like most others — ad-
dition of H2 would induce loss of free N2 with concomitant
formation of the zirconocene dihydride. Addition of 1 atm of
H2 (1 atm = 101.325 kPa) to [(η5-C5Me4H)2Zr]2(µ2, η2, η2-
N2) in benzene-d6 immediately bleached the intense purple
solution and resulted in a sparingly soluble white solid.
Much to our surprise (and delight!), we isolated not the ex-
Having hydrogenated atmospheric nitrogen to a hydrazido
ligand, we next explored the possibility of fixation to ammo-
nia. Mild thermolysis of [(η5-C5Me4H)2ZrH]2(µ2, η2, η2-
N2H2) resulted in loss of H2 along with N2 scission, forming
the zirconocene amido, nitrido complex, [(η5-C5Me4H)2Zr]2-
(µ2-N)(µ2-NH2) (eq. [7]). Protonolysis of this molecule with
mineral acids does indeed yield free ammonia. Although this
was an interesting result, N2 fixation where H2 was the sole
hydrogen source was desired. Warming heptane solutions of
© 2005 NRC Canada

Pool and Chirik
293
[(η5-C5Me4H)2ZrH]2(µ2, η2, η2-N2H2) to 85 °C under 1 atm
of H2 (1 atm = 101.325 kPa) yielded small quantities of free
ammonia, the zirconocene dihydride, and several Zr(IV)
products arising from the reaction of the two.
tions. The [C5Me4H]– ligand supports a side-on bound zir-
conocene dinitrogen complex in which two metallocene
wedges are twisted with respect to one another. This distor-
tion from the electronically preferred planar geometry im-
parts zirconium-imido character to the N2 ligand, resulting
in facile 1,2-addition of H2 at ambient temperature. Because
such subtle changes in ligand structure produce such signifi-
cant consequences, the prospects for stoichiometric and pos-
sibly even catalytic nitrogen functionalization reactions are
promising.
The remarkable hydrogenation activity observed with
[(η5-C5Me4H)2Zr]2(µ2, η2, η2-N2) prompted us to investigate
the reduction of (η5-C5Me5)(η5-C5Me4H)ZrCl2. Reaction of
the zirconocene dichloride or diiodide with KC8 afforded a
purple solution from which the end-on dinitrogen complex,
[(η5-C5Me5)(η5-C5Me4H)Zr(η1-N2)2](µ2,η1,η1-N2), was iso-
lated. Exposure to 1 atm of H2 (1 atm = 101.325 kPa) re-
sulted in loss of N2 and formation of the zirconocene
dihydride (47). The reactions of zirconocene dinitrogen
complexes again emphasize the sensitivity of the observed
chemistry to cyclopentadienyl substituent effects. Replacing
a single methyl group by a hydrogen atom completely alters
the course of N2 fixation. In the more substituted cases, triv-
ial N2 loss is observed whereas in the less-hindered example,
ammonia synthesis can be achieved under mild conditions in
solution.
Recent work in our laboratory has focused on the origin
of the unprecedented reactivity of [(η5-C5Me4H)2Zr]2(µ2, η2,
η2-N2). Both computational and isotopic labeling studies
(47) are consistent with the “twist” of the dimer as the origin
of N2 hydrogenation. Canting of the zirconium centers per-
mits increased π back donation to the π* molecular orbital of
the side-on bound N2 ligand, imparting significant Zr=N
character and allowing facile 1,2-addition of H2 across the
zirconium–nitrogen bond (48, 49). Thus, both side-on coor-
dination of N2 and twisting of the zirconium dimer appear to
be essential ingredients for nitrogen fixation.
Experimental
General considerations
All air- and moisture-sensitive manipulations were carried
out using standard vacuum line, Schlenk, or cannula tech-
niques or in an M. Braun inert-atmosphere drybox contain-
ing an atmosphere of purified nitrogen. The M. Braun
drybox was equipped with a cold well designed for freezing
samples in liquid nitrogen. All glassware was predried in an
oven at 150 °C. Solvents for air- and moisture-sensitive
manipulations were initially dried and deoxygenated using
literature procedures (50). Deuterated solvents for NMR
spectroscopy were distilled from sodium metal under an at-
mosphere of argon and stored over 4 Å molecular sieves. Ar-
gon and hydrogen gas were purchased from Airgas
Incorporated and passed through a column containing man-
ganese oxide on vermiculite and 4 Å molecular sieves before
admission to the high-vacuum line.
1H and 13C NMR spectra were recorded on a Varian Inova
400 spectrometer operating at 399.860 MHz (1H) and
100.511 MHz (13C). All chemical shifts are reported relative
1
to SiMe4 using H (residual) or 13C NMR chemical shifts of
Conclusions and future outlook
the solvent as a secondary standard. Elemental analyses were
performed by Robertson Microlit Laboratories, Inc. (Madi-
son, N.J.).
Over the past 3 years, our laboratory has observed the
profound consequences that cyclopentadienyl ligand manip-
ulations have on the activation, hapticity, and reactivity of
dinitrogen with group 4 metallocenes. For titanium sandwich
complexes, introduction of large substituents into [C5Me4R]–
ligands allowed observation of monomeric, bis-dinitrogen
complexes that are isoelectronic and isostructural to
titanocene di-carbonyl complexes. In zirconium chemistry,
access to the desired Zr(II) dinitrogen complex can be con-
trolled by manipulation of the electronic properties of the
cyclopentadienyl substituents. For electron-deficient metallo-
cenes, reductive elimination is facile, whereas for electron-
rich compounds, alkali-metal reduction is the preferred
synthetic route. Significantly, either alkali-metal reduction or
reductive elimination with the corresponding bis-indenyl
zirconocene complexes affords unusual zirconium sandwich
compounds in which one of the ligands is engaged in η9
bonding.
Single crystals suitable for X-ray diffraction were coated
with polyisobutylene oil in a drybox and were quickly trans-
ferred to the goniometer head of a Siemens SMART CCD
area detector system equipped with a molybdenum X-ray
tube (λ = 0.710 73 Å). Preliminary data revealed the crystal
system. A hemisphere routine was used for data collection
and determination of lattice constants. The space group was
identified and the data were processed using the Bruker
SAINT program and corrected for absorption using
SADABS. The structures were solved using direct methods
(SHELXS) completed by subsequent Fourier synthesis and
refined by full-matrix least-squares procedures. The crystal-
lographic data for complexes 1 and 2 are provided in the
supplementary data.4
Preparation of [(η5-C5H4SiMe3)2Zr(µ2-Cl)]2
A 250 mL round-bottomed flask was charged with
0.901 g (2.06 mmol) of (η5-C5H4SiMe3)2ZrCl2 (51) and ap-
proximately 100 mL of toluene. In a separate flask, 5 equiv.
Investigation into the substituent effects on the hapticity
of dinitrogen allowed observation of the first hydrogenation
of N2 to ammonia in solution under relatively mild condi-
4 Supplementary data may be purchased from the Directory of Unpublished Data, Document Delivery, CISTI, National Research Council
Canada, Ottawa, ON K1A 0S2, Canada (http://cisti-icist.nrc-cnrc.gc.ca/irm/unpub_e.shtml for information on ordering electronically).
CCDC 261388 and 261389 contain the crystallographic data for this manuscript. These data can be obtained, free of charge, via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ,
UK; fax +44 1223 336033; or deposit@ccdc.cam.ac.uk).
© 2005 NRC Canada

294
Can. J. Chem. Vol. 83, 2005
8. D.V. Yandulov and R.R. Schrock. Science, 301, 76 (2003).
9. V. Ritleng, D.V. Yandulov, W.W. Weare, R.R. Schrock, A.S.
Hock, and W.M. Davis. J. Am. Chem. Soc. 126, 6150 (2004).
10. M.E. Vol’pin and V.B. Shur. Nature (London), 209, 1236 (1966).
11. E.E. van Tamelen. Acc. Chem. Res. 3, 361 (1970).
12. H.H. Brintzinger. J. Am. Chem. Soc. 88, 4305 (1966).
13. J.E. Bercaw and H.H. Brintzinger. J. Am. Chem. Soc. 93, 2045
(1971).
14. J.M. Manriquez and J.E. Bercaw. J. Am. Chem. Soc. 96, 6229
(1974).
15. M.D. Fryzuk and S.A. Johnson. Coord. Chem. Rev. 200, 379
(2000).
16. (a) M. Mori, K. Hori, M. Akashi, M. Hori, Y. Sato, and M.
Nishida. Angew. Chem. Int. Ed. 37, 636. (1998); (b) M.
Akashi, M. Nishida, and M. Mori. Chem. Lett. 465 (1999);
(c) K. Hori and M. Mori. J. Am. Chem. Soc. 120, 7651
(1998); (d) M. Akashi, Y. Sato, and M. Mori. J. Org. Chem.
66, 7873 (2001).
of 0.5% (by weight) Na(Hg) was prepared. The solution of
the zirconocene dichloride was added to the amalgam and
the reaction mixture was stirred for 4 days. The resulting
purple solution was separated from the amalgam by pipette
and filtered through Celite. The solvent was removed in
vacuo, and the purple oil was triturated with pentane to yield
a purple solid. Recrystallization from pentane at –35 °C af-
1
forded 0.553 g (67% yield) of [(η5-C5H4SiMe3)2ZrCl]2. H
NMR (benzene-d6) δ: 0.40 (s, 36H, SiMe3), 5.31 (s, 8H, Cp),
5.44 (s, 8H, Cp). 13C NMR (benzene-d6) δ: 1.29 (SiMe3),
104.10, 107.31, 113.12 (Cp). Anal. calcd. for C32H52Si4Zr2Cl2:
C 47.89, H 6.53; found: C 48.20, H 6.31.
Preparation of (η5-C5H4SiMe3)2ZrI2
A 100 mL round-bottomed flask was charged with
1.445 g (3.32 mmol) of (η5-C5H4SiMe3)2ZrCl2 and approxi-
mately 50 mL of CH2Cl2. A CH2Cl2 solution containing
1.552 g (3.959 mmol) of BI3 was prepared and added to the
metallocene solution with stirring. The reaction mixture was
stirred for 2 days, after which time the solvent was removed
in vacuo. The resulting yellow solid was extracted into
pentane and the yellow solution filtered through Celite. The
pentane was removed in vacuo, leaving 0.586 g (29%) of a
yellow solid identified as (η5-C5H4SiMe3)2ZrI2. 1H NMR
(benzene-d6) δ: 0.29 (s, 18H, SiMe3), 6.07 (d, 4H, Cp), 6.68
(d, 4H, Cp). 13C NMR (benzene-d6) δ: 0.57 (SiMe3), 115.95,
123.74, 125.40 (Cp).
17. M.D. Fryzuk, J.B. Love, and S.J. Rettig. Science (Washington,
D.C.), 275, 1445 (1997).
18. L. Morello, J.B. Love, B.O. Patrick, and M.D. Fryzuk. J. Am.
Chem. Soc. 126, 9480 (2004).
19. T.E. Hanna, I. Keresztes, E. Lobkovsky, W.H. Bernskoetter,
and P.J. Chirik. Organometallics, 23, 3448 (2004).
20. J.M. De Wolf, R. Blaauw, A. Meetsma, J.H. Teuben, R. Gyepes,
V. Varga, K. Mach, N. Veldman, and A.L. Spek. Organo-
metallics, 15, 4977 (1996).
21. P.B. Hitchcock, F. Kerton, and G.A. Lawless. J. Am. Chem.
Soc. 120, 10 264 (1998).
22. (a) L. Lukesová, M. Horácek, P. Stepnicka, K. Fejfarová, R.
Gyepes, I. Cisorová, J. Kubista, and K. Mach. J. Organomet.
Chem. 663, 134 (2002); (b) M. Horácek, V. Kupfer, U.
Thewalt, P. Stepnicka, M. Polasek, and K. Mach. Organo-
metallics, 18, 3572 (1999).
Preparation of [(η5-C5H4SiMe3)2Zr(µ2-I)]2
This molecule was prepared in a similar manner to [(η5-
C5H4SiMe3)2ZrCl]2 with 0.143 g (0.234 mmol) of (η5-
C5H4SiMe3)2ZrI2 yielding 0.087 g (77% yield) of deep-
purple crystals identified as [(η5-C5H4SiMe3)2ZrI]2. H NMR
1
23. T.E. Hanna, E. Lobkovsky, and P.J. Chirik. J. Am. Chem. Soc.
(benzene-d6) δ: 0.41 (s, 36H, SiMe3), 5.21 (d, 8H, Cp), 5.36
(d, 8H, Cp). 13C NMR (benzene-d6) δ: 1.14 (SiMe3), 102.04,
103.44, 113.64 (Cp). Anal. calcd. for C32H52Si4Zr2I2: C
39.00, H 5.32; found: C 38.90, H 4.93.
126, 14 688 (2004).
24. I.F. Urazowski, V.I. Ponomaryev, I.E. Nifantev, and D.A.
Lemenovskii. J. Organomet. Chem. 368, 287 (1989).
25. P.B. Hitchcock, M.F. Lappert, G.A. Lawless, H. Oliver, and
E.J. Ryan. J. Chem. Soc. Chem. Commun. 474 (1992).
26. P.A. Deck, X. Cheng, E.J. Stoebenau, C. Slebonick, D.R.
Billodeaux, and F.R. Fronczek. Organometallics, 19, 5404 (2000).
27. J.A. Pool. Ph.D. thesis, Cornell University, Ithaca, N.Y. 2004.
28. Y. Wielstra, S. Gambarotta, A. Meetsma, and J.L. de Boer.
Organometallics, 8, 250 (1989).
Acknowledgements
We would like to thank the US National Science Founda-
tion, the US National Institutes of Health, the Petroleum Re-
search Fund administered by the American Chemical
Society, and the Research Corporation (Cottrell Scholar
award to PJC) for financial support for N2 chemistry. PJC
would also like to thank the team of dedicated graduate stu-
dents and postdocs who have made the chemistry happen.
29. Y. Wielstra, S. Gambarotta, A. Meetsma, and A.L. Spek.
Organometallics, 8, 2948 (1989).
30. R.L. DeKock, M.A. Peterson, L.E.L. Reynolds, L.-H. Chen,
E.J. Baerends, and P. Vernooijs. Organometallics, 12, 2794
(1993).
31. M. Bénard and M.-M. Rohmer. J. Am. Chem. Soc. 114, 4785
(1992).
References
32. C.E. Zachmanoglou, A. Docrat, B.M. Bridgewater, G.E. Parkin,
C.G. Brandow, J.E. Bercaw, C.N. Jardine, M. Lyall, J.C.
Green, and J.B. Kiester. J. Am. Chem. Soc. 124, 9525 (2002).
33. A. Antifiolo, M.F. Lappert, A. Singh, D.J.W. Winterborn, L.M.
Engelhardt, C.L. Raston, A.H. White, A.J. Carty, and N.J. Tay-
lor. J. Chem. Soc. Dalton Trans. 1463 (1987).
1. A.D. Allen and C.V. Senoff. J. Chem. Soc. Chem. Commun.
621 (1965).
2. R.E. Shepherd and H. Taube. Inorg. Chem. 12, 1392 (1973).
3. J.N. Armor and H. Taube. J. Am. Chem. Soc. 92, 2560 (1970).
4. F. Bottomley and S.C. Nyburg. J. Chem. Soc. Chem. Commun.
897 (1966).
5. J. Chatt, J.R. Dilworth, and R.L. Richards. Chem. Rev. 78, 589
(1978).
6. T.A. Betley and J.C. Peters. J. Am. Chem. Soc. 125, 10 782
(2003).
34. J. Jeffrey, M.F. Lappert, and P.I. Riley. J. Organomet. Chem.
181, 25 (1979).
35. M.D. Fryzuk, S.A. Johnson, and S.J. Rettig. J. Am. Chem.
Soc. 120, 11 024 (1998).
36. M.D. Fryzuk, B.A. MacKay, and B.O. Patrick. J. Am. Chem.
7. Y. Nishibayashi, S. Iwai, and M. Hidai. Science (Washington,
D.C.), 279, 540 (1998).
Soc. 125, 3234 (2003).
© 2005 NRC Canada

Pool and Chirik
295
37. P.J. Chirik, L.M. Henling, and J.E. Bercaw. Organometallics,
20, 534 (2001).
38. J.A. Pool, C.A. Bradley, and P.J. Chirik. Organometallics, 21,
1271 (2002).
44. C.A. Bradley, I. Keresztes, E. Lobkovsky, V.G. Young, and P.J.
Chirik. J. Am. Chem. Soc. 126, 16 937 (2004).
45. C. Janiak, U. Versteeg, K.C.H. Lange, R. Weiman, and E.
Hahn. J. Organomet. Chem. 501, 219 (1995).
39. J.A. Pool, E. Lobkovsky, and P.J. Chirik. J. Am. Chem. Soc.
46. J.A. Pool, E. Lobkovsky, and P.J. Chirik. Nature (London),
125, 2241 (2003).
427, 529 (2004).
40. J.A. Pool, E. Lobkovsky, and P.J. Chirik. Organometallics, 22,
2797 (2003).
47. J.A. Pool, W.H. Bernskoetter, and P.J. Chirik. J. Am. Chem.
Soc. 126, 14 326 (2004).
41. G.W. Coates. Chem. Rev. 100, 1223 (2000).
42. M.E. Rerek and F. Basolo. J. Am. Chem. Soc. 106, 5908
(1984).
43. C.A. Bradley, E. Lobkovsky, and P.J. Chirik. J. Am. Chem.
Soc. 125, 8110 (2003).
48. C.C. Cummins, S.M. Baxter, and P.T. Wolczanski. J. Am.
Chem. Soc. 110, 8731 (1988).
49. A.P. Duncan and R.G. Bergman. Chem. Rec. 2, 431 (2002).
50. A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, and
F.J. Timmers. Organometallics, 15, 1518 (1996).
© 2005 NRC Canada