Mechanistic studies of ammonia borane dehydrogenation catalyzed by iron pincer complexes

Article abstract of DOI:10.1021/ja5058423

A series of iron bis(phosphinite) pincer complexes with the formula of [2,6-(ⁱPr₂PO)₂C₆H ₃]Fe(PMe₂R)₂H (R = Me, 1; R = Ph, 2) or [2,6-(ⁱPr₂PO)₂-4-(MeO)C₆H ₂]Fe(PMe₂Ph)₂H (3) have been tested for catalytic dehydrogenation of ammonia borane (AB). At 60 °C, complexes 1-3 release 2.3-2.5 equiv of H₂ per AB in 24 h. Among the three iron catalysts, 3 exhibits the highest activity in terms of both the rate and the extent of H₂ release. The initial rate for the dehydrogenation of AB catalyzed by 3 is first order in 3 and zero order in AB. The kinetic isotope effect (KIE) observed for doubly labeled AB (k_NH3BH3/k _ND3BD3 = 3.7) is the product of individual KIEs (k _NH3BH3/k_ND3BH3 = 2.0 and k_NH3BH3/k _NH3BD3 = 1.7), suggesting that B-H and N-H bonds are simultaneously broken during the rate-determining step. NMR studies support that the catalytically active species is an AB-bound iron complex formed by displacing trans PMe₃ or PMe₂Ph (relative to the hydride) by AB. Loss of NH₃ from the AB-bound iron species as well as catalyst degradation contributes to the decreased rate of H₂ release at the late stage of the dehydrogenation reaction.

Full text of DOI:10.1021/ja5058423

Article
pubs.acs.org/JACS
Mechanistic Studies of Ammonia Borane Dehydrogenation Catalyzed
by Iron Pincer Complexes
Papri Bhattacharya, Jeanette A. Krause, and Hairong Guan*
Department of Chemistry, University of Cincinnati, P.O. Box 210172, Cincinnati, Ohio 45221-0172, United States
S
* Supporting Information
ABSTRACT: A series of iron bis(phosphinite) pincer
complexes with the formula of [2,6-(iPr2PO)2C6H3]Fe-
(PMe2R)2H (R = Me, 1; R = Ph, 2) or [2,6-(iPr2PO)2-4-
(MeO)C6H2]Fe(PMe2Ph)2H (3) have been tested for
catalytic dehydrogenation of ammonia borane (AB). At 60
°C, complexes 1−3 release 2.3−2.5 equiv of H2 per AB in 24
h. Among the three iron catalysts, 3 exhibits the highest
activity in terms of both the rate and the extent of H2 release.
The initial rate for the dehydrogenation of AB catalyzed by 3 is
first order in 3 and zero order in AB. The kinetic isotope effect
(KIE) observed for doubly labeled AB (kNH3BH3/kND3BD3 = 3.7) is the product of individual KIEs (kNH3BH3/kND3BH3 = 2.0 and
k
NH3BH3/kNH3BD3 = 1.7), suggesting that B−H and N−H bonds are simultaneously broken during the rate-determining step.
NMR studies support that the catalytically active species is an AB-bound iron complex formed by displacing trans PMe3 or
PMe2Ph (relative to the hydride) by AB. Loss of NH3 from the AB-bound iron species as well as catalyst degradation contributes
to the decreased rate of H2 release at the late stage of the dehydrogenation reaction.
Developing iron-based catalytic systems would be even more
desirable because iron is relatively inexpensive and the most
abundant transition metal.12 To date, only a handful of iron
catalysts have been reported for the dehydrogenation of AB
(Figure 1). Baker et al. have shown that FeH(CH2PMe2)-
(PMe3)3 (A), which is in an equilibrium with Fe(PMe3)4,13
catalyzes the release of H2 from AB at 20 °C.9 The low thermal
stability of this particular iron complex, however, results in
significant decomposition and loss of catalytic activity over
time. A more recent study by Baker, Gordon, and co-workers
has focused on a series of iron complexes bearing both amido
INTRODUCTION
■
Ammonia borane (AB = NH3BH3), a molecule first synthesized
and characterized by Shore and Parry in 1955,1 has been
studied extensively in recent years, largely due to its potential to
be utilized as a chemical hydrogen-storage material.2 Although
releasing H2 from AB takes place thermally without a catalyst,3
high temperatures and slow reaction rates are the main
drawbacks that would impede the practical use of AB as a
carrier for dihydrogen. In contrast, metal-catalyzed release of
H2 from AB could operate at a lower temperature and produce
H2 at a much faster rate. Additionally, the extent of H2 released
could be higher in a metal-catalyzed process, thereby taking
advantage of the high gravimetric capacity of hydrogen in AB
(19.6 wt %). Driven by these benefits, sufficient efforts have
been made to develop catalytic systems for AB dehydrogen-
ation.2 Some of the most efficient catalysts include Heinekey
and Goldberg’s iridium POCOP-pincer complex (0.25−1 mol
%, 1 equiv of H2 per AB, RT, 3−36 min),4 Fagnou’s ruthenium
phosphino-amine complexes (0.03−0.1 mol %, ∼1 equiv of H2
per AB, RT, 5−15 min),5 Schneider’s ruthenium PNP-pincer
complex (0.01−0.1 mol %, 0.8−1 equiv of H2 per AB, RT, 3−
20 min),6 and Kang’s cationic Pd(II) complexes (3 mol %, 2
equiv of H2 per AB, 25 °C, <1 min).7 For catalytic systems with
high extent of H2 release, Manners’ heterogeneous rhodium
system (1.2 mol %, 45 °C, 48−84 h),8 Baker’s nickel NHC
complexes (5 mol %, 60 °C, 4 h),9 Williams’ ruthenium
bis(pyridyl)borate complex (5 mol %, 70 °C, 0.5 h),10 and
Jagirdar’s (η6-C6H6)Cr(CO)3 (2 mol %, 0 °C, hν, 1.5 h)11 have
been shown to release >2 equiv of H2 per AB.
Figure 1. Iron-based catalysts for the dehydrogenation of AB and the
extent of H2 released in aprotic solvents.
Received: June 10, 2014
Published: July 18, 2014
© 2014 American Chemical Society
11153
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
and phosphine ligands (B−E).14 These compounds are active
at 60 °C and typically release 1−1.7 equiv of H2 per AB;
catalyst stability under the reaction conditions remains an issue,
though. The Manners group has utilized [CpFe(CO)2]2 (F) to
catalyze the dehydrogenation of AB15 and its related amine
boranes,16 and CpFe(CO)2I to catalyze the dehydrogenation of
Me2NHBH3.16 However, the catalysts need to be activated
through photoirradiation. Sonnenberg and Morris have recently
reported a highly efficient catalyst system involving iron
complexes supported by a tetradentate bis(iminophosphine)
ligand (G).17 In the presence of large excess of KOtBu, these
complexes have been shown to release H2 at a rate up to 3.66
s−1 (H2 per Fe). Interestingly, mechanistic investigation has
suggested that the active species are not homogeneous
molecular catalysts but zerovalent iron nanoparticles stabilized
by the tetradentate ligand.
We have been interested in developing more robust iron
catalysts for the activation of small molecules including AB. In
particular, we have chosen pincer ligands18 for the catalyst
design because of their high tunability and, more importantly,
the ability to minimize the degradation of iron complexes.19 In
our recent report,20 we have demonstrated that iron
bis(phosphinite) or POCOP-pincer complex 1 (Figure 2) is
the reaction. By comparison, synthesis of 1 from 1,3-
(iPr2PO)2C6H4 and Fe(PMe3)4 finishes within 3 h, probably
because PMe3 is less sterically bulky and more electron-
donating, resulting in more facile oxidative addition of the C−
H bond.
Compound 2 (in THF-d8) displays a characteristic hydride
resonance at −15.04 ppm as a triplet of doublet of doublets
(tdd). On the basis of our previous study of related iron pincer
complexes,20 the JP−H coupling constants of 76.8, 48.0, and 23.2
Hz are attributed to coupling with phosphorus nuclei from the
pincer ligand, cis- and trans-PMe2Ph, respectively. As expected,
the 31P{1H} NMR spectrum of the same solution shows a
doublet of doublets (dd) at 217.4 ppm and a pair of doublet of
triplets (dt) at 23.7 and 14.6 ppm. The NMR data of 3 are
similar to those of 2; the hydride resonance is located at −15.32
ppm (tdd, J = 77.5, 48.1, and 24.9 Hz) and phosphorus
resonances are found at 218.8 (dd), 23.9 (dt), and 15.0 (dt)
ppm. The presence of the hydride ligand in 2 and 3 was further
supported by IR spectroscopy, which exhibits a sharp band of
medium intensity at 1890 (for 2) or 1927 cm−1 (for 3) as
anticipated for the Fe−H stretch.
The structure of 3 was also established by X-ray
crystallography. As illustrated in Figure 3, the hydride was
Figure 2. Iron POCOP-pincer complexes.
an effective catalyst for the hydrosilylation of aldehydes and
ketones. Mechanistic studies have indicated that dissociation of
the trans-PMe3 (relative to the hydride) provides a vacant
coordination site for substrate activation. We surmised that the
coordinatively unsaturated intermediate might interact with AB
as well to release H2, and the rigidity of the pincer framework
could make the catalyst long-lived. Because phosphine
dissociation was thought to be the key to unlock the catalytic
activity of the pincer complex, two structural modifications
were made to expedite this process. In compound 2, a more
bulky ancillary ligand, PMe2Ph, is used in place of PMe3.21 In
the case of 3, an electron-donating MeO group is also
introduced at the para position to the ipso carbon for increased
electron density at the iron center.
In this paper, we will show that complexes 1−3 are active
catalysts for the dehydrogenation of AB, releasing more H2 per
AB than any of the iron-based catalytic systems reported to
date.22 Detailed NMR and kinetics studies including isotope
effect measurements support the homogeneous nature of our
catalytic system and the involvement of the pincer backbone in
the catalytic reactions. Deactivation pathways for the iron
catalysts at the late stage of the dehydrogenation reactions will
also be discussed.
Figure 3. ORTEP drawing of [2,6-(iPr2PO)2-4-(MeO)C6H2]Fe(H)-
(PMe2Ph)2 (3) at the 50% probability level. Selected bond lengths (Å)
and angles (deg): Fe−H 1.21(4), Fe−C(1) 2.015(4), Fe−P(1)
2.1903(11), Fe−P(2) 2.1917(11), Fe−P(3) 2.2210(11), Fe−P(4)
2.2789(11), C(1)−Fe−P(1) 76.61(11), C(1)−Fe−P(2) 77.64(11),
C(1)−Fe−P(4) 82.98(11), P(1)−Fe−P(4) 99.21(4), P(2)−Fe−P(4)
104.41(4), P(3)−Fe−P(4) 100.07(4), P(1)−Fe−P(3) 102.02(4),
P(2)−Fe−P(3) 102.23(4), P(1)−Fe−P(2) 142.32(5).
RESULTS
located directly from the difference map and shown to be cis to
the ipso carbon. The octahedral geometry around iron is
significantly distorted and more so than that of 1. To avoid the
steric congestion incurred by the trans-PMe2Ph, the pincer
phosphorus atoms (P1 and P2) are shifted toward the hydride
ligand (Figure 4) and displaced 0.5740(19) Å and 0.4510(20)
Å out of the least-squares plane defined by the C1, O1, O2, and
■
Synthesis and Structure of Iron Pincer Complexes.
Following a similar procedure used for the synthesis of 1,20 iron
POCOP-pincer complexes 2 and 3 were prepared from
“Fe(PMe2Ph)4”23 and an appropriate diphosphinite (eq 1). In
this case, cyclometalation of the diphosphinite ligand is
considerably slower, requiring as long as 48 h to complete
11154
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
(cyclodiborazanyl)aminoborohydride (BCDB).3e,26 The pre-
cipitate was characterized by IR, which shows a broad N−H
band at 3213 cm−1 and almost no bands for the B−H stretch
(Figure S2 in the Supporting Information [SI]).
Complex 2, which differs from 1 only by the ancillary
phosphine ligands, shows improved catalytic activity. When 2
and 20 equiv of AB were mixed in a similar fashion as the
reaction with 1, the color of the mixture changed from yellow
to red immediately, indicative of a faster generation of the
active species. The amount of H2 produced in the beginning of
the reaction was found to be higher than the reaction catalyzed
by 1. However, as time progressed, the solution became darkly
colored, likely due to catalyst degradation. The decompostion
of the catalyst was further substantiated by the fact that H2
evolution ceased after approximately 12 h.
Complex 3 is a superior catalyst in terms of both rate and the
extent of H2 released (Figure 5). After 6 h of mixing, the 11B
NMR spectrum of the reaction (Figure S1 in the SI) indicates
formation of borazine, polyborazylene, CTB, and BCDB. The
multiplet observed at −37.7 ppm is assigned to BH3·PMe2Ph
and integrated as 10 mol % of starting AB, suggesting that both
PMe2Ph ligands of 3 have been trapped by BH3. Measuring H2
evolution from this process gives 1.0 equiv, 2.0 equiv, and 2.5
equiv of H2 per AB in 1.5, 7, and 16 h, respectively. Initiation of
an active species from 3 appears to be as fast as 2; an immediate
color change from yellow to red was observed upon mixing 3
with AB. However, as illustrated by Figure 5, the amount of
released H2 still increases after 12 h, implying that the lifetime
of the catalytically active species in this case is longer than that
of 2.
Figure 4. Side view of compound 3 (hydride shown in green).
Fe atoms.24 The pincer aromatic ring (C1 through C6 plane) is
tilted 8.41(5)° from the C1−O1−O2−Fe plane while leaning
toward the trans PMe2Ph. The Fe−P(4) distance [2.2789(11)
Å] is longer than the Fe−P(3) distance [2.2210(11) Å],
presumably due to a stronger trans influence of the hydride
than the pincer aromatic ring. Compared to 1, all the Fe−P
bonds of 3 with the exception of the Fe−P3 bond are elongated
by 0.02 Å, further suggesting that the iron center is more
crowded.
Dehydrogenation of AB Catalyzed by 1−3. At room
temperature, complex 1 shows no catalytic activity for the
dehydrogenation of AB. At 60 °C, however, an immediate gas
evolution was observed when a solution of 1 in THF was mixed
with a solution of AB (20 equiv) in diglyme. The amount of H2
was measured by displacing water from an inverted buret. After
24 h, 2.5 equiv of H2 per AB was released (Figure 5). It should
Although the reaction profiles in Figure 5 show no evidence
for an induction period as commonly observed in heteroge-
neous catalysis,27 the color change during the aforementioned
dehydrogenation reactions, particularly in the case of 2,
cautioned us about the possibility of iron nanoparticles
catalyzing the reaction. A commonly used method to
distinguish between homogeneous and heterogeneous catalysis
is the mercury test. However, as Morris et al. have pointed
out,28 Hg(0) may not form an amalgam with iron, therefore a
negative mercury test is not sufficient to rule out a
heterogeneous mechanism. Nevertheless, excess elemental
mercury was added to the dehydrogenation of AB catalyzed
by 1−3, which resulted in no appreciable change in H2
evolution (Figures S3−S5 in the SI). More definitive evidence
supporting a homogeneous system for our iron POCOP-pincer
catalysts is the reproducibility of the kinetics experiments that
will be described in the next section.
While effective, all three iron complexes have a limited life
span under catalytic conditions, either being converted to a
catalytically less active species or being decomposed. For
instance, when the reaction between 3 and 20 equiv of AB in
THF/diglyme was complete (20 h at 60 °C), a second 20 equiv
of AB was added, resulting in only 75% of AB being reacted
after 24 h.
Figure 5. Amount of H2 released from AB (1.0 M solution in 1:4
THF/diglyme) at 60 °C with or without an iron catalyst (5 mol %).
be mentioned that thermal decomposition of AB also takes
place at 60 °C in THF/diglyme, but at a much slower rate. In
addition, without a catalyst, the amount of H2 released
plateaued at ∼1.3 equiv of H2 per AB. During the course of
the dehydrogenation reaction catalyzed by 1, the mixture
gradually changed color from yellow to red, accompanied by a
gray precipitate. The 11B NMR spectrum of the soluble
materials shows a doublet at 30.7 ppm and a broad singlet at
25.8 ppm, suggesting the formation of borazine and
polyborazylene, respectively. Other observable boron reso-
nances are a multiplet at −36.4 ppm for BH3·PMe3, and peaks
within the range from −5 to −25 ppm that have been
previously identified as cyclotriborazane (CTB)3e,25 and B-
Kinetic Study. Reaction profiles for iron-catalyzed dehy-
drogenation of AB were also generated by monitoring the
disappearance of AB by 11B NMR spectroscopy (Figure 6).
Treatment of the overall data shows a deviation from a first-
order reaction; rather, two different rate regions are identified.
At the beginning of the reaction, the decay of AB is fast and
linear as a function of time. Later on, the dehydrogenation
reaction becomes slower than what is expected for an
exponential decay of AB. Since the kinetics changes over the
11155
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
Figure 6. Disappearance of AB (1.0 M solution in 1:4 THF/diglyme)
at 60 °C with or without an iron catalyst (5 mol %).
Figure 7. Disappearance of NH3BH3 and its isotopomers (1.0 M in
1:4 THF/diglyme) in the presence of 5 mol % of 3 at 60 °C. The
concentrations of the substrates were determined from 11B NMR.
course of the reaction and the background reaction comes to
play as the reaction becomes slower, initial rates (20−25%
conversion for AB) from the linear region of the plots were
used for the kinetic study. The initial rates for the dehydrogen-
ation of AB were measured as (5.2 1.2) × 10−5, (2.3 0.5) ×
10−4, and (3.5 0.3) × 10−4 M/s for the reaction catalyzed by
1−3, respectively. Initial rates determined from gas evolution
measurements are similar except in the case of 1, where the H2
evolution rate is about twice that of the AB disappearance rate
(Table S1 in the SI). Compared to the uncatalyzed thermal
reaction, the rate for the dehydrogenation of AB catalyzed by 3
is about 50 times faster. Since 3 exhibits the highest catalytic
activity among the three iron pincer complexes and it shows
negligible decomposition at the early stage of the catalytic
reaction, it was chosen as the catalyst for further mechanistic
studies.
The dependence of dehydrogenation rate on substrate
concentration was determined by measuring the initial rates
νobs with variable [AB]0 (0.75−2.0 M) but a fixed catalyst
concentration of 0.050 M (Table S2 in the SI). Initial rates for
the uncatalyzed thermal reaction (νtherm) were also determined
under similar reaction conditions (Table S3 in the SI). Catalytic
rates νcat were then calculated by subtracting νtherm from νobs for
the same [AB]0. Within the experimental errors, the value for
νobs is constant, suggesting that the reaction has no dependence
on the concentration of AB (Figure S6 in the SI). Similarly, to
determine the reaction order for the catalyst, initial rates were
measured by varying the catalyst loading (2.5−7.5 mol %) with
a fixed initial AB concentration of 1.0 M. Plotting ln νcat against
ln[Fe] gives a straight line with a slope of 1.04 0.06 (Figure
S7 in the SI). Thus, the dehydrogenation of AB in the presence
of catalyst 3 is first order in 3 and zero order in AB (eq 2). The
lack of dependence on the concentration of AB suggests that 3
may rapidly interact with AB and fully convert to an
intermediate prior to the rate-determining step.
where N−H and B−H bonds are broken simultaneously in the
rate-determining step.
Stoichiometric and Catalytic Reactions of 3 with AB.
To gain a better understanding of how 3 releases H2 from AB, a
reaction of 3 with one equivalent of AB in 1:4 THF-d8/diglyme
1
was studied by 31P{1H} and H NMR spectroscopy. At room
temperature, a new iron species 4 started to emerge in the
31P{1H} NMR spectrum as a doublet at 235.4 ppm (JP−P = 13.0
Hz) and a triplet at 28.1 ppm (JP−P = 13.0 Hz) with an
integration ratio of 2:1. After 24 h, 15% of 3 was converted to 4
along with an equimolar amount of BH3·PMe2Ph, which was
characterized as a quartet at 2.6 ppm (JP−B = 56.8 Hz). At that
point, the free diphosphinite ligand (∼2% of the starting iron
complex) was also observed as a singlet at 145.8 ppm. The
same reaction performed at 60 °C for 10 min gave rise to a
similar mixture plus a new iron species 5, which was about 4%
of total iron species. The phosphorus resonances of 5 are
similar to those of 4, featuring a doublet at 233.8 ppm (JP−P
=
29.2 Hz) and a triplet at 38.5 ppm (JP−P = 29.2 Hz). The
1
hydride region of the H NMR spectrum displayed a broad
resonance at −9.09 ppm and a multiplet at −14.42 ppm
corresponding to 4. The integration ratio of these two
resonances was calculated to be 3:1. The hydride resonances
of 5 were, however, too weak to be unambiguously located.
Both 31P{1H} and 1H NMR data suggest that 4 is produced as a
result of one PMe2Ph ligand of 3 (most likely the one trans to
the hydride) being displaced by AB (Figure 8).30 The broad
resonance at −9.09 ppm can be assigned to the BH hydrogens.
A catalytic reaction of AB with 3 (5 mol % catalyst loading)
in 1:4 THF-d8/diglyme was also monitored by 31P{1H} and 1H
NMR spectroscopy. At 60 °C, 3 disappeared completely within
10 min, providing 4 as the major iron species (75% of total iron
species) as well as BH3·PMe2Ph (Figure 9). Of the three minor
iron species, one is consistent with 5 as described above. The
ν
= k[AB]0[3]1
(2)
cat
Despite the fact that the catalytic rate is independent of
[AB], kinetic isotope effects (KIEs) should tell if N−H or B−H
or both bonds are broken during the rate-determining step.29
As illustrated in Figure 7 and Table S6 in the SI, deuteration of
AB on either the nitrogen or the boron site results in a slower
rate. The product of the individual KIEs [(kNH3BH3/kND3BH3) ×
(kNH3BH3/kNH3BD3) = (2.0 0.3) × (1.7 0.3) = (3.4 1.1)]
equals the KIE observed for the doubly labeled substrate
[(kNH3BH3/kND3BD3) = (3.7
0.8)], supporting a mechanism
Figure 8. Proposed structure of 4.
11156
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
led to 60% conversion of 3 to 5 without the formation of 5′.31
The ratio between 3 and 5 stayed constant from 10 min to 24
h, suggesting that an equilibrium had been established.
Evidently, the noncoordinating solvent benzene not only favors
the dissociation of BH3 from BH3·THF, but also prevents the
isomerization of 5 to 5′.
Trapping of Aminoborane. Because more than 2 equiv of
H2 per AB were obtained from the catalytic reactions, it was
anticipated that most, if not all, aminoborane (NH2BH2) was
set free from the iron center. A metal-bound aminoborane
would favor the formation of insoluble oligomers, leading to
less H2 production.14,26 The presence of free aminoborane
could be probed by using cyclohexene as a trapping agent, a
method that was previously developed by Baker, Dixon, and co-
workers.26 Dehydrogenation of AB catalyzed by 5 mol % of 3
was therefore carried out in the presence of 20 equiv of
cyclohexene with respect to AB. The trapping product H2N−
B(C6H11)2 (47.8 ppm in 11B NMR) was observed from the very
beginning of the reaction. In addition to this boron species, a
small amount of borazine and polyborazylene was also detected
by 11B NMR.
Cross-Coupling of NMe3BH3 and NH3BEt3. To indirectly
determine whether H2 is released from the same AB molecule
or from intermolecular dehydrocoupling of two AB’s, complex
3 was mixed with 20 equiv of NMe3BH3 and 20 equiv of
NH3BEt3 in THF-d8/diglyme. At 60 °C after 24 h, a small
amount of H2 was produced. According to 11B NMR, only 29%
of the combined starting materials reacted to generate BH3·
PMe2Ph (5%) and some insoluble materials. This result
suggests that intermolecular dehydrocoupling between
NMe3BH3 and NH3BEt3 is possible, albeit at a rate that is
substantially slower than the rate for the dehydrogenation of
AB. These AB derivatives also trigger more rapid catalyst
decomposition, leading to the free phosphinite ligand (>15% in
4 h at 60 °C) as confirmed by 31P{1H} NMR.
Figure 9. 31P{1H} NMR spectra of (a) AB and 3 (5 mol % catalyst
loading) mixed in 1:4 THF-d8/diglyme at room temperature for 15
min, (b) heated at 60 °C for 10 min, (c) 60 °C for 1.5 h, and (d) 60
°C for 3 h (* for 4, • for 4′, + for 5, o for 6).
resonance at 238.2 ppm (represented by the open circle in
Figure 9) is a singlet, implying that both PMe2Ph ligands have
dissociated from iron. On the basis of our previous study of
ligand substitution reaction for the related compound 1, this
resonance is attributed to a pincer complex (6) bearing two AB
molecules. The resonances at 236.8 ppm (doublet, JP−P = 26.0
Hz) and 22.1 ppm (triplet, JP−P = 26.0 Hz) belong to one
compound 4′, which should contain only one PMe2Ph ligand.
This compound is likely to be the geometric isomer of 4 with
PMe2Ph bound trans rather than cis to the hydride. Compared
to 4, the more upfield shift of the PMe2Ph resonance of 4′ is
due to less shielding from the iron center induced by the
strongly trans-influencing hydride ligand. As the reaction
progressed, 4′ started to diminish, while the intensity of the
resonance for 5 and 6 increased gradually. After 1.5 h, which
was the time required to release 1 equiv of H2 from AB, a small
amount of other iron species became noticeable by 31P{1H}
NMR. They are likely to be some iron complexes bearing
partially dehydrogenated AB.
Compound 5 was thought to be a BH3-coordinated complex
resulting from 4 via the loss of NH3. To generate 5
independently, 3 was treated with 2 equiv of BH3·THF in
1:4 THF-d8/diglyme. Immediately following the mixing at
room temperature, 35% of 3 was reacted to yield 5 and its
geometric isomer 5′ (Scheme 1). After 1.5 h, the conversion of
3 was increased slightly to 40% with substantially more 5′ being
produced. Interestingly, replacing the mix-solvent with C6D6
DISCUSSION
■
Reactivity Difference Among 1−3. The space-filling
model of 1 and 3 (Figure 10) supports the notion that having
Scheme 1. Room Temperature Reaction of 3 with BH3·THF
in 1:4 THF-d8/Diglyme
Figure 10. Space-filling model of 1 (left) and 3 (right). The hydride
ligand is shown in green.
PMe2Ph instead of PMe3 as the ancillary ligand makes the iron
center more crowded. As predicted, the rate for substituting
CO for PMe2Ph from 2 and 3 is much faster than the rate for
substituting CO for PMe3 from 1.32 Our previous work on
complexes of this type shows that the kinetic substitution
products are those with the trans-PMe3 or PMe2Ph being
displaced by an incoming ligand.20,32 These results are
consistent with the trans effect order of H > Ar (aromatic
ring). In comparison with 1, the higher catalytic activity of 2 in
the dehydrogenation of AB can be rationalized by a more rapid
dissociation of the monophosphine. The reactivity of 3 is
11157
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
further enhanced by the p-methoxy group, which provides more
electron density to the iron center to facilitate ligand
dissociation. The resulting coordinatively unsaturated inter-
mediate, the presumed active species, is better stabilized by the
more electron-donating PMe3 or the methoxy group. The
lifetime of the active catalyst formed from 1 or 3 is therefore
longer than 2, which is consistent with the observation that 1
and 3 outperform 2 for a higher amount of H2 released (Figure
5).
Scheme 2. Proposed Catalytic Cycle for the
Dehydrogenation of AB Catalyzed by 3
Mechanism for the Early Stage of AB Dehydrogen-
ation. Compared to Baker’s iron systems,9,14 incorporating a
POCOP-pincer ligand in this work does show improved
stability for iron catalysts, providing the opportunity to
interpret the kinetic data at the molecular level. The
diamagnetic nature of our iron complexes also allowed us to
conveniently investigate the dehydrogenation reaction using
NMR spectroscopy. As illustrated by Figure 9, the reaction of 3
with AB generates a mixture of iron species and the ratios of
these species change over time, which inevitably complicates
the kinetics for the overall dehydrogenation process. However,
at the beginning of the reaction (<20 min, 20−25% conversion
for AB), the dominant iron species is the AB-bound complex 4.
Therefore, it is reasonable to assume that 4 is responsible for
the initial release of H2. The lack of dependence of the initial
rate on [AB] is due to a rapid conversion of 3 to 4 under the
catalytic conditions. If the rate-determining step follows
without the involvement of another AB molecule, the
dehydrogenation rate would be first-order dependent on [Fe]
as shown in eq 2. Williams and co-workers have also reported
that the dehydrogenation of AB catalyzed by Shvo’s catalyst is
zero order in [AB] during the first two stages of the reaction.33
They have attributed it to the dissociation of the starting
diruthenium complex and the loss of H2 from a hydrox-
ycyclopentadienyl ruthenium hydride being the rate-determin-
ing step for the first and second stage of the dehydrogenation
reaction. Schneider et al. have recently shown that dehydrogen-
ation of ND3BH3 and ND3BD3 catalyzed by a 5-coordinate
PNP-pincer ruthenium complex is zero order in the substrate
for more than two half-lives.34 Interestingly, dehydrogenation
of the protio AB under otherwise the same reaction conditions
exhibits first order in AB.
likely. To account for the observed rate law in eq 2 and the
KIEs, one would have to propose a rapid loss of the second
PMe2Ph or the dissociation of a pincer arm,40 followed by rate-
limiting protonation of the hydride by the NH3 moiety of AB.
The NMR experiments described earlier show no evidence for
the formation of a 5-coordinate intermediate. Dissociation of
the second PMe2Ph eventually takes place but at a rate too slow
to explain the rapid catalysis at the beginning of the
dehydrogenation process.
Catalyst Deactivation Pathways. Although the iron
catalysts described herein have reasonable thermal stability, a
number of pathways could potentially lead to a gradual loss of
catalytic activity over time. In the case of 3 (Scheme 3), the
Scheme 3. Reaction Pathways for 4 under the Catalytic
Conditions
A mechanism that is consistent with our kinetic data
including KIEs is outlined in Scheme 2. The proposed rate-
determining step involves simultaneously breaking the N−H
and B−H bonds of AB, which results in the protonation of the
ipso carbon and the transfer of a hydride ligand from B to Fe.
Such an activation pathway for AB or H2 has been more
commonly proposed in Noyori−Morris35 or Shvo-type36
bifunctional catalysts, where the metal center accepts a hydride
and the neighboring heteroatom (nitrogen or oxygen) accepts a
proton.5,6,14,33 Direct participation of a carbon-based ligand in
the activation of AB, H2, or related molecules is less common,
but not unprecedented. A computational study carried out by
Yang and Hall37 on Baker’s Ni(NHC)2 system9 has revealed
that dehydrogenation of AB starts with proton transfer from
nitrogen to the carbene carbon. A more recent DFT study by
Vanka et al. has further supported the feasibility of using a C−
M bond as a motif for designing ligand−metal bifunctional
catalysts.38 Experimental observation of the dihydride species
shown in Scheme 2 is, however, challenging. The analogous C−
H bond activation to restore the pincer framework has been
shown to be facile even at room temperature.39 Alternative
mechanisms without the protonation of the ipso carbon are less
catalytically active species 4 undergoes a geometric isomer-
ization to yield 4′, placing the AB molecule trans to the ipso
carbon. Under the catalytic conditions, a large excess of AB is
available for further substitution of PMe2Ph in 4′ to generate 6,
which contains two AB molecules. Alternatively, 4 can bypass
4′ and be converted to 6 directly. Perhaps 6 is capable of losing
H2 via a similar transition state as described for 4. The
formation of 4′ makes the protonation of the ipso carbon
impossible, and prevents H2 loss unless the NH3 moiety of the
11158
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
solvents were collected from an Innovative Technology solvent
purification system and used throughout all experiments. Methanol
was degassed by bubbling argon through it for 15 min and then dried
over molecular sieves. Tetrahydrofuran-d8 (99.5% D, packed in sealed
ampules) was purchased from Cambridge Isotope Laboratories, Inc.
and used without further purification. Ammonia borane (AB),
anhydrous diglyme, and NMe3BH3 were purchased from Sigma-
Aldrich and used as received. 1H, 13C{1H}, and 31P{1H} NMR spectra
were recorded on a Bruker Avance-400 MHz spectrometer. 11B NMR
spectra were recorded on a Bruker AMX 400 MHz wide-bore
spectrometer. Chemical shift values in 1H and 13C NMR spectra were
referenced internally to the residual solvent resonances. 31P{1H} and
11B NMR spectra were referenced externally to 85% H3PO4 (0 ppm)
and BF3·Et2O (0 ppm), respectively. Infrared spectra were recorded on
a Thermo Scientific Nicolet 6700 FT-IR spectrometer equipped with
smart orbit diamond attenuated total reflectance (ATR) accessory. 1,3-
(iPr2PO)2C6H3,42 1,3-(iPr2PO)2-5-OMe-C6H3,43 and complex 120
were prepared as described in the literature.
coordinated AB can protonate the hydride. A main catalyst
deactivation pathway is the formation of 5 through the
dissociation of NH3 from 4. As shown in Figure 9, its
concentration is built up significantly after 3 h at 60 °C, thereby
reducing the concentration of the catalytically active species.
Another catalyst deactivation pathway is the formation of the
free diphosphinite ligand, which was observed from the
stoichiometric reaction between 3 and AB. To our surprise,
under the catalytic conditions, the free diphosphinite ligand was
not detected by NMR. Decomplexation of the pincer ligand is
somehow inhibited when AB is present in large excess.
Nevertheless, the release of the diphosphinite ligand is expected
toward the late stage of the dehydrogenation when the
concentration of AB is substantially reduced. More pronounced
catalyst degradation was observed when 3 was treated with
NMe3BH3 and NH3BEt3. Upon interacting with the iron
center, these more bulky substrates probably cause more
structural perturbation to the complex, thereby facilitating aryl
hydride reductive elimination to generate the free diphosphinite
ligand.
Preparation of “Fe(PMe2Ph)4”. “Fe(PMe2Ph)4” was prepared
according to a similar method that was reported for the synthesis of
Fe(PMe3)4.20,44 In a Schlenk flask under an argon atmosphere, Mg
turnings (1.50 g, 61.7 mmol) were vigorously stirred for at least 24 h
while being heated by a 130 °C oil bath. Once the flask was cooled
down to room temperature, FeCl2 (500 mg, 3.94 mmol), THF (15
mL), and PMe2Ph (2.8 mL, 19.7 mmol) were added sequentially. The
resulting mixture was stirred vigorously at room temperature for 2−3 h
by which time the color of the mixture changed to yellowish brown.
The volatiles were removed under the vacuum, and the remaining
solid was treated with 30 mL of pentane. After filtration, removal of
pentane from the filtrate under vacuum produced “Fe(PMe2Ph)4” as a
sticky yellow solid (2.27g, 95% yield), which was used for the
subsequent synthesis without further purification. NMR spectra of this
material showed at least three different iron species, likely due to the
activation of both aliphatic and aromatic C−H bonds. 31P {1H} NMR
(162 MHz, C6D6, δ): −7.1 to −5.8 (br, 1P), 31.0−31.9 (br, 1P), 41.1−
42.3 (br, 2P) (species 1); −37.6 (dt, JP−P = 46.8 and 24.8 Hz, 1P), 29.3
(dd, JP−P = 35.6 and 24.8 Hz, 2P), 39.5 (dt, JP−P = 46.8 and 35.6 Hz,
1P) (species 2); −15.3 to −14.1 (m, 1P), 37.7 (m, 3P) (species 3).
Synthesis of [2,6-(iPr2PO)2-C6H3]Fe(H)(PMe2Ph)2 (2). 1,3-
(iPr2PO)2C6H4 (843 mg, 2.46 mmol) and “Fe(PMe2Ph)4” (1.5 g,
2.47 mmol) were mixed in 15 mL of THF at room temperature. The
resulting mixture was stirred for 48 h, at which point the color of the
solution changed from yellow to dark brown. The volatiles were
removed under vacuum, and the yellowish residue was treated with a
mixture of pentane and diethyl ether (20 mL each) and filtered
through a pad of Celite. The volatiles were removed under vacuum to
give a yellow oily residue, which was washed with cold methanol (5
mL × 3), and the remaining solid was dried under vacuum. The
resulting slightly brownish solid was redissolved in diethyl ether and
filtered through a pad of Celite. The filtrate was dried under vacuum to
produce a yellow powder (1.05 g, 63% yield). The pure product
should appear yellow in color. If needed, the steps from washing with
cold methanol to filtration through Celite and drying under vacuum
can be repeated to ensure the purity. 1H NMR (400 MHz, THF-d8, δ):
−15.04 (tdd, JP−H = 76.8, 48.0, and 23.2 Hz, FeH, 1H), 0.97−1.05 (m,
CHCH3, 12H), 1.08−1.13 (m, CHCH3, 12H), 1.17 (d, JP−H = 3.4 Hz,
PMe2Ph, 6H), 1.64 (d, JP−H = 3.8 Hz, PMe2Ph, 6H), 2.14−2.18 (m,
CHCH3, 2H), 2.24−2.32 (m, CHCH3, 2H), 6.28 (d, JP−H = 7.4 Hz,
ArH, 2H), 6.60 (t, JP−H = 7.4 Hz, ArH, 1H), 7.20−7.26 (m, ArH, 6H),
7.36 (t, JH−H = 6.9 Hz, ArH, 2H), 7.57 (t, JH−H = 7.4 Hz, ArH, 2H).
13C {1H} NMR (101 MHz, THF-d8, δ): 18.2 (s, CHCH3), 19.1 (s,
CHCH3), 19.9 (s, CHCH3), 20.0 (s, CHCH3), 21.2 (d, JP−C = 20.6 Hz,
PMe2Ph), 27.6 (dd, JP−C = 19.9 and 3.4 Hz, PMe2Ph), 35.4 (s,
CHCH3), 35.4−35.8 (m, CHCH3), 102.9−103.1 (m, ArC), 123.2 (d,
JP−C = 3.4 Hz, ArC), 127.9 (s, ArC), 128.1 (d, JP−C = 7.3 Hz, ArC),
128.3 (d, JP−C = 6.3 Hz, ArC), 128.5 (s, ArC), 131.0 (d, JP−C = 9.5 Hz,
ArC), 131.2 (d, JP−C = 8.0 Hz, ArC), 148.8 (d, JP−C = 24.8 Hz, ArC),
148.8−149.1 (m, ArC), 149.4 (d, JP−C = 17.4 Hz, ArC), 165.2 (td, JP−C
= 10.0 and 4.5 Hz, ArC). 31P {1H} NMR (162 MHz, THF-d8, δ): 14.6
(dt, JP−P = 32.2 and 25.0 Hz, PMe2Ph, 1P), 23.7 (dt, JP−P = 32.2 and
CONCLUSIONS
■
In recent years, the interest in AB has been driven by its
potential application as chemical hydrogen-storage materials.
Although significant progress has been made in developing
catalytic systems for the release of H2 from AB, many
challenges remain. In addition to identifying efficient and
cost-effective methods to regenerate AB from spent fuels,2f,41
there is also an important need for designing catalysts that are
derived from earth-abundant metals and can release most of the
available H2 from AB.
Herein we have described an iron-based catalytic system
releasing 2.3−2.5 equiv H2 per AB, which, to the best of our
knowledge, is the highest extent of H2 release among all the
iron systems known to date. Installing a POCOP-pincer ligand
has been demonstrated as a key to stabilize the iron complexes
and extend the lifetime of the iron catalysts for the release of H2
over a long period of time. The relatively high stability of iron
pincer complexes has also made mechanistic investigation
possible. At the early stage of the dehydrogenation process, the
catalytically active species is generated via a rapid dissociation
of the phosphine ligand trans to the hydride, followed by the
coordination of AB. The phosphine dissociation step can be
promoted by a more bulky ligand and a more donating pincer
ligand. The dehydrogenation reaction is first-order dependent
on the iron catalyst and zero-order dependent on AB. The rate-
determining step involves simultaneous transfer of H− and H+
from AB to the iron and pincer ipso carbon respectively, which
has been supported by kinetic isotope effect measurements. As
the dehydrogenation reaction proceeds, geometric isomer-
ization of the AB-bound complex, further substitution of the
phosphine by AB, dissociation of NH3 from the AB-bound
complex, and degradation of the iron complexes to the free
diphosphinite ligand take place. At the very least the latter two
processes are the culprits for the diminished catalytic activity
toward the late stage of the dehydrogenation process. Our
future efforts will be devoted to developing more robust and
efficient iron pincer catalysts.
EXPERIMENTAL SECTION
General Comments. All the organometallic compounds were
prepared and handled under an argon atmosphere using standard
Schlenk and inert-atmosphere box techniques. Dry and oxygen-free
■
11159
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
13.2 Hz, PMe2Ph, 1P), 217.4 (dd, JP−P = 25.0 and 13.2 Hz, OPiPr2,
2P). IR (solid): νFe−H =1890 cm−1. Anal. Calcd for C34H54FeO2P4: C,
60.54; H, 8.07. Found: C, 60.40; H, 8.07.
(4.7 g, 95% yield). About 1.5 equiv of 1,4-dioxane per NH3BEt3
remained present despite prolonged drying, and therefore this material
was used in further experiments. 1H NMR (400 MHz, C6D6, δ): 0.25−
0.35 (m, CH2, 6H), 0.88−0.95 (m, CH3, 9H), 1.83 (br, NH3, 3H),
3.34 (s, CH2 of 1,4-dioxane, 12 H). 13C{1H} NMR (101 MHz, C6D6,
δ): 10.2 (s, CH3), 14.8 (br, CH2), 67.0 (s, CH2 of 1,4-dioxane). 11B
NMR (128 MHz, C6D6, δ): −12.4 (s).
NMR Reaction of 3 with both NMe3BH3 and NH3BEt3. In a
flame-dried J. Young NMR tube, complex 3 (12.3 mg, 17.5 μmol),
NMe3BH3 (25.5 mg, 0.35 mmol), and NH3BEt3 (86.5 mg, 0.35 mmol)
were mixed in THF-d8 (70 μL) and diglyme (280 μL). BF3·Et2O
(sealed in a capillary tube) was added as an external standard. The
progress of the reaction was monitored by NMR at room temperature.
After 24 h, the NMR tube was connected to a Schlenk line, exposed to
an argon atmosphere, and heated at 60 °C. The reaction was further
monitored at this temperature by NMR.
Synthesis of {2,6-(iPr2PO)2-4-(OMe)C6H2}Fe(H)(PMe2Ph)2 (3).
This compound was prepared in 61% yield by a procedure similar to
that used for 2. X-ray quality crystals of 3 were obtained from
recrystallization of the iron complex in THF/diethyl ether at −30 °C.
1H NMR (400 MHz, THF-d8, δ): −15.32 (tdd, JP−H = 77.5, 48.1, and
24.9 Hz, FeH, 1H), 0.95−1.04 (m, CHCH3, 12H), 1.08−1.15 (m,
CHCH3, 12H), 1.18 (d, JP−H = 5.3 Hz, PMe2Ph, 6H), 1.62 (d, JP−H
=
5.5 Hz, PMe2Ph, 6H), 2.09−2.21 (m, CHCH3, 2H), 2.23−2.31 (m,
CHCH3, 2H), 3.65 (s, OCH3, 3H), 6.02 (s, ArH, 2H), 7.21−7.28 (m,
ArH, 6H), 7.38 (t, JH−H = 7.4 Hz, ArH, 2H), 7.54 (t, JH−H = 7.8 Hz,
ArH, 2H). 13C{1H} NMR (101 MHz, THF-d8, δ): 18.4 (s, CHCH3),
19.3 (s, CHCH3), 20.0 (s, CHCH3), 20.2 (s, CHCH3), 21.5 (d, JP−C
=
20.2 Hz, PMe2Ph), 27.9 (dd, JP−C = 20.2 and 4.0 Hz, PMe2Ph), 35.7 (s,
CHCH3), 35.6−36.0 (m, CHCH3), 55.2 (s, OCH3), 91.0 (d, JP−C = 1.4
Hz, ArC), 128.0 (s, ArC), 128.3 (d, JP−C = 7.4 Hz, ArC), 128.4 (d, JP−C
= 6.2 Hz, ArC), 128.6 (s, ArC), 131.1 (d, JP−C = 9.4 Hz, ArC), 131.3
(d, JP−C = 8.1 Hz, ArC), 134.5−135.3 (m, ArC), 149.2 (d, JP−C = 25.3
Hz, ArC), 149.7 (d, JP−C = 16.2 Hz, ArC), 159.8 (d, JP−C = 3.2 Hz,
ArC), 164.6 (td, JP−C = 10.0 and 3.6 Hz, ArC). 31P{1H} NMR (162
MHz, THF-d8, δ): 15.0 (dt, JP−P = 32.0 and 24.8 Hz, PMe2Ph, 1P),
23.9 (dt, JP−P = 32.0 and 13.4 Hz, PMe2Ph, 1P), 218.8 (dd, JP−P = 24.8
and 13.4 Hz, OPiPr2, 2P). IR (solid): νFe−H = 1927 cm−1. Anal. Calcd
for C35H56FeO3P4: C, 59.67; H, 8.01. Found: C, 59.57; H, 8.19.
General Procedure for Measuring the Amount of H2
Produced from Catalytic Dehydrogenation of AB. In a flame-
dried 5 mL Schlenk tube, a 1.25 M diglyme solution of AB (0.6 mL,
0.75 mmol) was mixed with a 0.25 M THF solution of an iron catalyst
(0.15 mL, 0.0375 mmol). The Schlenk tube was closed by a Teflon
valve, and its side arm was attached to a Sigma-Aldrich gas measuring
buret with a thin Tygon tubing. The entire system was purged with
argon before the Teflon valve of the Schlenk tube was opened to
equilibrate the system. The Schlenk tube was immersed in a 60 °C oil
bath while keeping the reaction mixture stirring at a fixed rate of 300
rpm. The amount of H2 evolved was measured at different time
intervals.
ASSOCIATED CONTENT
* Supporting Information
■
S
11B NMR spectra of catalytic dehydrogenation of AB, IR
spectrum of the insoluble materials obtained from AB
dehydrogenation, complete details of the crystallographic
study, kinetic data, and plots of these data. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
hairong.guan@uc.edu
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This paper is dedicated to the memory of Professor Sheldon G.
Shore (The Ohio State University), a great mentor and a
pioneer in studying boron hydrides. We thank the National
Science Foundation (CHE-0952083) and the Alfred P. Sloan
Foundation (research fellowship to H.G.) for supporting this
research. Crystallographic data were collected on a Bruker
SMART6000 diffractometer which was funded by an NSF-MRI
grant (CHE-0215950).
General Procedure for Kinetic Study of Catalytic Dehydro-
genation of AB. In a flame-dried J. Young NMR tube, a diglyme
solution of AB (0.4 mL) was mixed with an appropriate amount of 3
dissolved in THF (0.1 mL). BF3·Et2O (sealed in a capillary tube) was
added as an external standard. The NMR tube was then attached to a
Schlenk line under an argon atmosphere. The solution was heated at
60 °C allowing H2 to escape through the Schlenk line, which was
connected to a bubbler (reactions in a sealed NMR tube without
venting were much slower). 11B NMR spectrum of the reaction
mixture was recorded periodically to measure the change in [AB] over
time. The rates for uncatalyzed reaction were measured similarly.
NMR Reaction of 3 with AB. In a flame-dried J. Young NMR
tube, complex 3 (10 mg, 14.2 μmol) and AB (14.2 μmol for the
stoichiometric reaction or 0.284 mmol for the catalytic reaction) were
dissolved in THF-d8 and diglyme (1:4). The progress of the reaction
REFERENCES
■
(1) Shore, S. G.; Parry, R. W. J. Am. Chem. Soc. 1955, 77, 6084−6085.
(2) (a) Stephens, F. H.; Pons, V.; Baker, R. T. Dalton Trans. 2007,
2613−2626. (b) Marder, T. B. Angew. Chem., Int. Ed. 2007, 46, 8116−
8118. (c) Peng, B.; Chen, J. Energy Environ. Sci. 2008, 1, 479−483.
(d) Hamilton, C. W.; Baker, R. T.; Staubitz, A.; Manners, I. Chem. Soc.
Rev. 2009, 38, 279−293. (e) Staubitz, A.; Robertson, A. P. M.;
Manners, I. Chem. Rev. 2010, 110, 4079−4124. (f) Smythe, N. C.;
Gordon, J. C. Eur. J. Inorg. Chem. 2010, 509−521.
1
was monitored at room temperature by H, 31P{1H}, and 11B NMR.
(3) Thermal decomposition of AB in the solid state, ionic liquids, or
solutions: (a) Wolf, G.; Baumann, J.; Baitalow, F.; Hoffmann, F. P.
Thermochim. Acta 2000, 343, 19−25. (b) Baitalow, F.; Baumann, J.;
After 24 h, the NMR tube was connected to a Schlenk line filled with
argon and heated at 60 °C. The reaction was further monitored at this
temperature by NMR.
Wolf, G.; Jaenicke-Roßler, K.; Leitner, G. Thermochim. Acta 2002, 391,
̈
Procedure for the Trapping of Aminoborane. In a flame-dried
J. Young NMR tube, 3 (17.6 mg, 25.0 μmol) and AB (15.4 mg, 0.5
mmol) were dissolved in the mixture of THF (0.1 mL) and diglyme
(0.4 mL). Cyclohexene (1.01 mL, 10 mmol) was added to the NMR
tube, which was then exposed to an argon atmosphere and heated at
60 °C. The progress of this reaction was monitored by 11B NMR.
Synthesis of NH3BEt3. This compound was prepared according to
a slightly modified procedure from the one described by Schneider and
co-workers.34 In a flame-dried Schlenk flask, a 0.50 M solution of NH3
in 1,4-dioxane (50 mL, 25 mmol) was mixed with a 1.0 M solution of
BEt3 in hexane (20 mL, 20 mmol) at −78 °C. The mixture was
gradually warmed to room temperature and stirred for 30 min.
Volatiles were removed under vacuum, resulting in a clear oily material
159−168. (c) Stowe, A. C.; Shaw, W. J.; Linehan, J. C.; Schmid, B.;
Autrey, T. Phys. Chem. Chem. Phys. 2007, 9, 1831−1836. (d) Bluhm,
M. E.; Bradley, M. G.; Butterick, R., III; Kusari, U.; Sneddon, L. G. J.
Am. Chem. Soc. 2006, 128, 7748−7749. (e) Shaw, W. J.; Linehan, J. C.;
Szymczak, N. K.; Heldebrant, D. J.; Yonker, C.; Camaioni, D. M.;
Baker, R. T.; Autrey, T. Angew. Chem., Int. Ed. 2008, 47, 7493−7496.
(f) Al-Kukhun, A.; Hwang, H. T.; Varma, A. Int. J. Hydrogen Energy
2013, 38, 169−179.
(4) Denney, M. C.; Pons, V.; Hebden, T. J.; Heinekey, D. M.;
Goldberg, K. I. J. Am. Chem. Soc. 2006, 128, 12048−12049.
(5) Blaquiere, N.; Diallo-Garcia, S.; Gorelsky, S. I.; Black, D. A.;
Fagnou, K. J. Am. Chem. Soc. 2008, 130, 14034−14035.
11160
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161

Journal of the American Chemical Society
Article
(6) Kaß, M.; Friedrich, A.; Drees, M.; Schneider, S. Angew. Chem., Int.
Ed. 2009, 48, 905−907.
(7) Kim, S.-K.; Han, W.-S.; Kim, T.-J.; Kim, T.-Y.; Nam, S. W.;
(31) We cannot rule out the possibility of THF molecule remaining
coordinated to the boron in 5 and 5′. The 11B NMR of the reaction in
Scheme 1 shows a broad singlet at 18.4 ppm in addition to a quartet of
doublets at −37.5 ppm (J = 90.8 and 53.8 Hz) as expected for BH3·
PMe2Ph. This result suggests that BH3·THF undergoes rapid BH3
exchange with 5 and 5′.
̈
́
Mitoraj, M.; Piekos, Ł.; Michalak, A.; Hwang, S.-J.; Kang, S. O. J. Am.
Chem. Soc. 2010, 132, 9954−9955.
(8) Jaska, C. A.; Temple, K.; Lough, A. J.; Manners, I. J. Am. Chem.
Soc. 2003, 125, 9424−9434.
(32) Upon exposure to 1 atm of CO, displacing the trans-PMe2Ph
from 2 and 3 by CO completes in 5 min, whereas the similar reaction
with 1 requires 24 h to reach completion. For details, see:
Bhattacharya, P. Small Molecule Activation with Iron Pincer Complexes.
Ph.D. Thesis, University of Cincinnati: Cincinnati, 2014.
(33) (a) Conley, B. L.; Williams, T. J. Chem. Commun. 2010, 46,
4815−4817. (b) Lu, Z.; Conley, B. L.; Williams, T. J. Organometallics
2012, 31, 6705−6714.
(9) Keaton, R. J.; Blacquiere, J. M.; Baker, R. T. J. Am. Chem. Soc.
2007, 129, 1844−1845.
(10) Conley, B. L.; Guess, D.; Williams, T. J. J. Am. Chem. Soc. 2011,
133, 14212−14215.
(11) Bera, B.; Jagirdar, B. R. Inorg. Chim. Acta 2011, 372, 200−205.
(12) Plietker, B., Ed. Iron Catalysis in Organic Chemistry: Reactions
and Applications; Wiley-VCH: Weinheim, Germany, 2008.
(13) (a) Rathke, J. W.; Muetterties, E. L. J. Am. Chem. Soc. 1975, 97,
3272−3273. (b) Karsch, H. H.; Klein, H.-F.; Schmidbaur, H. Angew.
Chem., Int. Ed. Engl. 1975, 14, 637−638.
(34) Marziale, A. N.; Friedrich, A.; Klopsch, I.; Drees, M.; Celinski, V.
R.; Schmedt auf der Gunne, J.; Schneider, S. J. Am. Chem. Soc. 2013,
̈
135, 13342−13355.
(35) (a) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97−102.
(b) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40−73.
(c) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev.
2004, 248, 2201−2237. (d) Ikariya, T.; Blacker, A. J. Acc. Chem. Res.
2007, 40, 1300−1308. (e) Morris, R. H. Chem. Soc. Rev. 2009, 38,
2282−2291.
(14) Baker, R. T.; Gordon, J. C.; Hamilton, C. W.; Henson, N. J.; Lin,
P.-H.; Maguire, S.; Murugesu, M.; Scott, B. L.; Smythe, N. C. J. Am.
Chem. Soc. 2012, 134, 5598−5609.
(15) Vance, J. R.; Robertson, A. P. M.; Lee, K.; Manners, I. Chem.
Eur. J. 2011, 17, 4099−4103.
(16) Vance, J. R.; Schafer, A.; Robertson, A. P. M.; Lee, K.; Turner, J.;
̈
(36) Conley, B. L.; Pennington-Boggio, M. K.; Boz, E.; Williams, T. J.
Chem. Rev. 2010, 110, 2294−2312.
Whittell, G. R.; Manners, I. J. Am. Chem. Soc. 2014, 136, 3048−3064.
(17) Sonnenberg, J. F.; Morris, R. H. ACS Catal. 2013, 3, 1092−
1102.
(37) Yang, X.; Hall, M. B. J. Am. Chem. Soc. 2008, 130, 1798−1799.
(38) Ghatak, K.; Mane, M.; Vanka, K. ACS Catal. 2013, 3, 920−927.
(39) Xu, G.; Sun, H.; Li, X. Organometallics 2009, 28, 6090−6095.
(40) The mechanism involving the dissociation of the pincer arm
cannot be exclusively ruled out here. However, such a mechanism is
disfavored based on our mechanistic studies on the geometric
isomerization of related 18-electron iron POCOP-pincer complexes,
which argue against the dissociation of the pincer arm.32
(41) Sutton, A. D.; Burrell, A. K.; Dixon, D. A.; Garner, E. B., III;
Gordon, J. C.; Nakagawa, T.; Ott, K. C.; Robinson, J. P.; Vasiliu, M.
Science 2011, 331, 1426−1429.
(18) (a) Gossage, R. A.; van de Kuil, L. A.; van Koten, G. Acc. Chem.
Res. 1998, 31, 423−431. (b) Albrecht, M.; van Koten, G. Angew.
Chem., Int. Ed. 2001, 40, 3750−3781. (c) Singleton, J. T. Tetrahedron
2003, 59, 1837−1857. (d) van der Boom, M. E.; Milstein, D. Chem.
Rev. 2003, 103, 1759−1792. (e) Liang, L.-C. Coord. Chem. Rev. 2006,
250, 1152−1177. (f) Nishiyama, H. Chem. Soc. Rev. 2007, 36, 1133−
1141. (g) The Chemistry of Pincer Compounds; Morales-Morales, D.,
Jensen, C. M., Eds.; Elsevier: Amsterdam, 2007. (h) Choi, J.;
MacArthur, A. H. R.; Brookhart, M.; Goldman, A. S. Chem. Rev.
2011, 111, 1761−1779. (i) Selander, N.; Szabo,
́
K. J. Chem. Rev. 2011,
(42) Pandarus, V.; Zargarian, D. Organometallics 2007, 26, 4321−
4334.
(43) Vabre, B.; Spasyuk, D. M.; Zargarian, D. Organometallics 2012,
31, 8561−8570.
(44) Karsch, H. H. Chem. Ber. 1977, 110, 2699−2711.
111, 2048−2076.
(19) For a review on iron pincer complexes, see: Bhattacharya, P.;
Guan, H. Comments Inorg. Chem. 2011, 32, 88−112.
(20) Bhattacharya, P.; Krause, J. A.; Guan, H. Organometallics 2011,
30, 4720−4729.
(21) The Tolman cone angles for PMe2Ph and PMe3 are 122° and
118°, respectively. For details, see Tolman, C. A. Chem. Rev. 1977, 77,
313−348.
i
(22) In a protic solvent such as PrOH, iron complexes bearing a
PNNP ligand (G in Figure 1) can release 2.5−2.9 equiv of H2 per
AB.17 However, this process produces B(OiPr)3 as the spent fuel, and
the B−O bonds are too strong to allow the regeneration of AB. For
this reason, studies of AB dehydrogenation have been focused on
reactions carried out in aprotic solvents.
(23) See Experimental Section for details.
(24) The deviation from planarity for the C1−O1−O2−Fe plane was
calculated to be 0.0027 Å.
(25) Gaines, D. F.; Schaeffer, R. J. Am. Chem. Soc. 1963, 85, 3592−
3594.
(26) Pons, V.; Baker, R. T.; Szymczak, N. K.; Heldebrant, D. J.;
Linehan, J. C.; Matus, M. H.; Grant, D. J.; Dixon, D. A. Chem.
Commun. 2008, 6597−6599.
(27) (a) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003,
198, 317−341. (b) Crabtree, R. H. Chem. Rev. 2012, 112, 1536−1554.
(28) Mikhailine, A. A.; Maishan, M. I.; Lough, A. J.; Morris, R. H. J.
Am. Chem. Soc. 2012, 134, 12266−12280.
(29) Even though νcat is independent of [AB], the k value for
different isotopomers of AB could be different.
(30) Alternatively, AB may be σ-bonded to Fe via one of the B−H
bonds. For reviews on σ-borane complexes, see: (a) Alcaraz, G.; Sabo-
Etienne, S. Coord. Chem. Rev. 2008, 252, 2395−2409. (b) Lin, Z.
Struct. Bonding 2008, 130, 123−148.
11161
dx.doi.org/10.1021/ja5058423 | J. Am. Chem. Soc. 2014, 136, 11153−11161