Catalysis and Mechanism of H2 Release from Amine-Boranes by Diiron Complexes

Article abstract of DOI:10.1021/acs.inorgchem.5b02601

Studies focused on the dehydrogenation of amine-borane by diiron complexes that serve as well-characterized rudimentary models of the diiron subsite in [FeFe]-hydrogenase are reported. Complexes of formulation (μ-SCH₂XCH₂S)[Fe(CO)

Full text of DOI:10.1021/acs.inorgchem.5b02601

Article
pubs.acs.org/IC
Catalysis and Mechanism of H Release from Amine-Boranes by
2
Diiron Complexes
Allen M. Lunsford,‡ Jan H. Blank,†,§ Salvador Moncho, Steven C. Haas, Sohail Muhammad,
,§
†
‡
,†
†,∥
†
,‡
Edward N. Brothers, Marcetta Y. Darensbourg,* and Ashfaq A. Bengali*
†
Department of Chemistry, Texas A&M University at Qatar, Doha, Qatar
‡
Department of Chemistry, Texas A&M University, College Station, Texas 77843, United States
S Supporting Information
*
ABSTRACT: Studies focused on the dehydrogenation of amine-borane by diiron
complexes that serve as well-characterized rudimentary models of the diiron subsite
in [FeFe]-hydrogenase are reported. Complexes of formulation (μ-SCH XCH S)-
2
2
[
Fe(CO) ] , with X = CH , CMe , CEt , NMe, NtBu, and NPh, 1-CO through 6-
3 2 2 2 2
CO, respectively, were determined to be photocatalysts for release of H gas from a
solution of H B ← NHMe (B:A ), dissolved in THF. The thermal displacement of
the tertiary amine-borane, H B ← NEt (B:A ) from photochemically generated (μ-
SCH XCH S)[Fe(CO) ][Fe(CO) (μ-H)(BH −NEt )], 1-B:A through 6-B:A , by
P(OEt) was monitored by time-resolved FTIR spectroscopy. Rates and activation
2
s
3
2
t
3
3
t
t
2
2
3
2
2
3
3
barriers for this substitution reaction were consistent with a dissociative mechanism
for the alkylated bridgehead species 2-CO through 6-CO, and associative or
interchange for 1-CO. DFT calculations supported an intermediate [I] for the
dissociative process featuring a coordinatively unsaturated diiron complex stabilized
by an agostic interaction between the metal center and the C−H bond of an alkyl
s
group on the central bridgehead atom of the SRS linker. The rate of H production from the initially formed 1-B:A through 6-
2
s
t
t
B:A complexes was inversely correlated with the lifetime of the analogous 1-B:A through 6-B:A adducts. Possible mechanisms
are presented which feature involvement of the pendent nitrogen base as well as a separate mechanism for the all carbon
bridgeheads.
INTRODUCTION
The potential utility of amine-boranes as H storage materials
2
■
together with the quest for greener and lower cost catalysts has
led to the application of molecular Fe compounds for the
dehydrogenation process. Vance et al. have explored Fe(II)
The storage and transportation of H is an essential component
of any future economy based on this high-energy fuel. Among
various materials and compounds suggested for the storage
2
5
5
complexes of the type [(η -C H )Fe(CO) X] and [(η -C H )-
5
5
2
5
5
challenge of this lightest of gases are amine-boranes, [H B ←
3
+
Fe(CO) solv] as catalysts for H release from the secondary
2
2
NHnR
3−n
and are relatively stable.
(n = 0−3)], as they have a high gravimetric H content
2
s
1−4
amine-borane, H B ← NHMe (B:A ), concluding that the
3 2
In fact, the controlled H release,
2
binding of boron-hydride to an open site on the iron center was a
key step in the reaction mechanism that led to heterolytic
dehydrogenation, of these compounds typically uses transition
5
−10
metals as catalysts.
As most such complexes are based on
+
−
coupling of H /H and H release with formation of H BNR
2
2
2
precious metals, there is interest in developing first-row
transition metals as more sustainable catalysts.
2
3
11−14
followed by dimerization. Experimental evidence supported by
A mecha-
DFT calculations indicated the initial formation of a Fe-(μ-H)-B
nistic understanding of such catalysis and likely intermediates is
key not only for optimizing the dehydrogenation process but also
for offering insight into the even greater specific challenge of the
reverse reaction (i.e., the hydrogenation of the boron−nitrogen
28
species as a key reaction intermediate.
As mimics for the active site of hydrogenase enzymes which
catalyze the reversible oxidation of H , diiron model complexes
2
1
5−20
with key features of the diiron catalytic site, (μ-SCH XCH S)-
2
2
product).
It is therefore important to quantify the extent of
[
Fe(CO) ] , have been extensively characterized. Figure 1
3 2
amine-borane dehydrogenation by such catalysts and to identify
and study the reactivity of the intermediates that are formed
presents the complexes chosen for our current study, differing
from each other by the nature of X (X = CH , CMe , CEt , NMe,
2
2
2
during H release. The reaction is expected to proceed by way of
2
29−32
NtBu, and NPh; 1-CO through 6-CO, respectively).
Here
an initially formed M−H−B complex which, due to the hydridic
nature of the B−H bond, is characterized to be dominated by the
sigma donation from borohydride as ligand to the metal with a
structures from X-ray diffraction data are reproduced in order to
emphasize the steric properties of the bridgehead substituent.
2
1−24
low back-donating component.
Although there is spectro-
scopic and crystallographic evidence for several such complexes,
few studies have investigated the reactivity of such species.
Received: November 10, 2015
25−27
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 1. Molecular crystal structures from XRD of diiron complexes used as dehydrogenation catalysts. The structures of 4-CO depict the two different
orientations of the lone pair of electrons on the nitrogen i.e., pointed toward the iron center (left) or away (right). Both isomers are present in the crystal
structure. Abbreviations: pdt = propane dithiolate; dmpdt = dimethylpropane dithiolate; depdt = diethylpropane dithiolate. See Figure 2 for a chemdraw
representation of the structures.
Earlier photochemical experiments demonstrated efficient loss of
a single CO from complex 1-CO upon exposure to UV light and
established the propensity of the coordinatively unsaturated
diiron unit to scavenge and bind substrates such as olefins or
3
3
alkynes. The (μ-SRS)[Fe(CO) ] complexes therefore appear
3
2
to be ideally formulated for the binding of substrates including
hydrides (in the form of amine-boranes) and H2 upon
photochemical activation.
The six complexes in our study are well characterized in the
solid state and in solution. Dynamic NMR studies noted effects
of bridgehead steric bulk (comparing complexes 1-CO, 2-CO
and 3-CO) on the Fe(CO) rotor mobility; experimental
3
activation barriers were supported by DFT-calculated Fe(CO)
3
2
9−32
rotational barriers.
As the (μ-SRS)[Fe(CO) ] complexes
3 2
+
are known to be electrocatalysts for H reduction to H , the series
2
was also explored for bridgehead effects of the μ-SRS dithiolate
3
2
linker. Cyclic voltammetry experiments found the pendent
base (as in 4-CO, 5-CO, 6-CO) to have a small but positive effect
on proton uptake in reduced species. An increase in activity upon
introduction of the nitrogen atom in the dithiolate linker is
consistent with the established pendent base/proton shuttling
s
Figure 2. (a) Catalysts used in the dehydrogenation of B:A . (b)
t
t
Comparative kinetic study of the lability of the 1-B:A through 6-B:A
adducts.
34
t
effect in the well-known DuBois nickel catalyst. A similar
presence of the tertiary amine-borane, B:A , Figure 2b. A
prominent feature of the selection of diiron complexes is the
potential for a pendent base effect on the progress of the reaction.
We relate the lability of the initially formed, complexes 1-B:A
through 6-B:A , as affected by steric constraints at the binding
heterolytic oxidation of H was studied in a series of diphosphine-
2
substituted [FeFe]-hydrogenase complexes featuring a PNP
35
t
ligand analogous to the one used by DuBois et al. If the PNP
ligand is substituted with an all-carbon diphosphine, the activity
t
site, to the overall H release efficiency when a secondary amine-
of H oxidation is completely lost, an observation that highlights
2
2
s
borane, B:A , is used as the substrate.
the essential role played by the pendent base.
Herein we report the use of such complexes as photocatalysts
s
EXPERIMENTAL SECTION
Methods and Materials. Unless specifically stated, all syntheses
and manipulations were performed using standard Schlenk-line and
syringe/rubber septa techniques under N or in an Ar atmosphere
for H release from B:A , Figure 2a. Photolysis is required to
■
2
remove a CO and open a site for substrate binding; the quantity
of hydrogen released is monitored with time. As H release does
2
t
2
not occur in the tertiary amine analogue, H B ← NEt (B:A ), a
3
3
glovebox. Solvents were purified according to published procedures, and
unique opportunity exists for exploration of a likely intermediate
involving the binding of the borohydride. Hence we describe a
mechanistic study focused on the reactivity of the initial species
formed upon photolysis of several diiron complexes in the
freshly distilled under N prior to use or purified and degassed via a
2
Bruker solvent system. The following materials were of reagent grade
and used as received: H B ← NHMe (Alfa Aeasar) and d -THF
3
2
8
solvents (Cambridge Isotope Laboratories). The diiron complexes used
B
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 3. Solution 11B NMR of the reaction mixture from the dehydrogenation using 1-CO as the catalyst in d -THF. The triplet centered at 5.21 ppm
8
corresponds to the cyclic byproduct.
in this study were synthesized following previously published
procedures; 1-CO, 2-CO, 3-CO, 4-CO, 5-CO, 6-CO.
Physical Measurements. Identification of diiron reactants and
products was largely by solution FTIR spectroscopy, recorded on a
to the IR cell using a syringe pump and care was taken to minimize
exposure of the photolysis solution to ambient light by covering the
delivery syringe and flask in aluminum foil. After injection into the cell,
the sample was allowed to thermally equilibrate for a few minutes before
the experiment was initiated.
29
30
30
31
31
31
Bruker Tensor 37 FTIR spectrometer using a CaF solution cell of 0.2
2
11
mm path length. The boron-containing product was identified by
B
All experiments were conducted under pseudo first-order conditions
NMR spectroscopy, recorded on a Varian Unity Inova 400 FT NMR at
with the concentration of the P(OEt) ligand at least 10 times greater
than that of the reactant metal complex. Observed rate constants (kobs)
3
1
28.19 MHz and 21.0 °C referenced to BF ·OEt = 0 ppm. When
3
2
evolution of H ceased, a 0.35 mL sample of the reaction solution was
were obtained from single exponential fits to the temporal profile of
either reactant or product complexes. The stated errors in the rate
constants and activation parameters were obtained from a least-squares
fit to the data as reported by the data analysis program Kaleidagraph.
b. Computational Studies. DFT calculations were performed using
2
withdrawn and injected into an NMR tube, and the solvent level was
adjusted to 5 cm with d -THF. Gas identification was accomplished with
8
an Agilent Trace 1300 GC equipped with a thermal conductivity
detector and a custom-made 120 cm stainless steel column packed with
Carbosieve-II from Sigma-Aldrich. The carrier gas was Ar, and
throughout the entire separation, the column was kept at 200 °C,
while the detector was at 250 °C. Identification of gases from the
reactions was accomplished by withdrawing 400 μL of the headspace
using a 0.5 mL Valco Precision Sampling Syringe, Series A-2 equipped
with a Valco Precision Sampling syringe needle with a 5 point side port.
Figure S33 shows a standard G.C. trace with a reference sample
consisting of 0.1 mL of H , 0.1 mL of gas from the atmosphere (N /O ),
36
the development version of the Gaussian suite of programs. All
geometries were fully optimized using the ωB97XD functional and the
3
7,38
all-electron, triple-ζ basis set def2-TZVPP.
Reported geometries
were assigned to minima or transition states according to their number
of imaginary frequencies. Numerical accuracy was ensured by using a
large integration grid with 99 radial shells of 590 points (“ultrafine”).
Gibbs energies and enthalpies were calculated at 298 K and 1 atm with
fully optimized geometries. Vibrational frequencies were scaled by a
linear fitting of the frequencies of compound 5-CO with experimental
values. This methodology was chosen on the basis of the performance of
2
2
2
and 1 mL of CH . H is the first peak to elute from the column at 1.28
4
2
min, followed by N /O at 2.52 min, and finally CH at 4.09 min.
2
2
4
33
All infrared spectra in the kinetic study were recorded using a Bruker
Vertex 80 FTIR spectrometer equipped with both step-scan and rapid-
scan capabilities. A temperature controlled, 0.5 mm path length IR
our previous calculations of diiron complexes of this type. Figures of
computed geometries included in this work were rendered using
3
9
CYLview.
transmission cell (Harrick Scientific) with CaF windows was used to
c. H Release Studies. In all cases, solutions of the diiron catalyst and
2
2
acquire the spectra. The cell temperature was controlled with the use of a
circulating water bath (±0.1 °C). Photolysis of the solution was achieved
using a Quantel Brilliant B Nd:YAG laser operating at 355 nm (70 mJ/
pulse) with a 7 ns pulse width.
amine-borane were prepared in an Ar filled glovebox, placed in 2 and 10
mL volumetric flasks, respectively, capped with a rubber septa,
parafilmed, copper wired, removed from the glovebox, and filled to
the appropriate level with anhydrous THF. A degassed 5 mL Pyrex glass
vial with a small stir bar and rubber septum was loaded with 2 mL of the
amine-borane solution as well as 0.5 mL of the diiron solution
corresponding to a 10 mol % loading of diiron catalyst. Prior to
commencing the photolysis reaction, the vial was connected via PTFE
tubing to a 50 mL buret filled with water connected to a variable height
funnel. The pressure was equilibrated before irradiation by a Hg-arc
lamp from ORIEL Instruments operating at 200W with a spectral
a. Kinetic Studies. An appropriate amount of the diiron complex was
dissolved in methylcyclohexane solvent to yield a 2−4 mM solution that
was purged with N . The required amounts of triethylamine-borane
2
t
(
B:A ) and triethylphosphite (P(OEt) ) were added to this solution
3
using a gastight syringe. For each experiment, the concentration of
borane was kept constant (1.35 M), and the amount of P(OEt) was
3
varied between 0.05 and 0.60 M. The resulting solution was transferred
C
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
s
Figure 4. Plots of rates of H released from photolysis of THF solutions of B:A in the presence of diiron carbonyls as catalysts, 1-CO through 6-CO.
2
Each data point is the average of three separate runs and an alternate display with error bars is given in Figures S13−S24.
window from 250 to 2500 nm. Following a short induction period,
bubbles of a released gas were seen to displace water in the buret.
Volume measurements were taken every 5 min until further changes
ceased. A 200 μL sample of the headspace was analyzed by gas
over time are similar for all complexes, Figures S25−S30. On
average, 3 mol of CO were obtained per mol of diiron catalyst
over the course of ca. 2 h. Analysis of the final solution by FTIR
finds CO absorption peaks consistent with the remaining (μ-
SRS)[Fe(CO) ] complexes and a degradation product with an
chromatography (Figures S7−S12). Detection of H was confirmed by
2
3
2
comparison to standards composed of mixtures of H , CH , N and O
2
4
2
2
absorption pattern consistent with a monosubstituted penta-
carbonyl species, presumably solvated (μ-SRS)[Fe (CO) -solv],
(Figure S33). Data for triplicate runs of gas collection experiments can
be found in Figures S1−S6 including plots of ln (VMax/VMax − VTime) vs
2
5
time Figures S13−S24.
Figures S31−S32. A dark brown/black decomposition solid was
also deposited on the sides of the flask.
Control experiments were run in the same manner as the gas
collection experiments except the reaction flask was not connected to
the gas collection apparatus. To confirm the necessity of the diiron
The boron-containing product from an H2 production
11
experiment with 1-CO as the catalyst was investigated by
B
s
catalyst, B:A was dissolved in THF and irradiated for 2 h and the
NMR. The spent solution displays a triplet centered at 5.21 ppm,
headspace analyzed via gas chromatography (Figure S35). No peak for
referenced to BF ·Et O = 0.0 ppm, Figure 3, corresponding to
3
2
H was detected. To confirm the need for light, the vial consisting of the
2
s
the [H B-NMe ] cyclic dimer, as characterized in analogous
2 2 2
diiron catalyst and B:A was wrapped in aluminum foil and placed in the
40
metal carbonyl photocatalyzed H release reactions. A quartet
dark for 2 h before adding THF and stirring the reaction for 2 h. Analysis
2
for the amine-borane starting material is still present in low
abundance at −13.5 ppm.
The natural log plots of H2 collection from multiple
experiments are presented in Figure 4, demonstrating varying
of the headspace gas showed no H production (Figure S36). Irradiating
2
s
the diiron catalyst in THF lacking B:A for 2 h shows some CO but no
H release (Figure S34). Finally, to confirm the necessity for a hydrogen
2
t
s
atom on both the nitrogen and boron, B:A was used in place of B:A
from which no H was detected (Figure S37). In all control experiments,
2
rates of H release. The raw data obtained can be found in Figure
2
after 2 h of irradiation, 400 μL of the headspace was withdrawn followed
S1−S6. A linear correlation indicates first-order release of H
2
by injecting 200 μL into the GC. No peak corresponding to H could be
s
2
from B:A with the slope of the line of best fit used to rank the
detected.
catalysts. The final volume of gas collected averages to 20.54 ±
0
.13 mL over all the reactions. However, there are significant
differences in the length of time it takes to reach completion.
Complex 1-CO had the fastest rate of H production with gas
RESULTS AND DISCUSSION
■
s
a. H Release from B:A . We have examined the series of
2
2
complexes in Figure 1 for their mediation of H release from the
evolution ceasing after 85 min. Reflecting the need for an open
site on the catalyst, there is an initial induction period of ca. 10
min prior to the development of the plots above that analyze for
first order kinetics. Conclusions regarding the rate dependence
on μ-SRS bridgehead are discussed below.
2
s
secondary amine-borane, B:A according to Figure 2a. Gases
which were released during continuous photolysis of the reaction
mixture were collected and measured as described in the
Experimental Section. One mol of H per mol of B:A was
s
2
evolved over the course of the reaction. A minor amount of CO
was present in the measurements as expected from the known
photolability of CO from these complexes. This was further
confirmed by measuring the gas obtained from photolysis of the
b. Kinetic Studies (Single-Shot Laser Photolysis).
Previous studies of low valent metal carbonyls as photocatalysts
for H release from amine-boranes have suggested a boron-
2
hydride-M(CO) adduct as a key intermediate required for bond
X
s
diiron complex in the absence of B:A . Plots of the CO released
activation. This hypothesis is further supported by DFT
D
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
calculations and 11B NMR studies.41 Photolability of CO in
complexes 1-CO through 6-CO produces a key species which
may be scavenged by the amine-borane, as expressed in Figure
t
1
b. We have also characterized the reactivity of the 1-B:A
t
through 6-B:A adducts, the second step of the reaction in
Figure 1b, which unlike the 1-B:A through 6-B:A complexes, do
s
s
not proceed toward dehydrogenation. The aim was to correlate
the lability of these adducts with the rate of H release from the
2
s
s
analogous 1-B:A through 6-B:A complexes described above.
Previously, a similar kinetic method was successfully used to
investigate the reactivity and binding enthalpy of weakly
coordinated substrates such as alkenes, alkynes, and furans to
33
Figure 6. Calculated structure of 5-B:At.
the iron center in 1-CO.
Single-shot photolysis of 5-CO (2.2 mM) in methylcyclohex-
t
ane solution at room temperature in the presence of 0.47 M B:A
of P(OEt) to the photolysis solution results in the conversion of
3
t
results in the formation of a new complex absorbing at 2042,
985, 1965, and 1924 cm (Figure 5). In the presence of the
5-B:A to a new complex absorbing at 2049, 1990, 1969, and 1940
−1
−1
1
cm which grows in at the same rate (Figure 7). Because
Figure 5. Difference IR spectra obtained upon photolysis of 5-CO in the
t
presence of 0.47 M B:A at 298 K. The positive peaks are due to the
formation of an intermediate, assigned to 5-B:A (see Figure 6) and the
negative peaks are due to the loss of 5-CO.
t
Figure 7. Spectral changes observed upon photolysis of 5-CO in the
t
presence of 1.35 M B:A and 0.117 M P(OEt) at 283 K in
3
t
methylcyclohexane. Peaks due to 5-B:A and 5-P(OEt) decay and
3
grow, respectively, at the same rate.
s
secondary amine-borane, B:A , a species with almost identical
CO stretching frequencies (2043, 1988, 1971, 1939 cm ) and
pattern is observed suggesting that the structures of the initial
−1
photolysis of 5-CO in the presence of only P(OEt) also yields
3
the same product (Figures S41−S46 and Table S2), this species
t
is assigned to the 5-P(OEt) complex. The displacement of B:A
3
complexes in both cases are very similar, Figure S47. H gas
2
from the iron center was studied over a wide range of [P(OEt)3]
and temperatures.
generated upon photolysis in the IR cell precluded further
s
3,9,24,25,41
investigation of the reactivity of the FeFe-B:A adduct.
t
As shown in Figure 8, when the concentration of B:A is held
Because UV photolysis of metal carbonyls such as 1-CO in the
constant, kobs exhibits a nonlinear dependence on [P(OEt)3],
approaching a limiting value at high ligand concentration. The
observed saturation behavior is consistent with the dissociative
mechanism shown in Scheme 1. Application of the steady state
assumption to the intermediate complex, [I], results in the kobs
presence of classical coordinating ligands such as PR or P(OEt)3
3
results in CO loss and formation of the corresponding PR
3
3
3
t
ligated product, we assign this species to the 5-B:A complex.
Similar species are known to form upon photolysis of other metal
carbonyls in the presence of tertiary amine-boranes, and some
42−45
concentration dependence shown in eq 1. The limiting value of
have been isolated and fully characterized.
The calculated
t
kobs yields k , the rate constant for the dissociation of the 5-B:A
1
geometry of the B−H−M interaction finds a relatively large angle
of 135.2°, supporting our supposition that this interaction
derives from sigma donation of electron density on the hydride
adduct.
k k [P(OEt) ]
1
2
3
kobs
=
(
Figure 6); scaled computed νCO bands are in good agreement
t
(1)
k [B:A] + k [P(OEt) ]
−
1
2
3
with the experimental values (Table S1).
In the absence of other coordinating ligands, 5-B:A is
relatively stable with a t1/2 = 51 s at 298 K. However, addition
t
As discussed below, with the assistance of DFT calculations,
[I] in Scheme 1 is identified as a species in which the vacant site
E
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Unfortunately, the displacement reaction could not be probed
at high concentrations of [P(OEt) ] at temperatures >283 K
3
because of the rapid reaction rate and a reduction in the signal of
t
5
-B:A due to competitive binding of P(OEt) with 5. Therefore,
3
the reaction was studied at relatively low [P(OEt) ] concen-
3
trations and higher temperatures in order to obtain the relevant
activation parameters. The results are shown in Figure 10.
t
Figure 8. kobs dependence on [P(OEt) ] at 283 K with [B:A ] = 1.35 M.
3
Scheme 1. Dissociative Mechanism of Amine-Borane
Substitution
t
Figure 10. Plot of the kobs versus [P(OEt) ]/[B:A ] at several
3
temperatures. The nonzero intercepts are similar in magnitude to the
t
background decay rate constant of 5-B:A in the absence of P(OEt) . See
3
Table S3 for all rate constants.
t
Under these reaction conditions (k [B:A ] ≫ k [P(OEt) ]),
−1
2
3
a linear relationship between k and reactant concentrations is
obs
expected because eq 1 simplifies to that shown in eq 2.
t
generated by loss of B:A is stabilized by an agostic interaction
⎡
⎤
t
[P(OEt )]
[B:A]
k k
1 2
between the C−H bond of an alkyl group from the N Bu
bridgehead and the iron center (Figure 9). The bent geometry of
3
kobs = k′
⎢k′ =
⎥
⎦
t
(2)
⎣
k
−
1
t
A plot of kobs vs [P(OEt) ]/[B:A ] is linear and yields a slope of
3
k′ = k k /k . As shown in Figure 11, an Eyring plot obtained
1
2
−1
‡
from the temperature dependence of k′ yields ΔH = 15.1 ±
total
‡
1
.0 kcal/mol and ΔS = −1 ± 3 e.u. for the ligand exchange
total
‡
‡
‡
‡
‡
reaction where ΔH = ΔH + (ΔH − ΔH ) and ΔS
=
total
total
k
1
k
2
k
−1
‡
‡
‡
ΔS + (ΔS − ΔS ).
k
1
k
2
k
−1
The measured activation enthalpy is in good agreement with a
DFT-calculated value of 17.6 kcal/mol for the difference in
t
enthalpy between 5-B:A and [I] + H BNEt . The close
3
3
agreement between the experimental and calculated values
suggests that the transition state structure is similar to that of the
calculated geometry of the intermediate species and that reaction
t
of [I] with B:A or P(OEt) proceeds with similar activation
3
Figure 9. DFT-calculated structure for the intermediate complex, [I]
‡
‡
barriers (i.e., (ΔH − ΔH ) ≈ 0).
t
k
2
k
−1
generated from 5-B:A .
To study the influence of the bridgehead group on the reaction
t
rate, the displacement of B:A by P(OEt) from the respective
3
the M−H−C interaction, of angle 115.7°, contrasts to the B−H−
M interaction, as expected for the greater covalency of the C−H
bond relative to the polar-covalent B−H bond. Although the
diiron species was investigated under identical reaction
conditions. As shown in Table 1, the reaction rate constants
are largest for systems that have an alkyl group present on the
t
barrier to N-inversion in (μ-SCH N( Bu)CH S) of coordina-
bridgehead atom, regardless of whether it is carbon or nitrogen.
2
2
t
tively saturated 5-CO prohibits such agostic interaction, it is
expected to be much less in the coordinatively unsaturated
species.
The absence of alkyl groups as in 1-B:A results in a k for the
obs
displacement reaction that is a factor of 25−130 less than the
other complexes studied. This observation suggests an important
F
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
dependence on [P(OEt) ] even at the highest concentrations
3
used and at all temperatures studied. An Eyring plot yields
‡
‡
activation parameters of ΔH = 7.1 ± 0.6 kcal/mol and ΔS =
t
−
36 ± 7 e.u which, in dramatic contrast to the 5-B:A system, are
strongly indicative of a substitution mechanism with a
bimolecular rate-determining step. These results can be
explained with the support of precedence (vide infra) and with
t
the assistance of DFT calculations which predict a 1-B:A bond
dissociation enthalpy (BDE) of 25−30 kcal/mol (Table 2),
Table 2. DFT-Calculated 1-CO through 6-CO Bond
a
Dissociation Enthalpies for the Complexes Studied
CH2
1)
CMe2
(2)
CEt2
(3)
NMe
(4)
NtBu
(5)
NPh
(6)
(
apical-
proximal
27.6
21.8
18.0
20.1
17.6
16.7
apical distal
basal-distal
29.5
24.8
27.6
30.1
24.8
26.0
30.8
24.3
25.9
29.2
25.3
25.8
29.2
24.9
27.4
30.1
25.0
29.8
basal proximal
a
The bolded numbers refer to species in which the vacant site
generated upon loss of borane is stabilized by an alkyl C−H agostic
interaction with the metal center Scheme 2, eq 1.
Figure 11. Eyring plot obtained from the temperature dependence of k′.
Table 1. Observed Rate Constants (k ) for the Substitution
Reaction
obs
much higher than the experimental enthalpy of activation.
Without the ability to access a low-energy intermediate
analogous to [I], and a relatively large barrier for the dissociative
a
X =
t
channel, it is reasonable to expect that substitution of B:A from
T (K) CH (1) CMe (2) CEt (3) NMe (4) NtBu (5) NPh (6)
2
2
2
5.85
*
t
1
-B:A may follow an associative or interchange pathway if the
2
2
2
83.15
88.15
93.15
-
3.20
4.96
6.41
1.64
2.62
3.34
0.72
1.00
1.44
1.52
2.23
3.16
transition state energy lies below the dissociative limit.
0.04
0.05
Consistent with this analysis, the experimental activation
*
‡
parameters for the bimolecular reaction yield a ΔG2 of 18 ± 2
98
a
t
All experiments were conducted under identical conditions of [B:A ]
kcal/mol which is on the order of, or lower than, a calculated
*
(
1.35 M) and [P(OEt) ] (0.117 M). Too fast to reliably measure.
‡
3
value of ΔG2 = 19−23 kcal/mol for the dissociative step (the
98
energy range is due to the presence of different structural
4
6
t
role for the alkyl groups in stabilizing the transition state accessed
in the reaction.
To further probe the origin of the large reactivity difference
isomers). A bimolecular pathway for B:A displacement from
t
the 1-B:A complex is not without precedence. Previous studies
have demonstrated convincingly that thermal displacement of
−
47,48
between the alkylated bridgehead and X = CH complexes, the
2
CO from 1-CO by CN proceeds in a S 2 manner.
By
N
t
t
complete kinetics of B:A substitution from 1-B:A was studied.
analogy with this earlier study, it is expected that nucleophilic
t
As shown in Figure 12, in contrast to the 5-B:A system, the
attack of the P(OEt) ligand results in the displacement of the
3
t
t
observed rate of B:A displacement from 1-B:A exhibits a linear
Fe−Fe bond density with concomitant rotation of the Fe-
t
(
CO) (B:A ) group placing a CO below the Fe−Fe vector in
2
position to interact with the increased electron density of the
unrotated Fe. Reformation of the Fe−Fe bond results in ejection
t
of B:A to form the product complex (Scheme 2).
c. Computational Studies for the Intermediates.
Computational methods evaluated the energetics of the possible
t
t
products of H−B coordination of B:A in complexes 1-B:A
t
through 6-B:A , as well as the related unsaturated complexes. For
each complex, all possible coordination geometries were
analyzed. Geometry optimizations were done using the
ωB97XD/Def2TZVPP level of theory; detailed methodology
description is provided in the Methods and Materials section.
The experimental findings are consistent with DFT calcu-
t
lations which indicate that for all complexes except 1-B:A , the
vacant coordination site generated upon borane dissociation is
stabilized by an agostic interaction between the Fe center and the
C−H bond of an alkyl group on the bridgehead atom. For
complex 5, this intermediate is calculated to be 9.5 kcal/mol
more stable than the unstabilized species. An alternate structure
for [I] would involve solvent coordination to the vacant site to
form the solvated species. Indeed, DFT calculations predict a
solvation enthalpy of 8.8 kcal/mol for such a complex. While the
t
Figure 12. Plot of k vs [P(OEt) ] for the displacement of B:A from 1-
obs
3
t
B:A by P(OEt) . See Table S4 for all rate constants.
3
G
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
t
Scheme 2. Proposed Bimolecular Mechanism for the Substitution Reaction in Figure 2b of 1-B:A (i.e., the (μ-pdt) Diiron
Complex)
t
t
t
Figure 13. All possible geometries adopted by 1-B:A through 6-B:A . DFT-calculated geometries for 5-B:A are in color.
stabilization enthalpies are similar for the agostic and
methylcyclohexane solvated structures, the generation of the
former complex by an intramolecular interaction would make it
kinetically favored over the latter species.
originate from the isomer that has the lowest barrier to B:At
dissociation. For the alkylated bridgehead complexes, this isomer
is the apical-proximal structure because it places the alkyl group
in the correct position to stabilize the vacant coordination site
through a Fe−H−C agostic interaction. In cases where an alkyl
group is present on the bridgehead, this agostic stabilization is
calculated to be ≈3−10 kcal/mol and the lower energy of the
resulting transition state yields a faster substitution rate relative
to 1-CO which lacks this bonding possibility. As mentioned
t
t
The structures of 1-B:A through 6-B:A were calculated using
DFT. Due to the symmetry of the complex, four isomers are
possible (Figure 13). The geometry of the M−H−B interactions
for the complexes with different bridgeheads are in agreement
t
with the analysis for 5-B:A . A description of the fluxionality of
t
the bridgeheads and the Fe(CO) rotor activity in complexes (1-
earlier, the calculated 17.6 kcal/mol 5-B:A BDE is in good
3
CO through 6-CO) can be found in ref 32. The amine-borane
ligand can be either directly coordinated under the bridgehead
agreement with the experimental activation barrier of 15.1 kcal/
mol, suggesting that the substitution reaction proceeds by way of
the alkyl stabilized intermediate and that this agostic bond is
mostly formed in the transition state.
(proximal) or on the other metal center (distal), and it can
occupy either an apical or basal site. The modeling results rule
out the possibility of an electronic influence of the bridgehead
Correlation between Borane Substitution Rates and
t
atom, N or C, upon the Fe−B:A bond dissociation enthalpy
the Rate of H Formation. It is assumed in the following
2
t
(BDE). For similar isomers of the respective complexes without
analysis that the trend in the kinetic stabilities of 1-B:A through
6-B:A are similar to those of the 1-B:A through 6-B:A adducts
which yield H . The correlations for k (i.e., the rate constant for
t
s
s
agostic stabilization of the vacant site (e.g., the open site
t
generated from the basal-distal isomer, Figure 13), the Fe−B:A
2
obs
t
BDE’s remain essentially unchanged for 1-B:A (24.8 kcal/mol),
-B:A (25.3 kcal/mol), and 5-B:A (24.9 kcal/mol). The
distinguishing feature between 1-B:A and all other complexes
is the absence of alkyl groups on the bridge. The presence of a
nitrogen heteroatom does not appear to play a major role in the
the displacement of amine-borane by P(OEt) ) and the rate of
3
t
t
4
H release are then based on two assumptions: (a) the major B−
2
t
H to Fe bond in the two adducts that precede bond activation,
i.e., the Fe−Fe-(μ-H)-BH NEt and the Fe−Fe-(μ-H)-
2
3
BH NHMe , are largely the same; and (b) only in the N-
2
2
t
rate of substitution nor in determining the Fe−B:A BDE;
containing complexes 4-CO, 5-CO, and 6-CO are there possible
however, the alkyl substituent on the N is influential.
interactions of the amine-boranes with the N-bridgehead atoms
t
When referenced to the same state (i.e., unstabilized vacant Fe
when switching from the tertiary amine-borane (B:A ) in the
t
t
s
coordination site upon B:A dissociation), the calculated 5-B:A
P(OEt) kinetic studies to the secondary amine-borane (B:A ) in
3
BDE’s are within 3 kcal/mol for the four possible structures with
the apical-distal geometry resulting in the most stable arrange-
ment (Table 2). The similar stabilities together with the fact that
interconversion of these isomers is expected to be rapid at the
temperatures used, suggests that the substitution reactions will
the H release studies.
As indicated in Scheme 3, the H release ranking is largely
opposite that of the kobs obtained for borohydride displacement
ranking. We interpret this to mean that the stabilization of the
intermediate [I], the agostic interaction, correlates with slower
2
2
H
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Scheme 3. Correlation of Rates of B:A Substitution from 1-
Article
t
Hall and Weller that explored the mechanism of dihydrogen
release from a Rh(I) diphosphine complex was used to develop
the mechanism displayed in Scheme 4a. In contrast, the N-
t
t
B:A through 6-B:A , at Room Temperature, and Relative
s
a
41
Rates of H Release from B:A (See )
2
bridgehead derivatives, complexes 4-CO, 5-CO, and 6-CO, have
the possibility for increased interactions due to the presence of a
proton in the secondary amine-borane as expressed in Scheme
4
b. Similar mechanisms were computationally and experimen-
a
tally developed for two other studies, one featuring a mono Fe
complex and another focusing on a Ru complex both of which
contained a N-donor within a chelating ligand.
Complexes in red represent all carbon bridgeheads, and complexes in
blue feature a pendent amine.
4
9,50
The
s
heterolytic removal of both the proton and hydride from the
amine-borane requires the N-bridgehead and bears similarity
H release from the B:A substrate, indicating that the longer the
2
lifetime of the later secondary amine-borane adduct, the more
facile is the H release process. Thus, complex 1- B:A , the
nonalkylated (pdt) complex, displays the slowest rate for B:A
+
−
40
t
with the H /H production (or oxidation) of H in the [FeFe]-
2
2
t
H ase enzyme.
2
In summary, the deceptive simplicity of this diiron complex
belies a range of sophisticated maneuvers that serve to provide a
path for the dehydrogenation of amine-boranes. An apparent role
for bridgehead hydrocarbon substituents that place C−H bonds
in close proximity, agostic interactions, with an open site on iron
diminishes residence time of the substrate, and hence impedes
the dehydrogenation process. On the other hand, the amine
functionality assists in the process, implicating its positive effect
displacement, yet the fastest rates of H release as compared to
2
the dimethyl and diethyl derivatives. That is, the agostic C−H
interaction with the open site from the dmpdt and depdt
diminishes the residence time of the substrate. The only complex
that does not follow an inverse order of H production as
2
t
t
compared to the kobs for B:A displacement is 5-B:A . This
complex, which features the largest steric bulk compared to the
other two nitrogen bridgehead derivatives, places an alkyl group
close to the iron center similar to the carbon bridgehead species
on the heterolytic H release.
2
t
complexes 2-CO and 3-CO. A complexity of the 5-B:A involves
ASSOCIATED CONTENT
Supporting Information
an inversion barrier at the nitrogen which could account for this
discrepancy; as yet undefined features might be at play.
The mechanisms offered in Scheme 4 add to the myriad
possibilities already expressed for metal-catalyzed dehydrocou-
pling of amine-boranes. In particular, the computational study by
■
*
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.5b02601.
Scheme 4. Minimal Proposed Mechanisms for Dehydrogenation of Amine-Borane by (a) C-Bridgehead and (b) N-Bridgehead
a
Diiron Complexes
a
Note that [I] is as defined in Figure 9, and the curved arrow at the bridgehead represents the flipping process available to the dithiolate linker.
Experimental support is for the bracketed portion.
I
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Plots of gas evolution experiments, gas chromatogram
Article
(
1
(
20) Summerscales, O. T.; Gordon, J. C. Dalton Trans. 2013, 42,
0075−10084.
21) Lachaize, S.; Essalah, K.; Montiel-Palma, V.; Vendier, L.;
Chaudret, B.; Barthelat, J.; Sabo-Etienne, S. Organometallics 2005, 24,
traces of headspace gas, kinetic data, difference FTIR
spectrum, and calculated νCO values (PDF)
2
935−2943.
AUTHOR INFORMATION
(22) Kawano, Y.; Yamaguchi, K.; Miyake, S.; Kakizawa, T.; Shimoi, M.
■
Chem. - Eur. J. 2007, 13, 6920−6931.
23) Dallanegra, R.; Chaplin, A. B.; Weller, A. S. Angew. Chem., Int. Ed.
009, 48, 6875−6878.
24) Sohail, M.; Moncho, S.; Brothers, E. N.; Darensbourg, D. J.;
Corresponding Authors
E-mail: marcetta@chem.tamu.edu.
E-mail: ashfaq.bengali@qatar.tamu.edu.
(
2
(
*
*
Bengali, A. A. Dalton Trans. 2013, 42, 6720−6723.
Present Address
∥
(25) Lin, T.; Peters, J. C. J. Am. Chem. Soc. 2013, 135, 15310−15313.
(
(
S.M.) Currently at Qatar Energy and Environment Institute
QEERI), Doha, Qatar.
Author Contributions
(26) Darensbourg, M. Y.; Bau, R.; Marks, M. W.; Burch, R. R., Jr.;
Deaton, J. C.; Slater, S. J. Am. Chem. Soc. 1982, 104, 6961−6969.
(27) Kirtley, S. W.; Andrews, M. A.; Bau, R.; Grynkewich, G. W.;
Marks, T. J.; Tipton, D. L.; Whittlesey, B. R. J. Am. Chem. Soc. 1977, 99,
§
(A.M.L. and J.H.B.) Equally contributing authors.
7
(
154−7162.
28) Vance, J. R.; Schaf
Whittell, G. R.; Manners, I. J. Am. Chem. Soc. 2014, 136, 3048−3064.
29) Winter, A.; Zsolnai, L.; Huttner, G. Z. Naturforsch., B: J. Chem. Sci.
1982, 37, 1430.
30) Singleton, M. L.; Jenkins, R. M.; Klemashevich, C. L.;
Notes
̈
er, A.; Robertson, P. A. M.; Lee, K.; Turner, J.;
The authors declare no competing financial interest.
(
ACKNOWLEDGMENTS
■
(
(
(
The experimental work was made possible by an NPRP award
NPRP 6-1184-1-224) from the Qatar National Research Fund
a member of Qatar Foundation). The statements made herein
Darensbourg, M. Y. C. R. Chim. 2008, 11, 861−874.
(31) Li, H.; Rauchfuss, T. B. J. Am. Chem. Soc. 2002, 124, 726−727.
(32) Crouthers, D. J.; Denny, J. A.; Bethel, R. D.; Munoz, D. G.;
are solely the responsibility of the authors. M.Y.D. appreciates
Darensbourg, M. Y. Organometallics 2014, 33, 4747−4755.
33) Muhammad, S.; Moncho, S.; Brothers, E. N.; Darensbourg, M. Y.;
Darensbourg, D. J.; Bengali, A. A. Inorg. Chem. 2012, 51, 7362−7369.
34) Helm, M. L.; Stewart, M. P.; Bullock, R. M.; DuBois, M. R.;
DuBois, D. L. Science 2011, 333, 863−866.
35) Wang, N.; Wang, M.; Wang, Y.; Zheng, D.; Han, H.; Ahlquist, M.
S. G.; Sun, L. J. Am. Chem. Soc. 2013, 135, 13688−13691.
36) Frisch, M. J. et al. Gaussian 09; Gaussian, Inc.: Wallingford, CT,
009.
37) Chai, J.-D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10,
615−6620.
38) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297−
305.
39) Legault, C. Y. CYLview, version 1.0b, 2209; Universite
Sherbrooke. http://www.cylview.org.
40) Sloan, M. E.; Staubitz, A.; Clark, T. J.; Russell, C. A.; Lloyd-Jones,
G. C.; Manners, I. J. Am. Chem. Soc. 2010, 132, 3831−3841.
41) Douglas, T. M.; Chaplin, A. B.; Weller, A. S.; Yang, X.; Hall, M. B.
J. Am. Chem. Soc. 2009, 131, 15440−15456.
42) Shimoi, M.; Nagai, S.; Ichikawa, M.; Kawano, Y.; Katoh, K.;
Uruichi, M.; Ogino, H. J. Am. Chem. Soc. 1999, 121, 11704−11712.
43) Kakizawa, T.; Kawano, Y.; Naganeyama, K.; Shimoi, M. Chem.
Lett. 2011, 40, 171−173.
44) Buss, J. A.; Edouard, G. A.; Cheng, C.; Shi, J.; Agapie, T. J. Am.
Chem. Soc. 2014, 136, 11272−11275.
45) Muhammad, S.; Moncho, S.; Brothers, E. N.; Bengali, A. A. Chem.
Commun. 2014, 50, 5874−5877.
46) The energy range is due to the presence of different structural
financial support from the National Science Foundation (CHE-
(
1266097) and the Robert A. Welch Foundation (A-0924).
(
REFERENCES
■
(
(
2
1) Stephens, F. H.; Pons, V.; Baker, R. T. Dalton Trans. 2007, 25,
613−2626.
(
2
(
6
(
3
(
(
(
2) Smythe, N. C.; Gordon, J. C. Eur. J. Inorg. Chem. 2010, 4, 509−521.
3) Staubitz, A. A.; Robertson, P. M.; Manners, I. Chem. Rev. 2010, 110,
4
(
7
(
7
(
079−4124.
4) Li, H.; Yang, Q.; Chen, X.; Shore, S. G. J. Organomet. Chem. 2014,
51, 60−66.
5) Alcaraz, G.; Sabo-Etienne, S. Angew. Chem., Int. Ed. 2010, 49,
170−7179.
6) Abboud, J. M.; Nem
Nerut, E. R. Chem. - Eur. J. 2012, 18, 3981−3991.
7) Leitao, E. M.; Stubbs, N. E.; Robertson, A. P. M.; Helten, H.; Cox,
R. J.; Lloyd-Jones, G. C.; Manners, I. J. Am. Chem. Soc. 2012, 134,
6805−16816.
8) Al-Kukhun, A.; Hwang, H. T.; Varma, A. Int. J. Hydrogen Energy
013, 38, 169−179.
9) Stubbs, N. E.; Robertson, A. P. M.; Leitao, E. M.; Manners, I. J.
Organomet. Chem. 2013, 730, 84−89.
10) Johnson, H. C.; Hooper, T. N.; Weller, A. S. Top. Organomet.
Chem. 2015, 49, 153−220.
11) Keaton, R. J.; Blacquiere, J. M.; Baker, R. T. J. Am. Chem. Soc.
́
de
́
eth, B.; Guillemin, J.; Burk, P.; Adamson, A.;
(
(
(
1
(
2
(
(
(
(
(
(
(
2
(
(
007, 129, 1844−1845.
(
12) Yang, X.; Hall, M. B. J. Am. Chem. Soc. 2008, 130, 1798−1799.
isomers. Since entropy changes are always overestimated by DFT, the
13) Zimmerman, P. A.; Paul, A.; Zhang, Z.; Musgrave, C. B. Angew.
‡
calculated ΔS2 for the dissociation reaction was corrected by a factor
98
Chem., Int. Ed. 2009, 48, 2201−2205.
14) Vogt, M.; Bruin, B.; Berke, H.; Trincado, M.; Gru
Chem. Sci. 2011, 2, 723−727.
15) Sutton, A. D.; Burrell, A. K.; Dixon, D. A.; Garner, E. B., III;
‡
of 0.5 to yield the reported estimated ΔG
values. Li, H.; Wang, X.;
298
(
̈
tzmacher, H.
Huang, F.; Lu, G.; Jiang, J.; Wang, Z.-X. Organometallics 2011, 30,
233−5247.
47) Darensbourg, M. Y.; Lyon, E. J.; Zhao, X.; Georgakaki, I. P. Proc.
Natl. Acad. Sci. U. S. A. 2003, 100, 3683−3688.
48) Lyon, E. J.; Georgakaki, I. P.; Reibenspies, J. H.; Darensbourg, M.
Y. J. Am. Chem. Soc. 2001, 123, 3268−3278.
49) Baker, R. T.; Gordon, J. C.; Hamilton, C. W.; Henson, N. J.; Lin,
5
(
(
Gordon, J. C.; Nakagawa, T.; Ott, K. c.; Robinson, J. P.; Vasiliu, M.
Science 2011, 331, 1426−1429.
(
(
16) Davis, B. L.; Dixon, D. A.; Garner, E. B.; Gordon, J. C.; Matus, E.
H.; Scott, B.; Stephens, F. H. Angew. Chem., Int. Ed. 2009, 48, 6812−
816.
17) Tang, Z.; Chen, H.; Chen, X.; Wu, L.; Yu, X. J. Am. Chem. Soc.
012, 134, 5464−5467.
18) Ramachandran, P. V.; Gagare, P. D. Inorg. Chem. 2007, 46, 7810−
817.
19) Hausdorf, S.; Baitalow, F.; Wolf, G.; Mertens, F. O. R. L. Int. J.
Hydrogen Energy 2008, 33, 608−614.
(
6
(
2
(
7
(
P.; Maguire, S.; Murugesu, M.; Scott, B. L.; Smythe, N. C. J. Am. Chem.
Soc. 2012, 134, 5598−5609.
(50) Marziale, A. N.; Friedrich, A.; Klopsch, I.; Drees, M.; Celinski, V.
R.; Gu
̈
nne, J. S.; Schneider, S. J. Am. Chem. Soc. 2013, 135, 13342−
1
3355.
J
DOI: 10.1021/acs.inorgchem.5b02601
Inorg. Chem. XXXX, XXX, XXX−XXX