Hydrogen gas formation from the photolysis of rhenium hydrides-mechanistic and computational studies

Article abstract of DOI:10.1039/c9dt03364e

The photolysis of 4,4′-disubstituted, 2,2′-bipyridine fac-Re(bpy)(CO)₃H derivatives produces stoichiometric H₂ gas. The rate of production varies greatly depending on the electronic nature of the disubstituted bipyridine (bpy) with halogenated substituents increasing the rate. Isotope labeling studies along with B3LYP geometry optimization DFT modeling studies indicate a mechanism involving a Re-H-Re bridging complex that leads to a dimeric Re-Re(η²-H₂) state prior to dissociating H₂ gas.

Full text of DOI:10.1039/c9dt03364e

Dalton
Transactions
View Article Online
View Journal | View Issue
COMMUNICATION
Hydrogen gas formation from the photolysis
of rhenium hydrides – mechanistic and
computational studies†‡
Cite this: Dalton Trans., 2019, 48,
16148
Received 17th August 2019,
Accepted 7th October 2019
Alyssa A. Webster,a Jianqiang Huo, a Jenna Milliken,a Pat Sullivan,§a Jan Kubelka*b
a
DOI: 10.1039/c9dt03364e
and John O. Hoberg
*
rsc.li/dalton
The photolysis of 4,4’-disubstituted, 2,2’-bipyridine fac-Re(bpy) reported for photocatalytic H2 production.9 Intermolecular
(CO)3H derivatives produces stoichiometric H2 gas. The rate of pro- hydrogen bonding of rhenium hydrides with proton donors
duction varies greatly depending on the electronic nature of the has also been studied lending insight into the formation of
disubstituted bipyridine (bpy) with halogenated substituents ReH⋯HX bonds.23 Perhaps most closely related to this work is
increasing the rate. Isotope labeling studies along with B3LYP geo- the use of thermally unstable [Re(η2-H2)(CO)n(L)m]+ complexes
metry optimization DFT modeling studies indicate a mechanism to form [Re(CO)n(L)m]BPh4 complexes via the evolution of H2.24
involving a Re–H–Re bridging complex that leads to a dimeric Re– However given the extensive investigations with Re(bpy)CO3X,
Re(η2-H2) state prior to dissociating H2 gas.
where X = H or halides, an in-depth study on the mechanism
of X2 formation has not been conducted. In this communi-
cation, the photochemical production of H2 from bipyridyl tri-
carbonyl rhenium(I) hydrides is presented, along with a
mechanism based on computational studies.
Rising carbon dioxide (CO2) levels as a result of burning fossil
fuels continue to necessitate the development of alternative
carbon neutral energy sources.1,2 Efficient sources of carbon
neutral fuels, such as hydrogen gas (H2), could provide the
energy required to meet the rising global energy demands.
Rhenium(I) metal complexes have been investigated for
their role in the photocatalytic reduction of CO2 to CO,3–8 and
for the role they play as photosensitizers and photocatalysts in
the solar energy conversion of water to H2.9 The advantages of
utilizing rhenium(I) metal catalysts lie in their relative stability
and the ease of excited state studies performed at room
temperature.10–15 For example, the use of fac-Re(bpy)(CO)3Cl
(bpy = 2,2′-bipyridine) as an efficient homogeneous catalyst for
the selective, sustained photochemical, or electrochemical
reduction of CO2 to CO has been reported.3,16,17 The corres-
ponding rhenium(I) bipyridine carbonyl hydride analogues
have also been reported,18–20 in which electrocatalytic CO2
reduction and insertion and photodissociation of Re–H have
been investigated. Modifications to the bipyridine ligand have
also provided some insight into the properties and photoreac-
tivity of rhenium(I) hydrides during photolysis.21,22 Recently, a
three component system involving select rhenium hydrides,
cobalt complexes and triethanolamine (TEA) has been
Although there has been an extensive amount of studies on
Re(I) complexes, remarkably, the photochemical production of
H2 from fac-Re(bpy)(CO)3H does not appear to have been
reported.25 Given that the synthesis of 2,2′-bipyridine fac-Re
(bpy)(CO)3H is known, Scheme 1,26,27 and its 4,4′-disubstituted
derivatives are all accomplished by facile, high yielding reac-
tions, this system is ideal for studying this important reaction.
Photolysis, using a 200 W mercury–xenon lamp (λ =
380–750 nm), was initially performed in air on Re(bpy)(CO)3H
in a variety of solvents (CD3CN, THF-d8, acetone-d6 and
1
CD2Cl2) and monitored by H NMR, Fig. 1. Consistent results
were observed in all cases with the appearance of the H2 reso-
nance and the disappearance of the Re–H resonance. The dis-
aDepartment of Chemistry, University of Wyoming, USA. E-mail: hoberg@uwyo.edu
bDepartment of Petroleum Engineering, University of Wyoming, USA
†In memory of Pat Sullivan.
‡Electronic supplementary information (ESI) available. See DOI: 10.1039/
C9DT03364E
Scheme 1 The reaction pathway for the synthesis of substituted Re
§Deceased.
(bpy)(CO)3H complexes and hydrogen formation.
16148 | Dalton Trans., 2019, 48, 16148–16152
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Communication
appearance of the hydride resonance, 10 and 8 minutes
respectively, whereas complexes 1a–c needed appreciably
longer times of 20 (1a), 35 (1b) and 40 (1c) minutes to comple-
tely degrade at identical concentrations. This decrease in elec-
tron density in the bipyridine ring system allows for electronic
stabilization during the metal to ligand charge transfer occur-
ring during photolysis, which has also been observed and
reported for fac-Re(bpy)(CO)3Cl complexes.13
Photolysis was next carried out in an oxygen free environ-
ment, using 1e, which showed an accelerated rate of formation
of H2. The initial degradation of Re–H was observed prior to
photolysis; however, upon initiation of photolysis, the for-
mation of the degradation product ceased and hydrogen for-
mation began, also confirming the dependence of light. In the
absence of light and O2, degradation of the complex to an
unknown product occurred without the production of H2.
Additional photolysis studies in both the presence and the
absence of water were conducted in an effort to determine the
role water plays in the reaction pathway. Thus, complex 1e was
dissolved in nearly dry d6-acetone. Initial 1H NMR was per-
formed to determine the extent of water in the solvent, and
photolysis was then performed until the water peak and the
rhenium hydride resonances disappeared. Once the water
peak disappeared, photolysis was continued until no
additional H2 was produced. Additional complex 1e was added
to the solution (only the Re–H singlet was visible in the NMR
spectrum, with no H2 or H2O signals observed) and photolysis
was resumed. The formation of new H2 was again observed
demonstrating that H2 production is only dependent upon the
presence of Re–H and light. Finally, a dry d6-acetone/Re–H
sample was spiked with D2O and photolyzed, which produced
only an H2 signal and no HD triplet.
With these data in hand, the mechanism illustrated in
Scheme 2 is proposed. An excited state Re(bpy)(CO)3H 1*
undergoes initial dissociation of CO to produce 2*.
Photochemical dissociation of CO from fac-Re(bpy)(CO)3Cl has
been shown to occur by Ishitani and co-workers.8,31 In their
report, loss of CO occurs followed by rearrangement and then
re-coordination of CO to produce mer-Re(bpy)(CO)3Cl in both
the presence and the absence of a CO atmosphere. Following
CO dissociation 2* relaxes to the ground (singlet) state 2,
which forms bridged Re–H–Re 3 with subsequent hydride
transfer and Re–Re dimer formation to produce Re(η2-H2) 4.
Dihydrogen transfer between CpRe(PPh3)2H3 and IrBr(CO)
(dppe) has been reported,32 and stable cationic η2-H2 rhenium
complexes have been synthesized and characterized through
the addition of H2 to unsaturated rhenium complexes;33 thus
excellent precedence exists for this step. Loss of H2 with con-
current solvent coordination, which could also be water, leads
to the observed dimers 5 and/or 6. Solvent coordination in
these types of systems is also well precedented,4,28 however,
coordination of CO could also occur in this step.
Fig. 1 1H NMR in acetone-d6 of the photolysis of 1a. Chemical shift in
ppm.
appearance of the hydride resonance at 1.84 ppm along with
the appearance of the H2 resonance at 4.5 ppm is apparent.
Interestingly, the spectrum also shows the disappearance of
the water peak at 2.8 ppm and the degradation of the aromatic
bipyridine peaks from 7.4–9.2 ppm. This degradation is attrib-
uted to the formation of the known solid [Re(bpy)(CO)3]2
dimer (see ESI Fig. S2‡),4,28 which precipitates out of solution
as a green solid upon exposure to light. UV-vis spectral data for
the solid were consistent with previously published data.4
Given that no HD production was observed when using
deuterated solvents, a crossover experiment was performed to
identify the source of the H2 gas. Photolysis of an approxi-
mately 1 : 1 mixture of Re(bpy)(CO)3H and Re(bpy)(CO)3D com-
plexes resulted in the production of both HD and H2 peaks as
1
identified via H NMR spectroscopy, Fig. 2. The HD peak was
confirmed from both the chemical shift and the coupling con-
stant (δ = 4.54 ppm and JHD = 42.5 Hz).29,30 This indicates that
hydrogen production is directly related to the reaction of
rhenium hydride or rhenium deuteride with only each other
and not with the solvent.
2,2′-Disubstituted bipyridines containing electron with-
drawing groups 1d and 1e (R = Cl and Br) showed the fast dis-
The reaction pathway in Scheme 2, where after the initial
CO dissociation (1*→2*) the reaction proceeds in the elec-
tronic ground states, is the most likely pathway, as the excited
state lifetimes are on the order of tens to hundreds of
Fig. 2 The 1H NMR crossover experiment of a 1 : 1 ReH/ReD mixture in
acetone-d6. (a) 0 minutes and (b) 30 minutes.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans., 2019, 48, 16148–16152 | 16149

View Article Online
Communication
Dalton Transactions
(DFT). Previous computational studies found that the final
dimeric complex [Re(bpy)(CO)3]2 6 prefers a skewed trans
arrangement (rotation about the Re–Re bond)28 with the orien-
tation of the stacked Re(bpy)(CO)3 rotated by ∼140° over the
skewed cis arrangement, with an ∼40° angle. We found this to be
the case without the empirical correction for dispersion, however,
inclusion of the dispersion interaction34 reverses the order of
stability with the skewed cis conformation preferred by
∼3 kcal mol−1. The skewed cis arrangement is also consistent
with the crystal structure of the manganese biphosphine analogue
[Mn(tmbp)CO3]2 (tmpb = 4,4′-tetramethyl-2,2′-biphosphine).35
The initial structures of intermediates 3,
4 and 5
(Scheme 2) were therefore constructed to correspond to the
skew cis final product. These were fully optimized at the DFT
level, with the B3LYP density functional and the 6-31G(d) basis
set for C and H, 6-31+G(d) for N and O, LANL2DZ for Re,28
Grimme empirical dispersion correction34 and the polarized
continuum model for the solvent.36,37 We also modeled the
corresponding triplet excited states 3*, 4*, 5* and 6*.
The optimized geometry parameters are summarized in
Tables S1–S3 (ESI‡), along with the electronic energies and
thermal free energies (at 298 K, 1 atm) of the individual
species (Table S4‡). A schematic diagram of the Gibbs free
energies computed using the corrected SMD method36,37 for
acetone solvent is plotted in Fig. 3. (The corresponding
diagram for THF is shown in the ESI, Fig. S2.‡)
Scheme 2 The proposed mechanism for the formation of H2 from the
photolysis of Re(bpy)(CO)3H.
It is evident that multiple pathways with the triplet–singlet
relaxation occurring at different steps are feasible. The triplet
pathway (red in Fig. 3) is thermodynamically downhill after the
dissociation of CO from 1*, which requires a total of ∼86 kcal
mol−1 (the triplet state 1* is about 56 kcal mol−1 higher in
energy than the ground singlet state and the CO dissociation
requires an additional ∼30 kcal mol−1). By contrast, the singlet
pathway (highlighted in blue in Fig. 3) has an uphill step of
∼10 kcal mol−1 in the rearrangement of the hydrogen bridged
3 to Re(η2-H2) 4. The final product singlet [Re(bpy)(CO)3]2 is
thermodynamically more stable than the reactant singlet Re
nanoseconds,4,25 which are presumably significantly shorter
than the rate of the next (bimolecular) reaction step. However,
it is possible that alternative pathways exist with additional
steps in the excited (triplet) states, in particular since the reac-
tion is performed under continuous illumination. As the com-
putational modeling below demonstrates, all these pathways
are energetically feasible (Fig. 3) and the singlet and triplet
intermediate species are structurally very similar (ESI‡).
Support for the proposed mechanism was obtained from
computational modeling using density functional theory
(bpy)(CO2)3H, by ∼20 kcal mol−1
.
Structurally, the triplet state of monomer 1* is considerably
distorted from the octahedral geometry around the Re atom
(Table S2 and Fig. S2‡). 2* has a rather regular trigonal bipyra-
midal arrangement, in contrast to the ground state 2, which
again resembles an octahedral structure, but with one equator-
ial coordination site vacant. For the remaining reaction inter-
mediates (3–5) and the product complex (6) the singlet and
triplet structures are very similar (Tables S2 and S3, and
Fig. S2‡). The first proposed double-Re complex 3 has the two
Re centers bridged by hydrogen, however, a Re–Re bond is also
likely, since the Re–Re distance is less than 0.3 Å longer than
that in the fully formed complex 6 (for the triplet states of 3*
Fig. 3 The schematic energy diagram for the proposed reaction
mechanism (in acetone). Gibbs free energy calculated using the cor- the Re–Re distance is only 0.1 Å greater than that in 6*).
rected SMD method (see text for details) is plotted for each of the
chemical species relative to the reactant species: 1 (singlet ground state
Re(bpy)(CO2)3H), 1* (triplet excited state Re(bpy)(CO2)3H) and solvent
(acetone). The blue levels correspond to the singlet states and the red
Complexes 4 and 4* have the shortest Re–Re distances for all
the singlet and triplet species, respectively.
Note also that if only vibrational contributions to the free
energy are considered (the rightmost column of Table S4‡),
levels to triplet states. The numbering follows that in Scheme 2 text (see
also Fig. S2 in the ESI‡).
corresponding to the limit of no free translations or rotations
16150 | Dalton Trans., 2019, 48, 16148–16152
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Communication
which are generally significantly constrained in the condensed 11 R. Heydová, E. Gindensperger, R. Romano, J. Sýkora,
phase (solution), the differences between the individual
A. Vlček, S. Záliš and C. Daniel, J. Phys. Chem. A, 2012, 116,
11319–11329.
12 S. E. Hightower, R. C. Corcoran and B. P. Sullivan, Inorg.
Chem., 2005, 44, 9601–9603.
species are more pronounced. For example, the overall ΔG for
the reaction is about 30 kcal mol−1
.
13 A. Kumar, S.-S. Sun and A. J. Lees, in Photophysics of
Organometallics, ed. A. J. Lees, Springer Berlin Heidelberg,
Berlin, Heidelberg, 2009, vol. 29, pp. 37–71.
Conclusions
Photocatalytic hydrogen formation using the energy from the 14 S. Bolaño, J. Bravo, J. Castro, S. García-Fontán, M. C. Marín
sun may provide an alternative green energy source. In this
work, we have shown that rhenium hydrides can produce H2
and P. Rodríguez-Seoane, J. Organomet. Chem., 2005, 690,
4945–4958.
via photolysis and have further elucidated a working mecha- 15 L. A. Worl, R. Duesing, P. Chen, L. D. Ciana and
nism for this process. An alternative mechanism to the pro- T. J. Meyer, J. Chem. Soc., Dalton Trans., 1991, 848–858.
posed bridged monohydrogen Re–H–Re species would be a 16 J. Hawecker, J.-M. Lehn and R. Ziessel, J. Chem. Soc., Chem.
bridged dihydrogen complex (Re–H2–Re), which has been pos- Commun., 1983, 536–538.
tulated previously in both ruthenium and osmium 17 C. Pac, K. Ishii and S. Yanagida, Chem. Lett., 1989, 18, 765–
systems.38,39 However, modeling studies have indicated that an
768.
analogous bridged Re–H2–Re intermediate is of considerably 18 B. P. Sullivan and T. J. Meyer, Organometallics, 1986, 5,
higher energy than the Re–H–Re species proposed herein. 1500–1502.
Thus, we propose that the rhenium system reported follows a 19 B. P. Sullivan, J. Phys. Chem., 1989, 93, 24–26.
discreetly different mechanism.
20 I. Bruand-Cote and C. Daniel, Chem. – Eur. J., 2002, 8,
1361–1371.
21 M. Cattaneo, F. Fagalde, N. E. Katz, C. D. Borsarelli and
T. Parella, Eur. J. Inorg. Chem., 2007, 2007, 5323–5332.
22 Md. A. Masood, B. P. Sullivan and D. J. Hodgson, Inorg.
Chem., 1994, 33, 5360–5362.
Conflicts of interest
There are no conflicts to declare.
23 A. Messmer, H. Jacobsen and H. Berke, Chem. – Eur. J.,
1999, 5, 3341–3349.
24 G. Albertin, S. Antoniutti, S. Garcia-Fontán, R. Carballo and
F. Padoan, J. Chem. Soc., Dalton Trans., 1998, 2071–2082.
Acknowledgements
We thank the University of Wyoming School of Energy 25 R. N. Perutz and B. Procacci, Chem. Rev., 2016, 116, 8506–8544.
Resources for financial support of this work.
26 J. K. Hino, L. Della Ciana, W. J. Dressick and B. P. Sullivan,
Inorg. Chem., 1992, 31, 1072–1080.
27 D. H. Gibson and X. Yin, J. Am. Chem. Soc., 1998, 120,
11200–11201.
28 E. Fujita and J. T. Muckerman, Inorg. Chem., 2004, 43,
7636–7647.
Notes and references
1 N. S. Lewis, MRS Bull., 2007, 32, 808–820.
2 N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. U. S. A., 29 J. Y.-C. Chen, A. A. Martí, N. J. Turro, K. Komatsu,
2006, 103, 15729–15735.
3 J. Hawecker, J.-M. Lehn and R. Ziessel, Helv. Chim. Acta,
2004, 69, 1990–2012.
4 Y. Hayashi, S. Kita, B. S. Brunschwig and E. Fujita, J. Am.
Chem. Soc., 2003, 125, 11976–11987.
Y. Murata and R. G. Lawler, J. Phys. Chem. B, 2010, 114,
14689–14695.
30 H.-H. Limbach, T. Pery, N. Rothermel, B. Chaudret,
T. Gutmann and G. Buntkowsky, Phys. Chem. Chem. Phys.,
2018, 20, 10697–10712.
5 O. Ishitani, M. W. George, T. Ibusuki, F. P. A. Johnson, 31 Ishitani’s work was performed at 313 nm rather than the
K. Koike, K. Nozaki, C. Pac, J. J. Turner and J. R. Westwell,
Inorg. Chem., 1994, 33, 4712–4717.
6 F. P. A. Johnson, M. W. George, F. Hartl and J. J. Turner,
Organometallics, 1996, 15, 3374–3387.
7 P. Kurz, B. Probst, B. Spingler and R. Alberto, Eur. J. Inorg.
Chem., 2006, 2006, 2966–2974.
8 S. Sato, T. Morimoto and O. Ishitani, Inorg. Chem., 2007,
46, 9051–9053.
longer MLCT excitation used in this work. However, we
have performed the reaction with 1e at 315 nm and
observed the usual loss of the hydride.
32 W. D. Jones and J. A. Maguire, J. Am. Chem. Soc., 1985, 107,
4544–4546.
33 C. Bianchini, A. Marchi, L. Marvelli, M. Peruzzini,
A. Romerosa, R. Rossi and A. Vacca, Organometallics, 1995,
14, 3203–3215.
9 B. Probst, M. Guttentag, A. Rodenberg, P. Hamm and 34 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem.
R. Alberto, Inorg. Chem., 2011, 50, 3404–3412. Phys., 2010, 132, 154104.
10 K. Kalyanasundaram, J. Chem. Soc., Faraday Trans. 2, 1986, 35 F. Hartl, T. Mahabiersing, P. Le Floch, F. Mathey, L. Ricard,
82, 2401–2415.
P. Rosa and S. Záliš, Inorg. Chem., 2003, 42, 4442–4455.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans., 2019, 48, 16148–16152 | 16151

View Article Online
Communication
Dalton Transactions
36 J. B. Foresman and E. Frisch, Exploring Chemistry with 38 J. P. Collman, J. E. Hutchison, P. S. Wagenknecht,
Electronic Structure Methods, Gaussian Inc., Walingford, CT,
USA, 3rd edn, 2015.
N. S. Lewis, M. A. Lopez and R. Guilard, J. Am. Chem. Soc.,
1990, 112, 8206–8208.
37 A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. 39 J. P. Collman, P. S. Wagenknecht and N. S. Lewis, J. Am.
Chem. B, 2009, 113, 6378–6396.
Chem. Soc., 1992, 114, 5665–5673.
16152 | Dalton Trans., 2019, 48, 16148–16152
This journal is © The Royal Society of Chemistry 2019