Proton
–
hydride tautomerism in hydrogen
evolution catalysis
Luis M. Aguirre Quintana
a,b
, Samantha I. Johnson
a,b
, Sydney L. Corona
a,b
, Walther Villatoro
a,b
, William A. Goddard III
a,b
,
Michael K. Takase
a,b
, David G. VanderVelde
a,b
, Jay R. Winkler
a,b,1
, Harry B. Gray
a,b,1
, and James D. Blakemore
a,b,c,1
a
Beckman Institute, California Institute of Technology, Pasadena, CA 91125;
b
Division of Chemistry and Chemical Engineering, California Institute of
Technology, Pasadena, CA 91125; and
c
Department of Chemistry, University of Kansas, Lawrence, KS 66044
Contributed by Harry B. Gray, April 17, 2016 (sent for review February 4, 2016; reviewed by Alexander Miller and David Milstein)
Efficient generation of hydrogen
from renewable resources requires
development of catalysts that avoid deep wells and high barriers.
Information about the energy landscape for H
2
production can be
obtained by chemical characterizati
on of catalytic intermediates, but
few have been observed to date. We have isolated and characterized
a key intermediate in 2
e
–
+
2H
+
→
H
2
catalysis. This intermediate,
obtained by treatment of Cp*Rh(bpy) (Cp*,
η
5
-pentamethylcyclopen-
tadienyl; bpy,
κ
2
-2,2
′
-bipyridyl) with acid, is not a hydride species but
rather, bears [
η
4
-Cp*H] as a ligand. Delivery of a second proton to this
species leads to evolution of H
2
and reformation of
η
5
-Cp* bound to
rhodium(III). With suitable choices
of acids and bases, the Cp*Rh(bpy)
complex catalyzes facile and reversible interconversion of H
+
and H
2
.
rhodium
|
pentamethylcyclopentadienyl
|
catalysis
|
proton reduction
|
hydrogen evolution
S
ustainable and ec
onomically competitive production of hydro-
gen (H
2
) as a fuel depends on the development of new molecules
and materials that catalyze conver
sion of protons and electrons into
H
2
with high rates and minimal energy input (1). Platinum fulfills
these latter requirements but fails t
he economic test. Rational design
of new catalysts that overcome these challenges hinges on a detailed
understanding of the elementary chemical reaction steps involved in
H
–
H bond formation (2). Notably, catalysis with a molecular cobalt
complex, a [BF
2
]-bridged cobaloxime, has been shown by kinetics
and modeling to proceed by protonation of a Co
II
–
H complex to
yield H
2
(3). However, most intermediates in H
2
evolution elude
detection, limiting insight into the bond-breaking and -making pro-
cesses involved in catalysis (4). In recent work, we have descended to
the second row of group 9 in the periodic table to enable observation
and isolation of catalytic intermediates.
We report here on our investigations of [Cp*Rh
III
(diimine)L]
2
+
(Cp*,
η
5
-pentamethylcyclopentadienyl
) complexes that evolve hydro-
gen catalytically on reduction. In
early work, Kölle and Grätzel (5)
found that H
2
is evolved on reduction at pH
<
2inanaqueous
photochemical system, and Deronzier and coworkers (6) observed
electrochemical H
2
evolution at pH 1. Of interest here is our finding
that H
2
is evolved electrochemically in acetonitrile solutions of
[Cp*Rh
III
(diimine)L]
2
+
with tosic acid serving as the proton source
(7). The
η
5
-Cp* ligand imparts high stability and solubility to these
catalysts (8). Not surprisingly, then, they have been used widely for
catalysis of redox transformati
ons, including reduction of NAD
+
to
NADH (9
–
11), dehydrogenation of formic acid or alcohols (12, 13),
and hydrogenation of organic compounds (14). In these reactions,
Rh
III
–
H is believed to be a reactive intermediate (15).
Hydrogen evolution catalysts have been discovered recently in
which the metal center and associated ligands cooperate in un-
expected ways. A case in point featu
res ligand-centered protonation
and subsequent hydride-like reactivity of a nascent C
–
Hbondina
nickel phlorin system (16), although a nickel hydride was not im-
plicated in the cycle for hydrogen evolution. In other work of note,
Hull et al. (17) reported dehydrogenation of HCO
2
Hpromotedby
proton-responsive hydroxybipyridin
e-ligated catalysts, and Lacy et al.
(18) showed that ligand-centered protonation of a cobalt complex
couldleadtohydrogenevolution.
Much work has been done on Cp*Rh compounds. Of impor-
tance is that Maitlis and coworkers (19) reported Cp*Rh(Cp*H)
formation in low yield (14%) by reduction of [(Cp*)
2
Rh](PF
6
). In
other relevant work, Jones et al. (20) showed that the dihydride
complex Cp*Rh(PMe
3
)(H)
2
loses free Cp*H on treatment with
excess PMe
3
at elevated temperature. Also, several other rhodium
and iron complexes with Cp*H ligands are known (21
–
25).
Building on these earlier investigations, we report here an unusual
mode of metal
–
ligand cooperation that drives proton
–
hydride tauto-
merization during hydrogen evolution. We have found that
η
4
-Cp*H
is produced en route to H
2
production with Cp*Rh(diimine)
complexes, suggesting an important role for ligand participation in
hydrogen evolution catalysis.
Results
The preparation, isolation, and cryst
allographic characterization of
Cp*Rh
I
(diimine) complexes with
η
5
-Cp* and
κ
2
-diimine coordination
are all well-documented (26, 27). Treatment of the
κ
2
-2,2
′
-bipyridyl
[bpy; Cp*Rh(bpy);
1
] (Fig. 1) and 1,10-ph
enanthroline [phen;
Cp*Rh(phen);
2
] complexes with excess protonated dimethylforma-
mide ([DMF
·
H]
+
[OTf]
−
)inMeCN(pK
a
=
6.1) results in stoichio-
metric production of H
2
gas. Use of weaker acids lowers the
thermodynamic driving force for hydrogen evolution (28), likely
slowing the reaction and increasing the chance of trapping one or
more intermediates. Treatment of
1
with one to five equivalents of
[Et
3
NH]
+
Br
−
(pK
a
=
18.8) in CD
3
CN results in a rapid color change
Significance
The discovery of efficient hydrogen evolution catalysts for solar
fuels production continues to be an active research field. Catalyst
optimization depends on detaile
d knowledge of the elementary
chemical reaction steps involved in catalysis. Isolation of interme-
diates in catalytic processes is un
common owing to their necessarily
low stability. By using weak acids, we have isolated and charac-
terizedanintermediateinthe2
e
−
+
2H
+
→
H
2
reaction catalyzed
by
η
5
-pentamethylcyclopentadienyl (Cp*) Rh(
κ
2
-2,2
′
-bipyridyl)
[Rh(bpy)]. We find that the preferred site of Cp*Rh(bpy) protonation
is not the metal center but is the Cp* ligand. Despite the reputation
of Cp* as a stable ligand in organometallic chemistry, these results
suggest an important role for close metal
–
ligand cooperation in
promoting hydrogen
–
evolution catalysis.
Author contributions: S.I.J., D.G.V., J.R.W., H.B.G., and J.D.B. designed research; L.M.A.Q.,
S.I.J., S.L.C., W.V., M.K.T., and J.D.B. performed research; W.A.G. contributed new reagents/
analytic tools; L.M.A.Q., S.I.J., S.L.C., W.V., M.K.T., D.G.V., J.R.W., H.B.G., and J.D.B. analyzed
data; and S.I.J., J.R.W., H.B.G., and J.D.B. wrote the paper.
Reviewers: A.M., University of North Carolina; and D.M., The Weizmann Institute of
Science.
The authors declare no conflict of interest.
Data deposition: The atomic coordinates and structure factors have been deposited in the
Cambridge Crystallographic Data Centre (accession no. 1424707).
1
To whom correspondence may be addressed. Email: winklerj@caltech.edu, hbgray@
caltech.edu, or blakemore@ku.edu.
This article contains supporting information online at
www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1606018113/-/DCSupplemental
.
www.pnas.org/cgi/doi/10.1073/pnas.1606018113
PNAS
|
June 7, 2016
|
vol. 113
|
no. 23
|
6409
–
6414
CHEMISTRY
from purple to reddish brown but no gas evolution, consistent with
formation of compound
3
(Fig. 1). Similar spectral features are
encountered on treatment of
1
or
2
with [Et
3
NH]
+
OTf
–
, resulting
in formation of analogous compounds
4
(with bpy) and
5
(with
phen) bearing the [Cp*H] ligand.
Allowing a concentrated acetonitrile solution of
3
to stand for
several days yielded orange
–
yellow single crystals suitable for X-ray
diffraction (the structure that we obtained is shown in Fig. 2). The
geometry at the Rh
I
center is distorted square pyramidal, with a
mirror plane containing the Rh1
–
Br1 bond bisecting the molecule.
A proton (observable in the electron density map) is present on the
ring of the former Cp* ligand, with the [Cp*H] bound in an
η
4
mode. The bpy ligand retains typical
κ
2
coordination in accordance
with our interpretation of its
1
Hand
13
C NMR spectra. The Rh1
–
N1
bond length is 2.117(1) Å, whereas Rh1
–
C2 and Rh1
–
C3 are 2.133
(1) and 2.112(2) Å, respectively. The mirror plane symmetry element
requires identical bond distances from Rh1 to the corresponding
N1
′
,C2
′
,andC3
′
atoms.
†
In
1
,theC
–
C bond length of 1.422(4) Å between the pyridyl
rings of the bpy ligand is consistent with considerable bpy anion
character (29) caused by electron donation from Rh
I
augmented by
the strongly
π
-donating
η
5
-Cp* ligand. On protonation of Cp* to
form
3
,thisC
–
C bond lengthens to 1.475(3) Å, similar to the distance
found in free bpy and notably, an analogous [Cp*Rh
III
(bpy)Cl]Cl
complex (30). This marked structural change is attributable to
the conversion of the
π
-donating [
η
5
-Cp*]
−
ligand to a
π
-accepting
[
η
4
-Cp*H]
0
group. In the crystal structure of a free Cp*H analog,
Cp(CH
2
Ar)
5
H[Ar
=
−
(C
6
H
4
)CH
3
], the corresponding C2
–
C3
bond length is 1.343(3) Å, and the C3
–
C3
′
bond is 1.467(3) Å (31).
In
3
,theC2
–
C3 and C3
–
C3
′
bonds are nearly identical at 1.442(2)
and 1.442(3) Å, respectively, consistent with both
σ
-donation and
π
-backbonding between Rh1 and [Cp*H].
Electronic absorption spectra re
flect the marked change in elec-
tronic structure on protonation of
1
to form
3
. Bands in the visible
region (
e
∼
4
–
15
×
10
3
M
−
1
cm
−
1
) assigned to metal-to-ligand
charge-transfer(MLCT)transit
ions (27, 32) dominate the spectrum
of
1
, accounting for its purple color. Conversely, the spectrum of
3
displays intense bands (
e
∼
20
–
30
×
10
3
M
–
1
cm
–
1
)acrosstheUV
region with weaker absorption tailing into the visible (
e
∼
1.7
×
10
3
M
–
1
cm
–
1
at 507 nm), producing a reddish brown solution.
The UV absorption in
3
is attributable to both intraligand and
MLCT transitions. The observed blue shift of MLCT absorption
bands is consistent with a positive shift of the rhodium(II/I) for-
mal potential on shifting coordination from the electron-donating
[
η
5
-Cp*]
–
ligand to the
π
-accepting [
η
4
-Cp*H] group. Similarly,
Rh(norbornadienyl)(bpy)Cl is a reddish solid exhibiting only
weak visible absorption (
e
∼
0.85
×
10
3
M
–
1
cm
–
1
at 478 nm) (33),
and [Rh(CO)
2
(bpy)]ClO
4
is a yellow solid (34). Because most
[Cp*Rh
III
(diimine)L]
2
+
complexes are pale yellow to orange in ap-
pearance, the coloration of
3
more closely resembles that of an
Rh
III
compound, although it is formally Rh
I
(5, 7). This finding raises
the possibility that compounds protonated at Cp* may have escaped
detection in the past, because their UV-visible absorption profiles
are not markedly different from s
pectra of their hydride analogs.
Exposure of
3
,
4
,and
5
to excess Et
3
NH
+
does not lead to ad-
ditional protonation of the Rh complex or any other transformation.
Treatment with stronger acid ([DMF
·
H]
+
[OTf]
–
), however, triggers
quantitative H
2
evolution and generation of Rh
III
. For example,
addition of 1 eq [NEt
3
H]
+
[OTf]
–
to 50 mg
2
results in negligible H
2
Fig. 1.
(
A
) Structures of
1
,
3
,and
4
.(
B
) Proposed cycle for H
2
evolution cat-
alyzed by
1
through a [Cp*H]Rh complex.
Fig. 2.
Structure of
3
. Displacement ellipsoids are shown at 50% probability. One
cocrystallized acetonitrile molecule and all H atoms, except those bonded to C1
and C6, are omitted for clarity. Blue, nitrogen; purple, rhodium; red, bromide.
†
The distinctive
1
H NMR spectrum of
3
is characterized by two singlets in the alkyl region
(1.85 and 0.94 ppm) along with a doublet and quartet (
d
=
0.54;
q
=
2.54 ppm;
3
J
H
–
H
=
6.2 Hz).
The relative integrations (6:6:3:1, respectivel
y) of these peak groups are consistent with [Cp*H]
coordinated to Rh
I
. Resonances in the aromatic region shift on conversion of
1
to
3
but retain
the expected couplings for an intact bpy ligand. The remaining features in the
1
HNMR
spectrum of
3
are attributed to excess [Et
3
NH]
+
Br
–
and residual CH
3
CN. Coordination of
[Cp*H] to Rh is confirmed by comparison with the
1
H NMR of the free ligand (
s
=
1.74 and
1.79 ppm;
d
=
0.97;
q
=
2.45 ppm;
3
J
H
–
H
=
7.6 Hz) as well as the
13
C{
1
H} NMR of
3
[
d
=
94.4 ppm;
1
J
Rh
–
C
=
10.6 Hz;
d
=
55.4 ppm;
1
J
Rh
–
C
=
10.9 Hz;
d
=
57 ppm;
2
J
Rh
–
C
=
3.6 Hz;
103
Rh(I
=
1/2) is
100% abundant].
6410
|
www.pnas.org/cgi/doi/10.1073/pnas.1606018113
Quintana et al.
production. Addition of 1 eq [DMF
·
H]
+
[OTf]
–
to the solution
containing
5
produces a small yield of H
2
(9%); the predominant
reaction is protonation of NEt
3
to form Et
3
NH
+
.Additionofa
second 1 eq [DMF
·
H]
+
[OTf]
–
results in quantitative formation of
H
2
based on remaining Rh
I
(91%). The color of the solution at the
end of these reactions is pale yellow, consistent with the presence
of [Cp*Rh
III
]. NMR spectra of the reaction products indicate
production of H
2
along with resonances for [Cp*Rh
III
]complexes
with either bound acetonitrile or [OTf]
–
. Experiments carried out
with 1,3,5-(OMe)
3
C
6
H
3
as internal standard confirm
>
99% final
conversion to [Cp*Rh
III
(bpy)] complexes. Hydrogen production
can be made catalytic by chemical or electrochemical reduction of
[Cp*Rh
III
(diimine)] to regenerate [Cp*Rh
I
(diimine)].
With this knowledge of the observed intermediate in hand,
density functional theory (DFT) ca
lculations were carried out for
1
,
4H
, and two tautomeric forms of
4
.Fig.3
A
shows calculated free
energies for these structures and the overall H
2
evolution ther-
modynamics (values in braces are those calculated by DFT).
‡
The
upper pathway features protonation of
1
by triethylammonium to
form
4
, which is exergonic by
∼
1 kcal/mol, whereas formation of
the Rh
III
–
Htautomer(
4H
) is endergonic by 6.0 kcal/mol. Complex
4
bears a C
–
H bond on the Cp* ring, with an unremarkable C
–
H
bond distance computed to be 1.1 Å. Importantly, the Rh to H dis-
tance is computed to be 2.86 Å, thus excluding a bridging mode
between the metal center and the ring. Another tautomer
bearing
exo
-[
η
4
-Cp*H] (
4X
), in which the ring-bound H faces
away from Rh, is slightly less stable (
+
0.5 kcal/mol) than
4
.
The lower pathway features the same protonated structures
formed in reaction with [DMF
·
H]
+
. The difference between
4
and
4H
is still 6.9 kcal/mol, but the stronger acid increases the free energy
released on formation of
4
from
1
to 17.1 kcal/mol. In accordance
with experiment, this pathway l
eads to exergonic evolution of H
2
.
The free energy changes for the overall reactions are shown in the
last column, and they are determined from the experimental Rh
III/I
reduction potential [
−
1.05 V vs. a ferricenium/ferrocene (Fc
+
/0
)
reference] and acid pK
a
values of the acids in MeCN (7, 30, 35).
Consistent with experimental observations, H
2
evolution using 2 eq
[Et
3
NH]
+
is endergonic, whereas H
2
production with a combination
of 1 eq [Et
3
NH]
+
and 1 eq [DMF
·
H]
+
is a spontaneous process.
The frontier molecular orbitals for
1
,
4H
,and
4
are shown in Fig.
3
B
. The ground-state highest occu
pied molecular orbital (HOMO)
of
1
is delocalized onto the ligands (Fig. 3
B
), consistent with the
X-ray structure, suggesting significant back donation of electron
density from the Rh
I
center. On protonation to form
4
,theHOMO
shifts to a primarily Rh
I
-centered
d
z
2
orbital stabilized by the newly
formed
π
-accepting [Cp*H] ligand.
Discussion
Based on our experimental data and computational studies, we
postulate that reaction of
1
with protic acids initially produces
4H
,
which then rearranges to the more stable
4
. Miller and coworkers
(36) proposed an analogous pathway involving
4
in the de-
carboxylation of an Rh
III
formate complex and most importantly,
have directly observed the hydride intermediate by low-temperature
protonation of Cp*Rh(bpy). In the work by Miller and coworkers
(36), the hydride converts to an
η
4
-Cp*H complex on warming, a
sequence consistent with our observation of exclusive
endo
-pro-
tonation of the Cp* ligand. The computed structure of
4H
also
features [Cp*] character in the HOMO, providing electron density
that would attract H
+
to form
4
.
Deprotonation measurements suggest facile formation and
interconversion of
4H
and
4
(as well as analogous
5H
/
5
), but only
stronger acids effect the conversion of
3
,
4
,or
5
to H
2
and
[Cp*Rh
III
(diamine)]. With [NEt
3
H]
+
[OTf]
–
, hydrogen evolution
with
1
or
2
is slightly endergonic based on Rh reduction potentials
obtained from cyclic voltammetry (7, 15). However, dissolution of
[Cp*Rh
III
(bpy)MeCN][PF
6
]
2
in MeCN containing
∼
25 eq NEt
3
un-
der 1 atm H
2
at ambient temperature generates
1
within 30 min,
implying low barriers for the individual steps in our reaction sequence.
Two plausible pathways can be envisioned for hydrogen evolu-
tion from
4
(Fig. 4). Transient reformation of the monohydride
Cp*Rh
III
–
H (A in Fig. 4) could be followed by protonation at the
metal center to produce an Rh
III
(H
2
) complex. Indeed,
1
H NMR
spectral features suggest the existence of such a hydride species in the
case of a bulky
κ
2
-6,6
′
-dimethyl-bpy analog (10, 13) or in C
6
D
5
Cl
Fig. 3.
(
A
) Relative free energies of
1
, its protonated analogs, and net reaction products on H
2
evolution. The lowest energy protonated form
4
bears
endo
-
[
η
4
-Cp*H]. Relative energies are given in green, with modeled values shown in braces and experimental thermodynamic values shown without braces;
conjugate bases for the protonation events are given in blue. (
B
) Frontier molecular orbitals computed for compounds
1
,
4H
, and
4
.
‡
SI Appendix
and
Materials and Methods
contain detailed descriptions of computational
methods. Experimental thermodynamic values for H
2
evolution are based on published
redox potentials (6, 7) and experimental pK
a
values.
Quintana et al.
PNAS
|
June 7, 2016
|
vol. 113
|
no. 23
|
6411
CHEMISTRY
at 238 K (24). Our isolation and characterization of
4
, however, raise
the possibility of an alternative p
athway (B in Fig. 4), in which Cp*H
serves as a proton relay (37): protonation of
4
at the metal center
would form [(Cp*H)Rh(H)(bpy)]
2
+
, which then could evolve H
2
through intramolecular reductive elimination. Protonation of
4
at the
Rh center is consistent with the
d
z
2
-localized HOMO identified in
DFT calculations. The fact that Et
3
NH
+
cannot drive this reaction at
low concentrations is attributable to the decreased basicity of
4
;useof
stronger acid ([DMF
·
H]
+
) leads to rapid H
2
generation. The intra-
molecular reductive elimination mechanism is analogous to pathways
proposed to be operative in other highly active H
2
evolution catalysts;
the search for experimental evidence for these pathways remains an
active area of investigation (38
–
40).
In our case, [Cp*H] does not disso
ciate from the metal center, and
quantitative H
2
evolution proceeds from the reduced complex (21).
Importantly,
η
5
-Cp* coordination to Rh is restored after hydrogen
evolution. The unique properties of the Cp* ligand framework enable
π
-donor [
η
5
-Cp*]
–
to convert reversibly to
π
-acceptor [
η
4
-Cp*H]
0
,
stabilizing intermediates in the c
atalytic cycle. Furthermore, this
transformation places a proton tanta
lizingly close to the metal center
(41, 42). For [Cp*Rh] bearing
κ
2
-(
C,N
)-2-phenylpyridine (ppy),
Norton and coworkers (43) have shown that an Rh
III
–
Histhe
preferred form, likely because of the more electron-rich metal cen-
ter. However, the Rh
III/I
[Cp*Rh
III
(ppy)(CH
3
CN)
+
/Cp*Rh
I
(ppy)
–
]
potential in the ppy complex is
∼
500 mV more negative than that
of
1
, resulting in a barrier that prevents H
2
-driven production of
[Cp*Rh(ppy)]
–
with NEt
3
as base (44).
In Cp*Rh(diimine) complexes, the metal center seems to act as
the primary site of proton capture from solution, initially forming a
hydride species (
4H
) that we do not directly detect. This hydride
then tautomerizes by proton tran
sfer to the nearby basic [Cp*]
–
ligand, restoring the formal Rh
I
oxidation state on formation of
4
.
This proton
–
hydride tautomerism represents an interesting analog
to proton capture mechanisms in o
ther complexes, such as those
described by DuBois and coworkers (38), which rely on pendant
amine bases to shuttle protons to the metal center. Our case also
contrasts with older reports fro
m Davies et al. (45), Reger et al.
(46), and Norton and coworke
rs (47) involving direct
exo
-transfer
of [H
–
] to a Cp ligand bound to iron or tungsten complexes.
The pathway of H
2
generation involving initial protonation of
complex
4
and subsequent H
–
H bond formation is reminiscent of
bond activation by close metal
–
ligand cooperation (48), especially
those cases involving l
igand dearomatization
–
aromatization pro-
cesses in pincer-type systems (49). Milstein and coworkers (50) have
observed proton transfer to and from a pincer arm in an Ru
–
H
complex, resulting in aromatization
–
dearomatization of a pyridine
core on the ligand. Additionally, in a related example, an Ir com-
plex was suggested to react with D
2
through initial dearomatization
of a pincer ligand (51). It follows that the specific organic acid
–
base
used with the complexes described here could affect the energetics
and preference of the reaction channel through hydrogen-bonding
interactions. In the case of
4
, proton transfer from [Cp*H] to form
H
2
would result in aromatization of the [Cp*]; this energetically
favorable process lowers the barrier to H
–
H bond formation.
Hydrogen evolution catalysts must assemble two electrons and
two protons into an H
2
molecule. The order in which these four
particles are delivered to the catalyst impacts the speed and ener-
getics of the transformation. In the case of [Cp*Rh
III
(diimine)L]
2
+
catalysts, the sequence is
e
–
+
e
–
+
H
+
+
H
+
. Owing to the instability
of the Rh
II
oxidation state, the second electron is delivered at a
more positive potential than the fi
rst electron. The energetics of
subsequent proton delivery, however, follow the usual trend,
wherein the first proton can be delivered by a relatively weak acid,
but a stronger acid is required to deliver the second proton. After
the second proton has been delivered to
4
or
4H
,thecomplexis
poisedtofacilitateH
–
H bond formation followed by H
2
dissocia-
tion. Our findings provide key insight into the design elements re-
quired for a catalyst that avoids deep wells and high barriers in 2
e
–
+
2H
+
hydrogen evolution.
Materials and Methods
All manipulations were carried out in a dry N
2
-filled glovebox (Vacuum At-
mospheres Co.) or under N
2
atmosphere using standard Schlenk techniques
unless otherwise noted. All solvents were of commercial grade and dried over
activated alumina using a J. C. Meyer Solvent Purification System before use. All
chemicals were from major commercial suppliers and used as received without
additional purification.
1
H,
13
C, and
19
F NMR spectra were collected on 400- or
500-MHz Varian or Bruker Spectrometers and referenced to the residual pro-
tiosolvent signal (52) in the case of
1
Hand
13
C or the deuterium lock signal in
the case of
19
F. Chemical shifts (
δ
) are reported in units of parts per million, and
coupling constants (
J
) are reported in hertz.
Cp*Rh(bpy) and Cp*Rh(phen) were synthesized according to the literature
method using Na(Hg)
27
.[DMF
·
H]
+
OTf
–
was synthesized according to the
method by Favier and Duñach (53). [Et
3
NH]
+
OTf
–
was prepared according to the
literature (54). The rhodium complexes bearing Cp*H were not isolated in pure
form, but instead generated in situ by treatment with triethylammonium salts.
1
HNMR,
13
C{
1
H} NMR, and electronic absorption spectroscopy were used to
confirm clean conversion of the rhodium starting materials. In one case (
3
),
crystals suitable for X-ray diffraction were obtained as described below.
(Cp*)Rh(bpy)Br (3).
In a typical experiment,
∼
5mg
1
was dissolved in acetoni-
trile, and then,
∼
1
–
5eq[Et
3
NH]
+
Br
–
was added, resulting in a color change from
purple to brown and quantitative formation of
3
. The resulting solution con-
tains
3
plus the conjugate base NEt
3
and any excess [Et
3
NH]
+
Br
–
.ForNMR
analysis, the compound was prepared in
d
3
-MeCN and transferred to a J. Young
Tube. No decomposition of the compound over hours was detectable. How-
ever,
3
was prepared fresh for each experiment and typically not isolated as
asolid.
Peak multiplicities for the NMR spectra are listed, where d is doublet, q is
quartet, s is singlet, dd is doublet of doublets, td is triplet of doublets, t is
triplet, and ddd is doublet of doublet of doublets.
1
H NMR (400 MHz; MeCN-
d
3
)
δ
8.88 (d, 2H,
3
J
H
–
H
=
5.1 Hz), 8.28 (d, 2H,
3
J
H
–
H
=
8.2Hz),8.02(td,2H,
3
J
H
–
H
=
7.9 Hz,
4
J
H
–
H
=
1.6 Hz), 7.59 (t, 2H,
3
J
H
–
H
=
6.4 Hz),
Fig. 4.
Proposed pathways for hydrogen evolution starting from
4
.
6412
|
www.pnas.org/cgi/doi/10.1073/pnas.1606018113
Quintana et al.
2.54 (q, 1H,
3
J
H
–
H
=
6.2 Hz), 1.85 (s, 6H), 0.94 (s, 6H), 0.54 (d, 3H,
3
J
H
–
H
=
6.2 Hz).
13
C{
1
H} NMR (101 MHz; MeCN-
d
3
)
δ
154.7, 152.2, 138.2, 127.1, 123.2, 94.4 (d,
1
J
Rh
–
H
=
10.6 Hz), 57.0 (d,
2
J
Rh
–
H
=
3.6 Hz), 55.4 (d,
1
J
Rh
–
H
=
10.9 Hz), 19.9,
12.1, 10.7. Electronic absorption (MeCN-
d
3
):
λ
max
(
e
)
=
507 nm (1,710 L mol
–
1
cm
–
1
),
351 (8,950), 289 (33,400), 273 (31,500).
Crystals suitable for X-ray diffraction were obtained by allowing a so-
lution of
3
in MeCN to stand at room temperature under inert atmosphere
for several days.
(Cp*)Rh(bpy)OTf (4).
Compound
4
was prepared analogously to
3
but with
[Et
3
NH]
+
OTf
–
[;
4
is stable at room temperature for 1
–
2h.
1
H NMR (400 MHz;
MeCN-
d
3
)
δ
8.98 (d, 2H,
3
J
H
–
H
=
5.0 Hz), 8.36 (d, 2H,
3
J
H
–
H
=
8.1 Hz), 8.10 (td, 2H,
3
J
H
–
H
=
7.9 Hz,
4
J
H
–
H
=
1.7Hz),7.67(ddd,2H,
3
J
H
–
H
=
7.6 and 5.3 Hz,
4
J
H
–
H
=
1.2 Hz),
2.31 (q, 1H,
3
J
H
–
H
=
6.1 Hz), 1.92 (s, 6H), 0.79 (s, 6H), and 0.53 (d, 3H,
3
J
H
–
H
=
6.2 Hz).
19
F NMR (376 MHz; MeCN-
d
3
)
δ
−
79.4.
(Cp*)Rh(phen)OTf (5).
Compound
5
was prepared analogously to
3
but starting
with
2
and [Et
3
NH]
+
OTf
–
.
1
H NMR (400 MHz; MeCN-
d
3
) 9.14 (dd, 2H,
3
J
H
–
H
=
5.0 Hz,
4
J
H
–
H
=
1.0 Hz), 8.55 (dd, 2H,
3
J
H
–
H
=
8.1 Hz,
4
J
H
–
H
=
1.4 Hz), 8.05 (s, 2H), 7.90 (dd,
2H,
3
J
H
–
H
=
4.9 and 8.2 Hz), 2.47 (q, 1H,
3
J
H
–
H
=
6.1 Hz), 1.86 (s, 6H), 1.10 (s, 6H), 0.60
(d, 3H,
3
J
H
–
H
=
6.2 Hz).
X-Ray Crystallography.
Low-temperature diffraction data (
φ
-and
ω
-scans) were
collected on a Bruker AXS D8 VENTURE KAPPA Diffractometer coupled to a
PHOTON 100 CMOS Detector with Mo
K
α
radiation (
λ
=
0.71073 Å) from an I
μ
S
Microsource for the structure of compound
3
. The structure was solved by
direct methods using SHELXS (55) and refined against
F
2
on all data by full-
matrix least squares with SHELXL-2014 (56) using established refinement
techniques (57). All nonhydrogen atoms were refined anisotropically. Unless
otherwise noted, all hydrogen atoms were included in the model at geo-
metrically calculated positions and refined using a riding model. The isotropic
displacement parameters of all hydrogen atoms were fixed to 1.2 times the
U
value of the atoms to which they are linked (1.5 times for methyl groups).
Compound
3
crystallizes in the orthorhombic space group
Pnma
with one-half
a molecule in the asymmetric unit along with one-half a molecule of acetonitrile.
The coordinates for the hydrogen atom bound to C1 were located in the dif-
ference Fourier synthesis and refined semifreely with the help of a restraint on
the C
–
H distance [0.98(4) Å].
Cambridge Crystallographic Data Centre accession number 1424707 contains
the supplementary crystallographic data for compound
3
. These data can be
obtained free of charge at
https://summary.ccdc.cam.ac.uk/structure-summary-
form
or from the Cambridge Crystallographic Data Centre.
GC.
Gas analysis for determination of hydrogen evolution on acid addition was
performed with an Agilent 7890A Gas Chromatograph with separate columns for
analyses of hydrogen gas and nitrogen, oxygen, carbon dioxide, carbon mon-
oxide, hydrogen disulfid
e, methane, ethane, ethyl
ene, and ethyne. The in-
strument was calibrated with a standard gas mixture to obtain quantitative data.
For the hydrogen detection experiment, 50 mg
1
or
2
were loaded into an
N
2
-purged Schlenk flask and dissolved in
MeCN. A background ex
periment before
addition of acid confirmed only nitro
gen and minimal oxygen in the glovebox
atmosphere. Samples of headspace gas we
re taken after progressive additions of
1eq[Et
3
NH]
+
OTf
–
,1eq[DMF
·
H]
+
OTf
–
, and an additional 1 eq [DMF
·
H]
+
OTf
–
.
Computational Details.
Geometry, frequency, and sol
vation calculations were
performed with the B3LYP functional (58, 59) with 6
–
31G** basis set on organics
(60, 61). Rh was treated with the Los Alamos Small Core Potential and 2-
ζ
Basis
Set (62). Single-point energy calculat
ions were completed with the M06 func-
tional (63) with the 6
–
311G**
++
basis on organics. Rh was treated with the 3-
ζ
LACV3P**
++
basis set augmented with f and diffuse functions (64). Solvation in
acetonitrile was applied using the Poisson Boltzmann polarizable continuum
model with a dielectric constant and
proberadiusof37.5and2.19Å,re-
spectively. To calculate the free energ
y of acetonitrile, 1-atm ideal gas free
energy was calculated, and the empirical e
nergy of vaporization, 1.27 kcal/mol,
was subtracted (65). Free energies were computed from the following equation:
G
=
E
M
06
+
G
solv
+
E
ZPE
+
H
vib
+
H
TR
−
T
ð
S
vib
+
S
elec
Þ
,
where zero point energies, enthalpic, and entropic effects were provided by
frequency calculations at room temperature. To validate calculations with the
acids, the pK
a
values of triethylammonium and [DMF
·
H]
+
[OTf]
−
were calculated
in acetonitrile. For these calculations, the value of the proton free energy in
solution was calculated fro
m its gas-phase free energy (
G
H
+
=
H
−
TS
=
2.
5k
B
T
−
T
×
26.04
=
−
6.3 kcal/mol) plus the empirical free energy of solvation at a
concentraion of 1 M [
Δ
G
H
+
,solv
=
−
260.2
+
k
B
T
ln(24.5)]. This calculation yields a
value of
−
264.6 kcal/mol (66). Using this method, the pK
a
of triethylammonium
in acetonitrile was calculated as 17.8, which compares well with the experi-
mentally measured value of 18.8. To obtain a correct pK
a
of dimethylforma-
mide in acetonitrile, the van der Waals radii of the acidic proton and basic
oxygen were changed to 1.15 and 2.00 Å, respectively. Use of these values,
which were published previously (67), produced a calculated pK
a
of 5.9, which
again compares well with the experimentally measured value of 6.1. In calcu-
lations involving the Rh complexes, the acid complexes themselves were used
rather than the energy of the free proton. Calculations were completed in
Jaguar (68); similarly calculated systems have shown good agreement with
experiment (69, 70). In addition, calculated structures were in good agreement
with the crystal structures for
1
and
3
, with errors in the Rh
–
N bond length of
∼
0.05 Å and the C
–
C bond lengths of
∼
0.01 Å.
ACKNOWLEDGMENTS.
S.I.J. thanks Dr. Robert Nielsen for helpful discussions.
This research, which was carried out in pa
rt at the Molecular Materials Research
Center and the Laser Resource Center of
the Beckman Institute (California
Institute of Technology), was supported by NSF CCI Solar Fuels Program CHE-
1305124. S.I.J. and J.D.B. acknowledge fellowships from the Resnick Sustainability
Institute at Caltech.
1. Lewis NS, Nocera DG (2006) Powering the planet: Chemical challenges in solar energy
utilization.
Proc Natl Acad Sci USA
103(43):15729
–
15735.
2. Halpern J, Peters E (1955) Mechanism of the catalytic activation of molecular hydro-
gen by metal ions.
J Chem Phys
23(3):605.
3. Dempsey JL, Winkler JR, Gray HB (2010) Mechanism of H
2
evolution from a photo-
generated hydridocobaloxime.
J Am Chem Soc
132(47):16774
–
16776.
4. Lacy DC, Roberts GM, Peters JC (2015) The cobalt hydride that never was: Revisiting
Schrauzer
’
s
“
hydridocobaloxime.
”
J Am Chem Soc
137(14):4860
–
4864.
5. Kölle U, Grätzel M (1987) Organometallic rhodium(III) complexes as homogeneous
catalysts for the photoreduction of protons to hydrogen on colloidal TiO
2
.
Angew
Chem Int Ed Engl
26(6):567
–
570.
6. Cosnier S, Deronzier A, Vlachopoul
os N (1989) Carbon/poly {pyrrole-[(C
5
Me
5
)Rh(bpy)Cl]
+
}
modified electrodes; a molecularly-bas
ed material for hydrogen evolution (bpy
=
2,2 ́-
bipyridine).
J Chem Soc Chem Commun
(17):1259
–
1261.
7. Blakemore JD, Gupta A, Warren JJ, Brunschwig BS, Gray HB (2013) Noncovalent im-
mobilization of electrocatalysts on carbon electrodes for fuel production.
J Am Chem
Soc
135(49):18288
–
18291.
8. Brintzinger H, Bercaw JE (1971) Bis(pentamethylcyclopentadienyl)titanium(II). Iso-
lation and reactions with hydrogen, nitrogen, and carbon monoxide.
J Am Chem Soc
93(8):2045
–
2046.
9. Ruppert R, Herrmann S, Steckhan E (1987) Efficient indirect electrochemical in situ
regeneration of NADH: Electrochemically driven enzymatic reduction of pyruvate
catalyzed by D-LDH.
Tetrahedron Lett
28(52):6583
–
6586.
10. Steckhan E, et al. (1991) Analytical study of a series of substituted (2,2
′
-bipyridyl)(penta-
methylcyclopentadienyl)rhodium and -iridium c
omplexes with regard to their effectiveness
as redox catalysts for the indirect electro
chemical and chemical reduction of NAD(P)
+
.
Organometallics
10(5):1568
–
1577.
11. Cosnier S, Gunther H (1991) A polypyrrole [Rh
III
C
5
Me
5
(bpy)Cl]
+
modified electrode for
the reduction of NAD
+
cofactor: Application to the enzymatic reduction of pyruvate.
J Electroanal Chem Interfacial Electrochem
315(1
–
2):307
–
312.
12. Kölle U, Fränzl H (2000) Oxidation of Alcohols by [Cp*Rh(ppy)(OH)]
+
.
Monatsh Chem
131(12):1321
–
1326.
13. Fukuzumi S, Kobayashi T, Suenobu T (2008) Efficient catalytic decomposition of for-
mic acid for the selective generation of H
2
and H/D exchange with a water-soluble
rhodium complex in aqueous solution.
ChemSusChem
1(10):827
–
834.
14. Blacker AJ, et al. (2009) Synthesis and structure of
“
16-electron
”
rhodium(III) catalysts
for transfer hydrogenation of a cyclic imine: Mechanistic implications.
Chem Commun
(Camb)
44:6801
–
6803.
15. Kölle U, Kang BS, Infelta P, Comte P, Grätzel M (1989) Electrochemical and pulse-
radiolytic reduction of (pentamethylcyclopentadienyl)(polypyridyl)rhodium com-
plexes.
Chem Ber
122(10):1869
–
1880.
16. Solis BH, Maher AG, Dogutan DK, Nocera DG, Hammes-Schiffer S (2016) Nickel phlorin
intermediate formed by proton-coupled electron transfer in hydrogen evolution
mechanism.
Proc Natl Acad Sci USA
113(3):485
–
492.
17. Hull JF, et al. (2012) Reversible hydrogen storage using CO2 and a proton-switchable
iridium catalyst in aqueous media under mild temperatures and pressures.
Nat Chem
4(5):383
–
388.
18. Lacy DC, McCrory CCL, Peters JC (2014) Studies of cobalt-mediated electrocatalytic
CO2 reduction using a redox-active ligand.
Inorg Chem
53(10):4980
–
4988.
19. Gusev OV, et al. (1993) Electrochemical generation of 19- and 20-electron rhodoce-
nium complexes and their properties.
J Organomet Chem
452(1
–
2):219
–
222.
20. Jones WD, Kuykendall VL, Selmeczy AD (1991) Ring migration reactions of (C
5
Me
5
)Rh(PMe
3
)H
2
.
Evidence for
η
3
slippage and metal-to-ring hydride migration.
Organometallics
10(5):
1577
–
1586.
Quintana et al.
PNAS
|
June 7, 2016
|
vol. 113
|
no. 23
|
6413
CHEMISTRY
21. Macias R, et al. (1997) Effects of metal-centre orbital control on cluster character and
electron distribution between borane and hydrocarbon ligands; significance of the
structures of [
μ
-9,10-(SMe)-8,8-(PPh
3
)
2
-nido-8,7-IrSB
9
H
9
]and[
μ
-9,10-(SMe)-8-
(
η
4
-C
5
Me
5
H)-nido-8,7-RhSB
9
H
9
].
J Chem Soc Dalton Trans
(2):149
–
152.
22. Zamorano A, Rendón N, Valpuesta JEV, Álvarez E, Carmona E (2015) Synthesis and
reactivity toward H
2
of (
η
5
-C
5
Me
5
)Rh(III) complexes with bulky aminopyridinate li-
gands.
Inorg Chem
54(13):6573
–
6581.
23. Enders M, et al. (2002) Coordination chemistry of neutral quinolyl- and amino-
phenylcyclopentadiene derivatives.
J Organomet Chem
641(1
–
2):81
–
89.
24. Tsai W-M, Rausch MD, Rogers RD (1996) Improved synthesis of pentabenzylcyclo-
pentadiene and study of the reaction between pentabenzylcyclopentadiene and iron
pentacarbonyl.
Organometallics
15(11):2591
–
2594.
25. Liu L-K, Luh L-S (1996) Consecutive butylations on the cyclopentadienyl ring of the
[(
η
5
-C
5
H
5
)Fe(CO)
2
PPh
3
]
+
cation.
Organometallics
15(25):5263
–
5265.
26. Blakemore JD, et al. (2014) Pentamethylcyclopentadienyl rhodium complexes.
Polyhedron
84:14
–
18.
27. Nakai H, Jeong K, Matsumoto T, Ogo S (2014) Catalytic C-F bond hydrogenolysis of
fluoroaromatics by.
Organometallics
33(17):4349
–
4352.
28. Appel AM, Helm ML (2014) Determining the overpotential for a molecular electro-
catalyst.
ACS Catal
4(2):630
–
633.
29. Gore-Randall E, Irwin M, Denning MS, Goicoechea JM (2009) Synthesis and charac-
terization of alkali-metal salts of 2,2
′
- and 2,4
′
-bipyridyl radicals and dianions.
Inorg
Chem
48(17):8304
–
8316.
30. Dadci L, et al. (1995)
π
-Arene aqua complexes of cobalt, rhodium, iridium, and ru-
thenium: Preparation, structure, and kinetics of water exchange and water sub-
stitution.
Inorg Chem
34(1):306
–
315.
31. Schumann H, Sühring K, Weimann R, Hummert M (2000) Deca(4-methylbenzyl)fer-
rocene and -stannocene.
Z Naturforsch B
60(6):527
–
532.
32. Ladwig M, Kaim W (1992) Electronic structure of catalytic intermediates for pro-
duction of H
2
:(C
5
Me
5
)Ir(bpy) and its conjugated acid.
J Organomet Chem
439(1):
79
–
90.
33. Robertson JJ, Kadziola A, Krause RA, Larsen S (1989) Preparation and characterization
of four- and five-coordinate rhodium(I) complexes. Crystal structures of chloro(2-
phenylazo)pyridine)(norbornadiene)rhodium(I), (2,2
′
-bipyridyl)(norbornadiene)rhodium(I)
chloride hydrate, and chloro(2,2
′
-bipyridyl)(norbornadiene)rhodium(I).
Inorg Chem
28(11):
2097
–
2102.
34. Reddy GKN, Susheelamma CH (1970) 2,2
’
Bipyridyl and o-phenanthroline carbonyl
complexes of rhodium.
J Chem Soc D
(1):54.
35. Roberts JAS, Bullock RM (2013) Direct determination of equilibrium potentials for
hydrogen oxidation/production by open circuit potential measurements in acetoni-
trile.
Inorg Chem
52(7):3823
–
3835.
36. Pitman CL, Finster ON, Miller AJ (March 7, 2016) Cyclopentadiene-mediated hydride
transfer from rhodium complexes.
Chem Commun
, 10.1039/C6CC00575F.
37. Kefalidis CE, et al. (2015) Can a pentamethylcyclopentadienyl ligand act as a proton-
relay in f-element chemistry? Insights from a joint experimental/theoretical study.
Dalton Trans
44(6):2575
–
2587.
38. Wilson AD, et al. (2006) Hydrogen oxidation and production using nickel-based mo-
lecular catalysts with positioned proton relays.
J Am Chem Soc
128(1):358
–
366.
39. Jacques P-A, Artero V, Pécaut J, Fontecave M (2009) Cobalt and nickel diimine-dioxime
complexes as molecular electrocatalysts for hydrogen evolution with low overvoltages.
Proc Natl Acad Sci USA
106(49):20627
–
20632.
40. Solis BH, et al. (2014) Theoretical analysis of cobalt hangman porphyrins: Ligand
dearomatization and mechanistic implications for hydrogen evolution.
ACS Catal
4(12):4516
–
4526.
41. Helm ML, Stewart MP, Bullock RM, DuBois MR, DuBois DL (2011) A synthetic nickel
electrocatalyst with a turnover frequency above 100,000 s
⁻
¹forH
₂
production.
Science
333(6044):863
–
866.
42. Solis BH, Yu Y, Hammes-Schiffer S (2013) Effects of ligand modification and pro-
tonation on metal oxime hydrogen evolution electrocatalysts.
Inorg Chem
52(12):
6994
–
6999.
43. Hu Y, et al. (2012) Synthesis, electrochemistry, and reactivity of new iridium(III) and
rhodium(III) hydrides.
Organometallics
31(14):5058
–
5064.
44. Hu Y, Norton JR (2014) Kinetics and thermodynamics of H
-
/H
•
/H
+
transfer from a
rhodium(III) hydride.
J Am Chem Soc
136(16):5938
–
5948.
45. Davies SG, Hibberd J, Simpson SJ, Thomas
SE, Watts O (1984) Car
bon monoxide reduction.
Carbonyl(
η
5
-cyclopentadienyl)hydrido[1,2-bis(di
phenylphosphino)ethane]iron: Reactions
and formation by reduction of the complex [(
η
5
-C
5
H
5
)Fe(Ph
2
PCH
2
CH
2
PPh
2
)(CO)]PF
6
.
J Chem Soc Dalton Trans
(4):701
–
709.
46. Reger DL, Belmore KA, Atwood JL, Hunter WE (1983) The cis addition of hydride
to
η
2
-alkyne complexes by initial reaction at an
η
5
-cyclopentadienyl (
η
5
-C
5
H
5
)
ring. Crystal and molecular structure of the carbonyl-
η
5
-cyclopentadienyliron
complex (
η
5
-C
5
H
5
)FeCO(PPh
3
).
J Am Chem Soc
105(17):5710
–
5711.
47. Bullock RM, Headford CEL, Kegley SE, Norton JR (1985) Hydrogen exchange between
the methyl and hydride ligands of dicyclopentadienylhydridomethyltungsten prior to
methane elimination.
J Am Chem Soc
107(3):727
–
729.
48. Khusnutdinova JR, Milstein D (2015) Metal-ligand cooperation.
Angew Chem Int Ed
Engl
54(42):12236
–
12273.
49. Zhang J, Leitus G, Ben-David Y, Milstein D (2005) Facile conversion of alcohols into
esters and dihydrogen catalyzed by new ruthenium complexes.
J Am Chem Soc
127(31):10840
–
10841.
50. Zhang J, Gandelman M, Shimon LJW, Rozenberg H, Milstein D (2004) Electron-rich,
bulky ruthenium PNP-type complexes. Accep
torless catalytic alcohol dehydrogenation.
Organometallics
23(17):4026
–
4033.
51. Ben-Ari E, Leitus G, Shimon LJW, Milstein D (2006) Metal-ligand cooperation in C-H
and H
2
activation by an electron-rich PNP Ir(I) system: Facile ligand dearomatization-
aromatization as key steps.
J Am Chem Soc
128(48):15390
–
15391.
52. Fulmer GR, et al. (2010) NMR chemical shifts of trace impurities: Common laboratory
solvents, organics, and gases in deuterated solvents relevant to the organometallic
chemist.
Organometallics
29(9):2176
–
2179.
53. Favier I, Duñach E (2004) New protic salts of aprotic polar solvents.
Tetrahedron Lett
45(17):3393
–
3395.
54. Fumino K, Fossog V, Wittler K, Hempelmann R, Ludwig R (2013) Dissecting anion-
cation interaction energies in protic ionic liquids.
Angew Chem Int Ed Engl
52(8):
2368
–
2372.
55. Sheldrick GM (1990) Phase annealing in SHELX-90: Direct methods for larger struc-
tures.
Acta Crystallogr A
46(6):467
–
473.
56. Sheldrick GM (2008) A short history of SHELX.
Acta Crystallogr A
64(Pt 1):112
–
122.
57. Müller P (2009) Practical suggestions for better crystal structures.
Crystallogr Rev
15(1):57
–
83.
58. Becke AD (1993) Density-functional thermochemistry. III. The role of exact exchange.
J Chem Phys
98(7):5648
–
5652.
59. Lee C, Yang W, Parr RG (1988) Development of the Colle-Salvetti correlation-energy
formula into a functional of the electron density.
Phys Rev B Condens Matter
37(2):
785
–
789.
60. Francl MM, et al. (1982) Self-consistent molecular orbital methods. XXIII. A polariza-
tion-type basis set for second-row elements.
J Chem Phys
77(7):3654
–
3665.
61. Hehre WJ, Ditchfield R, Pople JA (1972) Self-consistent molecular orbital methods. XII.
Further extensions of Gaussian-type basis sets for use in molecular orbital studies of
organic molecules.
J Chem Phys
56(5):2257
–
2261.
62. Hay PJ, Wadt WR (1985) Ab initio effective core potentials for molecular calculations.
Potentials for potassium to gold including the outermost core orbitals.
J Chem Phys
82(1):299
–
310.
63. Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group
thermochemistry, thermochemical kinetics, noncovalent interactions, excited states,
and transition elements: Two new functionals and systematic testing of four M06-
class functionals and 12 other functionals.
Theor Chem Acc
120(1-3):215
–
241.
64. Martin JML, Sundermann A (2001) Correlation consistent valence basis sets for use
with the Stuttgart-Dresden-Bonn relativistic effective core potentials: The atoms Ga-Kr
and In-Xe.
JChemPhys
114(8):3408
–
3420.
65. Bridgeman OC, Aldrich EW (1964) Vapor pressure tables for water.
J Heat Transfer
86(2):279
–
286.
66. Kelly CP, Cramer CJ, Truhlar DG (2007) Single-ion solvation free energies and the
normal hydrogen electrode potential in methanol, acetonitrile, and dimethyl sulf-
oxide.
J Phys Chem B
111(2):408
–
422.
67. Lam YC, Nielsen RJ, Gray HB, Goddard WA (2015) A Mn bipyrimidine catalyst pre-
dicted to reduce CO2 at lower overpotential.
ACS Catal
5(4):2521
–
2528.
68. Bochevarov AD, et al. (2013) Jaguar: A high-performance quantum chemistry soft-
ware program with strengths in life and materials sciences.
Int J Quantum Chem
113(18):2110
–
2142.
69. Zhao Y, Truhlar DG (2009) Benchmark energetic data in a model system for Grubbs II
Metathesis Catalysis and Their Use for the Development, Assessment, and Validation
of electronic structure methods.
J Chem Theory Comput
5(2):324
–
333.
70. Young KJH, et al. (2011) Synthesis of osmium and ruthenium complexes bearing di-
methyl (S,S)-2,2
′
-(pyridine-2,6-diyl)-bis-(4,5-dihydrooxazol-4-carboxylate) ligand and
application to catalytic H/D exchange.
J Mol Catal A: Chem
339(1-2):17
–
23.
6414
|
www.pnas.org/cgi/doi/10.1073/pnas.1606018113
Quintana et al.