This document is the Accepted Manuscript version of a Published Work that appeared in final form in Chemistry of
Materials, copyright © American Chemical Society after peer review and technical editing by the publisher. To acce
the final edited and published work see https://pubs.acs.org/doi/10.1021/acs.chemmater.9b02910
Article

Cite This: Chem. Mater. 2019, 31, 8875−8885

pubs.acs.org/cm

Computational Exploration of NO Single-Site Disproportionation on
Fe-MOF‑5
Jesús Jover,*,†,‡ Carl K. Brozek,§ Mircea Dincǎ,∥ and Núria López‡

Downloaded via UNIV ROVIRA I VIRGILI on November 27, 2019 at 09:06:26 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

†

Secció de Química Inorgànica, Departament de Química Inorgànica i Orgànica and Institut de Química Teòrica i Computacional
(IQTC-UB), Universitat de Barcelona, Martí i Franquès 1-11, 08028 Barcelona, Spain
‡
Institute of Chemical Research of Catalonia, ICIQ, The Barcelona Institute of Science and Technology, BIST, Av. Països Catalans
16, 43007 Tarragona, Spain
§
Department of Chemistry and Biochemistry, University of Oregon, Eugene, Oregon 97403, United States
∥
Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139,
United States
S
* Supporting Information

ABSTRACT: Nitric oxide disproportionation at the site-isolated Fe
centers of the metal organic framework material known as Fe-MOF-5
has been explored with density functional theory (DFT). The
computed reaction sequence supports the mechanism suggested by
experiment that involves the formation of the monoanionic hyponitrite
radical. The validity of the computed reaction mechanism is bolstered
by impressive agreement between computed and experimental
vibrational spectroscopic evidence of each reaction step. Similarly the
analogous MnII-MOF-5 system indicates that the disproportionation of
NO should proceed smoothly with this single-site material. These
results, observed also for some homogeneous Mn(II) catalysts, indicate
that heterogeneous Mn-based materials could be employed as efficient biological and industrial catalytic systems in NO
disproportionation processes.

■

Mn,17−20 Ni,21,22 Cu,23−28 Ru,29 Rh,30 and Pd.31 As molecular
species these examples react under homogeneous conditions,
whereas ideal deNO x catalysts should perform under
heterogeneous conditions, in which the catalytic species
would have improved activity, stability, and reusability. Very
recently, crystalline materials have been targeted as single-site
heterogeneous catalysts that take advantage of the specific
steric and electronic advantages of molecular homogeneous
systems.32−36 In 2015, the first single-site heterogeneous
material to promote disproportionation of NO into N2O and
NO2− was reported. It was based on an Fe(II)-exchanged
variant of the ionic material known MOF-5, termed Fe-MOF5. The reaction (Scheme 1) proceeded over the course of
minutes under NO atmosphere at room temperature and could
be monitored in situ by infrared spectroscopy, with each
observed intermediate easily isolated for characterization by
complementary techniques.15
The suite of experimental data supported the proposal of a
reaction mechanism that proceeds through an unusual
monoanionic hyponitrite radical (N2O2•−). Although this
species has been invoked in a variety of biological and

INTRODUCTION
Nitric oxide (NO) is one of the key intermediates formed in
the denitrification processes of the global nitrogen cycle. In this
natural cycle nitrogen is converted into different chemical
species as it circulates through atmospheric, marine, and
terrestrial ecosystems. The generation and decomposition of
NO features prominently in diverse processes ranging from
immune response to neurotransmission to blood pressure
control.1 The chemistry underlying these processes typically
involves electron transfer between NO and transition metal
centers.2,3 Some of these redox reactions occur at the active
sites of metalloenzymes such as nitric oxide (NOR) and
flavodiiron protein (FDP) reductases.4−9 Among these biological transformations, disproportionation of NO is particularly appealing because NO is one of the key components of
NxOy species that pollute the atmosphere.10 Reproducing NO
disproportionation in synthetic systems for widespread deployment on an industrial scale for deNOx catalysis has been a
longstanding challenge.11,12 Fundamental research toward this
goal has focused on using molecular transition metalcontaining complexes to disproportionate NO into NO2−
and N2O or N2 as a way to glean the precise structural and
electronic factors key to maximizing catalytic activity in largescale practical systems. Thus, far, transition metals reported to
participate in NO disproportionation include Fe, 13−16
© 2019 American Chemical Society

Received: July 22, 2019
Revised: October 21, 2019
Published: October 22, 2019
8875

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

industrial transformations, it had never been stabilized
indefinitely for further characterization.37,38 Most plausible
reaction mechanisms of NO disproportion involve the
formation of a hyponitrite species, but whereas one route
could involve spontaneous conversion of N2O2− to N2O and a
metal oxo, the evidence suggested an alternative route
involving attack of NO at the N2O2− to form N3O3− that
then released N2O, leaving behind Fe−NO2. The present
computational work is focused on exploring the full reaction
mechanism of NO disproportionation promoted by Fe-MOF5, including the Fe-bound NxOy species not observed by
vibrational spectroscopy. We proceed by determining the most
favorable reaction sequence and then comparing calculated
spectroscopic features to experiment. To expand the
applicability of these results, we explore the ability of other
MOF-5 variants including divalent cations such as Mn(II) and
Co(II) for activity toward disproportionation of NO.

Scheme 1. NO Disproportionation Reaction, Including the
Proposed Mechanism and Plausible Species in the Fe-MOF5 Catalyzed NO Disproportionationa

■

RESULTS AND DISCUSSION
Computed Reaction Mechanism. The computed reaction sequence for the NO disproportionation on Fe-MOF-5
is shown in Scheme 2. Fe-MOF-5 is a derivative of the parent

a

Adapted with permission from J. Am. Chem. Soc. 2015, 137, 7495−
7501. Copyright 2015 American Chemical Society.

Scheme 2. Computed Reaction Pathway for the Fe-MOF-5 Catalyzed NO Disproportionationa

a

Energies are given in eV, the number of unpaired electrons of each species is given between parentheses. Color code of the Fe core: C = gray, N =
blue, O = red, Fe = purple, Zn = green; atoms represented as balls correspond to those actively participating in the reaction steps.
8876

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

compound MOF-5 that can be prepared via postsynthetic
partial ion metathesis of 25% of Zn with Fe.39 The starting
matrix MOF-5 shows a cubic crystal structure with [Zn4O(BDC)3]8 stoichiometry (BDC = 1,4-benzenedicarboxylate).
In this structure, corners are built up of zinc oxide tetrahedral
units, [Zn4O]6+, connected to each other by BDC ligands. The
reaction pathway shown in Scheme 2 includes most of the
structures proposed in the original experimental report and
allows the rationalization of a great number of the experimental
observations (vide inf ra). The reaction starts by the
coordination of a NO molecule to the high-spin iron(II)
center in Fe-MOF-5 (intermediate I).
There are two different possibilities for this process to take
place, NO could bind onto either an axial or an equatorial
position. The latter is slightly more favored, with an overall
energy gain of 0.35 eV. Thus, in intermediate II the Fe atom
displays a square pyramid structure in which the incoming NO
lies cis to the central oxo group of the MOF core. The Fe−NO
distance in this intermediate is as short as 1.83 Å and the Fe−
N−O angle is 148.0°, quite far from a linear arrangement.
These parameters agree with the expected one-electron
transfer between iron and the incoming nitric oxide molecule,
and therefore intermediate II should be classified as a FeIII−
NO− species.40 The addition of a second NO produces the
dinitrosyl intermediate III, which is slightly higher in energy
than the previous intermediate. In this case the iron center
adopts an octahedral arrangement in which the incoming nitric
oxide occupies the free axial coordination site. The Fe−NO
distances in this structure are 2.10 and 2.00 Å, for the axial and
equatorial NO groups, respectively. This seems to indicate that
the coordination of the second NO ligand entails a certain
degree of electron density transfer from the iron, which in turn
produces the elongation of the previously formed Fe−NO
bond. The reduction of the equatorial Fe−N−O angle to
135.4° seems also to support this fact. Overall, species III
seems to have a delocalized unpaired electron between the
iron(III) center and both NO groups. Once III has been
obtained, the reaction proceeds by the formation of the
hyponitrite species; this step involves the coupling of both
coordinated nitrogen atoms and, presumably, happens through
transition state TS_I. In this transition state the axial NO
slightly slips from the metal to come close to the equatorial
NO group. The final configuration shows that the Fe−Nax
distance increases from 2.10 to 2.98 Å while the N−N distance
shortens to 1.96 Å. The Fe−Neq distance in this species is 1.94
Å, slightly shorter than that found in the previous intermediate.
The energy requirement for this stage is 0.69 eV, and the
overall barrier up to this point is 0.79 eV. Other alternative
transition states were sought, e.g. the triangular −Fe−N−N−
species in which both N atoms remain attached to the iron
center; but those showed higher energy requirements and were
consequently discarded. After the N−N bond formation an
equatorial N-bound hyponitrite species is obtained (IVN,
Scheme 3, top); nevertheless, this compound easily evolves
into the η1-terminal O-bound hyponitrite (IV), which is more
stable by 0.12 eV. The Fe−O distance in IV is 1.98 Å, while
the N−N distance is 1.28 Å. This value is slightly longer than
that found in dianionic ONNO2− species in which the double
NN bond is found between 1.20 and 1.25 Å.41 In contrast,
in IVN the Fe−N distance is 2.07 Å while the N−N distance is
as long as 1.73 Å, much closer to that found in a single bond.41
A third option consists of the formation of the 5-membered
ring intermediate IVC (Scheme 3, top), enabled by the

Scheme 3. Possible Isomeric Species of Compounds IV
(top) and VII (bottom)a

a

Energies are given in eV and the number of unpaired electrons of
each species is given between parentheses.

chelating ability of the hyponitrite substituent. Although the
computed energy for this compound seems competitive, this
possibility can be ruled out based on several observations. First
of all, the direct formation of IVC entails a cyclic transition
state in which both O atoms are connected to the iron atom
while the N−N bond is formed. The intermediate previous to
this transition state, where both NO units are bound onto the
iron center through the oxygen atoms, is higher in energy than
TS_I, and consequently this pathway should be ruled out.
Alternatively the Fe−ONNO− species (IVC) could be formed
by isomerization of the hyponitrite in compound IV. This
process would imply a rearrangement of the ONNO moiety
from trans to cis, which is blocked by the double bond nature
of the NN unit. A hypothetical transition state for this
process was computed with the O−N−N−O dihedral angle
fixed at 90°, and the relative energy value found for such
species was as high as 0.91 eV, producing a much higher
energy barrier of 1.26 eV.
After IV has been formed, the third NO substrate binds to
the iron, occupying the free axial position and forming
intermediate V. In this case the Fe−NOax distance is 2.30 Å
while the Fe−ONNOeq distance slightly increases to 2.07 Å.
The coordination of the NO moiety onto IV is energetically
favored, and intermediate V is the most stable species (−0.52
eV) along the reaction pathway prior to the formation of the
disproportionation products. Those species are obtained by the
reaction between the coordinated hyponitrite and nitric oxide
as shown in TS_II (Scheme 2). In order to react, the
hyponitrite moiety rotates to coordinate to the iron center
through one of the nitrogen atoms; this process takes the
opposite terminal oxygen atom close to the axial NO group,
where the attack takes place. In the transition state the Fe−Neq,
Fe−Nax, and ON−ONNO distances are 2.07, 2.25, and 1.92 Å,
respectively, while the distance of the breaking N−O bond
slightly lengthens (1.33 Å vs 1.23 Å in intermediate V). This
apparently complicated process requires only 0.54 eV to occur.
Again, other possibilities were taken into consideration as
alternatives to this transition state but none of them were
compatible with getting to the final products. Once TS_II has
been overcome, NO2− and N2O are formed, and the system
greatly stabilizes in energy (intermediate VI). N2O is very
weakly bound, the Fe−ON2 distance is quite long (3.04 Å),
8877

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

Fe(III)-oxo compound (VO) is thermodynamically favored.
Nevertheless, in the presence of excess NO the reaction should
happen preferentially by the pathway described above, and in
that case, the barrier for producing VO from V would rise to
0.86 eV. The addition of NO to VOif formedautomatically
produces the final Fe−NO2 (VII); all the attempts to compute
a new species with an oxo group and nitric oxide on the iron
center at the same time have ended up delivering this
elimination product.
Connection between Experiments and Calculations.
A vast amount of experimental data is available in the original
report on the disproportionation reaction of NO with FeMOF-5.15 A great deal of these data are focused on identifying
the species produced during the reaction, and after an
extensive analysis the reaction was postulated to proceed
through the sequence shown in Scheme 1, where B should be
the iron nitrosyl {FeNO}7 adduct, C is a η1-terminal iron(III)
hyponitrite species (FeIII−ONNO), and D corresponds to an
N-bound Fe(III) nitro species. In this section the experimental
results are combined with the DFT study in order to assess the
computed reaction mechanism. In order to do this, the
available experimental data will be compared to parameters
directly obtained from the proposed species along the reaction
pathway, i.e. DRIFTS vs computed IR spectra, and EXAFS
(Figure S1) vs computed structural features. A summary of the
comparable experimental and calculated data of the proposed
intermediates formed along the reaction can be found in Table
1.
The starting Fe-MOF-5 species (A) can be easily identified
as I. A was characterized as a species containing isolated highspin Fe(II) centers; the calculations indicate that the S = 2
ferrous material is almost 1.5 eV lower in energy that the
corresponding low-spin S = 0 analog. This should not be
surprising because all the first row transition metals in a
tetrahedral coordination environment present a small d-orbital
splitting.
Species B is a bit harder to determine. In the experimental
report this compound was assigned to be an iron nitrosyl
{FeNO}7 adduct; however, computationally this species could
be either the nitrosyl (II) or dinitrosyl (III) derivative. The
computed energies of these species are quite close, and thus,
they could be considered to coexist under the reaction
conditions. The experimental DRIFTS for species B shows the
ν(N−O) broad band centered at 1793 cm−1 with a small
shoulder at lower wavenumbers (ca. 1760 cm−1). The
computed IR spectrum of II shows only one band close to
that region (1778 cm−1), but compound III displays two close
bands, one medium band at 1714 cm−1 and a strong band at
1793 cm−1, corresponding to the coupled symmetric and
asymmetric ν(N−O) vibrational modes of the NO groups,
respectively. The overlap of those signals, which probably
suffer some broadening, could produce the finally observed
spectrum of species B.
Identification of C would probably be the most difficult one
to undertake in this study; however, the outstanding
experimental characterization allows the assignment of C as
the FeIII−ONNO species. This statement can be confirmed
through the calculation of the different hyponitrite species
shown above. The EXAFS experiment rules out the possibility
of the chelating FeIII(−ONNO−) hyponitrite (IVC) to be
species C because the iron center in this compound should
have a coordination number of 5, and the distance between the
metal and the fifth substituent should be close to 1.98 Å.

and thus it can be easily released without a remarkable change
in the energy. The liberation of N2O produces the high-spin
FeIII−NO2 species as the final product of the reaction (VII). In
this intermediate, the NO2 substituent migrates to the
energetically more favored equatorial position and the Fe−
NO2 distance becomes 2.13 Å. Alternatively, the chelating
FeIII-nitrito (VIIO, Scheme 3, bottom) compound could be
obtained; this intermediate is, however, higher in energy than
the nitro isomer and is probably not formed under the reaction
conditions. Finally, it has to be noted that the recovery of the
initial Fe-MOF-5 catalyst from species VII should be possible
with a relatively low energy investment of 0.52 eV; nevertheless, this possibility was not addressed in the experimental
report.
In the original NO disproportionation with Fe-MOF-5
report the authors propose the reaction is promoted by a
mononuclear Fe(II) species, as shown here. Other alternatives
could involve the participation of Zn centers to produce Zn−
NO or Fe/Zn−NxOy chelating species. While a complete
exploration of these pathways has not been performed, some of
these Zn-active species have been computed and found to
produce higher energies than those shown in Scheme 2. In the
same line, thepotentiallyreversible terephthalate partial
dissociation, previously described in similar MOF systems,42−44 was taken into account. However, the calculations
indicate this process would not be energetically competitive
with the proposed catalytic cycle, and thus, this possibility was
finally discarded. As stated above, IV might undergo an
alternative conversion of the bound hyponitrite into N2O and
an iron(III)-oxo species. This pathway has been explored
(Scheme 4), and based only on the computed energies it could
not be discarded. Expelling N2O from IV requires 0.65 eV
(TS_O), which does not seem a large enough barrier to
prevent this process, especially because the formation of the
Scheme 4. Proposed Reaction Pathways from Intermediate
IVa

a
Black line = pathway in Scheme 2; blue line = formation of iron-oxo
intermediate (energies in eV).

8878

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

Table 1. Summary of Experimental and Calculated Data
Experimental data
Species

DRIFTS bands (cm−1)

Calculated data

EXAFS Fe−(NxOy) (Å)

B
ν(N−O) 1793
C
D

ν(N−N) 1178
ν(N−O) 868
ν(N−O) 1220
δ(O−N−O) 797

1.98
2.08

IR bands (cm−1)
ν(N−O) 1778
ν(N−O)a 1793
ν(N−O)s 1714
ν(N−N) 1111
ν(N−O) 1012
ν(N−O)s 1246
δ(O−N−O) 758

Computed Fe−(NxOy) (Å)
1.83
2.00(eq)2.10(ax)
1.98
2.13

Assignment
Fe(NO) (II)
Fe(NO)2 (III)
Fe-ONNO (IV)
Fe-NO2 (VII)

confirm the identity of this intermediate, and the nitrite
ν(N−O)a, ν(N−O)s, and δ(O−N−O) bands appear at 1416, 1246, and
758 cm−1. While the vibrations of the MOF lattice obscure the
first one, the other two are pretty close to those obtained in the
DRIFTS study (1220 and 797 cm−1). In the case of species
VIIO the IR spectrum shows the following bands: ν(N−O)a 1230,
ν(N−O)s 1139, and δ(O−N−O) 836 cm−1, the second of which
does not appear in the experiments although it should be
observed based on the computed intensity.
The calculation of the IR spectra of the intermediates along
the reaction allows the building of a simulated DRIFTS
spectrum, which can be compared to that obtained in the
original experiments (Figure 1). In order to build the
computed DRIFTS only species II, IV, and VII have been
taken into account since they have been identified as those
observed in the experiments (see the Computational Details
section for a description of this procedure). The experimental
and simulated spectra are quite similar; at first, they both show
the ν(N−O) stretching of intermediate B/II at around 1790
cm−1. This signal decays as the formation of the hyponitrite
intermediate proceeds (C/IV), and the main ν(N−N) band of
this compound rises at around 1100 cm−1; this value is slightly
lower than that found in experiments, where the band is
centered at around 1180 cm−1. When the third NO reacts with
the hyponitrite intermediate, the ν(N−O) and δ(O−N−O) bands of
compound D/VII appear close to 1250 and 760 cm−1 while
the signal of the former intermediate recedes to finally
disappear. Of course, the agreement between experiments
and calculations is not perfect and some discrepancies can be
found. Some of them are due to the inherent differences
between the experimental and computed systems; for instance,
the bands obtained for the computed structures IV and VII
around 1400−1500 cm−1 are not observed in the experimental
DRIFTS because of the strong phenyl C−C absorption bands.
A similar effect is observed in the ν(C−O) bands of the
carboxylate groups, where the experimental signals appear at
higher wavenumbers for the experimental study (1600−1700
cm−1) when compared to the smaller computed system
(1550−1600 cm−1). Additionally, in the computed spectra
the experimental band at 785 cm−1, corresponding to the
formation of C/IV, cannot be observed, probably because the
computed intensity is too low. Finally, in the experimental
spectra there is a band assigned to compound D/VII located at
1000 cm−1 that cannot be interpreted with the calculations
because the only computed signal to appear in that region
corresponds to the ν(N−O) signal of compound C/IV, which
displays a very poor intensity.
Finally, the experimental data indicates that the rate constant
of the disproportionation reaction, which is carried out at room
temperature, should be lower than 0.01 s−1 for the ratedetermining step. This value, when translated to an energy

Obviously, these requirements cannot be met by the chelating
hyponitrite and thus the search is reduced to the other two
options: the O- or N-bound η1-hyponitrite isomers, i.e. the
computed species IV and IVN. The calculations above state
that IVN is formed in the first place during the reaction but it
should end up evolving into IV, which is energetically more
stable by 0.12 eV. These energy values are quite similar, and
thus other criteria should be employed to computationally
ascertain the nature of C. The Fe−O distance found in IVN is
2.07 Å while in compound IV it is found to be 1.98 Å, a perfect
match to the distance determined by EXAFS. In addition, the
IR spectrum of IV resembles that obtained in the DRIFTS
experiment.
For this compound four IR bands related to the hyponitrite
substituent are found at 1453, 1111, 1012, and 690 cm−1. The
first, hidden in the DRIFTS spectra by the MOF vibrations,
and the third bands correspond to the terminal and Fe-bound
ν(N−O) vibrations, respectively. The band found at 1111 cm−1 is
related to the ν(N−N) motion, and finally, the lowest energy
band, which shows a very low absorption intensity,
corresponds to a δ(ONNO) wagging normal mode. The presence
of the band at 1111 cm−1 seems to confirm intermediate IV as
the experimental species C. On the other hand, the computed
band at 1012 cm−1 shows a quite large shift toward higher
energies for the Fe-bound ν(N−O) vibrational mode when
compared to the experimental DRIFTS. The computed IR
spectra found for the other possible isomers of C show
different features that do not comply with those observed in
the experiments. For instance, compound IVN shows three
main bands related to the hyponitrite vibrations at 1606, 1563,
and 701 cm−1; these bands, especially the second one, are quite
strong and correspond to the ν(N−O)a, ν(N−O)s, and ν(N−N)
vibrations, respectively. The IR spectrum for this species lacks
the band observed experimentally at 1178 cm−1, and hence,
species IVN can safely be discarded. The chelating hyponitrite
species IVC shows three hyponitrite-related IR stretching
bands: ν(N−N), ν(N−Oax), and ν(N−Oeq) at 1505, 965, and 775
cm−1, respectively. Again, this intermediate does not show any
signal near the DRIFTS band at 1178 cm−1 and, consequently,
should be discarded.
EXAFS and Mössbauer (Figure S2) measurements indicate
that species D is a high-spin pentacoordinated FeIII−NO2
species. The calculations of the plausible Fe−NO2 (VII) and
Fe−ONO− (VIIO) species indicate the species D has to be
assigned to VII because it complies with all the experimental
evidence. VII is lower in energy than VIIO, which also shows a
wrong coordination number of 6. In VII, the Fe−N distance is
slightly longer than that obtained in the experiments: 2.13 and
2.08 Å, respectively. In the case of VIIO both Fe−O distances
present larger values of 2.17 (equatorial) and 2.20 (axial) Å,
respectively. The computed IR spectrum of VII also serves to
8879

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

possible to find a stable electronic/geometric solution; that is,
optimizations often ended up in dissociated species. Consequently, the cobalt-doped system was finally discarded. In
contrast, the Mn-MOF-5 platform provides robust electronic
state solutions and the calculations indicate that the
disproportionation of NO should be possible with this catalyst.
The computed energies for the Mn-MOF-5 system are shown
in Scheme 5 along with those obtained for Fe-MOF-5. The
computed magnetization values and the Bader charges for Mn
and the NxOy fragments can be found in Table S2. As may be
observed, the potential energy profile for Mn(II) shows the
same trend as that obtained for Fe(II), although in the case of
the former the binding of the first two NO molecules is less
favored; that is, the formation of intermediates II and III in the
Mn-MOF-5 system requires a small investment of energy.
Nevertheless, the transition state leading to the hyponitrite
intermediate with Mn is not very demanding, and thus the
barrier to furnish the N−N bond, computed as the energy
difference between Mn-MOF-5 + 2NO and TS_I, is 0.70 eV,
in contrast with the 0.79 eV computed for Fe-MOF-5. In this
case, the most stable hyponitrite species corresponds to
intermediate IVN, with an overall computed energy of −0.45
eV and a Mn−N distance of 2.22 Å. The terminal Mn-ONNO
(IV) compound is higher in energy (−0.14 eV) while the
chelating IVC intermediate is found at −0.36 eV. The
formation of this latter compound, by rotating around the
N−N bond, should be plausible because the N−N distance in
IVN (1.69 Å) seems closer to a single bond character. After the
obtention of IVN the reaction proceeds by addition of the third
NO substrate, which attaches onto the free axial position of
manganese, while the hyponitrite remains connected to the
metal through the N atom in an equatorial site (VN). In this
case the Mn−Neq and Mn−Nax are very similar, 2.22 and 2.24
Å, respectively. The formation of VN takes the energy slightly
down to a relative value of −0.47 eV, which corresponds to the
lowest energy value so far. This architecture has also been
computed for the Fe-MOF-5 system and found to lie at a
slightly higher energy value than intermediate V. Finally, the
reaction between the Mn-bound NO and hyponitrite
substituents affords N2O and the final MnIII−NO2 species
(VII). This step requires 0.50 eV to occur, which is a very
similar value to that found for the Fe-MOF-5 system, and
states that the overall barrier of the disproportionation process
corresponds to the formation of the hyponitrite intermediate
with a value of 0.70 eV. As in the case of Fe-MOF-5, the MnONO- (VIIO) alternative nitrito species was also computed
and found to be slightly higher in energy than the nitro isomer:
−2.39 vs −2.25 eV for VII and VIIO, respectively. In
conclusion, the DFT calculations on the NO disproportion
promoted by Mn-MOF-5 indicate that the reaction should be
possible and probably faster than with the Fe-MOF-5 analog.
This mechanism should also be considered as a serious
candidate to explain the NO disproportion promoted by
flexible Mn homogeneous systems such as those reported by
Lippard in the late 1990s.17

Figure 1. Experimental (top) and simulated (bottom) DRIFTS
spectra of Fe-MOF-5 in the presence of NO. Spectra begin with
species B/II (blue), which transforms into C/IV (red) and into D/
VII (gold). The computed bands of Fe-MOF-5 are shown in gray.
The experimental DRIFTS spectrum adapted with permission from J.
Am. Chem. Soc. 2015, 137, 7495−7501. Copyright 2015 American
Chemical Society.

value by means of the Eyring−Polanyi equation, affords a
minimum activation barrier of 0.87 eV. The computed energy
barrier for the overall reaction, which corresponds to the
energy difference between intermediate II and TS_I, is 0.79
eV, slightly lower but still close to the minimum value inferred
from the experimental data.
Extension to Other MII-Doped MOF-5 Systems. Once
the exploration of the NO disproportionation reaction with FeMOF-5 has been completed, the computed mechanism has
been extended, carrying out the same set of calculations, to
other transition metal-doped MOF-5 analogues. At first, two
more metals were taken into consideration: Mn(II) and
Co(II); and thus the reaction sequence shown in Scheme 2
was computed for both of them. There are experimental
reports, most of those involving mononuclear manganese
species,17−20 showing that manganese should be able to engage
in the NO disproportionation; on the other hand, cobalt has
not been reported before to participate in such a process.
Additionally, Mn-MOF-5 has been already prepared and
employed in catalytic reactions,39,45 and thus it should be
possible to test experimentally in case the computed results are
encouraging. Unfortunately, the DFT study of the Co-MOF-5
system could not be achieved because many times it was not

■

CONCLUSIONS
The reaction mechanism of the NO disproportionation on FeMOF-5 has been successfully characterized by computational
means. The study allows the identification of the species
formed along the reaction course, which are in very good
agreement with those proposed in the original experimental
report. The computed mechanism states that the reaction
8880

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

Scheme 5. Relative Energy Profiles for the Fe- and Mn-MOF-5 Catalyzed NO Disproportionation

a

a

M: Fe = black, Mn = red, energies in eV, unpaired electrons are given between parentheses.

low coordinate divalent iron and manganese flexible MOFs,
which should be able to engage in similar NO abatement
reactions and hence open new research opportunities in
deNOx chemistry.

proceeds by the binding of two NO substrates onto the initial
catalyst, the formation of a η1-ONNO hyponitrite intermediate, and the subsequent coordination of a third NO substrate
that reacts with the O-bound hyponitrite to produce N2O and
an FeIII−NO2 bound species. The simulated IR spectra
obtained from the computed geometries show a very good
agreement with the bands observed during the experimental
DRIFTS monitoring, confirming the identity of the species
formed during the reaction. The computed structural features
and the experimental EXAFS data are also in very good
agreement, which positively contributes in the identification of
Fe-bound hyponitrite and N-nitrito species.
Subsequently, the computed mechanism has been extended
to explore whether other first-row divalent transition metals
could be employed in Mn-MOF-5 systems to promote the NO
disproportionation. The calculations involving Co(II) did not
provide any useful information; on the other hand, the MnMOF-5 catalyst should be able to disproportionate NO. The
computed barrier found for this system is lower than that
found for the original Fe-MOF-5 material, indicating that the
manganese analog should produce a faster reaction rate even
though its NO adsorption properties are slightly poorer.
Overall, the flexibility of the MOF system, combined with
the ability of the transition metal to produce stable structures
with different coordination numbers, seems to be the key for
producing an NO disproportionation operative platform. This
kind of chemical behavior could probably be extended to other

■

COMPUTATIONAL DETAILS

All calculations have been carried out at the DFT+U level as
implemented in the VASP plane wave code, version 5.2.46,47 The
exchange-correlation functional developed by Perdew−Burke−
Ernzerhof (PBE),48 in combination with an effective U term
(Ueff)49 of 3.5 for Fe and 3.0 eV Co and Mn, respectively, was
employed in the calculations. These settings have already been
reported to correctly reproduce important properties such as binding
energies.50−52 Inner electrons were replaced by all-electron frozen
cores PAW,53,54 whereas valence electrons were expanded in plane
waves with a cutoff energy of 450 eV. Geometries were optimized at
the Γ point until the force threshold of 0.015 eV/Å was reached.
Long-range dispersion contributions to the energies were included via
the Grimme DFT-D2 correction.55 Different spin states have been
evaluated for most of the intermediates along the reaction pathways,
but only the most stable ones are shown along the main text; however,
in practically all the cases the calculations show that the most stable
compounds are those in which the metal stays in a high-spin
configuration, regardless of its coordination environment. Transition
states were located with the climbing image nudged elastic band
method and successively refined via standard procedures.56 Initially,
an analogous DFT methodology (PBE/PAW) was employed to
model the reaction mechanism; however, these settings were found to
produce results that were in clear disagreement with the experimental
data; that is, species such as Fe-NO2 (VII) were found to be more
8881

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

stable in their low spin form, while experimentally VII has been
characterized as a high spin N-bound nitrito compound. These errors
can be related to the nature of the GGA functional employed, and we
decided that using the DFT+U methodology could compensate this
behavior. The U parameter was subsequently optimized by comparing
the relative energies of the high spin Fe-NO2 (VII) and chelating FeONO- (VIIO) species in calculations ranging from 3.0 to 5.0 eV in 0.5
eV steps. The relative energy of these intermediates switches when the
U parameter increases; that is, at higher U values VIIO is favored
while lower U values produce a more stable VII species. Since VII is
observed experimentally, we decided to stop the optimization
procedure at a point fulfilling this condition but close to the crossing,
which happened at U = 3.5 eV.
The DFT+U methodology is one of the suitable approaches to
reduce the electronic self-interaction error in DFT studies57−63 and
has been largely applied by the solid state community as a mimic for
hybrid functionals.64−68 However, the DFT+U methodology is not
free of errors and the quality of electron localization/delocalization64,68,69 and the adequacy of the energetics/densities balance when
compared to more accurate methods69−72 have been called into
question. Another possible source of error in DFT + U calculations
can be the lack of static correlation treatment.69 In principle,
multireferential character could be expected to appear in the studied
compounds, mainly due to the partially filled d-orbitals of the metal
center and the radical character of NO, which could produce mixed
low-energy states. However, the computed species herein are too large
to carry out a complete multirefential study with e.g. CASSCF/
CASPT2 calculations. Multireferential calculations using the Orca73
package were attempted for describing the electronic structure of the
cluster analog of the Fe(NO)2 (III) compound (see below);
nevertheless, the complex structure of this system, where unpaired
electrons can be found on both the metal and the NO ligands,
provided unconvincing results even when trying different active
spaces. Monodeterminantal calculations on molecular clusters of
different species were also computed with Gaussian0974 using the
def2-tzvp basis set75 and the PBE,48 PBE0,48,76,77 and B3LYP78−80
density functionals. These calculations show that, in all cases, the iron
center remains in a high spin state and in the case of the hybrid
functionals by a similar energy margin (ca. 1 eV) to that found in the
DFT+U periodic calculations. Thus, multireference effects should not
be expected to have a strong impact in the calculations.
The lattice parameter for MOF-5 was computed to be 13.070 Å
with the aforementioned DFT+U methodology, in excellent agreement with the experimental value of 12.909 Å available from X-ray
crystallography.81 In this optimized cell, one Zn atom was replaced
with Fe, and the structure was fully reoptimized. The final computed
system contains two tetrahedral cores [FeZn3O(BDC)3][Zn4O(BDC)3] (BDC = 1,4-benzenedicarboxylate) and has a Fe:Zn ratio
of 1:8 (12.5%). The replacement of any other Zn center by Fe is not
necessary since the experimental magnetic susceptibility measurements indicate that Fe-MOF-5 features isolated S = 2 ferrous ions.15
The energies of the molecular species, namely NO, N2O, and NO2,
have been corrected to include the corresponding zero-point energy
and the translational entropy at 298 K. Both these values have been
obtained from calculations with Gaussian09,74 employing the PBE
functional48 along the 6-31G* basis set.82−84 Additionally, the cyclic
intermediates (namely IVC, VIIO, TS_I, and TS_II) have been
corrected to include the entropy loss upon cyclization, which amounts
to 0.06 eV per bond at 298 K.85
The simulated IR spectra of the different Fe-MOF-5−NxOy have
been obtained using the optimized plane wave calculations as starting
point. The corresponding iron clusters have been extracted for each
compound, and the phenyl rings have been replaced by methyl
groups; hence, in these calculations the Fe-MOF-5 core structure is
[FeZn3O(OAc)6]. Finally the modified carboxylate groups have been
allowed to relax in a partial optimization of the geometry where the Fe
atom and the NxOy units have been kept frozen. The IR spectra have
been obtained with Gaussian09,74 using the PBE functional48 along
with the def2-tzvp basis set,75 and the final DRIFTS spectra have been

constructed by representing the characteristic bands of the involved
species in increasing and decreasing percentages in each reaction step.
Calculating quantitatively precise partial charges and electron
distributions in solids, which may be helpful in this study, has been
demonstrated to be a difficult process.86 However, we were able to
compute the magnetization and Bader partial charges for all the
species along the reaction sequence in order to ascertain the electron
transfer processes between, mainly, the iron center and the NxOy
moieties (Table S1). As expected, these computed quantities do not
show a quantitative agreement with their expected magnetization
values or formal oxidation states, preventing their usage as a clear
criterion for describing the electron transfer processes in the studied
reaction. Nevertheless, the computed parameters in Table S1 show a
relatively good qualitative trend for the studied disproportionation
reaction; that is, when iron(II) reacts with NO an evaluating increase
in the computed Bader charge for Fe is observed while nitric oxide
adopts an overall negative charge. This same behavior is also observed
in computed Bader charges obtained from cluster calculations of
reaction intermediates; on the other hand, other different methods for
obtaining the atomic partial charges from molecular species
(Mulliken, APT, and NBO) show a much worse agreement with
the expected formal oxidation state for iron.

■

ASSOCIATED CONTENT

S
* Supporting Information

The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.chemmater.9b02910.
Computed energy terms, magnetization values, Bader
charges, and fractional coordinates of the optimized
species (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: jesus.jover@qi.ub.es.
ORCID

Jesús Jover: 0000-0003-3383-4573
Núria López: 0000-0001-9150-5941
Notes

The authors declare no competing financial interest.
The computational data can be freely accessed through the
ioChem-BD repository: DOI: 10.19061/iochem-bd-1-130.

■

ACKNOWLEDGMENTS
This research has been supported by the Spanish Ministerio de
́
Economia y Competitividad (MINECO, grant CTQ201568770-R). M.D. acknowledges support from the National
Science Foundation through a CAREER award (Award
DMR1452612).

■

REFERENCES

(1) Ignarro, L.; Freeman, B. Nitric Oxide: Biology and Pathobiology,
3rd ed.; Academic Press: San Diego, CA, 2017.
(2) Hayton, T. W.; Legzdins, P.; Sharp, W. B. Coordination and
Organometallic Chemistry of Metal−NO Complexes. Chem. Rev.
2002, 102, 935−992.
(3) Lee, D.-H.; Mondal, B.; Karlin, K. D., Nitrogen Monoxide and
Nitrous Oxide Binding and Reduction. In Activation of Small
Molecules: Organometallic and Bioiniroganic Perspectives; 2006; pp
43−79.
(4) Wasser, I. M.; de Vries, S.; Moënne-Loccoz, P.; Schröder, I.;
Karlin, K. D. Nitric Oxide in Biological Denitrification: Fe/Cu
Metalloenzyme and Metal Complex NOx Redox Chemistry. Chem.
Rev. 2002, 102, 1201−1234.
8882

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

(5) Silaghi-Dumitrescu, R.; Ng, K. Y.; Viswanathan, R.; Kurtz, D. M.
A Flavo-Diiron Protein from Desulfovibrio vulgaris with Oxidase and
Nitric Oxide Reductase Activities. Evidence for an in Vivo Nitric
Oxide Scavenging Function. Biochemistry 2005, 44, 3572−3579.
(6) Collman, J. P.; Yang, Y.; Dey, A.; Decréau, R. A.; Ghosh, S.;
Ohta, T.; Solomon, E. I. A functional nitric oxide reductase model.
Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 15660.
(7) Hayashi, T.; Caranto, J. D.; Wampler, D. A.; Kurtz, D. M.;
Moënne-Loccoz, P. Insights into the Nitric Oxide Reductase
Mechanism of Flavodiiron Proteins from a Flavin-Free Enzyme.
Biochemistry 2010, 49, 7040−7049.
(8) Wang, J.; Schopfer, M. P.; Puiu, S. C.; Sarjeant, A. A. N.; Karlin,
K. D. Reductive Coupling of Nitrogen Monoxide (•NO) Facilitated
by Heme/Copper Complexes. Inorg. Chem. 2010, 49, 1404−1419.
(9) Hayashi, T.; Caranto, J. D.; Matsumura, H.; Kurtz, D. M.;
Moënne-Loccoz, P. Vibrational Analysis of Mononitrosyl Complexes
in Hemerythrin and Flavodiiron Proteins: Relevance to Detoxifying
NO Reductase. J. Am. Chem. Soc. 2012, 134, 6878−6884.
(10) Lelieveld, J.; Crutzen, P. J. Role of Deep Cloud Convection in
the Ozone Budget of the Troposphere. Science 1994, 264, 1759.
(11) Kim, M. H.; Nam, I.-S. In Catalysis; Spivey, J. J., Ed.; The Royal
Society of Chemistry: Cambridge, 2005; Vol. 18.
(12) Past and present in DeNOx catalysis. From molecular modelling to
chemical engineering; Granger, P., Pârvulescu, V. I., Eds.; Elsevier:
Oxford, 2007; Vol. 171.
(13) Franz, K. J.; Lippard, S. J. NO Disproportionation Reactivity of
Fe Tropocoronand Complexes. J. Am. Chem. Soc. 1999, 121, 10504−
10512.
(14) Patra, A. K.; Rose, M. J.; Olmstead, M. M.; Mascharak, P. K.
Reactions of Nitric Oxide with a Low-Spin Fe(III) Center Ligated to
a Tetradentate Dicarboxamide N4 Ligand: Parallels between Heme
and Non-heme Systems. J. Am. Chem. Soc. 2004, 126, 4780−4781.
(15) Brozek, C. K.; Miller, J. T.; Stoian, S. A.; Dincǎ, M. NO
Disproportionation at a Mononuclear Site-Isolated Fe2+ Center in
Fe2+-MOF-5. J. Am. Chem. Soc. 2015, 137, 7495−7501.
(16) Confer, A. M.; McQuilken, A. C.; Matsumura, H.; MoënneLoccoz, P.; Goldberg, D. P. A Nonheme, High-Spin {FeNO}8
Complex that Spontaneously Generates N2O. J. Am. Chem. Soc.
2017, 139, 10621−10624.
(17) Franz, K. J.; Lippard, S. J. Disproportionation of Nitric Oxide
Promoted by a Mn Tropocoronand. J. Am. Chem. Soc. 1998, 120,
9034−9040.
(18) Martirosyan, G. G.; Azizyan, A. S.; Kurtikyan, T. S.; Ford, P. C.
Low temperature NO disproportionation by Mn porphyrin.
Spectroscopic characterization of the unstable nitrosyl nitrito complex
MnIII(TPP)(NO)(ONO). Chem. Commun. 2004, 1488−1489.
(19) Filipović, M. R.; Duerr, K.; Mojović, M.; Simeunović, V.;
Zimmermann, R.; Niketić, V.; Ivanović-Burmazović, I. NO Dismutase
Activity of Seven-Coordinate Manganese(II) Pentaazamacrocyclic
Complexes. Angew. Chem., Int. Ed. 2008, 47, 8735−8739.
(20) Jiang, M.-X.; Liu, C.-G. Reduction of NO to N2O Catalyzed by
a Mn-Substituted Keggin-Type Polyoxometalate: A Density Functional Theory Study. J. Phys. Chem. C 2017, 121, 12735−12744.
(21) Wright, A. M.; Wu, G.; Hayton, T. W. Formation of N2O from
a Nickel Nitrosyl: Isolation of the cis-[N2O2]2− Intermediate. J. Am.
Chem. Soc. 2012, 134, 9930−9933.
(22) Wright, A. M.; Zaman, H. T.; Wu, G.; Hayton, T. W.
Mechanistic Insights into the Formation of N2O by a Nickel Nitrosyl
Complex. Inorg. Chem. 2014, 53, 3108−3116.
(23) Ruggiero, C. E.; Carrier, S. M.; Tolman, W. B. Reductive
Disproportionation of NO Mediated by Copper Complexes:
Modeling N2O Generation by Copper Proteins and Heterogeneous
Catalysts. Angew. Chem., Int. Ed. Engl. 1994, 33, 895−897.
(24) Schneider, J. L.; Carrier, S. M.; Ruggiero, C. E.; Young, V. G.;
Tolman, W. B. Influences of Ligand Environment on the
Spectroscopic Properties and Disproportionation Reactivity of
Copper−Nitrosyl Complexes. J. Am. Chem. Soc. 1998, 120, 11408−
11418.

(25) Shimokawabe, M.; Okumura, K.; Ono, H.; Takezawa, N. N2O
and NO2 Formation in the Disproportionation of NO over Ion
Exchanged Cu-ZSM-5 at Lower Temperature. React. Kinet. Catal. Lett.
2001, 73, 267−274.
(26) Lionetti, D.; de Ruiter, G.; Agapie, T. A trans-Hyponitrite
Intermediate in the Reductive Coupling and Deoxygenation of Nitric
Oxide by a Tricopper−Lewis Acid Complex. J. Am. Chem. Soc. 2016,
138, 5008−5011.
(27) Metz, S. N2O Formation via Reductive Disproportionation of
NO by Mononuclear Copper Complexes: A Mechanistic DFT Study.
Inorg. Chem. 2017, 56, 3820−3833.
(28) Wijeratne, G. B.; Hematian, S.; Siegler, M. A.; Karlin, K. D.
Copper(I)/NO(g) Reductive Coupling Producing a trans-Hyponitrite Bridged Dicopper(II) Complex: Redox Reversal Giving Copper(I)/NO(g) Disproportionation. J. Am. Chem. Soc. 2017, 139, 13276−
13279.
(29) Azizyan, A. S.; Kurtikyan, T. S.; Martirosyan, G. G.; Ford, P. C.
Tracking Reactive Intermediates by FTIR Monitoring of Reactions in
Low-Temperature Sublimed Solids: Nitric Oxide Disproportionation
Mediated by Ruthenium(II) Carbonyl Porphyrin Ru(TPP)(CO).
Inorg. Chem. 2013, 52, 5201−5205.
(30) Gaviglio, C.; Pellegrino, J.; Milstein, D.; Doctorovich, F. NȮ
disproportionation by a {RhNO}9 pincer-type complex. Dalton Trans.
2017, 46, 16878−16884.
(31) Stromnova, T. A.; Shishilov, O. N.; Churakov, A. V.; Kuz’mina,
L. G.; Howard, J. A. K. NO-disproportionation, promoted by Pdcluster: formation and X-ray structure of Pd8(μ-CO)4(μ-OOCCMe3)8[μ-N(O)O−]4. Chem. Commun. 2007, 4800−4802.
(32) Rogge, S. M. J.; Bavykina, A.; Hajek, J.; Garcia, H.; OlivosSuarez, A. I.; Sepúlveda-Escribano, A.; Vimont, A.; Clet, G.; Bazin, P.;
Kapteijn, F.; Daturi, M.; Ramos-Fernandez, E. V.; Llabrés i Xamena,
F. X.; Van Speybroeck, V.; Gascon, J. Metal−organic and covalent
organic frameworks as single-site catalysts. Chem. Soc. Rev. 2017, 46,
3134−3184.
(33) Chen, Z.; Vorobyeva, E.; Mitchell, S.; Fako, E.; Ortuño, M. A.;
López, N.; Collins, S. M.; Midgley, P. A.; Richard, S.; Vilé, G.; PérezRamírez, J. A heterogeneous single-atom palladium catalyst surpassing
homogeneous systems for Suzuki coupling. Nat. Nanotechnol. 2018,
13, 702−707.
(34) Cui, X.; Li, W.; Ryabchuk, P.; Junge, K.; Beller, M. Bridging
homogeneous and heterogeneous catalysis by heterogeneous singlemetal-site catalysts. Nat. Catal. 2018, 1, 385−397.
(35) Li, F.; Chen, Z. Cu dimer anchored on C2N monolayer: lowcost and efficient Bi-atom catalyst for CO oxidation. Nanoscale 2018,
10, 15696−15705.
(36) Tian, S.; Wang, Z.; Gong, W.; Chen, W.; Feng, Q.; Xu, Q.;
Chen, C.; Chen, C.; Peng, Q.; Gu, L.; Zhao, H.; Hu, P.; Wang, D.; Li,
Y. Temperature-Controlled Selectivity of Hydrogenation and Hydrodeoxygenation in the Conversion of Biomass Molecule by the Ru1/
mpg-C3N4 Catalyst. J. Am. Chem. Soc. 2018, 140, 11161−11164.
(37) Poskrebyshev, G. A.; Shafirovich, V.; Lymar, S. V. Hyponitrite
Radical, a Stable Adduct of Nitric Oxide and Nitroxyl. J. Am. Chem.
Soc. 2004, 126, 891−899.
(38) Halliwell, B.; Gutteridge, M. C. Free radicals in biology &
medicine, 5th ed.; Oxford University Press: Oxford, 2015.
(39) Brozek, C. K.; Dincǎ, M. Ti3+-, V2+/3+-, Cr2+/3+-, Mn2+-, and
Fe2+-Substituted MOF-5 and Redox Reactivity in Cr- and Fe-MOF-5.
J. Am. Chem. Soc. 2013, 135, 12886−12891.
(40) Enemark, J. H.; Feltham, R. D. Principles of structure, bonding,
and reactivity for metal nitrosyl complexes. Coord. Chem. Rev. 1974,
13, 339−406.
(41) Wright, A. M.; Hayton, T. W. Understanding the Role of
Hyponitrite in Nitric Oxide Reduction. Inorg. Chem. 2015, 54, 9330−
9341.
(42) Brozek, C. K.; Michaelis, V. K.; Ong, T.-C.; Bellarosa, L.;
López, N.; Griffin, R. G.; Dincǎ, M. Dynamic DMF Binding in MOF5 Enables the Formation of Metastable Cobalt-Substituted MOF-5
Analogues. ACS Cent. Sci. 2015, 1, 252−260.
8883

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

(43) Osadchii, D. Y.; Olivos-Suarez, A. I.; Szécsényi, Á .; Li, G.;
Nasalevich, M. A.; Dugulan, I. A.; Crespo, P. S.; Hensen, E. J. M.;
Veber, S. L.; Fedin, M. V.; Sankar, G.; Pidko, E. A.; Gascon, J. Isolated
Fe Sites in Metal Organic Frameworks Catalyze the Direct
Conversion of Methane to Methanol. ACS Catal. 2018, 8, 5542−
5548.
(44) Szécsényi, Á .; Li, G.; Gascon, J.; Pidko, E. A. Unraveling
reaction networks behind the catalytic oxidation of methane with
H2O2 over a mixed-metal MIL-53(Al,Fe) MOF catalyst. Chem. Sci.
2018, 9, 6765−6773.
(45) Stubbs, A. W.; Braglia, L.; Borfecchia, E.; Meyer, R. J.; RománLeshkov, Y.; Lamberti, C.; Dincǎ, M. Selective Catalytic Olefin
Epoxidation with MnII-Exchanged MOF-5. ACS Catal. 2018, 8, 596−
601.
(46) Kresse, G.; Hafner, J. ABINITIO MOLECULAR-DYNAMICS
FOR LIQUID-METALS. Phys. Rev. B: Condens. Matter Mater. Phys.
1993, 47, 558−561.
(47) Kresse, G.; Furthmuller, J. Efficient iterative schemes for ab
initio total-energy calculations using a plane-wave basis set. Phys. Rev.
B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186.
(48) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient
approximation made simple. Phys. Rev. Lett. 1996, 77, 3865−3868.
(49) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.;
Sutton, A. P. Electron-energy-loss spectra and the structural stability
of nickel oxide: An LSDA+U study. Phys. Rev. B: Condens. Matter
Mater. Phys. 1998, 57, 1505−1509.
(50) Patton, D. C.; Pederson, M. R. Application of the generalizedgradient approximation to rare-gas dimers. Phys. Rev. A: At., Mol., Opt.
Phys. 1997, 56, R2495−R2498.
(51) Zhang, Y.; Pan, W.; Yang, W. Describing van der Waals
Interaction in diatomic molecules with generalized gradient
approximations: The role of the exchange functional. J. Chem. Phys.
1997, 107, 7921−7925.
(52) Capdevila-Cortada, M.; Łodziana, Z.; López, N. Performance of
DFT+U Approaches in the Study of Catalytic Materials. ACS Catal.
2016, 6, 8370−8379.
(53) Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B:
Condens. Matter Mater. Phys. 1994, 50, 17953−17979.
(54) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the
projector augmented-wave method. Phys. Rev. B: Condens. Matter
Mater. Phys. 1999, 59, 1758−1775.
(55) Grimme, S. Semiempirical GGA-type density functional
constructed with a long-range dispersion correction. J. Comput.
Chem. 2006, 27, 1787−1799.
(56) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A climbing image
nudged elastic band method for finding saddle points and minimum
energy paths. J. Chem. Phys. 2000, 113, 9901−9904.
(57) Mori-Sánchez, P.; Cohen, A. J.; Yang, W. Many-electron selfinteraction error in approximate density functionals. J. Chem. Phys.
2006, 125, 201102.
(58) Ruzsinszky, A.; Perdew, J. P.; Csonka, G. I.; Vydrov, O. A.;
Scuseria, G. E. Density functionals that are one- and two- are not
always many-electron self-interaction-free, as shown for H2+, He2+,
LiH+, and Ne2+. J. Chem. Phys. 2007, 126, 104102.
(59) Cohen, A. J.; Mori-Sánchez, P.; Yang, W. Insights into Current
Limitations of Density Functional Theory. Science 2008, 321, 792.
(60) Haunschild, R.; Henderson, T. M.; Jiménez-Hoyos, C. A.;
Scuseria, G. E. Many-electron self-interaction and spin polarization
errors in local hybrid density functionals. J. Chem. Phys. 2010, 133,
134116.
(61) Zheng, X.; Liu, M.; Johnson, E. R.; Contreras-García, J.; Yang,
W. Delocalization error of density-functional approximations: A
distinct manifestation in hydrogen molecular chains. J. Chem. Phys.
2012, 137, 214106.
(62) Kim, M.-C.; Sim, E.; Burke, K. Understanding and Reducing
Errors in Density Functional Calculations. Phys. Rev. Lett. 2013, 111,
No. 073003.

(63) Schmidt, T.; Kümmel, S. One- and many-electron selfinteraction error in local and global hybrid functionals. Phys. Rev. B:
Condens. Matter Mater. Phys. 2016, 93, 165120.
(64) Gani, T. Z. H.; Kulik, H. J. Where Does the Density Localize?
Convergent Behavior for Global Hybrids, Range Separation, and DFT
+U. J. Chem. Theory Comput. 2016, 12, 5931−5945.
(65) Verma, P.; Truhlar, D. G. Does DFT+U mimic hybrid density
functionals? Theor. Chem. Acc. 2016, 135, 182.
(66) Gani, T. Z. H.; Kulik, H. J. Unifying Exchange Sensitivity in
Transition-Metal Spin-State Ordering and Catalysis through Bond
Valence Metrics. J. Chem. Theory Comput. 2017, 13, 5443−5457.
(67) Janet, J. P.; Zhao, Q.; Ioannidis, E. I.; Kulik, H. J. Density
functional theory for modelling large molecular adsorbate−surface
interactions: a mini-review and worked example. Mol. Simul. 2017, 43,
327−345.
(68) Zhao, Q.; Kulik, H. J. Where Does the Density Localize in the
Solid State? Divergent Behavior for Hybrids and DFT+U. J. Chem.
Theory Comput. 2018, 14, 670−683.
(69) Gaggioli, C. A.; Stoneburner, S. J.; Cramer, C. J.; Gagliardi, L.
Beyond Density Functional Theory: The Multiconfigurational
Approach To Model Heterogeneous Catalysis. ACS Catal. 2019, 9,
8481−8502 and references therein. .
(70) He, Y.; Gräfenstein, J.; Kraka, E.; Cremer, D. What correlation
effects are covered by density functional theory? Mol. Phys. 2000, 98,
1639−1658.
(71) Bochevarov, A. D.; Friesner, R. A. The densities produced by
the density functional theory: Comparison to full configuration
interaction. J. Chem. Phys. 2008, 128, No. 034102.
(72) Medvedev, M. G.; Bushmarinov, I. S.; Sun, J.; Perdew, J. P.;
Lyssenko, K. A. Density functional theory is straying from the path
toward the exact functional. Science 2017, 355, 49.
(73) Neese, F. The ORCA program system. WIREs Comput. Mol. Sci.
2012, 2, 73−78.
(74) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.;
Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.;
Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.;
Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi,
R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar,
S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox,
J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;
Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A.
D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.
Gaussian09, Revision D.01; Gaussian, Inc.: Wallingford CT, 2009.
(75) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence,
triple zeta valence and quadruple zeta valence quality for H to Rn:
Design an assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7,
3297−3305.
(76) Perdew, J. P.; Burke, K.; Ernzerhof, M. Errata: Generalized
gradient approximation made simple. Phys. Rev. Lett. 1997, 78, 1396.
(77) Adamo, C.; Barone, V. Toward reliable density functional
methods without adjustable parameters: The PBE0 model. J. Chem.
Phys. 1999, 110, 6158−6169.
(78) Lee, C.; Yang, W.; Parr, R. G. Development of the ColleSalvetti correlation-energy formula into a functional of the electron
density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789.
(79) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Results obtained
with the correlation-energy density functionals of Becke and Lee,
Yang and Parr. Chem. Phys. Lett. 1989, 157, 200−206.
(80) Becke, A. D. Density-functional thermochemistry. III. The role
of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652.
(81) Turner, S.; Lebedev, O. I.; Schroeder, F.; Esken, D.; Fischer, R.
A.; Van Tendeloo, G. Direct imaging of loaded metal-organic
8884

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885

Chemistry of Materials

Article

framework materials (metal@MOF-5). Chem. Mater. 2008, 20, 5622−
5627.
(82) Ditchfield, R.; Hehre, W. J.; Pople, J. A. Self-Consistent
Molecular-Orbital Methods. IX. An Extended Gaussian-Type Basis for
Molecular-Orbital Studies of Organic Molecules. J. Chem. Phys. 1971,
54, 724−728.
(83) Hehre, W. J.; Ditchfield, R.; Pople, J. A. 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. 1972, 56, 2257−2261.
(84) Hariharan, P. C.; Pople, J. A. The influence of polarization
functions on molecular orbital hydrogenation energies. Theor. Chem.
Acc. 1973, 28, 213−222.
(85) Page, M. I.; Jencks, W. P. Entropic Contributions to Rate
Accelerations in Enzymic and Intramolecular Reactions and the
Chelate Effect. Proc. Natl. Acad. Sci. U. S. A. 1971, 68, 1678.
(86) Walsh, A.; Sokol, A. A.; Buckeridge, J.; Scanlon, D. O.; Catlow,
C. R. A. Oxidation states and ionicity. Nat. Mater. 2018, 17, 958−964.

8885

DOI: 10.1021/acs.chemmater.9b02910
Chem. Mater. 2019, 31, 8875−8885