Ruthenium Metal–Organic Frameworks with Different Defect Types: Influence on Porosity, Sorption, and Catalytic Properties

Article abstract of DOI:10.1002/chem.201602641

By employing the mixed-component, solid-solution approach, various functionalized ditopic isophthalate (ip) defect-generating linkers denoted 5-X-ipH₂, where X=OH (1), H (2), NH₂(3), Br (4), were introduced into the mixed-valent ruthenium analogue of [Cu₃(btc)₂]_n(HKUST-1, btc=benzene-1,3,5-tricarboxylate) to yield Ru-DEMOFs (defect-engineered metal–organic frameworks) of the general empirical formula [Ru₃(btc)_2?x(5-X-ip)_xY_y]_n. Framework incorporation of 5-X-ip was confirmed by powder XRD, FTIR spectroscopy, ultrahigh-vacuum IR spectroscopy, thermogravimetric analysis,¹H NMR spectroscopy, N₂sorption, and X-ray absorption near edge structure. Interestingly, Ru-DEMOF 1 c with 32 % framework incorporation of 5-OH-ip shows the highest BET surface area (≈1300 m²g^?1, N₂adsorption, 77 K) among all materials (including the parent framework [Ru₃(btc)₂Y_y]_n). The characterization data are consistent with two kinds of structural defects induced by framework incorporation of 5-X-ip: modified paddlewheel nodes featuring reduced ruthenium sites (Ru^δ+, 0<δ<2, type A) and missing nodes leading to enhanced porosity (type B). Their relative abundances depend on the choice of the functional group X in the defect linkers. Defects A and B also appeared to play a key role in sorption of small molecules (i.e., CO₂, CO, H₂) and the catalytic properties of the materials (i.e., ethylene dimerization and the Paal–Knorr reaction).

Full text of DOI:10.1002/chem.201602641

DOI: 10.1002/chem.201602641
Full Paper
&
Metal–Organic Frameworks
Ruthenium Metal–Organic Frameworks with Different Defect
Types: Influence on Porosity, Sorption, and Catalytic Properties
[
a]
[a]
[a]
[b]
[c]
Wenhua Zhang, Max Kauer, Olesia Halbherr, Konstantin Epp, Penghu Guo,
[d]
[d]
[a]
[e]
Miguel I. Gonzalez, Dianne J. Xiao, Christian Wiktor, Francesc X. LIabrꢀs i Xamena,
Christof Wçll, Yuemin Wang,* Martin Muhler, and Roland A. Fischer*
[f]
[f]
[c]
[b]
2
ꢀ1
Abstract: By employing the mixed-component, solid-solu-
tion approach, various functionalized ditopic isophthalate
highest BET surface area (ꢁ1300 m g , N adsorption, 77 K)
2
among all materials (including the parent framework
[Ru (btc) Y ] ). The characterization data are consistent with
(
ip) defect-generating linkers denoted 5-X-ipH , where X=
2
3
2 y n
OH (1), H (2), NH2 (3), Br (4), were introduced into the
two kinds of structural defects induced by framework incor-
poration of 5-X-ip: modified paddlewheel nodes featuring
mixed-valent ruthenium analogue of [Cu (btc) ] (HKUST-1,
3
2 n
d+
btc=benzene-1,3,5-tricarboxylate) to yield Ru-DEMOFs
defect-engineered metal–organic frameworks) of the gener-
al empirical formula [Ru (btc) (5-X-ip) Y ] . Framework in-
reduced ruthenium sites (Ru , 0<d<2, type A) and miss-
(
ing nodes leading to enhanced porosity (type B). Their rela-
tive abundances depend on the choice of the functional
group X in the defect linkers. Defects A and B also appeared
to play a key role in sorption of small molecules (i.e., CO2,
3
2ꢀx
x y n
corporation of 5-X-ip was confirmed by powder XRD, FTIR
spectroscopy, ultrahigh-vacuum IR spectroscopy, thermogra-
1
vimetric analysis, H NMR spectroscopy, N sorption, and X-
CO, H ) and the catalytic properties of the materials (i.e., eth-
2
2
ray absorption near edge structure. Interestingly, Ru-DEMOF
ylene dimerization and the Paal–Knorr reaction).
1
c with 32% framework incorporation of 5-OH-ip shows the
[
6]
Introduction
tant role in the material properties. The diversity of MOF
structures and the associated issues of intrinsic and intentional
structural defects, nonstoichiometry, and heterogeneity sug-
gest a much broader approach for understanding and tailoring
Metal–organic frameworks (MOFs), constructed from metal ion
nodes and organic linkers, have been studied in a number of
[
1]
[2]
[3]
[7]
fields including gas sorption, gas separation, catalysis,
their chemical and physical properties. Ravon and co-workers
[
4]
[5]
sensing, and drug delivery. Recently, it has been recognized
that MOFs, similar to other (crystalline) solid-state materials,
are typically not perfect crystals with infinite periodic repeti-
tion or ordering of identical groups of atoms in space. In fact,
defects of various types and length scales (even more general-
ly: structural heterogeneity) cannot be rigorously avoided
during synthesis and crystal growth and often play an impor-
reported Zn-MOFs featuring defects as acid catalytic centers
(generated by fast precipitation or partial linker substitution)
that are responsible for shape selectivity in alkylation of large
[8]
aromatic molecules. Vermoortele et al. demonstrated genera-
tion of defects in UiO-66(Zr) by the coordination modulation
[9]
approach. As a result, better catalytic performance in several
Lewis acid catalyzed reactions was observed for the defect-en-
[a] W. Zhang, M. Kauer, Dr. O. Halbherr, Dr. C. Wiktor
[e] Dr. F. X. LIabrꢁs i Xamena
Chair of Inorganic Chemistry II
Ruhr University Bochum
Instituto deTecnologꢂa Quꢂmica (ITQ)
Universitat Politꢃcnica de Valꢃncia
Universitꢀtsstrasse 150, 44801 Bochum (Germany)
Consejo Superior de Investigaciones Cientꢂficas
Avenida de los Naranjos s/n, 46022 Valencia (Spain)
[
b] K. Epp, Prof. Dr. R. A. Fischer
Chair of Inorganic and Metal-Organic Chemistry
Technical University Munich
Lichtenbergstrasse 4, 85748 Garching (Germany)
E-mail: roland.fischer@tum.de
[f] Prof. Dr. C. Wçll, Dr. Y. Wang
Institute of Functional Interfaces (IFG)
Karlsruhe Institute of Technology (KIT), Karlsruhe (Germany)
E-mail: yuemin.wang@kit.edu
[c] P. Guo, Prof. Dr. M. Muhler
Supporting information for this article can be found under
Laboratory of Industrial Chemistry
Ruhr University Bochum
http://dx.doi.org/10.1002/chem.201602641. It contains detailed experimen-
tal protocols, elemental analysis, PXRD, TGA, NMR, IR, EDX, XANES, XPS,
Universitꢀtsstrasse 150, 44801 Bochum (Germany)
2 2 2
UHV-IR, and gas-sorption data (N , CO, H , CO ).
[
d] M. I. Gonzalez, D. J. Xiao
Department of Chemistry
University of California
Berkeley 94720 California (USA)
Chem. Eur. J. 2016, 22, 1 – 12
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
gineered MOF (DEMOF). Recently, a direct correlation was
found to exist between the concentration of missing-linker de-
fects in UiO-66 and its catalytic activity for the acid-catalyzed
[
10]
esterification of levulinic acid with alcohols. Controlled intro-
duction and characterization of specific defects into MOFs is
quite a challenge in comparison with the parent (more or less)
“
defect-free” reference materials.
Cu (btc) ] (also known as HKUST-1) as well as the iso-
[
11]
12]
[
3
2 n
[
[12b,13]
[12b,14]
structural family [M (btc) ] (M=Mo,
Cr,
Ni,
3
2 n
[
12b,15]
[12b,16]
Ru,
Zn;
btc=benzene-1,3,5-tricarboxylate) has at-
tracted considerable recent attention, mainly due to coordina-
tively unsaturated metal sites (CUS), which could be readily
available after appropriate post-synthetic treatment. Intention-
al defect generation in [Cu (btc) ] was initially reported by
3
2 n
Baiker et al., who studied partial replacement of btc by the ho-
mologous pyridine-3,5-dicarboxylate (pydc) linker and revealed
enhanced catalytic activity of the modified MOF in the oxida-
tion of benzene derivatives related to modified coordinatively
Figure 1. Design concept of Ru-DEMOFs by the mixed-component/frag-
mented-linker solid-solution approach and sketch of the two most impor-
tant possible paddlewheel-related defects.
[
17]
unsaturated sites (MCUS) at the nodes.
Subsequently, we
have demonstrated that not only pydc, but also a series of
other defect-generating linkers, can be incorporated in the
[
7,20]
gands Y.
In defects of type B, whole paddlewheel nodes
[
18]
[
Cu (btc) ] framework.
Interestingly, functionalized meso-
are eliminated from the structure, as has been demonstrated
3
2 n
[19]
pores appeared to be generated along with MCUS in such
DEMOFs. In parallel, Hupp et al. gave an account of enhanced
in the case of ip introduction into [Cu (btc) ] . It is conceiva-
3 2 n
ble that such missing-node defects could be formed in a corre-
lated fashion during the crystal growth process of the MOF,
and in such a case larger regions of missing nodes may be
generated to yield mesopores. Understanding of the nature
and formation of different types of isolated local and larger-
[19]
porosity by introducing isophthalate (ip) into [Cu (btc) ] .
3
2 n
The concept of defect engineering by a mixed-component ap-
proach using fragmented linkers and parent linkers was also
[
20]
expanded to the isostructural Ru analogue.
Ru-DEMOFs
ꢀ
ꢀ
ꢀ
[
Ru (btc) (pydc) Y ] (Y=counterion=Cl , OH , OAc , etc.;
scale correlated defect sites in DEMOFs of the [M (btc) ] family
3 2 n
3
2ꢀx
x
y n
0
<yꢂ1.5) were obtained by direct mixing of H pydc and
and MOFs in general is very limited, and this research is still in
2
[7,18–20]
H btc in one-step solvothermal synthesis. The generation of re-
its infancy.
Herein, we follow up our previous investiga-
3
d+
duced Ru (0<d<2) sites at the paddlewheel nodes as a con-
tions on defect engineering in [Ru (btc) Y ] using H pydc as
3
2
y n
2
[
20]
sequence of defect linker incorporation was correlated with
enhanced CO uptake and improved catalytic activity in olefin
hydrogenation. Remarkably, low-temperature (100 K) dissocia-
(single-) defect-generating linker.
Similar to our previous
[18]
more elaborate study on defect-engineered Cu-HKUST-1, we
now selected a series of 5-X-isophthalic acids (5-X-ipH , X=OH,
2
tive chemisorption of CO leading to coordinated CO was ob-
H, NH , Br) and mixed them with H btc to yield Ru-DEMOFs of
2
2
3
served by ultrahigh-vacuum (UHV) FTIR spectroscopy in the
dark. However, regardless of the obviously more advanced
properties compared with the parent material, the actual com-
positions of these Ru-DEMOFs are very difficult to establish
and are not entirely clear. This stems from the structural com-
plexity of such multicomponent solid-solution MOFs and
a range of possible compositional variations that must be
taken into account, which are not mutually exclusive (counter-
ions, residual synthesis components such as linkers, coordina-
tion modulators, etc.).
the general formula [Ru (btc) (5-X-ip) Y ] (Figure 1). A
3 2ꢀx x y n
number of complementary characterization methods were ap-
plied to obtain compositional and microstructure data of the
materials. The aim was to at least qualitatively assess the ten-
dency of type A and type B defect formation as a function of
synthetic conditions and the chosen 5-X-ip linker. Several cata-
lytic reactions, such as ethylene dimerization and the Paal–
Knorr reaction (usually taking place under Lewis-acidic condi-
tions) as well as low-temperature CO !CO dissociative chemi-
2
sorption in the dark under UHV conditions, were chosen as
test reactions to monitor the influence of defects on the per-
formance of the Ru-DEMOFs.
Figure 1 summarizes the concept of defect introduction into
mixed-valent [Ru (btc) Y ] by using divalent isophthalate link-
3
2 y n
ers as fragmented variants of trivalent btc. Mainly, two types of
defects can be distinguished: Type A defects are structurally
and electronically modified paddlewheel units (missing partial
carboxylate ligator function), in which (at least) one out of the
four bridging carboxyl groups of the parent btc linker in the
regular paddlewheels has been partly substituted by the (neu-
tral) functional group of the defect linker. Consequently, for
charge compensation, the metal sites can be partly reduced
Results and Discussion
In the following, we provide a comprehensive account on the
conducted experiments and obtained analytical and characteri-
zation data for elucidation of the defect structure of Ru-
DEMOFs of the general formula [Ru (btc) (5-X-ip) Y ] . The
3
2ꢀx
x y n
main goal of this study was to substantiate the hypothesis of
(mixed-valent state) or carry additional anionic (monovalent) li-
type A and type B defect formation in these systems
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
(
Figure 1). Due to the large amount of information collected
in which the incorporation of 5-OH-ip stays the same (or
a little less) than the feed concentration of the starting reac-
tants. This result might be attributable to the slightly different
coordination equilibria correlated with the Brønsted acid disso-
ciation constants of the used linkers: pK (H btc)=3.12, 3.89,
on the systems, only the most significant data have been se-
lected for display in figures and tables in the main text. To
save space for concise discussion, most of the systematic data
documentation is given in the Supporting Information.
a
3
[22]
4
.70 and pK (H ip)=3.46, 4.46. However, the pK differences
a
2
a
are small and a conclusive explanation based on pK is not
a
Synthesis and characterization of Ru-DEMOFs 1a–d, 2a–d,
straightforward. Dinc a˘ et al. argued that deprotonated btc
could serve as a counterion to provide the charge balance of
3
a–d, and 4a–c
5
+
[12b]
All Ru-DEMOFs were obtained according to the previously
[Ru2] units in the Ru analogue of Cu-HKUST-1.
From our
[
20]
published procedure, by using [Ru (OOCCH ) Cl] , H btc, and
earlier investigation we concluded that H btc is present in the
2
3 4
n
3
3
defect-generating 5-X-ipH as starting materials in acetic acid/
pores of [Ru (btc) Y ] in the form of fully protonated species,
2
3
2 y n
[23]
water solution under solvothermal conditions. The numbering
scheme is based on the framework-incorporated 5-X-ip linker:
rather than as deprotonated counterions. Both studies indi-
cate that the linker used for the synthesis could still be present
in the obtained materials. Hence, it is conceivable that the
X=OH (1), H (2), NH (3), or Br (4). Further labeling (a, b, c,
2
etc.) refers to various degrees of defect-linker incorporation.
The powder (P) XRD patterns of the as-synthesized and activat-
ed samples indicated that all materials (except 2d) are crystal-
line and isostructural with the parent Ru-MOF that was ob-
defect-generating H ip linker can also reside as a counterion or
2
as a guest molecule in the pores of Ru-DEMOFs 2a and 2b.
On the basis of these arguments, it can be understood why
the amount of incorporated ip can be higher than expected
from the H ip:H btc feed ratio. Indeed, very weak IR bands in
tained by employing the same method with H btc alone (Sup-
3
2
3
2
+/3+
porting Information, Figure S1 and S2). Since Ru
ions can
the range of n(C=O), n(C=C), and n (COO) vibrations (1736–
as
ꢀ
1
be easily reduced to eventually yield ruthenium metal, the sol-
vothermal synthesis conditions must be carefully monitored.
However, no PXRD peaks in the 2q range of 40–458 were
found, neither for the as-synthesized nor for the activated sam-
ples, and this suggests the absence of a substantial amount of
1521 cm ) were observed for 2a–2d. However, the small sep-
aration of the n(C=O) bands of uncoordinated acetic acid
ꢀ
1 [24]
(1715–1640 cm ),
protonated/uncoordinated
H ip
2
ꢀ
1
ꢀ1
(1670 cm ), and H btc (1680 cm ; see Figure S8 in the Sup-
3
porting Information) makes it hard to unambiguously deter-
mine the exact origin of these weak bands.
0
Ru nanoparticles (Ru NPs). Reflections at 42.1 and 448 are usu-
1
ally assigned to (001) and (101) faces of hexagonal close-
In all H NMR spectra of the digested Ru-DEMOF samples
[
21]
packed metallic Ru particles or Ru NPs. We describe further
characterization, by X-ray absorption near edge structure
a resonance at about 1.91 ppm, originating from the CH3
groups of acetate, was detected. This is in line with our previ-
ous investigations of the parent single-linker Ru-MOFs ob-
tained from various Ru precursors and the Ru-DEMOFs with
pydc defect linker, for which the same resonance due to ace-
tate was always observed in the digested samples with
a molar ratio n(AcO):n(btc) in the range from 0.7 to 1.8 as
(XANES) in particular, below. The corresponding results provide
additional evidence for the absence of Ru/RuO NPs in the ob-
x
tained Ru-DEMOFs.
With the exception of 2d, thermogravimetric analysis (TGA)
showed no decomposition of the prepared materials for tem-
peratures below 2508C, and this suggests that the Ru-DEMOFs
preserve the thermal stability of the parent MOF (Supporting
Information, Figure S3). The incorporation and quantification of
defect linkers was initially performed by NMR spectroscopy of
the digested samples after activation and assessment of poros-
[20,23]
well.
We assume three main origins of acetate in the dis-
cussed Ru-MOFs: 1) it may act as a counterion to compensate
the overall charge of the framework (Supporting Information,
Figure S7a); 2) owing to the competition with btc and 5-X-ip, it
could be residual/unsubstituted and paddle-wheel-coordinated
acetate groups of the Ru precursor [Ru (OOCCH ) Cl] (Support-
1
ity (Supporting Information, Table S2). Thus, in the H NMR
2
3 2
n
spectra of 1a–d, the appearance of two additional peaks (com-
pared with the parent Ru-MOF) at 7.51 and 7.89 ppm is clearly
seen (Supporting Information, Figure S4). These two resonan-
ces can be assigned to the aromatic protons of 5-OH-ip linkers.
As expected, on increasing the initial amount of defect linker
ing Information, Figure S7b); 3) it may reside as AcOH in the
pores of the framework through weak interactions and there-
fore cannot be easily removed by washing and subsequent
thermal activation (i.e., heating at 1508C under dynamic
vacuum). Note that the glacial acetic acid/water mixture
(
5-OH-ipH ), the intensities of its characteristic resonances also
(AcOH/H O 0.05:1) was found to be essential for successful
2
2
increase in the solution spectrum of a digested sample (Sup-
porting Information Figure S4). A similar scenario was also ob-
served for the materials with ip defect linker (series 2a–c). Res-
onances corresponding to the protons of ip are seen at 8.4,
synthesis and was used for all reported Ru-(DE)MOFs (see Ex-
perimental Section). Besides, the employed linkers may also
serve as guest molecules, as was mentioned above. Attempts
to evaluate the presence of acetic acid and linkers from the IR
8.1, and 7.6 ppm (Supporting Information, Figure S5). The
spectra is not quite feasible due to the overlapping n (COO)
as
degree of framework incorporation of ip in 2a–c increases
with increasing concentration of the defect linker in the reac-
tant solutions. However, the incorporated amount of ip was
found to be higher than expected from the H btc:H ip feed
and n (COO) bands (for bidentate bridging and free carboxyl-
s
[25]
ate groups) and low intensity of the respective bands. Al-
though one can monitor acetate groups IR-spectroscopically
ꢀ
by employing F -labeled acetate (e.g., perfluoroacetate instead
3
2
[26]
ratio, which is in contrast with the case of Ru-DEMOF series 1,
of acetate) for synthesis, the crucial synthesis parameters of
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[
15a,20]
Ru-MOFs make this method impossible. Earlier efforts to use
various acids as reaction solvents failed to give the crystalline
linker and other mixed-linker Ru-DEMOFs reported so far).
This indicates that 5-OH-ip is a good candidate for defect engi-
neering in Ru-MOFs. For comparison of the obtained SBET
values for Ru-MOFs with those of the homologous Cu-MOFs, it
[
23]
parent Ru-MOF.
Consequently, the presence of acetate
(mainly) and unconsumed free linkers in Ru-DEMOFs raises
2
ꢀ1
questions about the actual sample compositions and local
structures, and thus causes more complications with ambigu-
ous interpretation of the analytical data than in case of the ho-
may be more informative to use m mmol units, due to the
substantial difference in the atomic weights of Cu and Ru.
Thus, as can be seen from Table S6 of the Supporting Informa-
[
18]
2
ꢀ1
mologous Cu-DEMOFs. The compositions of the Ru-DEMOFs
discussed in this work were derived by taking into account the
complete set of experimental results, and a full account is
given in the Supporting Information (Tables S3 and S4). As
quantitative digestion of 3a–d is not sufficient, detection and
tion, the SBET value of the parent Ru-MOF (964 m mmol ) is
quite similar to that of typical reference material [Cu (btc) ]
3
2 n
2
ꢀ1
(1049 m mmol ). More importantly, the S values of the Ru-
BET
DEMOFs are slightly higher than those of both parent ana-
logues.
determination of the defect-linker contents (5-NH -ip) from
In contrast to Ru-DEMOF series 1, the integration of the ip
defect linker generally results in lower porosities, as demon-
strated by the results for 2a–2d. However, the SBET values of
2a and 2b are still higher than that of the parent Ru-MOF.
Nonetheless, with increasing ip content, the surface area de-
creases. Thus, with regard to porosity, the optimum degree of
incorporation of defect linkers in this series is 28% (2b). In the
2
1
H NMR measurements was not possible. Nevertheless, the IR
spectra of these samples provide qualitative evidence for 5-
NH -ip incorporation. Indeed, the bands observed at 1651 and
2
ꢀ
1
1
264 cm stem from d(NꢀH) and n(CꢀN) vibrations of 5-NH -
2
ip (Supporting Information, Figure S8). In case of the 5-Br-ip
(
4a–c), the gradual increase in intensity of the respective sig-
1
nals in both H NMR and TEM-EDX (EDX=energy-dispersive X-
ray spectroscopy) spectra suggests the presence of defect link-
ers in these Ru-DEMOFs (see Figures S6 and S9 in the Support-
ing Information).
case of 5-NH -ip defect linkers, the porosity is fully maintained
2
in Ru-DEMOFs 3a–c, whereas increasing the feed content to
50% leads to a decrease in porosity instead (3d, S
781 m g ). These observations on the 2a–d and 3a–d series
=
BET
2
ꢀ1
To further support incorporation of the defect linker in the
framework and rule out its extra-framework inclusion in the
are probably due to the accommodation of unconverted 5-X-
ipH in the pores. With increasing feed concentration of 5-X-
2
pores in a large quantity, we carried out N sorption measure-
ipH , the competition between btc and 5-X-ip increases. Thus,
2
2
ments on the activated materials. All materials (1a–d, 2a–d,
it is more difficult to rule out the presence of unconverted 5-X-
3
a–d, and 4a–c) show type I N sorption isotherms at 77 K
ipH as the reason for the decreasing S . Along the same line
2
2
BET
without any hysteresis loop (Supporting Information Fig-
ure S10), which indicate that all Ru-DEMOFs are microporous
of reasoning, the S values of 4a–c dramatically decreased
BET
ꢀ1
2
from 996 to 693 m g . Hence, optimized incorporation of 5-
Br-ip can only reach 17% (4a). Notably, in our previously stud-
ies on defect-engineered Cu-HKUST-1, the generation of meso-
[
27]
materials. In general, by comparing the BET surface areas S
BET
2
ꢀ1
with that of the parent Ru-MOF (998 m g ), one can consider
Ru-DEMOFs to be analogous if the SBET value is in/above the
range of that of the parent MOF. In other words, the considera-
bly high SBET values of the discussed Ru-DEMOFs rule out sub-
stantial guest inclusion and pore blocking, for example, by un-
[18]
pores was observed. Curiously, in the present case of the ho-
mologous Ru-DEMOFs, we notice an opposite trend that may
be attributed to the preference for isolated defects rather than
correlated defects in the framework (Supporting Information,
Figure S11).
converted defect linkers, acetate, or Ru/RuO NPs. Interestingly,
x
when 5-OH-ipH is used as defect linker, S of the Ru-DEMOFs
2
BET
gradually increases until 5-OH-ip is incorporated up to 32%
1c) (Figure 2). Notably, the SBET value of 1c is the highest
among all [Ru (L) Y ] isostructural MOFs (including the single-
XANES, XPS, and UHV-FTIR studies: variation of ruthenium
oxidation state as an indication of type of defect formation
(
3
2 y n
In our previous studies on doping of the Ru-MOF with pydc
defect linkers, we assigned the introduced local defects, ac-
[17]
cording to Baiker et al., as modified Ru paddlewheels with
three btc and one pydc linker, in which the pyridyl N site acts
as a weakly binding ligator site toward the Ru dimer. Such
structural irregularities are associated with (partially) missing
linker functionality (i.e., carboxylate) and consequently steric
and electronic modification of the Ru secondary building units
(
SBUs; see Figure 1). Additionally, partial ruthenium reduction
2
+
3+
was observed, that is, apart from the expected Ru and Ru
as in the parent Ru MOF), we also observed distinctly reduced
(
d+
framework Ru species (Ru , 0<d<2) in the pydc-doped Ru-
[20]
DEMOFs. Since pydc is a divalent (anionic) linker with one
less charge than btc, charge compensation by reduction of
some Ru ions may take place (in the absence or presence of
other charge-compensating effects such as accommodation of
Figure 2. Variation of SBET derived from measured N
77 K) with increasing degree of defect-linker doping. “Parent” stands for
parent single-linker Ru-MOF (with only btc).
2
sorption isotherms
(
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
additional Y anions). We also reported significant formation of
frameworks. Further information on the local environment of
Ru can be obtained from extended X-ray absorption fine struc-
ture (EXAFS). The analysis of the experimental data is, however,
a formidable challenge, since many different models for Ru-
DEMOFs must be considered. Therefore, a thorough analysis of
the EXAFS data is outside the scope of this paper.
2
+
+
mixed-valent Cu /Cu paddlewheels for the related pydc-
[
18]
doped Cu-HKUST-1. Interestingly, in parallel studies on iso-
structural Cu-HKUST-1, doped with ip (rather than pydc and
other 5-X-ip), predominantly defects B (see Figure 1) and no
significant change of the copper oxidation state were report-
[
19]
ed. Thus, the choice of the defect-generating linker in com-
bination with the type of metal and the synthetic conditions
The different materials fabricated in this study were also
characterized by high-resolution X-ray photoelectron spectros-
copy (XPS) to confirm the assignments based on XANES and
for a more quantitative estimation of the relative abundance
(doping level) most likely defines the types of defects induced
in a particular type of MOF structure. Therefore, we were inter-
ested in investigating the oxidation state of the Ru centers in
3
+
2+
d+
of the different Ru species (Ru , Ru , and Ru ) in the Ru-
DEMOFs. The MOFs 1a, 1c, 1d, 2a, and 2b with 5-OH-ip and
ip defect linkers, respectively, were selected. Remarkably, simi-
lar to the related pydc-doped Ru-DEMOFs we elaborated earli-
the series with 5-X-ip defect linkers (X=OH, H, NH , and Br) in
2
[20]
comparison with our previous findings for the pydc case,
and to further confirm the assignment of the defect types.
[
20]
d+
The XANES spectra (Supporting Information, Figure S12)
show deviation of the edge-jump derivatives of 1a–d from
that of the parent Ru-MOF. For 1a–c, the gradual shift of posi-
tion of the edge jump to lower energies suggests reduced Ru
sites with oxidation states lower than 2+ or 3+. We relate
this observation to the generation of defects A in 1a–c. How-
ever, in the case of 1d, the position of the edge jump moved
again to higher energies, that is, the amount of reduced Ru
species is now decreased with respect to 1a–c. The different
trend suggests that defects of type B may also have been cre-
ated when 5-OH-ip linkers were incorporated into the frame-
work in fractions higher than 32%. The same trend was ob-
er, reduced Ru species were found in Ru-DEMOFs 1a, 1c,
2
+
3+
and 1d in addition to the Ru and Ru species. Two Ru 3d
doublets (3d5/2 and 3d3/2) at about 281.6 and 285.8 eV as well
2
+
3+
as at 282.6 and 286.8 eV, attributed to the Ru and Ru spe-
cies, respectively, were observed in the deconvoluted XP spec-
tra of all studied samples (Figure 3). In addition, another Ru 3d
doublet appears at about 280.5 and 284.7 eV in the spectra of
Ru-DEMOFs 1a, 1c, and 1d. On the basis of quantitative esti-
3
+
2+
mations of the XP spectra, some variation of the Ru /Ru
d+
ratio and, in particular, reduction to Ru in 1a, 1c, and 1d
can be clearly seen (Supporting Information, Table S6). From
3
+
2+
d+
the parent Ru-MOF to the Ru-DEMOF 1a, the Ru /Ru /Ru
served for the 5-NH -ip linker (3a–c). Indeed, the XANES spec-
2
d+
tra of 3a,b suggest again the presence of reduced Ru (0<
d<2) sites, but the oxidation state of the Ru centers in 3c is
the same as in the parent Ru-MOF (Supporting Information,
Figure S14). Furthermore, when the 5-Br-ipH defect linker is
2
employed, the oxidation state of Ru varies as well, in particular
for 4a (Supporting Information, Figure S15). These data sup-
port the assignment of locally modified node defects A for
these samples. Interestingly, the position of the edge jump in
the case of the ip-doped materials 2a–b does not change
compared with that of the parent Ru-MOF (Supporting Infor-
mation, Figure S13). The Ru oxidation state is not affected by
the incorporation of ip, and this could be taken as an indica-
tion of a predominance of defects of type B for 2a,b. When
the functional group of the defect linkers changes from small,
coordinatively inert H to the somewhat larger and coordina-
tively more suitable ligator sites such as OH, NH , and Br, the
2
defects of type A are favored at relatively low degrees of
doping. However, with increasing degree of doping, increasing
creation of defects B besides some remaining defects A results
in less pronounced variation of the oxidation state of Ru in the
defect-engineered samples compared with the parent material.
The curves of the XANES spectra (normalized absorption vs.
energy, Figure S16 of the Supporting Information) of all the
Ru-DEMOFs and the parent Ru-MOF are strikingly different
from those of Ru NPs, Ru metal, and RuO . These observations
2
support the PXRD results (see above and Figures S1 and S2 of
the Supporting Information) and also allow one to rule out the
presence of other Ru phases as significant impurities in the Ru-
Figure 3. Deconvoluted XP spectra of the Ru-DEMOFs A) 1a (8% 5-OH-ip),
1
c (32% 5-OH-ip), and 1d (37% of 5-OH-ip) and B) 2a (15% ip) and 2b
28% ip) in the Ru 3d and C 1s regions in comparison with the parent Ru-
MOF.
[
21]
DEMOFs including the parent Ru-MOF. Thus, we assume that
it is valid to assign all Ru species to the metal nodes of the
(
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
ratio varies from 1/1.24/0 to 1/1.75/0.08, and this indicates that
the incorporation of defect linkers induces partial ruthenium
In our earlier report on pydc-doped Ru-DEMOFs, we report-
ed on low-temperature CO !CO dissociative chemisorption
2
2
+
d+
reduction (i.e., relative concentrations of Ru and Ru in-
crease). Moreover, on increasing the degree of incorporation of
under UHV conditions in a dark environment, a feature which
[20]
is not inherent to the parent Ru-MOF. We concluded that
d+
5
-OH-ip to 32% for 1c (feed level 30%), the amount of re-
CO !CO reduction might be triggered by the reduced Ru
2
d+
duced Ru
value among all studied samples (Ru /Ru /Ru =1/1.69/
species considerably increases to the highest
sites and the pyridyl N sites in the proximity of the modified
Ru centers at the nodes. To further confirm these assumptions,
we studied the same reaction with Ru-DEMOFs that feature re-
3
+
2+
d+
d+
0
.59). Importantly, the Ru -related doublet decreases signifi-
cantly in intensity with further increase of the incorporation to
7% for 1d (feed level 50%). This is in contrast to the observa-
d+
duced Ru sites. As these sites were created by the incorpora-
3
tion of the 5-OH-ip linkers in 1a–1d, we selected 1c and 1d
tion for pydc-doped Ru-DEMOFs, for which a continuous rise
as two representative samples that both exhibit high degrees
d+
2+
of the Ru signals was detected with increasing degree of
of 5-OH-ip incorporation but quite different contents of Ru
[
20]
2+
d+
pydc incorporation. The lower concentrations of Ru and
and Ru . As shown in Figure 4, on CO dosing for both the
2
d+
3+
2+
d+
Ru in 1d (Ru /Ru /Ru =1:1.23:0.09) compared to 1c can
be explained by our assumption of simultaneous generation of
both defects A and B with distinct preference for B over A at
higher feed concentrations of 5-OH-ip. Interestingly, for the ip-
d+
doped materials 2a,b no highly reduced Ru was observed in
the XP spectra (no Ru 3d doublets at ca. 280.5 and 284.7 eV;
Figure 3). These data confirm the assumption deduced from
XANES data suggesting that application of defect linkers with
coordinatively inactive functional group H (2a, b) is likely to
more selectively induce missing-node defects B in the Ru-
DEMOFs without Ru reduction, as this is indicative of a type A
defect. According to the XANES and XPS results, we thus con-
clude that the Ru-DEMOFs with 5-X-ip (X=OH, NH , and Br)
2
contain predominantly defects of type A, especially at moder-
ate degrees of incorporation. However, the simultaneous gen-
eration of defects B in these samples is also likely, as indirectly
deduced from the nonlinear dependence of the XANES and
XPS data on the doping level.
Consequently, UHV-IR spectroscopic investigations employ-
ing CO as a probe molecule were carried out on two represen-
tative samples 1c and 1d to gain complementary evidence on
the valence states of framework Ru sites and the associated
defect types. Parent Ru-MOF was reported to show bands at
Figure 4. UHV-IR spectra of 1c (right) and parent Ru-MOF (left) on CO
dosing at 90–95 K after annealing the sample at 475–500 K. The bands at
2
ꢀ
1
2
041 and 1993 cm are assigned to vibrations of CO bound to reduced
d+
2
Ru -sites and indicate dissociative chemisorption of CO .
ꢀ
1
3+
2
172 and 2127 cm , arising from CO interactions with Ru
2
+
[20]
and Ru , respectively (Supporting Information, Figure S18).
Apart from these two bands, the spectra of samples 1c and
1
be attributed to CO bound to reduced Ru (0<d<2) and
again indicate the presence of defects A (i.e., modified paddle-
wheels). Notably, the relative intensity of the band at
parent Ru-MOF and Ru-DEMOF 1c, an intense and broad IR
ꢀ
1
ꢀ1
d show additional bands at 2041 and 1998 cm , which can
band appears at 2336 cm , which is assigned to the asymmet-
d+
ric stretching mode n (COO) of physisorbed CO binding line-
as
2
[15b]
ꢀ1
arly at various Ru sites.
The weak band at 2273 cm indi-
cates the presence of a minority CO species originating proba-
2
ꢀ
1
3+
13
2171 cm (Ru ) decreases greatly compared to that of the
bly from the adsorption of a small amount of CO with an ex-
2
parent Ru-MOF, and this is associated with the generation of
pected isotopic shift of 1.03. Apart from the CO -related bands,
2
2
+/d+
ꢀ1
reduced Ru
species. On the other hand, increasing the
two other bands at 2041 and 1993 cm have been observed
degree of 5-OH-ip doping from 32 (1c) to 37% (1d) leads to
a significant attenuation of the CO bands associated with Ru
for 1c, which are assigned to vibrations of CO bound to the re-
2
+
d+
duced Ru sites, in accordance with our previous data on
ꢀ
1
d+
ꢀ1
[20]
(
ca. 2127 cm ) and Ru (ca. 2041 and 1998 cm ), whereas
pydc-doped Ru-DEMOF.
This result indicates dissociative
d+
for pydc-doped Ru-DEMOFs the abundance of reduced Ru
chemisorption of CO to CO (reduction) in the Ru-DEMOF with
2
[
20]
increases gradually with increasing amount of pydc. These
findings confirm again the coexistence of defects A and B in
Ru-DEMOFs at relatively high degrees of 5-OH-ip incorporation
5-OH-ip defect linker (1c), whereas the parent Ru-MOF is es-
sentially inactive. However, in comparison to the extremely
high reactivity of pydc-doped Ru-DEMOFs, with which CO is
2
(
1d). The formation of defects B, that is, missing Ru paddle-
nearly totally converted to CO at higher concentrations of
2
+
d+
[20]
wheels, is accompanied by lower abundance of Ru and Ru
centers in the framework, which explains the observed de-
pydc defect linkers (e.g., 30%),
the production of CO is
rather limited for 5-OH-ip-doped Ru-DEMOFs (Figure 4). Inter-
2
+
d+
creased intensity of Ru and Ru in 1d.
estingly, Ru-DEMOF 1d is almost inactive for CO dissociation
2
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
(
Supporting Information, Figure S19). We correlate this obser-
lower uptake of 1d to less abundant Ru sites as a consequence
of missing metal nodes (i.e., type B defects). Hence, greater in-
corporation of the defect linker in 1d does not promote the
d+
vation to the much lower Ru content and in turn to the
more significant formation of missing-node defects (type B) for
1
d as compared to 1c. Overall, the presented IR data demon-
increase in CO uptake.
2
strate that the 5-OH-ip-doped Ru-DEMOFs show much lower
Furthermore, CO adsorption of the parent Ru-MOF, 1a, and
1c at 298 K also illustrates the tendency of enhanced uptake
in the case of the Ru-DEMOF samples (Supporting Information,
Figure S22). The small difference between 1a and the parent
Ru-MOF on CO adsorption is attributed to the low degree of
incorporation of the 5-OH-ip defect linker (8%). With increas-
ing degree of defect-linker incorporation (e.g., in 1c), the CO
uptake rises even at very low pressure (ca. 4 mbar). This in-
crease in CO capacity of the Ru-DEMOF is probably due to the
generation of the modified paddlewheels of type A, which
expose more open sites, because one carboxylate ligator site
of btc is substituted by the hydroxyl group of 5-OH-ip. Besides,
the electronic density varies as the oxidation state of rutheni-
um at the CUS is changed. Both could contribute to the en-
hancement of the adsorption properties.
reactivity for CO activation than pydc-doped materials. This
2
cannot be attributed only to the relatively low concentration
d+
[20]
of reduced Ru . In our previous study, we suggested that
CO activation is promoted by pydc linkers due to the pres-
2
d+
ence of basic pyridyl N sites in proximity to the reactive Ru
.
This hypothesis is indirectly supported by the properties of 1c
and 1d.
The combination of spectroscopic characterization data
(
XANES, XPS, UHV-FTIR with CO and CO as probes) presented
2
above suggests a more or less consistent, but complicated,
picture of the abundance of the two kinds of defects. Overall,
we assume that both types A and B are simultaneously gener-
ated in the Ru-DEMOFs in the case of coordinative weakly
binding ligator sites on the fragmented linkers. Type A appears
to be favored at low degrees of incorporation, for example, 1a
The H adsorption isotherms (at 77 K) of the parent Ru-MOF
2
(8%), and becomes more abundant up to a certain threshold,
and Ru-DEMOFs 1a and 1c follow nearly the same trend as in
for example, 1c (32%). Along with a further increase of incor-
poration level, for example, 1d (37%), defects of type B
appear to dominate over type A. The possibilities for formation
and distribution of various defect types is rather diverse and
may not be restricted to those mentioned here (see other pos-
sible defects and a summary of possible defect combinations
in Table S8 and Figure S20 of the Supporting Information). Ad-
ditional characterization and quantitative determination of the
defect types require additional work and support by theoreti-
cal modeling.
the cases of CO and CO adsorption (Supporting Information,
2
Figure S23a). Both defect-engineered materials 1a and 1c ex-
hibit higher H uptake at 1 bar than the parent intact frame-
2
work. However, H capacity is slightly lower when the incorpo-
2
ration of 5-OH-ip exceeds 8%, as in the case of 1c and 1d.
This could be due to the more significant generation of type B
defects in these Ru-DEMOFs. The calculated isosteric heats of
adsorption ꢀQ follow the order 1c>1a>parent Ru-MOF
st
when the respective values are compared at 1 mmol(H )/Ru
2
2
paddlewheel, which can be regarded as saturation adsorption
of H at the regular paddlewheel (Supporting Information, Fig-
2
ure S23b). This suggests stronger binding affinity of H for Ru-
2
CO , CO, and H sorption properties of Ru-DEMOFs
2
2
DEMOF with 32% of incorporated 5-OH-ip (1c), which again is
d+
The above characterization suggests the existence of two
types of defects (A and B) that are present in different absolute
and relative amounts as a consequence of 5-X-ip incorporation
in the framework. Both are likely to influence the gas-sorption
properties of the Ru-DEMOFs. Hence, we studied the adsorp-
tion of CO , CO, and H in a range of samples. In general, the
attributed to the larger amount of Ru sites in 1c.
For the 5-NH -ip defect linker, the respective Ru-DEMOFs
2
3a–c show the same tendencies in their absorption behavior
of CO and H as in their S (N ) values. The results of the
2
2
BET
2
XANES studies already revealed that defects B are likely to be
generated at the lower doping levels of 3a–c (5-NH -ip) as
2
2
2
results obtained and discussed in the following are not very
conclusive, as the observed changes in gas uptake as a function
of defect-linker incorporation are small or even close to the
error of the measurements. Nevertheless, we present the data
and provide a tentative discussion with respect to the possible
counteracting effects of the two types of defects in the materi-
als.
compared to 1a–d (5-OH-ip). This difference in abundance of
defects A/B leads to variation of the CO and H adsorption in
the 3a–c series. Even though 3a–c show higher uptake of CO2
2
2
and H than the parent Ru-MOF, the uptake of CO and H
2
2
2
does not further increase after the feed level of 5-NH -ip is
2
raised above 10% (3a; Supporting Information, Figure S26 and
S27). Similar to 1d, we assign the decrease of the uptake of
CO and H at 1 bar observed for 3c compared with 3a,b to
CO adsorption isotherms (298 K) of the parent reference Ru-
2
2
2
MOF, 1a, and 1c show gradual enhancement of the adsorption
capacity in the order Ru-MOF<1a<1c with increasing degree
of incorporation of the defect linker (Supporting Information,
more abundant type B defects.
Another indication of the suggested counteracting effects of
the defect sites is the comparison of the CO and H uptakes
2
2
Figure S21). However, the CO uptake of the 1d proved to be
of 2a, 2b, and 4a with those of the parent Ru-MOF (Support-
ing Information, Figure S24, S25, S28, and S29). However, the
variations are too small to be regarded as significant and are
not discussed here.
2
ꢀ1
2.9 mmolg at 1 bar, and lies in the order Ru-MOF<1d<1c
(Supporting Information, Table S9). According to our studies
on different defects in the series 1a–d, these results are not
unexpected. The increase of the CO uptake is attributed to
2
the reduced Ru sites at the type A defects, and the somewhat
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Catalytic test reactions
the transformation of ethylene to butene is the highest (TOF=
ꢀ
1
4.36 h ) when 1c is employed as catalyst. This value is still
Ethylene dimerization
much lower than that of zeolite-supported Ru-complex cata-
ꢀ1
Ethylene dimerization with MOF-supported catalysts has re-
lysts in the presence of H (4.36 vs. 216 h ). However, it is
2
[
28]
cently attracted a lot of attention. Most of the reported MOF
catalysts contain isolated single-metal (e.g., Ni, Ir) active sites.
Ru-MOFs have not yet been studied as heterogeneous cata-
lysts for ethylene dimerization, although chemisorbed Ru com-
plexes on dealuminated zeolite Y and an Ru complex that itself
catalyzes conversion of ethylene to butene have been already
higher than the TOF of the reported reaction without the pres-
ꢀ
4
ꢀ1
ꢀ1 [29b]
ence of H2 (1ꢂ10
s
=0.36 h ).
Distinction between
butene isomers, optimization of the reaction conditions for Ru-
DEMOF materials (e.g., use of H ), and the mechanism of this
2
reaction are currently under investigation. Nevertheless, the
preliminary results give a rather clear hint that the obtained Ru
materials could be utilized as catalysts for the dimerization of
ethylene. Due to the diverse modifiabilty of MOF materials
themselves, our current investigations on the ethylene dimeri-
zation with Ru-DEMOFs can provide a platform for further
studies on MOFs as catalysts in this field.
[
29]
reported. The redox activity of Rh in RhCa-X Zeolite plays an
[
30]
important role in catalyzing ethylene dimerization. Having
n+
distinct Ru CUS available in the Ru-DEMOFs (i.e., different Ru
species potentially involved into the redox process(es) of the
reaction flow), materials 1a (8% 5-OH-ip) and 1c (32% 5-OH-
ip) were selected for testing as catalysts in the selected test re-
action. Parent Ru-MOF was used as reference. Thus, activated
Ru-DEMOFs were loaded into a reactor to catalyze ethylene di-
merization in toluene at different temperatures under a pres-
sure of 800 psi (ꢁ55 bar). In general, all reactions led to the
formation of butene without producing any other a-olefins
Paal–Knorr pyrrole synthesis
Paal–Knorr pyrrole synthesis has attracted great attention due
to the huge synthetic variety of pyrroles and their derivatives,
[31]
which are key intermediates for various pharmaceuticals.
(
Supporting Information, Table S10). Addition of Et AlCl as co-
The reaction is typically performed under protic or Lewis acidic
conditions, by using primary amines and diketones. Parent Ru-
MOF [Ru (btc) Y ] is constructed from Ru paddlewheels in
2
catalyst accelerates the conversion (Supporting Information,
Table S10, entry 2 vs. entry 3). When the reaction was conduct-
ed at 268C, a lower yield was obtained than at 808C
3
2
1.5 n
2
which the axial Ru positions are partly occupied by strongly
binding Y or by other, weakly binding guest molecules. After
thermal treatment, a fraction of the axial positions can be ex-
posed as open Lewis acidic sites (e.g., useful as catalytic sites).
The structures of the Ru-DEMOFs discussed herein contain two
kinds of defects. The modified paddlewheel units (A) could
provide the materials with additional sites around the partially
reduced Ru centers. On the other hand, the eliminated entire
Ru paddlewheel clusters (B) can form vacant sites to play
a role in affecting the adsorption properties, especially when
the X functional group at the defect-generating linker is as
small as H. Hence, Ru-DEMOFs 1a–c, 2a, and 2b were utilized
for catalyzing the reaction of 2,5-hexadione and phenylamine
in toluene at 908C to form dimethylphenyl-1H-pyrrole
(
Table S10, entry 4 vs. entry 5). However, reactions conducted
for 1 and 24 h gave almost the same turnover frequency
ꢀ1
(
TOF=0.88 vs. 0.92 h ), which suggests that Ru-DEMOF 1c
does not suffer deactivation as a function of on-stream time.
Therefore, parent Ru-MOF and Ru-DEMOFs 1a and 1c were
employed to study the dimerization of ethylene in toluene
under optimized conditions (800 psi, 808C, 2 h) in the presence
of Et AlCl. The obtained product was analyzed by gas chroma-
2
tography. As shown in entries 5–7 of Table S10 (Supporting In-
formation) and Figure 5, the TOFs of 1a and 1c increase with
(
Scheme 1).
Scheme 1. The Paal–Knorr catalytic reaction.
Figure 5. TOFs for parent Ru-MOF and Ru-DEMOFs 1a and 1c utilized as cat-
alysts for transformation of ethylene into toluene (800 psi, 808C, 2 h) in the
2
presence of Et AlCl (0.81 mL, 1m in heptane). TOF=(moles product)(moles
ꢀ
1
ꢀ1
All of the selected Ru-DEMOFs show apparently enhanced
conversion of phenylamine to the pyrrole in the initial 4 h in
comparison with the parent Ru-MOF (Figure 6). All the materi-
als tested were found to be stable under the reaction condi-
tions, as confirmed by comparing the PXRD patterns of fresh
and used materials.
metal)
h .
increasing degree of incorporation of defect linkers in compari-
son with the parent Ru-MOF. Since the contents of reduced
d+
Ru sites in 1c is the highest among the 1a–d series (see
above), there should be more M-CUSs in 1c and therefore the
chance of enhanced catalytic activity in the reaction. Indeed,
Among the tested materials, Ru-DEMOFs 1a and 2a show
the highest catalytic activity, although the other Ru-DEMOFs
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
into the Ru-DEMOFs, depending on the nature of the function-
al groups X of the defect linkers 5-X-ip (see Figure 1). Defect
linkers with sterically less demanding X groups probably still
exist, but weaker coordinative binding properties as a carboxyl-
ate ligator are likely to favor type A defects, defined as modi-
d+
fied paddlewheel nodes exhibiting reduced Ru sites. With in-
creasing defect-linker contents, defects of type B, that is, miss-
ing-node defects, can be created simultaneously in the
DEMOFs. When the functional groups X in the defect linker 5-
X-ip are much smaller than carboxylate and noncoordinating
(
e.g., H), defects of type B, defined as missing-node defects,
Figure 6. Time–yield plot of the Paal–Knorr reaction of phenylamine with
,5-hexadione in the presence of Ru-DEMOFs 1a–1d (8, 20, 32, and 37% 5-
OH-ip incorporation, respectively) as catalysts in comparison with the parent
Ru-MOF (parent). “Blank” stands for the experiment in which no catalyst was
added to the reaction mixture.
apparently become more dominant even in the case of low
doping levels (2a and 2b). The simultaneous presence of two
kinds of defects leads to the synergetic and counteracting ef-
fects on the porosity, sorption, and catalytic properties. In fact,
Ru-DEMOF 1c (32% 5-OH-ip incorporation), in which defects
of type A are more dominant, has the highest BET surface area
2
2
ꢀ1
(
1b, 1c, 1d and 2b) incorporate more defect linkers (Support-
(1302 m g ) among the all materials, including the parent Ru-
[20]
ing Information, Table S11). Steric hindrance and distinct defect
types (A vs. B) could be the main reason for this phenomenon.
When Ru-DEMOFs with 5-OH-ip defect linkers (1a–d) were em-
ployed as catalysts, the yield of pyrrole increased from 48
MOF and its DEMOF derivatives, reported so far. Concerning
enhanced gas-sorption properties, type A defects are more im-
portant than type B defects, since reduced and more accessi-
d+
ble Ru sites are produced in type A defects. Regarding the
d+
(parent Ru-MOF) to 58 (1d), 65 (1c), 73 (1b), and 77% (1a).
catalytic activity, Ru sites in Ru-DEMOF 1c are suggested to
This is attributed to the presence of reduced Ru centers (de-
be responsible for the significant increase in TOF of ethylene
dimerization, as is clearly seen in comparison with the parent
d+
fects of type A) in these materials. Reduced, softer Ru bind-
d+
ing sites can more easily undergo coordination of the O atom
of the carbonyl group, which favors nucleophilic attack by the
lone-pair of the amino group of phenylamine. However, with
increasing incorporation of 5-OH-ip, the gradual dominance of
defects B in these materials probably eliminates part of the re-
active metal centers, and thus affects the catalytic activity of
Ru-MOF and 1a, in which no or fewer Ru sites are present.
However, in the case of the Paal–Knorr pyrrole synthesis, the
possible steric hindrance between the functional groups (OH)
at the defect linker and 2,5-hexadione substrate is suggested
to be responsible for the reduced activity of Ru-DEMOFs 1a–d,
although they all feature enhanced conversion compared with
the parent Ru-MOF. All in all, the described Ru-DEMOFs ob-
tained by the solid-solution approach with various defect link-
ers turned out to be rather complex materials. Nevertheless,
the spectroscopic evidence and observed effects on sorption
properties and catalytic performance reasonably fit into
a model of the two types of defects, that is, modified and
missing nodes (see Figure 1).
1
b–d. On the other hand, Ru-DEMOF 2a shows a higher yield
of pyrrole compared to the parent Ru-MOF (82 vs. 48%), re-
gardless of the partial absence of metal centers (defects B;
Supporting Information, Figure S30). This could be a result of
the smaller steric hindrance surrounding the Ru-CUSs in the
case of the H ip defect linkers. When the contents of the incor-
2
porated defect linkers are rather low (i.e., as in 2a), the nega-
tive influence caused by missing metal centers is relatively in-
significant compared with the positive effect of the smaller
steric hindrance. Therefore, 2a shows as high a catalytic activi-
ty in the present reaction as 1a. Similar observations were
made for 1b,c because of the somewhat larger steric hin-
drance of OH groups in spite of the type A defects in these
materials.
Experimental Section
Materials
[32]
n
Ruthenium SBU precursor [Ru (OOCCH ) Cl]
single-linker Ru-MOF
and the parent
were prepared by following procedures
2
3 4
[15]
previously described in the literature. All other reagents were avail-
able commercially and used without further purification. Before
further manipulations, all activated samples were stored in a glove-
box under inert Ar atmosphere.
Conclusion
By applying the solid-solution approach, a range of Ru-
DEMOFs [Ru (btc) (5-X-ip) Y ] , isostructural to HKUST-1, were
3
2ꢀx
x
y n
Methods
obtained with the framework-incorporated defect linkers 5-OH-
PXRD data for all as-synthesized samples and activated materials
ip (1a–d), ip (2a and 2b), 5-NH -ip (3a–c), and 5-Br-ip (4a).
2
(2a–2d, 4a–4c) were collected with an X’Pert PRO PANalytical
Comparably high S values were measured for Ru-DEMOFs
BET
equipment (Bragg–Brentano geometry with automatic divergence
slits, position-sensitive detector, continuous mode, room tempera-
ture, CuKa radiation, Ni filter, in the range of 2q=5–508, step size
0.018). Activated samples were prepared on a silicon wafer in an
Ar-filled glovebox right before the measurement started. For the
2
ꢀ1
(
(
947–1302 m g ) relative to the parent Ru-MOF [Ru (btc) Y ]
3 2 1.5 n
2
ꢀ1 [15a,23]
704–998 m g ).
The highest degree of defect-linker in-
corporation appeared to be 37% (1d, 5-OH-ip). The data are
consistent with introduction of two kinds of defects (A and B)
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
measurements on activated samples (Ru-DEMOF series 1 and 3),
the data were collected at beam line BL9 of the synchrotron radia-
tion facility DELTA at a wavelength of 0.4592 ꢃ by using a two-di-
mensional MAR345 image-plate detector. The sample was filled
into standard capillaries (0.5 mm diameter) in an Ar-filled glovebox
and measured. The data were integrated by using the program
was set at 0.3 mm and the pass energy was fixed at 200 eV for all
the measurements. The overall energy resolution was better than
0.5 eV. A flood gun was used to compensate for the charging ef-
fects. All spectra reported here are calibrated to the C 1s core-level
binding energy of 285 eV. The XP spectra were deconvoluted by
using the CASA XPS program with a mixed Gaussian–Lorentzian
function and Shirley background subtraction. Gas chromatographic
measurements for the Paal–Knorr reaction were performed with an
Agilent Technologies 7890A with flame-ionization detector by
using an HP-5 capillary column (5% phenylmethylpolysiloxane) of
30 m length and 0.32 mm internal diameter and a BP20(WAX)
column of 15 m length and 0.32 mm internal diameter. Thereby,
the samples were measured at high dilution in volatile organic sol-
vents (usually ethanol or acetone).
[33]
package Fit2D
and transformed to the CuKa radiation (l=
1
.54178 ꢃ) used for all other PXRD measurements on reported ma-
terials for convenient comparison. Elemental analysis data were ob-
tained from the Mikroanalytisches Laboratorium Kolbe in Mꢄlheim
an der Ruhr (http://www.mikro-lab.de/index.html). Samples were
filled into finger Schlenk flasks in an Ar-filled glovebox and mea-
sured under Ar. FTIR spectra were collected with a Bruker Alpha
FTIR instrument in the ATR geometry with a diamond ATR unit in
ꢀ1
the range 400–4000 cm inside a LABStar MB 10 compact MBraun
glovebox (argon atmosphere). (TEM-)EDX spectra for the composi-
Synthesis of Ru-DEMOFs (1a–1d, 2a–2d, 3a–3d, 4a–4c)
2
tional determination of 4a–4c were recorded with a Tecnai G F20
equipped with a Schottky field-emission gun operated at an accel-
eration voltage of 200 kV at the department of Mechanical Engi-
neering, Ruhr-University Bochum. Liquid-phase H NMR spectra
The materials were synthesized followed by the reported con-
[20]
trolled SBU approach. Mixtures of parent linker H btc and defect
3
1
linker 5-X-ipH [X=OH (1), H (2), NH (3), and Br (4)] together with
2
2
were measured with a Bruker Avance DPX-200 spectrometer at
1.5 molar equivalents of [Ru (OOCCH ) Cl] (0.36 mmol,170 mg)
2 3 4 n
2
93 K in DCl/[D ]DMSO on digested activated MOF samples. CO ad-
were placed in 20 mL Teflon vessels. Subsequently, 4 mL of HPLC-
grade water and 0.7 mL of glacial acetic acid were added. The
molar ratios and amounts of the linkers used in the starting materi-
als are listed in Table S1 of the Supporting Information. The Teflon
reaction vessels were then placed in Teflon-lined stainless steel au-
toclaves. The autoclaves were further sealed and placed in a pre-
heated oven at 433 K for 72 h. Afterwards the autoclaves were
taken out and allowed to cool to room temperature. The resulting
powder products were collected by centrifugation. Afterwards,
they were soaked in HPLC-grade water (ca. 20 mL) and the solvent
was refreshed three times with the same amount of water every
24 h. Finally, the solvent was removed and collected powders were
dried in air at room temperature. The solids were activated by
6
sorption (298 K) and the majority of N (99.999%) sorption (77 K)
2
measurements were performed with a Micromeritics 3Flex instru-
ment. The N2 sorption isotherms of 3a–3d were collected with
a Quantachrome Autosorp-1 MP instrument, optimized protocols,
and N of 99.9995% purity. Experiments on sorption of CO (298 K)
and H (77 K) were conducted with a Micromeritics ASAP 2020 gas-
adsorption analyzer. TGA was performed with a TG/DSC NETZSCH
2
2
2
ꢀ
1
STA 409 PC instrument at a heating rate of 5 Kmin in the tem-
perature range 30–6008C at atmospheric pressure (sample weight
ꢀ
1
5
–10 mg, N (99.999%) gas flow (20 mLmin ). XANES spectra were
2
recorded at the Ru K-edge (22117 eV) by using an Si(311) mono-
chromator at Beamline BL8 of the synchrotron radiation facility
DELTA, TU Dortmund. The samples were filled into 1 mm capillaries
in an Ar-atmosphere glovebox before the measurement. The
XANES data were acquired by using 15 cm ion chamber filled with
Ar as I0, a 15 cm ion chamber filled with Xe as I1, and a 30 cm ion
chamber filled with Xe as I2. Approximately 30 mm above the
sample, a PIN diode captured a relatively wide solid angle of fluo-
rescence radiation. The energy was calibrated by measuring a met-
allic Ru foil as reference simultaneously with each sample scan.
The data processing and analysis were done with the program
ꢀ3
heating at 423 K for 24 h under dynamic vacuum (ca. 10 mbar).
Catalytic reactions
Ethylene dimerization
In a typical run (Supporting Information, Table S10, entry 5), 2.1 mg
of activated Ru-MOF 1c was placed in a 10 mL steel reactor, after
which 0.81 mL of Et
0.09 mL of undecane (0.01m in toluene) were added. After degass-
ing, the reactor was flushed with C (800 psi) and placed in a pre-
AlCl (1m in heptane), 2.1 mL of toluene, and
2
[34]
ATHENA.
For UHV-FTIR measurements, powder samples were
first pressed into a stainless steel grid covered by gold and then
mounted on a sample holder, which was specially designed for
FTIR transmission measurements under UHV conditions. The grid
was cleaned by heating to 850 K to remove all contaminants
formed its during preparation. The base pressure in the measure-
ment chamber was 5ꢂ10 mbar. The optical path inside the IR
spectrometer and the space between the spectrometer and UHV
chamber were also evacuated to avoid adsorption of atmospheric
moisture and thus ensure high sensitivity and stability. The MOF
samples were cleaned in the UHV chamber by heating to 500 K to
remove the contaminants from synthesis and all adsorbed species
such as water and hydroxyl groups. Prior to each exposure, a spec-
trum of the clean sample was recorded as a background reference.
H
2 4
heated oil bath at 808C with stirring for 2 h. Subsequently, the re-
action was stopped and the reactor was cooled at ꢀ78C in a mix-
ture of dry ice and ethylene glycol. Cold distilled water was added
to quench the reaction. The organic layer, kept cold, was then
quickly analyzed by gas chromatography (SRI 8610V GC, 60 mꢂ
0.54 mm internal diameter, 5.0 mm MXT-1 capillary columns) by
using the undecane signal as reference. For other entries, the
amount of catalyst, the reaction temperature, and time were
varied (Supporting Information, Table S10). In case of no additive,
the volume of toluene was changed to 2.91 mL. Other parameters
are the same as described above for Supporting Information, Table
S10, entry 5.
ꢀ
11
The exposure of the sample to CO and CO was carried out by
2
backfilling the measurement chamber through a leak valve. All
UHV-FTIR spectra were collected with 512 scans at a resolution of
Paal–Knorr reaction
ꢀ1
4
cm in transmission mode. XPS measurements were performed
In a typical run, 5 mg of activated Ru-MOF/DEMOF in 0.5 mL of tol-
uene was introduced into 125 mg (1.1 mmol) of 2,5-hexadione and
95 mg (1 mmol) of phenylamine and the mixture stirred at 908C
for 24 h under air. The reaction was monitored by taking aliquots
in a UHV setup equipped with a high-resolution Gammadata-Scien-
ta SES 2002 analyzer. A monochromatic AlKa X-ray source (energy
1
486.6 eV) was used as incident radiation. The analyzer slit width
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
10
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
every hour and analyzing the products by GC-MS. The reactions
were performed in closed (pressurized) reaction vials.
[9] F. Vermoortele, B. Bueken, G. Le Bars, B. Van de Voorde, M. Vandichel, K.
Houthoofd, A. Vimont, M. Daturi, M. Waroquier, V. Van Speybroeck, C.
Kirschhock, D. E. De Vos, J. Am. Chem. Soc. 2013, 135, 11465–11468.
10] F. G. Cirujano, A. Corma, F. X. Llabrꢀs i Xamena, Chem. Eng. Sci. 2015,
[
[
1
24, 52–60.
Acknowledgements
11] S. S. Y. Chui, S. M. F. Lo, J. P. H. Charmant, A. G. Orpen, I. D. Williams, Sci-
ence 1999, 283, 1148–1150.
This research was partially supported by the DFG project FI-
[12] a) M. Kramer, U. Schwarz, S. Kaskel, J. Mater. Chem. 2006, 16, 2245–
248; b) C. R. Wade, M. Dinca, Dalton Trans. 2012, 41, 7931–7938.
2
502/32-1. W.Z. is thankful for a PhD fellowship from the China
[
13] L. J. Murray, M. Dinca, J. Yano, S. Chavan, S. Bordiga, C. M. Brown, J. R.
Scholarship Council (CSC). W.Z. is also grateful to the Research
Long, J. Am. Chem. Soc. 2010, 132, 7856–7857.
Plus
School
at Ruhr-University Bochum for the support of her
[14] P. Maniam, N. Stock, Inorg. Chem. 2011, 50, 5085–5097.
[15] a) O. Kozachuk, K. Yusenko, H. Noei, Y. Wang, S. Walleck, T. Glaser, R. A.
Fischer, Chem. Commun. 2011, 47, 8509–8511; b) H. Noei, O. Kozachuk,
S. Amirjalayer, S. Bureekaew, M. Kauer, R. Schmid, B. Marler, M. Muhler,
R. A. Fischer, Y. Wang, J. Phys. Chem. C 2013, 117, 5658–5666.
[16] J. I. Feldblyum, M. Liu, D. W. Gidley, A. J. Matzger, J. Am. Chem. Soc.
2011, 133, 18257–18263.
PhD project and funding of an internship at UC Berkeley at the
group of Prof. Jeffrey. R. Long and collaboration with D.J.X.
and M.I.G. including Douglas Reed for the collection of CO iso-
therms (298 K). W.Z. also thanks Dr. Raghavender Medishetty
for the fruitful discussions. P.G. acknowledges the support of
the EU innovative Training Network “DEFect NETwork materials
science and engineering” (DEFNET). The authors further thank
the team at DELTA synchrotron facility at the TU Dortmund for
the support with the X-ray absorption spectroscopy experi-
ments performed at beam lines BL8 and the PXRD data collec-
tion at beam lines BL9.
[
[
17] S. Marx, W. Kleist, A. Baiker, J. Catal. 2011, 281, 76–87.
18] Z. Fang, J. P. Dꢄrholt, M. Kauer, W. Zhang, C. Lochenie, B. Jee, B. Albada,
N. Metzler-Nolte, A. Pçppl, B. Weber, M. Muhler, Y. Wang, R. Schmid,
R. A. Fischer, J. Am. Chem. Soc. 2014, 136, 9627–9636.
[19] G. Barin, V. Krungleviciute, O. Gutov, J. T. Hupp, T. Yildirim, O. K. Farha,
Inorg. Chem. 2014, 53, 6914–6919.
[
20] O. Kozachuk, I. Luz, F. X. Llabrꢀs i Xamena, H. Noei, M. Kauer, H. B.
Albada, E. D. Bloch, B. Marler, Y. Wang, M. Muhler, R. A. Fischer, Angew.
Chem. Int. Ed. 2014, 53, 7058–7062; Angew. Chem. 2014, 126, 7178–
7182.
Keywords: defect engineering · heterogeneous catalysis ·
[
21] F. Schrçder, D. Esken, M. Cokoja, M. W. E. van den Berg, O. I. Lebedev, G.
Van Tendeloo, B. Walaszek, G. Buntkowsky, H.-H. Limbach, B. Chaudret,
R. A. Fischer, J. Am. Chem. Soc. 2008, 130, 6119–6130.
hydrothermal synthesis
ruthenium
·
metal–organic frameworks
·
[
[
[
[
[
[
[
22] E. A. Braude, F. C. Nachod, Determination of Organic Structures by Physi-
cal Methods, Academic Press, New York, USA, 1955.
[
1] a) M. Latroche, S. Surblꢀ, C. Serre, C. Mellot-Draznieks, P. L. Llewellyn, J.-
H. Lee, J.-S. Chang, S. H. Jhung, G. Fꢀrey, Angew. Chem. Int. Ed. 2006, 45,
227–8231; Angew. Chem. 2006, 118, 8407–8411; b) X. Lin, J. Jia, X.
23] W. Zhang, O. Kozachuk, R. Medishetty, A. Schneemann, R. Wagner, K.
Khaletskaya, K. Epp, R. A. Fischer, Eur. J. Inorg. Chem. 2015, 3913–3920.
24] F. Genin, F. Quiles, A. Burneau, Phys. Chem. Chem. Phys. 2001, 3, 932–
8
Zhao, K. M. Thomas, A. J. Blake, G. S. Walker, N. R. Champness, P. Hub-
berstey, M. Schrçder, Angew. Chem. Int. Ed. 2006, 45, 7358–7364;
Angew. Chem. 2006, 118, 7518–7524; c) K. Sumida, D. L. Rogow, J. A.
Mason, T. M. McDonald, E. D. Bloch, Z. R. Herm, T.-H. Bae, J. R. Long,
Chem. Rev. 2012, 112, 724–781.
942.
25] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination
Compounds, Wiley, Hoboken, 2008, pp. 1–273.
26] Z. Wang, J. Liu, B. Lukose, Z. Gu, P. G. Weidler, H. Gliemann, T. Heine, C.
Wçll, Nano Lett. 2014, 14, 1526–1529.
27] K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rou-
querol, T. Siemieniewska, Pure Appl. Chem. 1985, 57, 603–619.
28] a) S. T. Madrahimov, J. R. Gallagher, G. Zhang, Z. Meinhart, S. J. Garibay,
M. Delferro, J. T. Miller, O. K. Farha, J. T. Hupp, S. T. Nguyen, ACS Catal.
[
2] a) A. Schneemann, E. D. Bloch, S. Henke, P. L. Llewellyn, J. R. Long, R. A.
Fischer, Chem. Eur. J. 2015, 21, 18764–18769; b) E. D. Bloch, M. R.
Hudson, J. A. Mason, S. Chavan, V. Crocellꢅ, J. D. Howe, K. Lee, A. L.
Dzubak, W. L. Queen, J. M. Zadrozny, S. J. Geier, L.-C. Lin, L. Gagliardi, B.
Smit, J. B. Neaton, S. Bordiga, C. M. Brown, J. R. Long, J. Am. Chem. Soc.
2
015, 5, 6713–6718; b) D. Yang, S. O. Odoh, J. Borycz, T. C. Wang, O. K.
Farha, J. T. Hupp, C. J. Cramer, L. Gagliardi, B. C. Gates, ACS Catal. 2016,
, 235–247; c) J. Canivet, S. Aguado, Y. Schuurman, D. Farrusseng, J.
2014, 136, 10752–10761; c) E. D. Bloch, W. L. Queen, R. Krishna, J. M. Za-
drozny, C. M. Brown, J. R. Long, Science 2012, 335, 1606–1610; d) T. Ro-
denas, I. Luz, G. Prieto, B. Seoane, H. Miro, A. Corma, F. Kapteijn, F. X.
Llabrꢀs i Xamena, J. Gascon, Nat. Mater. 2015, 14, 48–55.
6
Am. Chem. Soc. 2013, 135, 4195–4198; d) E. D. Metzger, C. K. Brozek,
R. J. Comito, M. Dinc a˘ , ACS Cent. Sci. 2016, 2, 148–153.
[3] a) J. S. Seo, D. Whang, H. Lee, S. I. Jun, J. Oh, Y. J. Jeon, K. Kim, Nature
[
29] a) I. Ogino, B. C. Gates, Chem. Eur. J. 2009, 15, 6827–6837; b) I. Ogino,
B. C. Gates, J. Am. Chem. Soc. 2008, 130, 13338–13346; c) G. Laurenczy,
A. E. Merbach, J. Chem. Soc. Chem. Commun. 1993, 187–189.
30] J. S. Bass, L. Kevan, J. Phys. Chem. 1990, 94, 1483–1489.
2000, 404, 982–986; b) A. Corma, H. Garcꢆa, F. X. Llabrꢀs i Xamena,
Chem. Rev. 2010, 110, 4606–4655.
[
[
4] E. Biemmi, C. Scherb, T. Bein, J. Am. Chem. Soc. 2007, 129, 8054–8055.
5] P. Horcajada, C. Serre, M. Vallet-Regꢆ, M. Sebban, F. Taulelle, G. Fꢀrey,
Angew. Chem. Int. Ed. 2006, 45, 5974–5978; Angew. Chem. 2006, 118,
[
[
31] a) A. L. Harreus, in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-
VCH, Weinheim, 2000, DOI: 10.1002/14356007.a22 453; b) R. J. Sund-
berg, in Comprehensive Heterocyclic Chemistry II (Ed.: Scriven), Perga-
mon, Oxford, 1996, pp. 119–206.
6120–6124.
[
6] a) R. A. Beyerlein, C. Choi-Feng, J. B. Hall, B. J. Huggins, G. J. Ray, Top.
Catal. 1997, 4, 27–42; b) C. Baerlocher, D. Xie, L. B. McCusker, S.-J.
Hwang, I. Y. Chan, K. Ong, A. W. Burton, S. I. Zones, Nat. Mater. 2008, 7,
[
[
[
32] R. W. Mitchell, A. Spencer, G. Wilkinson, J. Chem. Soc. Dalton Trans.
1
973, 846–854.
6
31–635; c) B. Bonelli, L. Forni, A. Aloise, J. B. Nagy, G. Fornasari, E. Gar-
rone, A. Gedeon, G. Giordano, F. Trifirꢇ, Microporous Mesoporous Mater.
007, 101, 153–160.
7] Z. Fang, B. Bueken, D. E. De Vos, R. A. Fischer, Angew. Chem. Int. Ed.
015, 54, 7234–7254; Angew. Chem. 2015, 127, 7340–7362.
33] A. P. Hammersley, S. O. Svensson, M. Hanfland, A. N. Fitch, D. Hauser-
mann, High Pressure Res. 1996, 14, 235–248.
34] B. Ravel, M. Newville, J. Synchrotron Radiat. 2005, 12, 537–541.
2
[
[
2
8] U. Ravon, M. Savonnet, S. Aguado, M. E. Domine, E. Janneau, D. Farrus-
Received: June 3, 2016
seng, Microporous Mesoporous Mater. 2010, 129, 319–329.
Published online on && &&, 0000
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
11
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
Engineered defects: Various 5-X-iso-
phthalate linkers were introduced into
the Ru analogue of HKUST-1 to yield
isoreticular, defect-engineered (DE) Ru-
MOFs. The characterization data are
consistent with two kinds of defects
&
Metal–Organic Frameworks
W. Zhang, M. Kauer, O. Halbherr, K. Epp,
P. Guo, M. I. Gonzalez, D. J. Xiao,
C. Wiktor, F. X. LIabrꢁs i Xamena, C. Wçll,
Y. Wang,* M. Muhler, R. A. Fischer*
(
type A: modified paddlewheels (PWs)
&& – &&
and type B: missing PWs; see figure),
and their relative abundance is con-
trolled by nature of the functional
group X of the defect-generating link-
ers.
Ruthenium Metal–Organic
Frameworks with Different Defect
Types: Influence on Porosity, Sorption,
and Catalytic Properties
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
12
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!