Extraordinary Separation of Acetylene-Containing Mixtures with Microporous Metal-Organic Frameworks with Open O Donor Sites and Tunable Robustness through Control of the Helical Chain Secondary Building Units

Article abstract of DOI:10.1002/chem.201505107

Acetylene separation is a very important but challenging industrial separation task. Here, through the solvothermal reaction of CuI and 5-triazole isophthalic acid in different solvents, two metal-organic frameworks (MOFs, FJU-21 and FJU-22) with open O donor sites and controllable robustness have been obtained for acetylene separation. They contain the same paddle-wheel {Cu₂(COO₂)₄} nodes and metal-ligand connection modes, but with different helical chains as secondary building units (SBUs), leading to different structural robustness for the MOFs. FJU-21 and FJU-22 are the first examples in which the MOFs' robustness is controlled by adjusting the helical chain SBUs. Good robustness gives the activated FJU-22 a, which has higher surface area and gas uptakes than the flexible FJU-21 a. Importantly, FJU-22 a shows extraordinary separation of acetylene mixtures under ambient conditions. The separation capacity of FJU-22 a for 50:50 C₂H₂/CO₂ mixtures is about twice that of the high-capacity HOF-3, and its actual separation selectivity for C₂H₂/C₂H₄ mixtures containing 1 % acetylene is the highest among reported porous materials. Based on first-principles calculations, the extraordinary separation performance of C₂H₂ for FJU-22 a was attributed to hydrogen-bonding interactions between the C₂H₂ molecules with the open O donors on the wall, which provide better recognition ability for C₂H₂ than other functional sites, including open metal sites and amino groups.

Full text of DOI:10.1002/chem.201505107

DOI: 10.1002/chem.201505107
Full Paper
&
Metal–Organic Frameworks
Extraordinary Separation of Acetylene-Containing Mixtures with
Microporous Metal–Organic Frameworks with Open O Donor Sites
and Tunable Robustness through Control of the Helical Chain
Secondary Building Units
Zizhu Yao,[a] Zhangjing Zhang,[a] Lizhen Liu,[a] Ziyin Li,[a] Wei Zhou,[b] Yunfeng Zhao,[c, f]
Yu Han,[c] Banglin Chen,[d] Rajamani Krishna,*[e] and Shengchang Xiang*[a]
Abstract: Acetylene separation is a very important but chal-
lenging industrial separation task. Here, through the solvo-
thermal reaction of CuI and 5-triazole isophthalic acid in dif-
ferent solvents, two metal–organic frameworks (MOFs, FJU-
21 and FJU-22) with open O donor sites and controllable ro-
bustness have been obtained for acetylene separation. They
contain the same paddle-wheel {Cu2(COO2)4} nodes and
metal–ligand connection modes, but with different helical
chains as secondary building units (SBUs), leading to differ-
ent structural robustness for the MOFs. FJU-21 and FJU-22
are the first examples in which the MOFs’ robustness is con-
trolled by adjusting the helical chain SBUs. Good robustness
gives the activated FJU-22a, which has higher surface area
and gas uptakes than the flexible FJU-21a. Importantly, FJU-
22a shows extraordinary separation of acetylene mixtures
under ambient conditions. The separation capacity of FJU-
22a for 50:50 C2H2/CO2 mixtures is about twice that of the
high-capacity HOF-3, and its actual separation selectivity for
C2H2/C2H4 mixtures containing 1% acetylene is the highest
among reported porous materials. Based on first-principles
calculations, the extraordinary separation performance of
C2H2 for FJU-22a was attributed to hydrogen-bonding inter-
actions between the C2H2 molecules with the open O
donors on the wall, which provide better recognition ability
for C2H2 than other functional sites, including open metal
sites and amino groups.
Introduction
Removal of acetylene from C2H2/C2H4 mixtures is an important
and highly challenging industrial process.[1–5] Ethylene, the larg-
est volume organic chemical, is widely used for the production
of polymers. Ethylene produced in steam crackers typically
contains on the order of 1% of acetylene,[6] which should be
reduced to an accepted level in the ethylene feed before poly-
merization, because acetylene can poison Ziegler–Natta cata-
lysts and also lower the resulting product quality.[7] Current ap-
proaches to remove acetylene include partial hydrogenation
over a noble metal catalyst[8] and the solvent extraction of
cracked olefins,[9] but both of these are cost and energy con-
sumptive.
[a] Z. Yao, Prof. Dr. Z. Zhang, L. Liu, Z. Li, Prof. Dr. S. Xiang
Fujian Provincial Key Laboratory of Polymer Materials
Fujian Normal University, 32 Shangsan Road
Fuzhou 350007 (P. R. China)
E-mail: scxiang@fjnu.edu.cn
[b] Prof. Dr. W. Zhou
NIST Center for Neutron Research, Gaithersburg
Maryland 20899-6102 (USA)
[c] Dr. Y. Zhao, Prof. Dr. Y. Han
Advanced Membranes and Porous Materials Center
Physical Sciences and Engineering Division
King Abdullah University of Science and Technology
Thuwal 23955-6900 (Saudi Arabia)
The emerging microporous metal–organic frameworks
(MOFs)[10–13] based on physical adsorption are promising as
cost-effective and efficient materials for gas separation, which
has been a topic of interest because the ability to rationally
design and chemically tune their architecture of the MOFs
allows chemists to establish various methods to achieve highly
selective gas adsorption.[10–29] Since the pioneered work of Kita-
gawa and co-workers,[14] some MOFs with high acetylene stor-
age have been realized by using immobilized open metal sites
(OMS).[30,31] For C2H2/C2H4 separation, after the first work realiz-
ed by using the flexible MOFs on the basis of metalloligands in
2011,[2] the series of MOF-74[3] with high densities of open
[d] Prof. Dr. B. Chen
Department of Chemistry, University of Texas at San Antonio
One UTSA Circle, San Antonio, Texas 78249-0698 (USA)
[e] Prof. Dr. R. Krishna
Van’t Hoff Institute for Molecular Sciences, University of Amsterdam
Science Park 904
Amsterdam 1098 XH (The Netherlands)
[f] Dr. Y. Zhao
Institute for New Energy Materials & Low-Carbon Technologies
School of Materials Science and Engineering
Tianjin University of Technology, Tianjin 300384 (P. R. China)
Supporting information and ORCID for the author for this article are avail-
able on the WWW under http://dx.doi.org/10.1002/chem.201505107.
Chem. Eur. J. 2016, 22, 5676 – 5683
5676
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
metal sites and NOTT-300[4] with multiple weak supramolecular
interactions were also employed. However, these three kinds
of MOFs are not the ideal materials; the former exhibit very
low acetylene uptake and the latter two show very low selec-
tivity toward acetylene. A recent MOF, UTSA-100, containing
amino groups[5] is a unique example for actual column break-
through experiments of C2H2/C2H4 mixtures, giving moderate
selectivity and moderate acetylene uptake. Additionally, owing
to the similar fluid properties of acetylene and CO2, the effi-
cient separation of C2H2/CO2 mixture is another technologically
interesting issue.[14,28] A few MOFs and zeolites with relevant
differences in capacity for C2H2/CO2 have been developed,
mainly through tuning the cross-section size of the pore.[32] To
date, no MOFs or other porous materials have been found
with actual C2H2/CO2 breakthrough experiments, except one
hydrogen-bonded organic framework (HOF-3),[33] however, this
has low separation capacity. The MOFs for the separation of
C2H2 mixtures have not been fully explored. It is still desirable
to explore new ways to construct the MOFs with excellent per-
formance for the challenging separations of C2H2/CO2 and
C2H2/C2H4 mixtures.
In this work, through solvothermal reactions of 5-triazole iso-
phthalic acid (H2L), CuI, and various solvents, two MOFs
{[Cu(L)](DMF)(H2O)1.5}n (FJU-21, DMF=N,N’-dimethylformamide)
and {[Cu(L)]·(DMA)(H2O)1.5}n (FJU-22, DMA=N,N’-dimethylacet-
amide) with open O donor sites and controllable robustness
have been obtained for the highly selective separation of both
C2H2/CO2 and C2H2/C2H4 mixtures. This was done on the base
of the following considerations. (1) The remarkable difference
between C2H2 and other gases including CO2 and ethylene is
the acidic hydrogen atoms at both ends of acetylene. Thus,
the different framework flexibilities, different OMS and amino
groups, and the open O donors can endow the MOFs with ex-
clusive recognition ability for C2H2 through extra hydrogen-
bonding interactions. (2) The use of 5-triazole isophthalic acid
(H2L) containing one heterocyclic and two carboxylate groups
to construct the MOFs may generate rich open O donor sites
(without H riding) standing on the framework wall to recog-
nize acetylene. (3) Solvent-induced structure diversity with con-
trollable robustness may be observed to further control the
acetylene separation capacity. As expected, the open O donors
can provide the MOFs better recognition ability for C2H2 than
other functional sites, including the OMS and amino groups.
The activated FJU-22a with open O donors and good robust-
ness exhibits extraordinary separation performances for both
C2H2/CO2 and C2H2/C2H4 mixtures at ambient conditions as
demonstrated by gas adsorption, breakthrough simulations,
actual column breakthrough experiments, and first-principles
calculations.
Figure 1. The structures of FJU-21 (a–g) and FJU-22 (h–m) showing: the co-
ordination environment of FJU-21 (a) and FJU-22 (h); the twist angle be-
tween the triazole ring and the plane of the benzene ring of the ligand in
FJU-21 (b) and FJU-22 (i); the two kinds of helical chain SBUs of FJU-21 (c
and d) along the b axis; the unique kind of helical chain SBUs with various
pitches along the a and c axes in FJU-22 (j and k); the triazole-pillared
[Cu2(isophthalate)4] bilayers in the orientation of the (200) plane of FJU-
21 (e); the [Cu2(isophthalate)4] monolayers in the orientation of the (100)
plane of FJU-21 (f); 3D framework of FJU-21 along the b axis (g); and 3D
framework of FJU-22 view along the a and c axes (l and m). Color code: Cu,
green; C, gray; O, red; N, blue.
except that DMF was replaced by DMA. Single-crystal X-ray dif-
fraction analyses reveal that FJU-21 and FJU-22 have the same
metal nodes and metal–ligand connection modes (Figure 1). In
each crystallographically independent unit, there is one CuII
ion and one ligand. The CuII ion is fivefold coordinated by four
oxygen atoms and one triazolyl N donor from five ligands and
has a square-pyramidal coordination environment. Two CuII
ions are bridged by four carboxylate groups to give a paddle-
wheel node (PWN). The axial site of the PWN is occupied by tri-
azolyl N donors. The ligand L employs its one triazole N and
four carboxylate O atoms to link three PWNs. In FJU-22, the
twist angle between the triazole ring and the plane of the ben-
zene ring of the ligand L (46.98) is higher than that of FJU-21
(41.38), as shown in Figure 1i and b. The non-planarity of
ligand L with various twist angles endows the two MOFs with
various helical chains that act as secondary building units
Results and Discussion
Solvent-induced structure diversity
Green needle-like crystals of FJU-21 could be obtained in
DMF/H2O (3:2, v/v) mixed solvent at 858C for 1 day, and bulk
crystals of FJU-22 could be obtained by the same method
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5677
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(SBUs). FJU-22 has a unique type of helical chain SBU with dif-
ferent pitches composed of the PWNs and N-benzene triazole
linkers along the a and c axes (Figure 1j and k). The linkage
between adjacent SBUs sharing the PWNs results in a three-di-
mensional (3D) skeleton (Figure 1l and m). FJU-21 has two
types of helical chain SBUs along the b axis. One has the
same composites as that in FJU-22 (Figure 1c), which connects
with the neighboring SBUs to give triazole-pillared
[Cu2(isophthalate)4] bilayers in the orientation of the (200)
plane (Figure 1e), whereas the other is made up of the PWNs
and isophthalate linkers (Figure 1d), and connects with the ad-
jacent SBUs to produce [Cu2(isophthalate)4] monolayers orient-
ed at the (100) plane (Figure 1 f). The bilayers and monolayers
stack on each other by layer-sharing along the a axis to form
the 3D framework (Figure 1g). By considering the PWN as
a six-connected octahedral node and the ligand as a three-con-
nected trigonal linker, the whole frameworks of FJU-21 and
FJU-22 can be simplified to a (3,6)-connected net with rutile
(rtl) and a-PbO2 (apo) topology, respectively. FJU-21 shows
a 1D channel along the a axis (5.00”9.60 Š2), whereas FJU-22
also has a 1D channel, but along the c axis (7.10”7.10 Š2; Fig-
ure S1 in the Supporting Information). PLATON calculations[34,35]
Gas adsorption
To assess the permanent porosity, the N2 sorption isotherms of
the activated FJU-21a and FJU-22a materials were examined
at 77 K (Figure 2), which yielded a reversible type I isotherm for
the microporous nature of the samples with Brunauer–
Emmett–Teller (BET) surface areas of 369.10 and 828.19 m2gÀ1,
respectively. FJU-21a shows a bimodal pore size distribution
centered at 5.2 and 8.7 Š, and FJU-22a has a distribution cen-
tered at 8.0 Š, as calculated by the non-local (NL)-DFT method;
these values are close to the pore sizes determined from the
crystal structures (Figure S1 in the Supporting Information). Al-
though their void volumes from the Platon calculations are
close, the BET surface area for FJU-21a, with the dynamic
framework, is only about half that for FJU-22a. The flexible
character of FJU-21 is further confirmed by a hysteresis loop in
the N2 adsorption isotherm at 77 K.
The unique pore structures encouraged us to examine the
capacities of the two MOFs for gas adsorption. The low-pres-
sure sorption isotherms of CO2, C2H2, and C2H4 were collected
at 273 and 296 K (Figure 2 and Figure S3 in the Supporting In-
formation). At 296 K and 1 bar, FJU-22a can adsorb 111.3,
114.8, and 85.8 cm3gÀ1 of CO2, C2H2, and C2H4, respectively. The
adsorption isotherms for C2H2, CO2, and C2H4 on FJU-21a are
very similar to those for FJU-22a and the adsorption capacity
follows the same hierarchy: C2H2 >CO2 >C2H4. This phenomen-
on may be attributed to the same pore surface structure re-
sulting from the same metal node and ligand connection
mode. However, the halved BET surface area for FJU-21a
makes its various gas uptakes fall to half the corresponding
values of FJU-22a. Furthermore, it is worth noting that the
acetylene uptake isotherms for FJU-21a and FJU-22a at 296 K
show a very sharp uptake at low pressure, whereas carbon di-
oxide uptake is much lower at this pressure. This discovery
motivated us to examine their feasibility for the industrially im-
portant C2H2/CO2 separation.
3
of FJU-21 and FJU-22 indicate their void volumes are 923.7 Š
3
(52.1% of the unit cell volume of 1773.3 Š3) and 1908 Š
(52.8% of the unit cell volume of 3614.2 Š3), respectively.
Owing to the same metal node and metal–ligand connec-
tion mode, FJU-21 and FJU-22 have similar pore surface struc-
tures. Nevertheless, it is worth noting that the solvent-induced
structural diversity gives the two MOFs distinct robustness
properties. The two as-synthesized MOFs were exchanged with
CH3OH and CH2Cl2, respectively, several times, then heated to
608C, and evacuated under high vacuum to obtain the desol-
vated frameworks FJU-21a and FJU-22a. FJU-21a is flexible,
whereas FJU-22a shows good robustness, as proved by
powder X-ray diffraction experiments (PXRD; Figure S2 in the
Supporting Information). For FJU-22, with the one unique type
of helical chain SBU, the 2q values are not shifted for the acti-
vated sample compared to the as-synthesized sample pattern.
However, for FJU-21, which contains one more type of helical
chain SBU, the values of 2q for the (100) and (002) planes are
shifted to higher angles for the activated sample, and no shift
for the (020) plane is seen, indicating that the dynamic fea-
tures are down to the [Cu2(isophthalate)4] monolayers oriented
at the (100) plane and constructed from the helical chains ex-
clusively in FJU-21 and not observed in FJU-22. In addition, if
exposed to air or water vapor, the values of 2q for FJU-21 are
shifted, whereas for FJU-22 there is no obvious change under
the same conditions, further indicating that FJU-22 has better
stability than FJU-21. Although several methods including
those using high-valent metal ions,[36] modulated synthesis,[37]
N-donor ligands,[38] and superhydrophobic ligands[39] have
been proposed to enhance MOF stability, FJU-21 and FJU-22
are the first examples to demonstrate control of MOF stability
and robustness by adjusting the helical chain SBUs.
C2H2/CO2 column breakthrough experiments
We first performed breakthrough simulations for a 50:50 (v/v)
C2H2/CO2 mixture on FJU-21a and FJU-22a by using the estab-
lished methodology.[40] As shown in Figure S4 (in the Support-
ing Information), FJU-21a and FJU-22a are able to separate
C2H2 from the C2H2/CO2 mixture at room temperature. Clearly,
FJU-22a, with good robustness, is more effective than FJU-
21a for the C2H2/CO2 separation. Thus, we only studied the
actual performance of FJU-22a in the experimental column
breakthrough.
In the actual column breakthrough experiment, an equimo-
lar C2H2/CO2 mixture was flowed over a packed column of the
FJU-22a solid with a total flow of 5 cm3 minÀ1 at 296 K
(Figure 3). CO2 was detected after the gas mixture has been in-
troduced into the column for about 12 min, whereas C2H2 was
not detected until a breakthrough time of 23 min was reached.
Thus, the separation of C2H2/CO2 mixture gases through
a column packed with FJU-22a solid can be achieved efficient-
ly. The breakthrough times of CO2 and C2H2 on the unique
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5678
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. Experimental column breakthrough curve for an equimolar C2H2/
CO2 mixture in an adsorber bed packed with FJU-22a (c) or HOF-3a
(a) at ambient conditions (296 K, 1 bar). The total flow is of 5 cm3 minÀ1
.
pore material HOF-3 were 5.5 and 9 min, respectively.[33] From
the breakthrough curve, the separation selectivity, a=(q1y2)/
(y1q2), for FJU-22a is 1.9, and close to 2.0 for HOF-3. However,
FJU-22a has a separation capacity of 44.13 cm3gÀ1, which is
much higher than the value of 25.54 cm3gÀ1 for HOF-3. At the
same separation conditions, FJU-22a, with the open O donors,
shows better performance than HOF-3, containing functional
amino groups. FJU-22a is the first example of a MOF whose
separation of a C2H2/CO2 mixture has been clearly established
by column breakthrough experiments.
First-principles calculations
To further understand the C2H2 and CO2 adsorption on FJU-
22a, detailed computational investigations were performed.
We first optimized the bare FJU-22a structure by first-princi-
ples DFT-D (dispersion-corrected density-functional theory) cal-
culations,[41] and then carried out Grand Canonical Monte Carlo
(GCMC) simulations by using the classical force-field method.
Based on the probability distribution of adsorbed gas mole-
cules generated from the GCMC simulations, we introduced
C2H2/CO2 accordingly into the FJU-22a channel pore, and fur-
ther optimized the structures by using DFT-D. We found that
the guest molecules are associated with particular adsorption
sites. Upon adsorption, both C2H2 and CO2 sit right at the small
cage connecting the two adjacent channel pores (Figure 4 and
Figure S5 in the Supporting Information). Although the linker
triazole ring has van der Waals (vdW) interactions with C2H2,
the relatively strong binding between C2H2 and FJU-22a clear-
ly comes from the hydrogen-bonding interactions between
the C2H2 and the framework O (d[O(-CO2)···H(C2H2)]=2.290 Š);
this interaction does not exist between CO2 and FJU-22a and
CO2 adsorption in the structure is mainly vdW-type in nature.
The static C2H2 and CO2 binding energies for FJU-22a, derived
from the DFT-D calculations, are 33.3 and 22.6 kJmolÀ1, respec-
tively. This difference in the gas/MOF framework interaction
strength is likely the reason why the performance of separa-
tion for C2H2/CO2 is outstanding in FJU-22a at room tempera-
ture.
Figure 2. N2 sorption isotherm and pore-size distributions at 77 K of FJU-
21a (a) and FJU-22a (b). Adsorption isotherms for C2H2, CO2, and C2H4 on
FJU-21a (c) and FJU-22a (d) at 296 K (solid symbols: adsorption; open sym-
bols: desorption).
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5679
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 4. The C2H2 (a) and CO2 (b) molecules sit right at the small cage con-
necting the two adjacent channel pores. Multiple-point interactions of the
C2H2 with framework: d[O(-CO2)···H(C2H2)]=2.290 Š, d[center(-triazolyl)-
···C(C2H2)]=3.563 Š; multiple-point interactions between CO2 molecule and
the framework: d[O(-CO2)···C(CO2)]=3.316 Š, d[C(-benzene)-
Figure 6. Selectivity for C2H2/C2H4 mixtures of some representative MOFs.
series of MOF-74 with high acetylene uptakes, but their selec-
tivities for C2H2/C2H4 separation are systemically quite low[3] as
the open metal sites have quite strong interactions toward
both ethylene and acetylene molecules. NOTT-300 has multiple
weak supramolecular interactions aligned within the host to
form an optimal geometry for the selective binding of hydro-
carbons, but the selectivity toward C2H2/C2H4 is still very low.[4]
The multiple supramolecular interactions in NOTT-300 include
the hydrogen-bonding interactions between C(C2H2 or C2H4)
and H-O(framework), which are clearly distinct from those be-
tween O(framework) and H-C(C2H2) in FJU-22a. The former
cannot recognize acetylene and ethylene very well, whereas
the latter endows FJU-22a with exceptionally high selectivity.
UTSA-100a,[5] containing amino groups with dual functionali-
ties to simultaneously bind acetylene and sieve ethylene, has
a relative high selectivity, about five times that for MOFs con-
taining OMS. It is worth noting that FJU-22 is isostructural
with UTSA-100. The enthalpies of acetylene adsorption on
both MOFs are almost same, but the separation selectivity, a,
of FJU-22a is more than twice that for UTSA-100a, indicating
that even in UTSA-100a (Figure S6 in the Supporting Informa-
tion), the main contribution for its high selectivity toward acet-
ylene may also come from the open O donor sites on the
framework wall, rather than the amino groups. Conversely, the
extra amino groups of UTSA-100a, to a certain extent, de-
crease its static C2H2 uptake. Despite its moderate static acety-
lene uptake, FJU-22a has the highest separation selectivity for
the C2H2/C2H4 mixture among the reported porous materials.
Based on the discussion above, open O donors on the pore
wall can provide MOFs with better recognition for C2H2/C2H4
mixtures than the OMS and amino groups. FJU-22a is thus su-
perior to the other MOFs, exhibiting highly efficient removal of
acetylene from ethylene/acetylene mixtures containing 1%
acetylene.
···O(CO2)]=3.389 Š. Color code: Cu, green; C, gray; H, yellow; O, red; N, blue.
C2H2/C2H4 breakthrough experiments
The performance of FJU-22a for the removal of acetylene
from C2H2/C2H4 mixtures containing 1% acetylene was exam-
ined through experimental column breakthrough in which
a C2H2/C2H4 (1:99, v/v) mixture was flowed over a packed
column of the FJU-22a solid with a total flow of 1.8 cm3 minÀ1
at 296 K. As shown in Figure 5, the separation of the C2H2/C2H4
Figure 5. Experimental column breakthrough curve for an C2H2/C2H4 (1:99,
v/v) mixture in an adsorber bed packed with FJU-22a (c) and UTSA-100a
(a) at ambient conditions (296 K, 1 bar). The total flows are 1.8 cm3 minÀ1
for FJU-22a and 2 cm3 minÀ1 for UTSA-100a.
mixture gases through a column packed with FJU-22a solid
can be efficiently achieved, and the separation selectivity, a,
for C2H2/C2H4 is 25.8. The adsorption and separation data for
C2H2 and C2H4 gases on FJU-22a and some representative
MOFs are given in Figure 6 and Table S4 (in the Supporting In-
formation). M’MOF-3a, with a flexible framework, exhibits rela-
tive high separation selectivity (24.0), but very low acetylene
uptakes, because of narrow pores and high sieving effects.[2]
High densities of open metal sites can significantly endow the
Conclusions
We have observed solvent-induced topological diversity ena-
bling controllable robustness in two metal–organic frameworks
(FJU-21 and FJU-22) and have demonstrated that microporous
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5680
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
{[Cu(L)]·(DMA)(H2O)1.5
}
(FJU-22): FJU-22 was obtained by the
MOFs with open O sites are highly selective for the separation
of C2H2/CO2 and C2H2/C2H4 at ambient conditions. Control of
the MOFs’ robustness through tuning the helical chain SBUs is
shown to be an efficient design approach for the first time.
The good robustness of the activated FJU-22a, with open O
donors, results in the extraordinary separation performance for
mixtures of C2H2/CO2 and C2H2/C2H4; this performance is supe-
rior to all the reported MOFs including the flexible FJU-21a.
The separation capacity of FJU-22a for 50:50 C2H2/CO2 mix-
tures is about twice that of HOF-3, the unique example for
separation before our experiments. The actual selectivity of
FJU-22a for C2H2/C2H4 mixtures containing 1% C2H2 is highest
among the reported porous materials. Such preferential ad-
sorption for C2H2 by FJU-22a rather than CO2 and C2H4 is at-
tributed to the open O donor sites on its framework wall.
Open O donors can provide MOFs with better recognition abil-
ity for C2H2 than other functional sites, including open metal
sites (OMS) and amino groups. It is expected that extensive re-
search endeavors on porous MOFs will facilitate the discoveries
of better C2H2 separation materials.
n
same procedure used for preparation of FJU-21 except that the
DMF was replaced with DMA. Green bulk crystals of FJU-22 were
obtained (37% yield with regard to H2L). Elemental analysis calcd
(%) for FJU-22: C 41.08, H 4.16, N 13.69; found: C 41.78, H 4.21, N
14.05.
Single-crystal X-ray structure determination
Data collection and structural analysis of the crystals were collect-
ed on an Agilent Technologies SuperNova Single Crystal Diffrac-
tometer equipped with graphite monochromatic Cu Ka radiation
(l=1.54184 Š). The crystal was kept at 293(10) K during data col-
lection. Using Olex2,[44] the structure was solved with the Superflip
structure solution program by using charge flipping and refined
with the ShelXL refinement package by using least-squares minimi-
zation. The hydrogen atoms on the ligands were placed in ideal-
ized positions and refined by using a riding model. We employed
PLATON[34]/SQUEEZE[35] to calculate the diffraction contribution of
the solvent molecules in FJU-21 and FJU-22 and thereby produce
a set of solvent-free diffraction intensities. The formulae of the
crystals were calculated by elemental analysis. The detailed crystal-
lographic data and structure refinement parameters for these com-
pounds are summarized in the Supporting Information, Table S1.
CCDC 1421052 (FJU-21) and 1421054 (FJU-22) contain the supple-
mentary crystallographic data for this paper. These data are provid-
ed free of charge by The Cambridge Crystallographic Data Centre.
Experimental Section
General
All reagents and solvents were used as received from commercial
suppliers without further purification. Thermogravimetric analyses
(TGA) were performed with a Mettler Toledo TGA/SDTA851e ana-
lyzer with a nitrogen flow and a heating rate of 10 KminÀ1 from
308C to 6008C. Elemental analysis was collected with a Vario EL III
elemental analyzer to give a ratio of C/H/N. Powder X-ray diffrac-
tion (PXRD) was carried out with a PANalytical X’Pert3 powder dif-
fractometer equipped with a Cu sealed tube (l=1.54178 Š) at
40 kV and 40 mA over the 2q range 5–258.
Gas adsorption measurements
A Micromeritics ASAP 2020 surface area analyzer was used to mea-
sure the gas adsorption. The sorption measurements were per-
formed at 77 K with liquid nitrogen and at 273 K with an ice/water
bath (slush). A water bath was used for adsorption isotherms at
296 K. DFT pore size distributions and pore sizes were calculated
from the N2 adsorption at 77 K.
Synthesis
Transient breakthrough simulations
N,N-Dimethylformamide azine dihydrochloride (DMAz): The
DMAz used in this study was synthesized by a modified version of
the method in refs. [42,43]. Thionyl chloride (SOCl2, 28.6 mL,
0.4 mol) was added with stirring to DMF (150 mL) at 58C. After ad-
dition, this mixture was kept at 58C for 24 h and then aqueous hy-
drazine hydrate (5 mL, 0.1 mol) in DMF (20 mL) was added slowly.
After addition, the mixture was stirred at room temperature for
48 h and the white precipitate of N,N-dimethylformamide azine di-
hydrochloride was collected by filtration and washed with DMF
and diethyl ether. Yield: 19.1 g; m.p.: 2518C.
The performance of industrial fixed-bed adsorbers is dictated by
a combination of adsorption selectivity and uptake capacity. For
a proper comparison of various MOFs, we performed transient
breakthrough simulations by using the simulation methodology.[40]
For the breakthrough simulations, the following parameter values
were used: length of packed bed, L=0.3 m; voidage of packed
bed=0.4; superficial gas velocity at inlet, u=0.04 msÀ1. The frame-
work density of FJU-21 is 1104 kgmÀ3 and for FJU-22 it is
1083 kgmÀ3
.
5-Triazole isophthalic acid (H2L): A mixture of N,N-dimethylform-
amide azine dihydrochloride (4.0 g, 18.66 mmol) and 5-amino iso-
phthalic acid (3.38 g, 18.66 mmol) was heated at reflux in 1,2-dime-
thylbenzene (50 mL) for 16 h to obtain a white solid. The solid was
filtered and washed with ethanol (2”15 mL) and diethyl ether (1”
17 mL). Yield: 1.62 g, 37.3%.
Column breakthrough test set-up, procedures, and measure-
ments
The mixed-gas breakthrough separation experiment was conduct-
ed at 296 K by using a laboratory-scale fix-bed reactor. In a typical
breakthrough experiment (Figure S7 in the Supporting Informa-
tion) for the C2H2/CO2 mixture, FJU-22 powder (1.2 g) was packed
into a quartz column (5.8 mm I.D.”150 mm) with silica wool filling
the void space. The sorbent was activated in situ in the column
{[Cu(L)](DMF)(H2O)1.5
}
n
(FJU-21): A mixture of CuI (0.1 mmol,
0.0191 g), H2L (0.1 mmol, 0.0223 g), DMF (3 mL), and H2O (2 mL)
was stirred for 10 min. Then, the solution was transferred to
a 23 mL glass reactor and heated to 858C. After 24 h, the system
was cooled to room temperature and green needle-like crystals
were obtained (32% yield with regard to H2L). Elemental analysis
calcd (%) for FJU-21: C 39.49, H 3.80, N 14.18; found: C 40.23, H
3.87, N 14.42.
with
a vacuum pump at 333 K for 24 h. A helium flow
(5 cm3 minÀ1) was introduced after the activation process to purge
the adsorbent. The flow of He was then turned off while a gas mix-
ture of C2H2/CO2 (50:50, v/v) at 5 cm3 minÀ1 was allowed to flow
into the column. The effluent from the column was monitored by
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5681
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
using a mass spectrometer (MS). The absolute adsorbed amount of
gas i (qi) is calculated from the breakthrough curve by Equation (1):
Acknowledgments
This work was financially supported by the National Natural
Science Foundation of China (21207018, 21273033, 21203024,
and 21573042) and the Fujian Science and Technology Depart-
ment (2014J06003 and 2014H6007). S.X. gratefully acknowl-
edges the support of the Recruitment Program of Global
Young Experts, Program for New Century Excellent Talents in
University (NCET-10-0108), and the Award ‘MinJiang Scholar
Program’ in Fujian Province.
t0
R
Fi xt0ÀVdead
À
FeDt
ð1Þ
0
qi ¼
m
in which Fi is the influent flow rate of the specific gas (cm3 minÀ1),
t0 is the adsorption time (min), Vdead is the dead volume of the
system (cm3), Fe is the effluent flow rate of the specific gas
(cm3 minÀ1), and m is the mass of the sorbent (g). The separation
factor (a) of the breakthrough experiment is determined as:
Keywords: column
breakthrough
·
metal–organic
frameworks · open O donors · selective gas adsorption ·
q1y2
y1q2
structural diversity
a ¼
ð2Þ
in which yi is the molar fraction of gas i in the gas mixture. The
column breakthrough measurement for C2H2/C2H4 (1:99, v/v) mix-
tures was similar to the measurements for the C2H2/CO2 mixture,
[1] F. Studt, F. Abild-Pedersen, T. Bligaard, R. Z. Sørensen, C. H. Christensen,
J. K. Nøskov, Science 2008, 320, 1320–1322.
[2] S. C. Xiang, Z. J. Zhang, C. G. Zhao, K. L. Hong, X. B. Zhao, D. R. Ding,
M. H. Xia, C. D. Wu, M. C. Das, R. Gill, K. M. Tomoas, B. L. Chen, Nat.
Commun. 2011, 2, 204.
except that the helium flow was changed to 1.8 cm3 minÀ1
.
[3] a) E. D. Bloch, W. L. Queen, R. Krishna, J. M. Zadrozny, C. M. Brown, J. R.
Long, Science 2012, 335, 1606–1610; b) Y. B. He, R. Krishna, B. L. Chen,
Energy Environ. Sci. 2012, 5, 9107–9120.
[4] S. H. Yang, A. J. Ramirez-Cuesta, R. Newby, V. Garcia-Sakai, P. Manuel,
S. K. Callear, S. I. Campbell, C. C. Tang, M. Schrçder, Nat. Chem. 2014, 7,
121–129.
[5] T. L. Hu, H. L. Wang, B. Li, R. Krishna, H. Wu, W. Zhou, Y. F. Zhao, Y. Han,
X. Wang, W. D. Zhu, Z. Z. Yao, S. C. Xiang, B. L. Chen, Nat. Commun.
2015, 6, 7328.
[6] B. M. Collins, US Patent 4126645, 1978.
[7] W. Huang, J. R. McCormick, R. F. Lobo, J. G. Chen, J. Catal. 2007, 246,
40–51.
[8] a) B. M. Choudary, M. Lakshmi Kantam, N. Mahender Reddy, K. Koteswar-
a Rao, Y. Haritha, V. Bhaskar, F. Figueras, A. Tuel, Appl. Catal. A 1999,
181, 139–144; b) N. A. Khan, S. Shaikhutdinov, H. J. Freund, Catal. Lett.
2006, 108, 159–164.
[9] K. Weissermel, H. J. Arpe, Industrial Organic Chemistry, 4th ed., 91–98,
Wiley-VCH, Weinheim, 2003.
Details of DFT-D calculations and GCMC simulations
First-principles calculations based on density-functional theory
were performed by using the PWSCF package.[41] A semiempirical
addition of dispersive forces to conventional DFT[45] was included
in the calculation to account for van der Waals interactions. We
used Vanderbilt-type ultrasoft pseudopotentials and the general-
ized gradient approximation (GGA) with the Perdew–Burke–Ernzer-
hof (PBE) exchange correlation. A cutoff energy of 544 eV and a 2”
2”2 k sampling were sufficient for the total energy to converge
within 0.5 meV per atom. We first optimized the bare FJU-22 struc-
ture. The optimized structure is fairly close to the experimental
structure determined from XRD. C2H2 or CO2 molecules were then
introduced to the optimized HOF structure (guided by the GCMC
result), followed by a full structural relaxation. To obtain the gas
binding energies, a free gas molecule placed in a supercell with
the same cell dimensions was also relaxed as a reference. The
static binding energy was then calculated by using: EB =[E(MOF)+
nE(gas)ÀE(MOF+ngas)]/n. Grand Canonical Monte Carlo (GCMC)
simulations[46] were performed for C2H2/CO2 adsorption on FJU-22,
with the gas molecules and the MOF frameworks both treated as
rigid bodies. A 2”2”2 supercell was used as the simulation box to
ensure the simulation accuracy. A total of 2”107 steps were used
for equilibration and an additional 2”107 steps were used to calcu-
late the ensemble average of gas adsorption sites and thermody-
namic properties. We used the standard universal force field
(UFF)[47] to describe the gas–framework interaction and the gas–
gas interaction. Atomic partial charges derived from first-principles
calculations were included in the simulation to account for electro-
static interactions. The cut-off radius used for the Lennard–Jones
interactions is 12.8 Š. The long-range electrostatic interactions
were treated by using the Ewald summation technique with tin-foil
boundary condition. Simulations were performed at various tem-
peratures and pressures. The probability distributions of adsorbed
C2H2 and CO2 were generated from the simulation after the equilib-
rium stage.
[10] a) H. Furukawa, K. E. Cordova, M. O’Keeffe, O. M. Yaghi, Science 2013,
341, 1230444; b) J. R. Long, O. M. Yaghi, Chem. Soc. Rev. 2009, 38, 1213–
1214.
[11] a) H. C. Zhou, S. Kitagawa, Chem. Soc. Rev. 2014, 43, 5415–5418; b) H.
Sato, W. Kosaka, R. Matsuda, A. Hori, Y. Hijikata, R. V. Belosludov, S.
Sakaki, M. Takata, S. Kitagawa, Science 2014, 343, 167–170.
[12] G. FØrey, C. Serre, Chem. Soc. Rev. 2009, 38, 1380–1399.
[13] a) S. J. Bao, R. Krishna, Y. B. He, J. S. Qin, Z. M. Su, S. L. Li, W. Xie, D. Y. Du,
W. W. He, S. R. Zhang, Y. Q. Lan, J. Mater. Chem. A 2015, 3, 7361–7367;
b) S. L. Li, Y. Q. Lan, H. Sakurai, Q. Xu, Chem. Eur. J. 2012, 18, 16302–
16309; c) Q. L. Zhu, Q. Xu, Chem. Soc. Rev. 2014, 43, 5468–5512; d) J. K.
Sun, Q. Xu, Energy Environ. Sci. 2014, 7, 2071–2100; e) J. S. Li, Y. J. Tang,
S. L. Li, S. R. Zhang, Z. H. Dai, Z. H. Si, Y. Q. Lan, CrystEngComm 2015, 17,
1080–1085; f) D. Y. Du, J. S. Qin, S. L. Li, Z. M. Su, Y. Q. Lan, Chem. Soc.
Rev. 2014, 43, 4615–4632.
[14] R. Matsuda, R. Kitaura, S. Kitagawa, Y. Kubota, R. V. Belosludov, T. C. Ko-
bayashi, H. Sakamoto, T. M. Takata, Y. Kawazoe, T. Chiba, Y. Mita, Nature
2005, 436, 238–241.
[15] a) J. B. Lin, W. Xue, J. P. Zhang, X. M. Chen, Chem. Commun. 2011, 47,
926–928; b) P. Q. Liao, H. Y. Chen, D. D. Zhou, S. Y. Liu, C. T. He, Z. B. Rui,
H. B. Ji, J. P. Zhang, X. M. Chen, Energy Environ. Sci. 2015, 8, 1011–1016.
[16] a) A. Samanta, A. Zhao, G. K. H. Shimizu, P. Sarkar, R. Gupta, Ind. Eng.
Chem. Res. 2012, 51, 1438–1463; b) R. Vaidhyanathan, S. S. Iremonger,
G. K. H. Shimizu, P. G. Boyd, S. Alavi, T. K. Woo, Angew. Chem. Int. Ed.
2012, 51, 1826–1829; Angew. Chem. 2012, 124, 1862–1865; c) R. Vaid-
hyanathan, S. S. Iremonger, K. W. Dawsona, G. K. H. Shimizu, Chem.
Commun. 2009, 5230–5232.
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5682
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[17] a) X. Zhao, X. H. Bu, Q. G. Zhai, H. Tran, P. Y. Feng, J. Am. Chem. Soc.
2015, 137, 1396–1399; b) Q. G. Zhai, Q. Lin, T. Wu, L. Wang, S. T. Zheng,
X. H. Bu, P. Y. Feng, Chem. Mater. 2012, 24, 2624–2626; c) Q. Lin, T. Wu,
S. T. Zheng, X. H. Bu, P. Y. Feng, J. Am. Chem. Soc. 2012, 134, 784–787.
[18] a) J. G. Duan, W. Q. Jin, R. Krishna, Inorg. Chem. 2015, 54, 4279–4284;
b) Y. P. Chen, Y. Y. Liu, D. H. Liu, M. Bosch, H. C. Zhou, J. Am. Chem. Soc.
2015, 137, 2919–2930; c) C. Wang, D. M. Liu, W. B. Lin, J. Am. Chem. Soc.
2013, 135, 13222–13234; d) W. J. Wang, D. Q. Yuan, Sci. Rep. 2014, 4,
5711.
[19] a) J. Y. Lee, L. Pan, X. Y. Huang, T. J. Emge, J. Li, Adv. Funct. Mater. 2011,
21, 993–998; b) H. Liu, Y. G. Zhao, Z. J. Zhang, N. Nijem, Y. J. Chabal,
H. P. Zeng, J. Li, Adv. Funct. Mater. 2011, 21, 4754–4762; c) Y. Chen, H.
Wang, J. Li, J. V. Lockard, J. Mater. Chem. A J. Mater. Chem. A. 2015, 3,
4945–4953; d) H. Wang, K. Yao, Z. Zhang, J. Jagiello, Q. Gong, Y. Han, J.
Li, Chem. Sci. 2014, 5, 620–624.
[20] a) X. Duan, J. F. Cai, J. C. Yu, C. D. Wu, Y. J. Cui, Y. Yang, G. D. Qian, Micro-
porous Mesoporous Mater. 2013, 181, 99–104; b) T. F. Xia, J. F. Cai, H. Z.
Wang, X. Duan, Y. J. Cui, Y. Yang, G. D. Qian, Microporous Mesoporous
Mater. 2015, 215, 109–115.
[21] M. Fischer, F. Hoffmann, M. Froeba, ChemPhysChem 2010, 11, 2220–
2229.
[22] J. W. Yoon, J. S. Lee, S. Lee, K. H. Cho, Y. K. Hwang, M. Daturi, C. H. Jun,
R. Krishna, J. S. Chang, Chem. Eur. J 2015, 50, 18431–18438.
[23] H. M. Wen, B. Li, H. L. Wang, R. Krishna, B. L. Chen, Chem. Commun.
2016, 52, 1166–1169.
[24] Y. B. He, S. C. Xiang, B. L. Chen, J. Am. Chem. Soc. 2011, 133, 14570–
14573.
[32] a) Y. B. He, S. C. Xiang, Z. J. Zhang, S. S. Xiong, F. R. Fronczek, R. Krishna,
M. O’Keeffe, B. L. Chen, Chem. Commun. 2012, 48, 10856–10858; b) H.
Xu, Y. B. He, Z. J. Zhang, S. C. Xiang, J. F. Cai, Y. J. Cui, Y. Yang, G. D. Qian,
B. L. Chen, J. Mater. Chem. A 2013, 1, 77–81; c) J. P. Zhang, X. M. Chen,
J. Am. Chem. Soc. 2009, 131, 5516–5521; d) M. Bülow, C. J. Guo, D. G.
Shen, F. R. Fitch, A. I. Shirley, A. I. L. Cava, J. P. Brooks, US Patent
6024781, 2000.
[33] P. Li, Y. B. He, Y. F. Zhao, L. H. Weng, H. L. Wang, R. Krishna, H. Wu, W.
Zhou, M. O’Keeffe, Y. Han, B. L. Chen, Angew. Chem. Int. Ed. 2015, 54,
574–577; Angew. Chem. 2015, 127, 584–587.
[34] A. L. Speka, J. Appl. Crystallogr. 2003, 36, 7–13.
[35] A. L. Spek, Acta Crystallogr. Sect. A 1990, 46, C34.
[36] a) H. Chevreau, T. Devic, F. Salles, G. Maurin, N. Stock, C. Serre, Angew.
Chem. Int. Ed. 2013, 52, 5056–5060; Angew. Chem. 2013, 125, 5160–
5164; b) L. Valenzano, B. Civalleri, S. Chavan, Chem. Mater. 2011, 23,
1700–1718.
[37] A. Schaate, P. Roy, A. Godt, J. Lippke, F. Waltz, M. Wiebcke, P. Behrens,
Chem. Eur. J. 2011, 17, 6643–6651.
[38] V. Colombo, S. Galli, H. J. Choi, G. D. Han, A. Maspero, G. Palmisano, N.
Masciocchic, J. R. Long, Chem. Sci. 2011, 2, 1311–1319.
[39] C. Yang, U. Kaipa, Q. Z. Mather, X. P. Wang, V. Nesterov, A. F. Venero,
M. A. Omary, J. Am. Chem. Soc. 2011, 133, 18094–18097.
[40] R. Krishna, RSC Adv. 2015, 5, 52269–52295.
[41] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D.
Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. Fabris, G.
Fratesi, S. de Gironcoli, R. Gebauer, U. Gerstmann, C. Gougoussis, A.
Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Mazzarello, S.
Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero,
A. P. Seitsonen, A. Smogunov, P. Umari, R. M. Wentzcovitch, J. Phys. Con-
dens. Matter 2009, 21, 395502.
[25] M. C. Das, Q. S. Guo, Y. B. He, J. Kim, C. G. Zhao, K. Hong, S. C. Xiang,
Z. J. Zhang, K. M. Thomas, R. Krishna, B. L. Chen, J. Am. Chem. Soc. 2012,
134, 8703–8710.
[26] S. J. Lee, J. W. Yoon, Y. K. Seo, M. B. Kim, S. K. Lee, U. H. Lee, Y. K. Hwang,
Y. S. Bae, J. S. Chang, Microporous Mesoporous Mater. 2014, 193, 160–
165.
[42] T. Panda, T. Kundu, R. Banerjee, Chem. Commun. 2012, 48, 5464–5466.
[43] L. Wen, W. Shi, X. T. Chen, H. H. Li, P. Cheng, Eur. J. Inorg. Chem. 2012,
22, 3562–3568.
[27] H. M. Wen, B. Li, H. L. Wang, C. Wu, K. Alfooty, R. Krishnad, B. L. Chen,
Chem. Commun. 2015, 51, 5610–5613.
[44] a) G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122; b) A. V.
Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard, H. Puschmann, J.
Appl. Crystallogr. 2009, 42, 339–341.
[45] V. Barone, M. Casarin, D. Forrer, M. Pavone, M. Sambi, A. J. Vittadini,
Comput. Chem. 2009, 30, 934–939.
[46] D. Frenkel, B. Smit, Computational Science Series, Academic Press, San
Diego, 2002.
[47] A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard III, W. M. Skid, J.
Am. Chem. Soc. 1992, 114, 10024–10035.
[28] Z. J. Zhang, S. C. Xiang, B. L. Chen, CrystEngComm 2011, 13, 5983–5992.
[29] a) Z. J. Zhang, Z. Z. Yao, S. C. Xiang, B. L. Chen, Energy Environ. Sci. 2014,
7, 2868–2899; <lit b>S. C. Xiang, Y. B. He, Z. J. Zhang, H. Wu, W.
Zhou, R. Krishna, B. L. Chen, Nat. Commun. 2012, 3, 954.
[30] a) S. C. Xiang, W. Zhou, Z. J. Zhang, M. A. Green, Y. Liu, B. L. Chen,
Angew. Chem. Int. Ed. 2010, 49, 4615–4618; Angew. Chem. 2010, 122,
4719–4722; b) S. C. Xiang, W. Zhou, J. M. Gallegos, Y. Liu, B. L. Chen, J.
Am. Chem. Soc. 2009, 131, 12415–12419; c) C. L. Song, J. Y. Hu, Y. J.
Ling, Y. L. Feng, D. L. Chen, Y. B. He, Dalton Trans. 2015, 44, 14823–
14829.
Received: December 21, 2015
[31] J. D. Pang, F. L. Jiang, M. Y. Wu, C. P. Liu, K. Z. Su, W. G. Lu, D. Q. Yuan,
M. C. Hong, Nat. Commun. 2015, 6, 7575.
Published online on March 2, 2016
Chem. Eur. J. 2016, 22, 5676 – 5683
www.chemeurj.org
5683
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim