Ethylene/ethane separation in a stable hydrogen-bonded organic framework through a gating mechanism

Article abstract of DOI:10.1038/s41557-021-00740-z

Porous materials are very promising for the development of cost- and energy-efficient separation processes, such as for the purification of ethylene from ethylene/ethane mixture—an important but currently challenging industrial process. Here we report a microporous hydrogen-bonded organic framework that takes up ethylene with very good selectivity over ethane through a gating mechanism. The material consists of tetracyano-bicarbazole building blocks held together through intermolecular CN···H–C hydrogen bonding interactions, and forms as a threefold-interpenetrated framework with pores of suitable size for the selective capture of ethylene. The hydrogen-bonded organic framework exhibits a gating mechanism in which the threshold pressure required for guest uptake varies with the temperature. Ethylene/ethane separation is validated by breakthrough experiments with high purity of ethylene (99.1%) at 333 K. Hydrogen-bonded organic frameworks are usually not robust, yet this material was stable under harsh conditions, including exposure to strong acidity, basicity and a variety of highly polar solvents. [Figure not available: see fulltext.]

Full text of DOI:10.1038/s41557-021-00740-z

Articles
https://doi.org/10.1038/s41557-021-00740-z
Ethylene/ethane separation in a stable
hydrogen-bonded organic framework through a
gating mechanism
Yisi Yangꢀ ꢀ1,4, Libo Liꢀ ꢀ2,4, Rui-Biao Linꢀ ꢀ3,4, Yingxiang Yeꢀ ꢀ1,3, Zizhu Yao1, Ling Yang2, Fahui Xiangꢀ ꢀ1,
1
1
ꢀ✉
1
ꢀ✉
3
ꢀ✉
Shimin Chen , Zhangjing Zhangꢀ ꢀ , Shengchang Xiangꢀ ꢀ and Banglin Chenꢀ ꢀ
Porous materials are very promising for the development of cost- and energy-efficient separation processes, such as for the puri-
fication of ethylene from ethylene/ethane mixture—an important but currently challenging industrial process. Here we report a
microporous hydrogen-bonded organic framework that takes up ethylene with very good selectivity over ethane through a gat-
ing mechanism. The material consists of tetracyano-bicarbazole building blocks held together through intermolecular CN···H–C
hydrogen bonding interactions, and forms as a threefold-interpenetrated framework with pores of suitable size for the selective
capture of ethylene. The hydrogen-bonded organic framework exhibits a gating mechanism in which the threshold pressure
required for guest uptake varies with the temperature. Ethylene/ethane separation is validated by breakthrough experiments
with high purity of ethylene (99.1%) at 333ꢀK. Hydrogen-bonded organic frameworks are usually not robust, yet this material
was stable under harsh conditions, including exposure to strong acidity, basicity and a variety of highly polar solvents.
eparations are critical processes in chemical industry1,2. The hydrocarbon separations are particularly challenging as they involve
processes to separate and purify industrial commodities, molecules with similar dimensions. For example, the dimensions of
S
such as distillation for gases and water, account for 10–15% ethylene are about 3.28×4.18×4.84A3, whereas those of ethane are
of the global energy consumption1. The development of alternative about 3.81×4.08×4.82A3 (refs. 32,33).
and energy-efficient separation technologies is therefore highly in
Furthermore, even for porous MOFs that seem to lend them-
demand3,4. Physical adsorption in porous adsorbents that enable selves well to high sieving separations from a structural point of view
separation through pressure swing adsorption (PSA) and thermal (considering their pore structures and window sizes), it is still very
swing adsorption4 is very promising in terms of energy saving. In challenging to fulfil high separation performance due to their flex-
fact, some adsorbents have already been implemented in industrial ible nature: the pores are typically gradually enlarged under slightly
separations, for example, the titanosilicate molecular sieve ETS-4 higher pressures and therefore they also entrap the larger hydro-
for the industrial-scale separation of natural gas4,5.
carbon, leading to co-adsorption. This effect was observed in the
However, challenges and huge energy consumptions exist for the UTSA-200 MOF we recently developed for the removal of propyne
major industrial gas separations. For example, separating ethylene (C3H4) from a mixture with propylene (C3H6)34, which substantially
from ethane requires cryogenic distillation, a very energy-intensive diminished the separation performance and the purity of the sep-
but important step involving repeated distillation–compression arated product. We also reported very good sieving performance
cycling, which consumes up to about 800PJ per year6, more than for UTSA-280 for ethylene/ethane separation29. We attributed the
0.3% of annual global energy consumption1. In a common process high sieving separation to the rigidity of UTSA-280 constrained
for ethylene production, which is based on the cracking of heavier by the squarate linkers, which almost completely prevented ethane
hydrocarbon fractions, followed by dehydrogenation reactions, the molecules from entering its pores. Given that most organic linkers
conversion yield of the latter step is only around 50 to 60%. Other within MOFs will undergo—at least to some extent—rotation and
processes such as catalytic dehydrogenation also give an equimo- distortion under different stimulus such as temperatures and pres-
lar mixture of C2H4 and C2H6. The upgrading of ethylene from the sures35–37, it has remained difficult to prepare rigid MOFs for high
C2H4/C2H6 mixture through the energy-efficient adsorption separa- sieving separations of hydrocarbons.
tion would thus reap enormous global benefits.
Gate pressure adsorption phenomena are well established in
Porous metal–organic frameworks (MOFs)7–15, covalent organic porous materials38. In general, adsorption/desorption is an equilib-
frameworks (COFs)16,17 and hydrogen bonded-organic frameworks rium physical and exothermic process (Fig. 1), which means that
(HOFs)18,19 have been explored for gas separation and purifica- increasing the adsorption temperature will favour the equilibrium to
tion20–22. The fact that MOFs are so well-suited to rational pore tun- the left (less gas uptake), whereas increasing the gas pressure (con-
ing and straightforwards pore functionalization has led to good centration of gas) will favour the equilibrium to the right (more gas
performances as adsorbents for several gas separations23–28, but uptake) (Fig. 1). This means that the flexibility of porous materials is
only a few examples of MOFs with very high selectivity have been not necessarily a hindrance when it comes to the selective uptake of
realized, particularly for the hydrocarbon separations29–31. Most guests—we can make use of mechanisms such as gating phenomena
1Fujian Provincial Key Laboratory of Polymer Materials, College of Chemistry and Materials Science, Fujian Normal University, Fuzhou, China. 2College
of Chemistry and Chemical Engineering, Shanxi Key Laboratory of Gas Energy Efficient and Clean Utilization, Taiyuan University of Technology, Taiyuan,
China. 3Department of Chemistry, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX, United States. 4These authors contributed
✉
equally: Yisi Yang, Libo Li, Rui-Biao Lin. e-mail: zzhang@fjnu.edu.cn; scxiang@fjnu.edu.cn; banglin.chen@utsa.edu
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
Results and discussion
Host ⊃ G
Host + G
The tetracyano-bicarbazole organic building block (3) was obtained
from the cyanolization of 3,3′,6,6′-tetrabromo-9,9′-bicarbazole (2)
through the oxidative coupling reaction of 3,6-dibromocarbazole
Host, porous material; G, guest molecules
A
A
A
1
(Fig. 2a), which was confirmed by infrared, H NMR, 13C NMR,
B
elemental analysis and thermogravimetric analysis (TGA).
The high-quality single crystals were recrystallized from a hot
N,N′-dimethylformamide (DMF)-saturated solution of HOF-FJU-1
powders. Single-crystal X-ray diffraction studies showed that
HOF-FJU-1 crystallizes in orthorhombic space group Pnn2
(Supplementary Fig. 1 and Supplementary Table 1). In this HOF,
there are three crystallographically unique bicarbazole molecules in
its asymmetric units. Each bicarbazole unit is linked to four adja-
cent bicarbazoles by four pairs of CN···H–C hydrogen bonds with
distances of 3.431‒3.536A along the ac plane to form a single dia
network (Fig. 2b,c, and Supplementary Figs. 2 and 6).
B
B
Pressure
Fig. 1 | tuning gate-pressures mechanism for highly selective gas
separation. An illustration of the representative variations observed in
typical gas adsorption isotherms of a flexible-robust adsorbent for smaller
gas A and larger gas B at three different temperatures. The sorption
equilibrium between the host (porous material) and the guest (G) is
shown above. Blue, low temperature; green, medium temperature; red,
high temperature; solid lines, smaller gas (A); dotted lines, larger gas (B).
The black arrow indicates that the gate-opening pressure for the larger
gas B gradually increases as the temperature increases. Although A is
adsorbed at all temperatures, uptake of B is only at lower temperatures
and its threshold pressure increases (uptake decreases) with increasing
temperature. This means that the selectivity of the separation of an A/B
mixture can be tuned by the conditions (temperature, pressure).
The void space in a single network allows the final structure to
exhibit a threefold-interpenetrated array, showing distinct offset
π···π interactions along the a axis with distances of 3.799‒4.794A
that may further enhance the rigidity of this type of flexible porous
materials (Supplementary Figs. 3 and 4). Following triple inter-
penetration, there are still one-dimensional pore channels run-
ning along the crystallographic [100] direction with a void pore
volume by ~20% of the total crystal cell, in which the large cages
(3.4×6.2×10.6A3) and the small necks (3.4×3.8×5.3A3) are alter-
nating (cages I and II in Supplementary Fig. 5). The pore window
size of the channels is about 3.4×5.3A2, calculated from projec-
tion of pore channels by taking account the van der Waals radius
of atoms (Fig. 2d and Supplementary Fig. 5). The pore windows fit
well to the molecular size of ethylene (Fig. 2e), which promotes our
further investigation on potential sieving separation.
by which the opening of pores to fulfil gas separations depend
The thermal stability of HOF-FJU-1 was investigated by TGA
on guest pressures and sometimes also temperatures (Fig. 1). The and variable-temperature powder X-ray diffraction (PXRD) in an
feature of adsorbate-dependent structural flexibility in a flexible N2 atmosphere (Supplementary Figs. 7 and 8). This HOF retained
Ca(H2tcpb) MOF has been harnessed for temperature-dependent its framework following heating above 523K. Notably, after vacuum
sieving separation of linear alkanes from the branched isomers— outgassing of the as-synthesized sample at 423K for 12h, the crystal
with molecular sizes of a large difference15. The open pore space in structure of guest-free HOF-FJU-1a was directly obtained on the
flexible MOFs with permanent porosity (so-called flexible-robust basis of X-ray diffraction analysis in a single-crystal to single-crystal
MOFs39) enables them to take up smaller gas components before transformation manner, indicating the robustness of HOF-FJU-1
transforming into large-pore phases for the larger gases at relatively during desolvation. The framework with weak intermolecular
high pressures. This is different from fully flexible MOFs, in which interactions and pore structures are well preserved after guest
the condensed phases show no gas adsorption under low pressures removal, with only slight changes in parameters of crystal unit-cell
(closed pores), and then exhibit structural transformations to open (Supplementary Table 1). The minimum dimensions of window size
phases (open pores) under gate-opening pressures, typically lead- in HOF-FJU-1a (3.4‒3.8A) match well with the minimum molecu-
ing to significant co-adsorption of the counterpart gases. If such a lar dimension of C2H4 (3.28A) over C2H6 (3.81A). Accordingly, the
material exhibits permanent small pores, it should be able to take relatively robust pore space with appropriate size rendered this HOF
up a large amount of the smaller hydrocarbon (guest A in Fig. 1), a potential adsorbent with a substantial sieving effect for ethylene/
yet the co-adsorption of the larger hydrocarbon molecules (guest B) ethane separation.
through the various temperatures, at correspondingly different gate
Sorption of N2 was performed at 77K to verify the porosity
pressures, can be controlled to minimize uptake of the larger mol- of HOF-FJU-1a (Supplementary Fig. 9). Analysis of N2 adsorp-
ecules. The flexible-robust characteristic is the prerequisite to real- tion isotherms gives an experimental Brunauer–Emmett–Teller
ize such unique tunable separation.
(Langmuir) surface area of 385 (431)m2 g−1 (Supplementary Fig. 10,
Compared with MOFs and COFs, HOFs are assembled by organic and Supplementary Tables 7 and 8). The pore size distribution of
molecular building blocks through weak hydrogen bonding; HOF HOF-FJU-1a has been analysed based on Hovath–Kawazoe model,
materials can thus be easily purified through simple recrystalliza- showing peaks centred at 6.0 and 11.4A. For comparison, another
tion, showing great solution processability, recyclability and healing pore size distribution has also been calculated using the Poreblazer
capability. This type of organic material can be straightforwardly program40, indicating two types of pore cavities of 6.2 and 10.8A, in
processed into different forms such as spheres and membranes for line with the sorption results and crystallographic structure dimen-
the column and membrane gas separations, respectively.
sions (Supplementary Figs. 5 and 11). According to IUPAC clas-
Here we report a microporous HOF with tunable, temperature- sification41, HOF-FJU-1a is a microporous material with pore size
dependent gate pressures for the separation of ethylene from eth- below 20A. The amount of adsorbed N2 in the first step is 75cm3 g−1
ane at 333K. The framework, termed HOF-FJU-1, is a micropo- at P/P0 =0.001, corresponding to a pore volume of ~0.12cm3 g−1.
rous hydrogen-bonded organic framework self-assembled from the The total N2 uptake at 77K and 1bar is 100cm3 g−1, which corre-
simple organic molecule 3,3′,6,6′-tetracyano-9,9′-bicarbazole (3). sponds to a total pore volume of ~0.15cm3 g−1, consistent with the
The recovered ethylene can reach the purity of 99.1% at 333K, as theoretical value calculated from crystal structure (0.17cm3 g−1).
established by the experimental breakthrough curve.
Similar stepwise adsorption isotherm was also observed for CO2 in
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
a
Br
Br
NC
CN
H
N
N
N
KMnO4/acetone
Reflux 5 h
CuCN/dry DMF
Reflux 48 h
N
N
Br
Br
Br
Br
NC
CN
2
3
1
b
c
a
b
c
d
e
b
c
Fig. 2 | Crystal structure of hOF-FJu-1. a, A synthetic scheme of 3, the organic building block for HOF-FJU-1. b, Adjacent building blocks—and, in
turn, the extended network—are held together by intermolecular hydrogen bonding (yellow) involving the cyano groups. Grey, carbon; blue, nitrogen;
white, hydrogen. c, The threefold-interpenetrated framework in dia topology; the three independent networks are shown in blue, red, and green, both in
ball-and-stick representation and as a simplified structure (nods and rods). d, The framework (with the three interpenetrated networks shown in blue,
red and green) with pore channels along the crystallographic [100] direction. e, A schematic diagram of the size-dependent separation for C2H4 and C2H6
molecules; the ethylene (top) and ethane (bottom) guest molecules are shown in sphere representation with carbon in red and hydrogen in white. In d and
e, the pore surface is shown in yellow, framework surface in grey.
HOF-FJU-1a at 195K and 1bar (Supplementary Fig. 12). The step- is accompanied by the flexibility of the HOF framework, in con-
wise adsorption isotherms for N2 and CO2 might indicate a flex- trast to those from capillary condensation of mesopore in adsor-
ible robust feature of HOF-FJU-1a (Supplementary Figs. 9 and 12). bents with typical type-iv isotherms. On the other hand, there is no
In situ PXRD patterns of HOF-FJU-1a following CO2 loading at shifting on other diffraction peaks, which reveals certain robustness
195K were collected to understand potential structural changes of existing in HOF-FJU-1.
the stepwise sorption behaviour. Only slight peak shift at (020) corre-
We further investigated the potential of HOF-FJU-1a for C2H4/
sponding to minor expansion of pore window was identified during C2H6 separation by collecting single-component adsorption iso-
CO2 loading (Supplementary Fig. 12), which indicates the flexibil- therms for C2H4 and C2H6 at ambient conditions. At 298 K, the
ity of this HOF. We note that the stepwise adsorption behaviour C2H4 sorption isotherm of HOF-FJU-1a showed a distinct sharp
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
a
b
c
1.0
1.0
0.8
0.6
0.4
0.2
0
1.0
0.8
0.6
0.4
0.2
0
0.8
0.6
0.4
0.2
0
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
Pressure (bar)
Pressure (bar)
Pressure (bar)
d
e
f
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
0
10
20
30
40
50
60
0
10
20
30
40
50
60
0
10
20
30
40
50
60
Time (min)
Time (min)
Time (min)
g
h
i
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
0
2
4
6
8
10
0
2
4
6
8
10
0
2
4
6
8
10
Time (min)
Time (min)
Time (min)
j
k
l
100
99
100
80
1.0
0.8
0.6
0.4
0.2
0
98
Cycling test
Run 1
89
88
87
60
40
20
0
H2 5%
C3H6 5%
CH4 5%
C3H8 5%
C2H6 40%
C2H4 40%
Run 2
Run 3
Run 4
Run 5
C2H6
C2H4
74
73
72
0
5
10
15
20
25
0
5
10
15
20
25
298 K
318 K
333 K
Time (min)
Time (min)
Fig. 3 | C2h4 and C2h6 sorption and separation in hOF-FJu-1a. a–c, Gas adsorption isotherms for ethylene and ethane in HOF-FJU-1a at 298ꢀK (a), 318ꢀK
(b) and 333ꢀK (c). Filled and open symbols represent adsorption and desorption, respectively. d–f, Breakthrough curves for C2H4/C2H6 mixture (50:50, v/v)
in a fixed bed packed with HOF-FJU-1a at 298ꢀK (d), 318ꢀK (e) and 333ꢀK (f). g–i, Concentration curves of the desorbed C2H4 from HOF-FJU-1a during the
regeneration process. Desorption was carried out by applying vacuum at 298ꢀK (g), 318ꢀK (h) and 333ꢀK (i). j, Multiple cycling tests of C2H4/C2H6 (50:50,
v/v) mixtures. k, Breakthrough curves of HOF-FJU-1a for the H2/C3H6/CH4/C3H8/C2H6/C2H4 mixture (5:5:5:5:40:40, v/v/v/v/v/v) at 333ꢀK and 1ꢀbar. CA
and C0 are the concentrations of each gas at the inlet and outlet, respectively. l, The purities of the C2H4 generated from HOF-FJU-1 during the regeneration
processes of the fixed bed at different temperatures. The error bars denote s.d. of the mean for nꢀ=ꢀ3 independent experiments.
step at relatively low pressures (Fig. 3a), indicating that it was
These results indicate that HOF-FJU-1a showed high separation
well accommodated in the compact pore space of this HOF and potential for C2H4/C2H6 at relatively low pressures. Given that the
could undergo equilibrium adsorption even at low pressures and gate pressures will be increased at elevated temperatures, we further
gas concentrations. The total C2H4 uptake at 1 bar and 298 K was investigated the sorption behaviour of HOF-FJU-1a for C2H4 and
47 cm3 g−1 (2.1 mmol g−1). Surprisingly, HOF-FJU-1a showed neg- C2H6 at 318K and 333K (Fig. 3b,c). Indeed, due to the increased
ligible C2H6 uptake when the dosing pressure was below ~0.58 bar, gate-opening pressures at 333K and 1bar, HOF-FJU-1a shows a neg-
followed by a steep rise after gate-opening with a total C2H6 uptake ligible uptake of 1cm3 g−1 (0.04mmolg−1) for C2H6, which was lower
of 38 cm3 g−1 at 1 bar.
than that of UTSA-280 (0.1mmolg−1) following sieve separation29,
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
a
b
D1
D1
D2
D2
b
c
D1: 3.849(4) Å
D2: 3.979(4) Å
c
D3
D3
D4
D4
D5
D6
D6
D5
b
c
c
D3: 4.124(4) Å
D4: 3.701(4) Å
D5: 3.786(4) Å
D6: 4.001(4) Å
a
Fig. 4 | Single-crystal structure of hOF-FJu-1∙0.94 C2h4. a, Top views of the packing diagram of the C2H4-loaded structure. The framework and pore
surface are shown in grey and yellow. The HOF is shown in ball-and-stick representation; building blocks of each of the three interpenetrating networks are
highlighted in blue, green and red, and the guests are shown in spherical representation. b,c, Preferential binding sites for C2H4 molecules and their close
C-H···π (ethylene) contacts with the framework in the small neck (b) and in the large cage (c), respectively. The white spheres represent hydrogen atoms
of C2H4, and the three crystallographically unique C2H4 molecules are shown with their respective carbon atoms in green, red and blue.
whereas steady C2H4 adsorption was still observed with a total of the benchmark UTSA-280 (15) and NOTT-300 (14.5) under
uptake of 36cm3 g−1 (1.6mmolg−1). The loading-dependent isosteric the same operation conditions (Supplementary Figs. 20–22, and
heats of C2H4 adsorption in HOF-FJU-1a were calculated based on Supplementary Tables 11 and 12).
the adsorption isotherms at different temperatures to give a moder-
Although the high C2H4/C2H6 selectivity of HOF-FJU-1a is suffi-
ate value of 32kJmol−1, which is consistent with physical adsorp- cient for high-purity ethylene capture, corresponding breakthrough
tion17,42 (Supplementary Figs. 13 and 14). To further understand tests in a fixed-bed column are prerequisite to validate the produc-
different adsorption behaviours for C2H4 and C2H6 in HOF-FJU-1a, tion purity of C2H4. We conducted fixed-bed breakthrough tests at
in situ loading-dependent PXRD patterns of HOF-FJU-1a under 298K to 333K, in which the mixture of C2H4/C2H6 (50:50, v/v) was
atmosphere of C2H4 or C2H6 at related temperatures have been col- flowed over a packed column of activated HOF-FJU-1a sample, with
lected. As shown in Supplementary Figs. 15 and 16, at 298K, both a total flow of 1.25mlmin–1 at different temperatures (Fig. 3d–f). At
the C2H4- and C2H6-loading PXRD patterns show similar shifting of 298K, although C2H4 capture was achieved, there was distinct C2H6
the (020) peaks, corresponding to slight structural transformation of co-adsorption in the packed column of HOF-FJU-1a, as indicated
pore window expansion. We note that the pressure where peak shift- by a moderate retention time for C2H6 in the breakthrough curves.
ing started is consistent with corresponding gas sorption isotherms, Co-adsorption occurred due to the partial pressure of C2H6 in the
which indicates the stepwise sorption is accompanied by expansion gas mixture reached the borderline of gate-opening, and the pro-
of pore window, referring to gate-opening. For C2H4-loading, simi- duction purity (the integration proportion of both concentration
lar gate-opening at higher pressure-threshold can be still observed at curves of desorbed C2H4 and C2H6 from the regeneration of HOF
318K. By contrast, there is no noticeable structural transformation material) of C2H4 at 298K was only 73.1 0.1% (Fig. 3g).
was observed for C2H6-loading indicated by corresponding PXRD
In contrast, by tuning the gate-pressures to minimize
patterns, consistent with negligible C2H6 uptake at 318K. These co-adsorptionofC2H6, theneatbreakthroughcurvesofHOF-FJU-1a
results reveal that by applying different stimulus such as controlling at 333K showed an immediate C2H6 elution without any detectable
the temperatures, flexible-robust HOFs can show corresponding C2H4, and very high C2H4 retention before the breakthrough of C2H4;
response namely gate-opening to accommodate different gas mol- thus, the C2H4 purity increased to 99.1 0.1% at 333K (Fig. 3h,i,l),
ecules, thus realizing highly efficient separation.
which is higher than those performed by ITQ-5543 and Cu(OPTz)36,
Ideal adsorbed solution theory (IAST) calculations were then and almost identical to that of UTSA-280 (99.2%)29. The captured
used to estimate the C2H4/C2H6 (50:50, v/v) separation selectiv- C2H4 amount from the equimolar C2H4/C2H6 breakthrough curve at
ity at different temperatures. At 298K, HOF-FJU-1a showed a 333K was calculated to be 0.407molkg–1 (Supplementary Fig. 25),
moderate C2H4/C2H6 selectivity of 10.5 due to the co-adsorption higher than those of UTSA-280 (0.326molkg–1) and NOTT-300
of C2H6. After increasing the temperatures to 333K where the (0.305molkg–1) under the same conditions (Supplementary Figs. 23
co-adsorption of C2H6 was suppressed, the C2H4/C2H6 selectivity of and 24, and Supplementary Table 13). These results demonstrate
HOF-FJU-1a increased to 42.3 at 1bar (Supplementary Figs. 17–19, that HOF-FJU-1a can realize high C2H4/C2H6 separation perfor-
and Supplementary Tables 9 and 10), which was higher than those mance under mild conditions.
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
To evaluate durability, multiple cycling breakthrough experi- 1 to 14, even in solutions of 12M HCl and 10M NaOH. These
ments were performed under the same conditions, showing that results reveal the great application potentials of HOF-based materi-
HOF-FJU-1a can retain 97% of its captured C2H4 productivity after als for realizing important gas separations.
five cycles relative the first cycle (Fig. 3j and Supplementary Fig. 26).
The effect of potential gas impurity in realistic conditions was also Online content
investigated. We performed a breakthrough study on a gas stream Any methods, additional references, Nature Research report-
of H2/C3H6/CH4/C3H8/C2H6/C2H4 (5:5:5:5:40:40, v/v/v/v/v/v) and ing summaries, source data, extended data, supplementary infor-
found that HOF-FJU-1a can still exclusively capture C2H4 from the mation, acknowledgements, peer review information; details of
mixture (Fig. 3k). Overall, these results demonstrate that high sepa- author contributions and competing interests; and statements of
ration performance of C2H4 from relevant mixtures can be realized data and code availability are available at https://doi.org/10.1038/
in HOF-FJU-1a.
s41557-021-00740-z.
We performed single-crystal X-ray diffraction measurements of
C2H4-loaded HOF-FJU-1a sample to determine the binding confor-
mations of C2H4. The data for HOF-FJU-1∙0.94C2H4 were collected
at 100K, from which the location of C2H4 molecules was identified
by an increase in residual electron intensity (Supplementary Fig. 27
and Supplementary Table 1). Three crystallographically unique C2H4
molecules were identified in the compact pore channel (Fig. 4a).
The C2H4 molecules were well dispersed along the one-dimensional
pore channel with only host–guest interactions observed, show-
ing multiple intermolecular interactions mainly C–H···π with the
framework (3.701–4.124A) (Figs. 4b,c). A single-crystal X-ray
diffraction study of HOF-FJU-1b (following ethane exposure)
shows that there is no electron density in the open pore space
(Supplementary Table 1), indicating that no ethane molecule was
trapped. Modelling simulations indicate that—without taking into
account the framework flexibility—the diffusion of both C2H4 and
C2H6 molecules into HOF-FJU-1 is impossible (Supplementary Fig.
28). The substantial C2H4/C2H6 separation here might involve the
dynamics of organic linkers.
Apart from the excellent C2H4/C2H6 separation performance, it is
notable that (unusually for a HOF) HOF-FJU-1 exhibited excellent
chemical stability, as validated by PXRD of samples following expo-
sure to different harsh conditions. Although heating to 573K led
to framework decomposition, HOF-FJU-1 retained its crystallinity
and porosity in aqueous solutions at pH values ranging from 1 to 14,
and even in solutions of 12M HCl and 10M NaOH (Supplementary
Figs. 29 and 30). Similarly, it was stable when immersed in vari-
ous organic solvents including CH2Cl2, hexane, 1,4-dioxane
and acetonitrile, although not in N,N′-diethylformamide nor
N,N′-dibutylformamide (Supplementary Figs. 30 and 31). The
PXRD patterns showed peaks shift slightly as a result of the flex-
ibility of the frameworks. Furthermore, HOF-FJU-1 was quickly
regenerated from the N,N′-DMF (Supplementary Fig. 32). Many
have reported that stable HOFs such as HOF-TCBP44, PCF-145 and
ZJU-HOF-146 are labile to basic solution and undergo deprotonation
or hydrolysis. The strong intermolecular π–π stacking and hydro-
gen bonding interactions, compact structure for minor exposure of
hydrogen-bonding motifs, and weak intrinsic acidity and basicity of
cyano groups jointly endowed HOF-FJU-1 with excellent thermal
and chemical stability in contrast to those HOFs with groups such
as amides and carboxylic acids (Supplementary Table 2).
Received: 9 June 2020; Accepted: 28 May 2021;
Published: xx xx xxxx
References
1. Sholl, D. S. & Lively, R. P. Seven chemical separations to change the world.
Nature 532, 435–437 (2016).
2. Lin, J. Y. S. Molecular sieves for gas separation. Science 353, 121–122 (2016).
3. Chu, S., Cui, Y. & Liu, N. Te path towards sustainable energy. Nat. Mater.
16, 16–22 (2017).
4. Kuznicki, S. M. et al. A titanosilicate molecular sieve with adjustable pores for
size-selective adsorption of molecules. Nature 412, 720–724 (2001).
5. Kuznicki, S. M. & Bell, V. A. Olefn separations employing CTS molecular
sieves. US Patent 6517611 (2003).
6. Mohamed, M. H. et al. Designing open metal sites in metal–organic
frameworks for parafn/olefn separations. J. Am. Chem. Soc. 141,
13003–13007 (2019).
7. Furukawa, H., Cordova, K. E., O’Keefe, M. & Yaghi, O. M. Te chemistry
and applications of metal–organic frameworks. Science 341, 1230444 (2013).
8. Li, L. et al. Ethane/ethylene separation in a metal–organic framework with
iron-peroxo sites. Science 362, 443–446 (2018).
9. Liao, P.-Q., Zhang, W.-X., Zhang, J.-P. & Chen, X.-M. Efcient purifcation of
ethene by an ethane-trapping metal–organic framework. Nat. Commun. 6,
8697 (2015).
10. Qazvini, T., Babarao, R., Shi, Z. L., Zhang, Y. B. & Telfer, S. G. A robust
ethane-trapping metal–organic framework with a high capacity for ethylene
purifcation. J. Am. Chem. Soc. 141, 5014–5020 (2019).
11. Yang, H. et al. Pore-space-partition-enabled exceptional ethane uptake and
ethane-selective ethane–ethylene separation. J. Am. Chem. Soc. 142,
2222–2227 (2020).
12. Zeng, H. et al. Cage-interconnected metal–organic framework with tailored
apertures for efcient C2H6/C2H4 separation under humid conditions. J. Am.
Chem. Soc. 141, 20390–20396 (2019).
13. Hao, H.-G. et al. Simultaneous trapping of C2H2 and C2H6 from a ternary
mixture of C2H2/C2H4/C2H6 in a robust metal–organic framework for the
purifcation of C2H4. Angew. Chem. Int. Ed. 130, 16299–16303 (2018).
14. Lin, L. et al. Fundamental insights into the reactivity and utilization of open
metal sites in Cu (i)-MFU-4l. Organometallics 38, 3453–3459 (2019).
15. Wang, H. et al. One-of-a-kind: a microporous metal–organic framework
capable of adsorptive separation of linear, mono- and di-branched alkane
isomers via temperature- and adsorbate-dependent molecular sieving. Energy
Environ. Sci. 11, 1226 (2018). Based on the 77 K N2 isotherms, fexible-robust
phenomena have not been clearly observed in this calcium MOF.
16. Huang, N., Wang, P. & Jiang, D. Covalent organic frameworks: a materials
platform for structural and functional designs. Nat. Rev. Mater. 1, 16068
(2016).
17. Li, B. et al. Introduction of π-complexation into porous aromatic framework
for highly selective adsorption of ethylene over ethane. J. Am. Chem. Soc.
136, 8654–8660 (2014).
In summary, we have successfully realized a flexible robust
hydrogen-bonded organic framework (HOF-FJU-1) with high siev-
ing separation for ethylene from ethane by tuning the adsorption
gate-pressures at different temperatures. This novel approach has
successfully minimized the co-adsorption of ethane while retained
a significant amount of ethylene uptakes, leading to considersbly
enhanced ethylene/ethane separation performance at the elevated
temperatures. This HOF material can readily produce high-purity
C2H4 (>99% pure) from C2H4/C2H6 mixture in a fixed-bed col-
umn at 333K, a temperature close to the real gas mixture streams
in practical processes, indicating the bright promise of this
flexible-robust porous material for the industrial C2H4/C2H6 sepa-
ration. Furthermore, HOF-FJU-1 is not only thermally stable but
also stable in common organic solvents and pH values ranging from
18. He, Y., Xiang, S. & Chen, B. A microporous hydrogen-bonded organic
framework for highly selective C2H2/C2H4 separation at ambient temperature.
J. Am. Chem. Soc. 133, 14570–14573 (2011).
19. Lin, R.-B. et al. Multifunctional porous hydrogen-bonded organic framework
materials. Chem. Soc. Rev. 48, 1362–1389 (2019).
20. Liao, P.-Q., Huang, N.-Y., Zhang, W.-X., Zhang, J.-P. & Chen, X.-M.
Controlling guest conformation for efcient purifcation of butadiene. Science
356, 1193–1196 (2017).
21. Chen, K.-J. et al. Synergistic sorbent separation for one-step ethylene
purifcation from a four-component mixture. Science 366, 241 (2019).
22. Zhang, X. et al. Selective ethane/ethylene separation in a robust microporous
hydrogen-bonded organic framework. J. Am. Chem. Soc. 142, 633–640
(2020).
23. Vaidhyanathan, R. et al. Direct observation and quantifcation of CO2 binding
within an amine-functionalized nanoporous solid. Science 330, 650–653
(2010).
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
24. Bloch, E. D. et al. Hydrocarbon separations in a metal–organic framework
with open iron(ii) coordination sites. Science 335, 1606–1610 (2012).
25. Yang, S. et al. Supramolecular binding and separation of hydrocarbons within
a functionalized porous metal–organic framework. Nat. Chem. 7, 121–129
(2014).
26. Cui, X. et al. Pore chemistry and size control in hybrid porous materials for
acetylene capture from ethylene. Science 353, 141–144 (2016).
27. Yoon, J. W. et al. Selective nitrogen capture by porous hybrid materials
containing accessible transition metal ion sites. Nat. Mater. 16, 526–531
(2016).
28. Wang, H. et al. Tailor-made microporous metal–organic frameworks for the
full separation of propane from propylene through selective size exclusion.
Adv. Mater. 30, 1805088 (2018).
29. Lin, R.-B. et al. Molecular sieving of ethylene from ethane using a rigid
metal–organic framework. Nat. Mater. 17, 1128–1133 (2018).
30. Cadiau, A., Adil, K., Bhatt, P. M., Belmabkhout, Y. & Eddaoudi, M. A
metal–organic framework-based splitter for separating propylene from
propane. Science 353, 137–140 (2016).
31. Bavykina, A. & Gascon, J. An efcient nanosieve. Nat. Mater. 17, 1057–1058
(2018).
38. Kitaura, R., Seki, K., Akiyama, G. & Kitagawa, S. Porous coordination-
polymer crystals with gated channels specifc for supercritical gases. Angew.
Chem. Int. Ed. 42, 428–431 (2003).
39. Li, L. et al. Flexible-robust metal–organic framework for efcient removal of
propyne from propylene. J. Am. Chem. Soc. 139, 7733–7736 (2017).
“Flexible–robust MOF” has been defned for the frst time in this literature, as
shown in Fig. 1b. Te “fexible-robust” has been also clarifed in illustration
on the N2 and CO2 adsorption of HOF-FJU-1a (Supplementary Figs. 9 and
12) at cryogenic temperatures.
40. Sarkisov, L. & Harrison, A. Computational structure characterisation tools in
application to ordered and disordered porous materials. Mol. Simul. 37,
1248–1257 (2011).
41. Tommes, M. et al. Physisorption of gases, with special reference to the
evaluation of surface area and pore size distribution. Pure Appl. Chem. 87,
1051–1069 (2015).
42. Aguado, S., Bergeret, G., Daniel, C. & Farrusseng, D. Absolute molecular
sieve separation of ethylene/ethane mixtures with silver zeolite A. J. Am.
Chem. Soc. 134, 14635–14637 (2012).
43. Bereciartua, P. J. et al. Control of zeolite framework fexibility and pore
topology for separation of ethane and ethylene. Science 358, 1068–1071
(2017).
32. Reid, C. R. & Tomas, K. M. Adsorption kinetics and size exclusion
properties of probe molecules for the selective porosity in a carbon molecular
sieve used for air separation. J. Phys. Chem. B 105, 10619–10629 (2001).
33. Webster, C. E., Drago, R. S. & Zerner, M. C. Molecular dimensions for
adsorptives. J. Am. Chem. Soc. 120, 5509–5516 (1998).
44. Hu, F. L. et al. An ultrastable and easily regenerated hydrogen-bonded
organic molecular framework with permanent porosity. Angew. Chem. Int. Ed.
56, 2101–2104 (2017).
45. Yin, Q. et al. An ultra-robust and crystalline redeemable hydrogen-bonded
organic framework for synergistic chemo-photodynamic therapy. Angew.
Chem. Int. Ed. 57, 7691–7696 (2018).
34. Li, L. et al. A metal–organic framework with suitable pore size and specifc
functional sites for the removal of trace propyne from propylene. Angew.
Chem. Int. Ed. 57, 15183–15188 (2018).
46. Zhang, X. et al. A rod-packing hydrogen-bonded organic framework with
suitable pore confnement for benchmark ethane/ethylene separation.
Angew. Chem. Int. Ed. 60, 10304–10310 (2021).
35. Krause, S. et al. A pressure-amplifying framework material with negative gas
adsorption transitions. Nature 532, 348–352 (2016).
36. Gu, C. et al. Design and control of gas difusion process in a nanoporous sof
crystal. Science 363, 387 (2019).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
37. Yang, H. et al. Lock-and-key and shape-memory efects in an unconventional
synthetic path to magnesium metal–organic frameworks. Angew. Chem. Int.
Ed. 58, 11757–11762 (2019).
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2021
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
The single crystal of HOF-FJU-1 was fixed inside of a glass capillary and
pretreated in a manner similar to the gas adsorption measurement to obtain
guest-free HOF-FJU-1a. After the single-crystal data collection for guest-free
HOF-FJU-1a in N2 atmosphere, the capillary with one open end was placed in a
desiccator to backfill with C2H4 or C2H6 for 12h to get the gas loaded single crystal.
The desiccator was filled with C2H4 or C2H6 through ASAP 2020 HD, and the
pressure inside of the desiccator was measured to be 1atm. The capillary was then
sealed with plasticine and kept at 100K during data collection for HOF-FJU⊃C2H4
or HOF-FJU-1b. The non-hydrogen atoms were refined anisotropically and the
non-hydrogen atoms, expect for C2H4 molecules, were fixed at calculated positions.
The hydrogen atoms of the C2H4 molecules were added and refined by using
a riding model. The restraints (DFIX) were used for the C2H4 molecules (C–C
distance). The occupancy of the guest C2H4 molecule was refined by introducing
partial occupancy. For the C2H4 molecules, the carbon atoms were refined at a
fixed occupancy factor of 0.76, 0.59 and 0.52 for C43, C44 and C45, respectively.
Both checkCIF results of HOF-FJU-1a and HOF-FJU-1b indicate there is one
A-level alert in each structure regarding solvent accessible voids. For HOF-FJU-1a,
the solvent accessible voids (270A3 per unit cells) can be attributed to the removal
of solvent molecules on heating before data collection. For HOF-FJU-1b, the
solvent-accessible voids (253A3 per unit cells) occur as there are not any guest
molecules that can be recognized from the very weak residual electron density
peaks during crystallographic refinement.
methods
Materials. 3,6-Dibromocarbazole (99%, HWG), potassium permanganate (99.5%,
SCRC), acetone (99.5%, SCRC), cuprous cyanide (99.5%, SCRC), anhydrous ferric
chloride (98%, Sigma-Aldrich) and anhydrous DMF (99%, Sigma-Aldrich) were
purchased and used without further purifcation. N2 (99.999%), C2H4 (99.99%),
C2H6 (99.99%), He (99.999%), C2H4/C2H6 =50/50 (v/v) and H2/C3H6/CH4/C3H8/
C2H6/C2H4 (5/5/5/5/40/40, v/v/v/v/v) were purchased from Beijing Special Gas
(China).
Synthesis of 3,3′,6,6′-tetrabromo-9,9′-bicarbazole (2). Potassium permanganate
(2.371g, 15mmol) was added at 323K to a solution of 3,6-dibromocarbazole (1)
(1.625g, 5mmol) in 25ml acetone and then the solution was stirred for 5h at
333K with a reflux condenser and cooled to room temperature. After removal of
the organic solvents, the residue was extracted with CHCl3 (250ml) for 12h with
stirring. The filtrate was washed three times with CHCl3. The residue was purified
by recrystallization from chloroform/hexane to give colourless crystals (0.83g, 51%
yield). 1H NMR (400MHz, DMSO-d6): δ=8.72 (d, 4 H, J=1.9Hz), 7.55 (dd, 4 H,
J=8.6, 1.9Hz), 6.91 (d, 4 H, J=8.7Hz).
Synthesis of 3,3′,6,6′-tetracyano-9,9′-bicarbazole (3). A mixture of CuCN (2.782g,
31.06mmol) and compound 2 (2g, 3.086mmol) in dry DMF (50ml) was added
to a 120ml Schleck flask charge with stir bar and reacted at 423K for 48h under
N2 atmosphere. After cooling to room temperature, concentrated HCl (40ml)
and iron(iii) trichloride (30g, 184.9mmol) were added to the mixture and stirred
under 273K for 2h. The reaction mixture was diluted with water (200ml),
filtered and the grey coloured solid was collected (1.3g, 97% yield); melting
point=574.5K. 1H NMR (400MHz, DMSO-d6): δ 9.10 (s, 4 H), 7.91 (d, 4 H,
J=8.5Hz), 7.32 (d, 4 H, J=8.7Hz); 13C NMR (400MHz, DMSO-d6) δ=141.45,
131.69, 127.12, 121.38, 119.22, 110.51, 105.12; IR (KBr): 2,225 (νCN), 1,602, 1,485,
1,451, 1,365, 1,292, 1,238, 1,186, 1,137, 1,028, 893, 815, 587cm–1. The compound
was best formulated as HOF-FJU-1·DMF·10H2O. TGA data: calculated weight loss
for 1 DMF and 10 H2O molecules: 33.44%; found: 35.60%; calculated elemental
analysis (found for C48H52N11O12): C (59.13%, 59.31%), H (5.37%, 5.34%), N
(15.80%, 15.82%).
In-situ gas-loaded PXRD. In situ gas-loaded PXRD patterns were recorded by
using a PANalytical X’Pert3 powder diffractometer equipped with a copper sealed
tube (λ=1.54056A) at 40kV and 40mA. An Anton Paar TTK 600 stage coupled
with an Anton Paar CCU 100 Temperature Control Unit was used to control the
temperature. In a typical experiment, 20mg of sample was activated and loaded
on a zero-background sample holder made for an Anton Paar TTK 600 chamber.
The data were collected from 5° to 30° (2θ) with a step-size of 0.0131303° and a
scan time of 40s per step. The sample was cooled from the room temperature to
195K by using liquid nitrogen set up for Anton Paar TTK 600. Once the chamber
and the sample were cooled to 195K, then CO2 was dosed in the chamber under
different pressures corresponding to the CO2 adsorption isotherm. Ethylene and
ethane were tested in the same way at 298K and 318K. The PXRD patterns are
recorded after equilibration (30min) at selected points of the isotherm, as show in
Supplementary Figs. 12, 15 and 16.
Crystallization of the organic building block (HOF-FJU-1). The organic building
block (3) (0.1g, 0.23mmol) was dissolved in DMF (2ml) under 403K in a glass
flask. The resulting solution was cooled to room temperature (296K) for 12h.
Colourless needle-like crystals were obtained. The PXRD results showed that the
synthesized HOF-FJU-1 is a pure phase (Supplementary Fig. 1).
The isosteric enthalpies of adsorption. The isosteric enthalpy of adsorption for
C2H4 and C2H6 were calculated using the data collected at 318K and 333K. The
data were fitted using a virial-type expression composed of parameters ai and bi
(Equation (1)),
Sample characterization. The crystallinity and phase purity of the samples were
measured using PXRD with a Rigaku Ultima IV X-ray diffractometer with copper
Kα radiation (λ=1.54184 A), under N2 atmosphere, scanning over the 5° to
30° range. The Fourier-transform infrared (KBr pellets) spectra were recorded
in the 400–4,000 cm–1 range on Thermo Nicolet 5700 FT-IR instruments.
1H NMR and 13C NMR experiments were performed on Bruker Advance III
400 MHz. Thermogravimetric analyses were performed with METTLER Q50
under N2 atmosphere with a heating rate of 10 K min–1 from 313 K to 973 K.
A micromeritics ASAP 2020 HD surface area analyser was used to measure
gas adsorption isotherms. Guest-free HOF-FJU-1 (0.120 g) was made from
fresh sample (0.133 g), vacuumed at room temperature for 24 h and further at
423 K until the outgas rate was 5 mm Hg min–1 before measurements. The N2
adsorption was tested at 77 K using liquid nitrogen to maintain the temperature,
ethane and ethylene were tested at 273 K, 298 K, 318 K and 333 K, respectively.
The single-crystal X-ray was performed with Agilent Technologies SuperNova
A diffractometer and the structures were solved by direct methods and refined
by full matrix least-squares methods with the SHELXT program package. The
pore size of HOF-FJU-1 was calculated by Poreblazer47 program on the basis of
the crystal structure. The probe in the application to the pore size distribution
calculations was a nitrogen atom 3.31 A in collision diameter. Elemental analyses
were performed on a Vario EL III analyser.
m
n
∑
∑
1
lnp = lnN +
a N +
i i
b N
i
(1)
i
T
i=0
i=0
where p is the pressure (mmHg), T is the temperature (kelvin), R is the universal
gas constant (8.314Jmol–1 K–1), N is the amount adsorbed (mgg–1), and m and
n determine the number of terms required to adequately describe the isotherm.
The isosteric enthalpies of adsorption Qst (kJ mol–1) were then calculated from
the fitting parameters using (Equation (2)), only data point at zero or near-zero
coverage is claimed to illustrate the adsorption enthalpy:
m
∑
Q = −R
st
a N
i
(2)
i
i=0
Prediction of the gas adsorption selectivity by IAST. Fitting details: the adsorption
data for C2H4 and C2H6 in HOF-FJU-1 at 273, 298, 318 and 333K were fitted with
single-site Langmuir–Freundlich equation.
1/n
bp
max
N =N
N
(3)
1/n
1 + bp
where p is the pressure of the bulk gas in equilibrium with the adsorbed phase
(kPa), N is the amount adsorbed per mass of adsorbent (mmolg–1), Nmax is
the saturation capacities of site 1 (mmolg–1), b is the affinity coefficient of
site 1 (1/kPa) and n represents the deviations from an ideal homogeneous surface;
the corresponding isotherm fit parameters for C2H4 and C2H6 in UTSA-280 and
NOTT-300 at 318 and 333K are provided in Supplementary Figs. 20 and 21.
IAST calculation: the adsorption selectivity based on IAST for mixed
C2H4/C2H6 is defined by the following equation:
Single-crystal X-ray diffraction studies. Data collection and structural analysis of
the crystals were collected on an Agilent Technologies SuperNova single-crystal
diffractometer equipped with graphite monochromatic copper Kα radiation
(λ=1.54184A). Using Olex2, the structure was solved with the ShelXT structure
solution program using intrinsic phasing and refined with the SHELXT refinement
package using least-squares minimization48.
The HOF-FJU-1 single crystal was kept at 293 K during data collection. The
highest peak (1.68 eA–3) in the structure of HOF-FJU-1 was located 2.70 A from
a hydrogen atom in bicarbazole. The solvent molecules in HOF-FJU-1 were
highly disordered and no satisfactory disorder model could be accomplished. A
region of solvent electron density was thus treated with the SQUEEZE program
in PLATON. After squeeze, the highest peak (0.26 eA–3) in the structure of
HOF-FJU-1 was located 0.23 A from a hydrogen atom in bicarbazole. All
non-hydrogen atoms were refined with anisotropic displacement parameters. The
hydrogen atoms on the framework were placed in idealized positions and refined
using a riding model.
q y
A
B
S
=
(4)
A/B
q y
B
A
where qi and yi are the mole fractions of component i (i=A, B) in the adsorbed and
bulk phases, respectively.
Breakthrough experiments. The breakthrough experiments for C2H4/C2H6
(50:50, v/v) gas mixtures were carried out at a flow rate of 1.25 ml min–1
.
Activated HOF-FJU-1a (1.10 g), UTSA-280 (1.17 g) and NOTT-300 (1.12 g)
NatuRE ChEmiStRY | www.nature.com/naturechemistry

Nature Chemistry
Articles
powder were packed into ф 3 ×300 mm stainless steel column. The experimental
set-up consisted of two fixed-bed stainless steel reactors. One column was
loaded with the adsorbent and kept at different temperatures (298 K, 318 K and
333 K), whereas the other reactor was used as a blank control group to stabilize
the gas flow. The flow rates of all gas mixtures were regulated by mass flow
controllers, and the effluent gas stream from the column is monitored by a gas
chromatography (Thermal Conductivity Detector, detection limit 0.1%). Before
the breakthrough experiment, we flushed the activated sample in adsorption bed
with helium gas (100 ml min–1) for 30 min at 333 K to ensure the total removal
of adsorbed gas. The production purity is calculated based on the integration
proportion of both concentration curves of desorbed C2H4 and C2H6 from the
regeneration of HOF material. By contrast, as C2H6-selective adsorbents can
directly produce pure C2H4 in one single separation step, their recovered purity
of ethylene is calculated based on the concentration ratio of C2H4 and C2H6 in the
outlets of the column.
48. Sheldrick, G. M. SHELXT-Integrated space-group and crystal-structure
determination. Acta Crystallogr. A 71, 3–8 (2015).
acknowledgements
We gratefully acknowledge the financial support from the National Natural Science
Foundation of China (grant nos. 21975044, 21971038 and 21922810), the Fujian
Provincial Department of Science and Technology (grant nos. 2019H6012 and
21019L3004) and the Welch Foundation (grant no. AX-1730).
author contributions
Y.S.Y., L.L., R.-B.L., Z.J.Z., S.C.X. and B.L.C. conceived the research idea and designed
the experiments. Y.S.Y. performed most of the experiments and analysed the data. L.L.
and L.Y. measured the laboratory-scale fixed-bed breakthrough tests of HOF-FJU-1.
Y.S.Y., L.L, R.-B.L., Z.J.Z., S.C.X. and B.L.C wrote the paper. All authors discussed the
results and commented on the manuscript. Y.S.Y., L.L. and R.-B.L. contributed equally
to this work.
Data availability
All data supporting the finding of this study are available within this article and
its Supplementary Information. Crystallographic data for the structures in this
article have been deposited at the Cambridge Crystallographic Data Centre
under deposition nos. CCDC 1878390 (HOF-FJU-1), 1871845 (HOF-FJU-
1⊃H2O), 1999088 (HOF-FJU-1a), 1942488 (HOF-FJU-1⊃C2H4) and 1999090
(HOF-FJU-1b). Copies of the data can be obtained free of charge from https://
www.ccdc.cam.ac.uk/structures/. Source data are provided with this paper.
Competing interests
The authors declare no competing interests.
additional information
Supplementary information The online version contains supplementary material
available at https://doi.org/10.1038/s41557-021-00740-z.
Correspondence and requests for materials should be addressed to Z.Z., S.X. or B.C.
Peer review information Nature Chemistry thanks Ashleigh Fletcher, Ichiro Hisaki,
Claire Hobday and the other, anonymous, reviewer(s) for their contribution to the peer
review of this work.
References
47. Sarkisov, L. & Harrison, A. Computational structure characterisation tools in
application to ordered and disordered porous materials. Mol. Simul. 37,
1248–1257 (2011).
Reprints and permissions information is available at www.nature.com/reprints.
NatuRE ChEmiStRY | www.nature.com/naturechemistry