Content uploaded by Emilio Pardo
Author content
All content in this area was uploaded by Emilio Pardo on Sep 02, 2021
Content may be subject to copyright.
1
DOI: 10.1002/((please add manuscript number))
Article type: Communication
Bioinspired Metal-Organic Frameworks in Mixed Matrix Membranes for Efficient
Static/Dynamic Removal of Mercury from Water
Rosaria Bruno, Marta Mon, Paula Escamilla, Jesus Ferrando-Soria,* Elisa Esposito, Alessio
Fuoco, Marcello Monteleone, Johannes C. Jansen,* Rosangela Elliani, Antonio Tagarelli,
Donatella Armentano,* Emilio Pardo*
R. Bruno, Dr. R. Elliani, Dr. A. Tagarelli, Dr. D. Armentano,
Dipartimento di Chimica e Tecnologie Chimiche, Università della Calabria, 87036, Rende,
Italy
E-mail: donatella.armentano@unical.it
Dr. E. Esposito, Dr. A. Fuoco, Dr. M. Monteleone, Dr. Johannes Carolus Jansen
Institute on Membrane Technology, CNR-ITM, Via P. Bucci 17/C, 87036 Rende, Italy
E-mail: johannescarolus.jansen@cnr.it
Dr. Marta Mon, Paula Escamilla, Dr. J. Ferrando-Soria, Dr. E. Pardo
Departamento de Química Inorgánica, Instituto de Ciencia Molecular (ICMOL), Universidad
de Valencia, 46980 Paterna, Valencia, Spain
E-mail: emilio.pardo@uv.es
Keywords: metal-organic frameworks, mixed matrix membranes, water remediation,
mercury(II), capture device
Abstract: We present the mercury removal efficiency of a novel MOF derived from the amino
acid S-methyl-L-cysteine and characterize the process by single crystal X-ray crystallography.
We further present a feasibility study on the performance of this MOF –and also that of another
MOF derived from the amino acid L-methionine– when used as the sorbent in mixed matrix
membranes (MMMs). These MOF-based MMMs exhibit high efficiency and selectivity –in
both static and dynamic regimes– in the removal of Hg2+ from aqueous environments, due to
the high density of thioalkyl groups decorating MOF channels. Both MMMs are capable to
reduce different concentration of the pollutant to acceptable limits for drinking water (< 2 ppb).
In addition, a novel device, consisting of the recirculation and adsorption of contaminated
solutions through the MOF-MMMs, has been designed and successfully explored in the
selective capture of Hg2+. Thus, filtration of Hg2+ solutions with multiple passes through the
permeation cell shows a gradual decrease of the pollutant concentration. These results suggest
2
the implementation of MOF-based MMMs in water remediation, after further optimization of
their morphology, sorption capacity and permeability, with a large potential benefit on natural
aquatic ecosystems helping to reduce either contaminants from accidental unauthorized or
deliberate metal industrial dumping and to ensure access for clean and potable freshwater.
Contamination of aquatic systems[1] constitutes one of the most serious environmental problems
faced by our modern society.[2] In this context, the United Nations (UN) agenda 2030[3] for
sustainable development has set access to safe drinking water as one of its main objectives. At
present, about half of the world population lives in areas with severe water stress, and this is
expected to worsen in the next decades due to the unrestrained consumerism of current society
and the concomitant global warming. In particular, among the plethora of contaminants arising
from human activities, heavy metals represent one of the most important environmental
concerns,[4,5] as a result of their toxicity, high persistence in the environment, great mobility in
water and high tendency to accumulate in living beings.[6] Therefore, the development of
efficient methods for the selective removal of heavy metals from aquatic environments attracts
great interest. In contrast to the current methods for eliminating the organic contaminants –
mainly based on the degradation by advanced oxidation processes (AOPs)[7], and in minor
degree on the adsorption in porous materials[8]– the removal of heavy metals requires their
physical capture. Existing methods include precipitation, coagulation/flocculation, membrane
technology, and the adsorption by porous materials.[9] However, such currently available
technologies, working separately, exhibit various shortcomings in terms of their still low
efficiency, selectivity and, especially, processability and recyclability.[10]
A relatively new class of porous materials, defined as metal-organic frameworks (MOFs),[11–14]
has emerged as one of the most promising candidates to overcome the aforementioned
shortcomings.[15] When properly chosen, water-stable MOFs offer tunable microporosity, and
the possibility to tailor their channels with the appropriate functionalities to enhance affinity for
3
contaminants in a selective manner.[15–17] For example, numerous MOFs and MOF-based
materials have been reported for the successful removal of heavy metals like lead (Pb),[18–23] or
mercury (Hg).[23–34] However, they exhibit important disadvantages such as difficult handling
or losses/leakages of fine powders,[35] that need to be improved for their exploitation in
industrial processes. In this context, with the aim to overcome these issues and make a step
forward on MOF-based technologies for water remediation, we propose the use of MOF-based
mixed matrix membranes (MMMs) –composite materials made up of homogeneously dispersed
MOF-fillers in a porous polymer film– for mercury removal from water. A number of
membrane-based processes for mercury removal from aqueous waste streams has been
proposed in the literature, based on micro-, ultra- and nano-filtration or reverse osmosis,[36]
supported liquid membranes,[37–39] liquid membranes with crown ethers[40] or
benzoylthiourea[41] carriers, or ion exchange membranes.[42] However, MOF-based MMMs
have mainly focused on industrially-relevant gas separation applications,[43–49] and their
application for water remediation has been barely explored.[50–52] This includes removal of
organic dyes,[53] arsenate,[54], humic acid,[55] for antifouling properties,[56] or simply to increase
flux while rejecting impurities.[57] To the best of our knowledge, no MOF-based MMMs have
been reported for the removal of mercury species from water.
We anticipate that the incorporation of the suitable MOF in the appropriate polymeric matrix
with a specifically designed morphology, should yield a MMM that combines the superior
capture properties of the MOF with improved handling and applicability of a polymer film. The
scope of this work is to demonstrate the feasibility to immobilize task-specific MOFs, with
exceptionally high selectivity for mercury, inside a tailor-made highly porous membrane, in
order to guarantee high-water fluxes, i.e. productivity, and an efficient separation process, i.e.
purity of the final products (Figure 1).
4
Figure 1. (A) Conceptualization of the filtration/adsorption process by means of a mixed matrix membrane in low
pollutant concentration regimes. B) Detail of the interaction between contaminant and functional groups decorating
MOFs channels.
In the search for highly performant materials for heavy metals removal, we have prepared two
MMMs containing a previously reported MOF derived from the amino acid L-methionine[58]
(Figure 2A) with formula {CaIICuII6[(S,S)-methox]3(OH)2(H2O)}·16H2O (1), and a novel MOF
derived from the amino acid S-methyl-L-cysteine with formula {CaIICuII6[(S,S)-
Mecysmox]3(OH)2(H2O)}·16H2O (2) (Figure 2B), where methox and Mecysmox ligands are
bis[(S)-methionine]oxalyl diamide and bis[S-methylcysteine]oxalyl diamide, respectively. The
high affinity of sulfur towards inorganic pollutants such as mercury is a well-known
phenomenon, and makes these two MOFs promising candidates for the preparation of MMMs
efficient for mercury removal from water.
5
Figure 2. Perspective views along the direction of channel propagation for MOFs 1 (A) and 2 (B) crystal structures.
Polycrystalline samples of 1 and 2 were obtained in multi-gram scale (see Experimental Section
in Supporting Information) and characterized by ICP-MS, CHNS and TGA (Figure S1)
analyses and PXRD measurements (Figure S2). Single crystals, suitable for X-ray diffraction,
of 2 were also obtained by a slow diffusion method (see Experimental Section, Supporting
Information) and its structure solved (vide infra) by Single-Crystal X-ray Diffraction (SCXRD)
under synchrotron radiation. Both 1 and 2 were embedded in a porous polyimide matrix
Matrimid®5218 by solution casting and nonsolvent-induced phase inversion, yielding two novel
mixed matrix membranes named 1@Matrimid® and 2@Matrimid® (Figures 3 and S3).
6
Figure 3. A,B) Optical Photograph of 1@Matrimid® and 2@Matrimid® after drying of the coagulated films
show a very even distribution of the greenish MOFs. Cross‑section C,E) and upper-layer D,F) SEM images of the
membranes 1@Matrimid® and 2@Matrimid®, respectively.
For comparison with the MMMs, the efficiency of 1 and 2 in water remediation was first tested
in their neat form. In this sense, the methionine-derived 1 had previously demonstrated very
high Hg(II) removal efficiency in powder and pellet form.[25] In turn, the methylcysteine-
derived MOF 2 is reported here for the first time. Thus, we evaluated the capture properties of
2 by soaking 10 mg of a polycrystalline sample of the MOF in an aqueous solution containing
10 ppm of the highly toxic Hg2+ cations and also other ions commonly present in drinking water,
like Na+, K+, Ca2+, Mg2+, HCO3-, Cl-, NO3- and SO42-, in order to simulate real conditions.
Quantification of the capture process via analysis of the ion concentration (Figure S4 and Table
S1) as a function time for 72 h showed a very efficient and selective capture of Hg2+ by 2 under
static conditions. In fact, 2 it is capable to reduce [Hg2+] from 10 ppm to 4.6 ppb, close enough
to acceptable limits for drinking water.[59] The capture process has much faster kinetics
compared with the results reported previously for 1,[25] which is in agreement with the higher
accessible surface of 2 (see Figure 2B and structural description).
7
Taking advantage of the robustness and crystallinity of MOF 2 and aiming at unveiling the
mechanisms governing its efficiency in Hg2+ capture, single-crystals of MOF 2 were soaked in
a saturated HgCl2 water/methanol solution for 48 h observing no degradation of the crystals.
The crystal structure of the hybrid material HgCl2@{CaIICuII6[(S,S)-mecysmox]3(OH)2(H2O)}
8H2O (HgCl2@2) could be resolved by single crystal X-ray diffraction under synchrotron
radiation. The structures of 2 and HgCl2@2, are isoreticular to 1,[58] crystallizing in the chiral
P63 space group of the hexagonal system (Table S2). They consist of a uni-nodal acs chiral net
built by calcium(II) vertexes and trans oxamidato-bridged dicopper(II) units, {CuII2[(S,S)-
mecysmox]} (Figure 2B, 4, S5 S8), which act as linkers between the CaII ions through the
carboxylate groups. In the resulting porous net, the functional flexible dimethyl thioether chains
of the methylcysteine amino acid remain confined in hexagonal channels of ca. 0.3 nm (Figure
B). The crystal structure of HgCl2@2 (Figure 4, S7 and S8) demonstrates that the CuIICaII three-
dimensional (3D) network of 2 (Figure 2B, S5 and S6) remains unaltered after the post-
synthetic (PS) insertion process[60–62] (Table S2). The guest HgCl2 molecules situate in the
hexagonal channels of the MOF, being anchored to sulfur atoms from the dimethyl thioether
groups decorating the walls of the pores (Figure 4). It confirms that the S…Hg interaction lies
at the origin of the material receptor properties of 2 for the mercury capture process.
MOF 2 shows intrinsic flexibility, with one of the two crystallographically distinct dimethyl
thioether chains exhibiting a distended conformation inward the pores. In this way, the MOF is
able for grasping a fraction of the guest molecules, and forcing the other chain in an extremely
bent conformation of the methyl groups to pinpoint a 50% of total Hg2+ ions in accessible
interstitial sites, pointing along c axis (Figures 4C-D and S8). The Hg-S bond distances
[2.225(3) and 2.23(4) Å] are in agreement with those found in the literature.[25,26]
8
Figure 4. Crystal structure of adsorbate HgCl2@2: A) Perspective views along the crystallographic c axis of the
HgCl2@2 structure; B) Details of a single channel along c and C) a crystallographic axis: D) Highlight of the
Hg···S coordinative interaction of the captured mercury species. Polyhedra: copper-cyan; calcium-blue. Spheres:
mercury-purple; chlorine-green; sulfur-yellow; carbon from methyl groups-grey. Sticks: carbon, nitrogen and
oxygen atoms from the ligand. Free water solvent molecules are omitted for clarity.
The formula of the adsorbate HgCl2@2 as (HgCl2){CaIICuII6[(S,S)-methox]3(OH)2(H2O)} .
8H2O were further confirmed by ICP-MS, CHNS and TGA analyses (Figures S1 and
Experimental Section in the Supporting Information).
9
Figure 5. A) Photos of the circular samples of the three types of membranes tested with a diameter of 4.7 cm.
Scheme of the static adsorption set-up B) and dynamic adsorption process C).
Based on the removal efficiency of 1 and 2, both were embedded in mixed matrix membranes
using Matrimid®5218 as polymer matrix, i.e. 1@Matrimid® and 2@Matrimid® with a loading
of 30 wt%. 300 mg of polycrystalline samples of both MOFs were firstly activated at 80 °C
under vacuum for 5 hours, and then homogeneously suspended in DMA (dimethylacetamide)
using ultrasounds (30 minutes). Then, a solution containing 1 g of Matrimid® (20 wt% in DMA)
was added to each MOF-suspension to form the casting solutions stirred for 24h. The
membranes were prepared via non-solvent induced phase inversion (NIPS) by casting the
solutions (using an Elcometer®3570 casting knife; casting gap = 250 µm) on a glass plate, and
exposing for a 1 minute to room condition (relative humidity 35%) before their immersion into
10
the coagulation bath of distilled water. The membranes were washed with distilled water to
remove traces of residual solvent, thoroughly dried and, finally, coated with a Hyflon® AD60x
protective layer. The detailed procedure is given in the supporting information.
Figure 6. Kinetic profiles of the mercury (II) capture by 1@Matrimid®, 2@Matrimid® and neat Matrimid®
membranes during soaking the corresponding circular flat membrane (surfaces of 17.34 cm2) in aqueous solutions
(oligo mineral water) of HgCl2 in the 0-72 h interval. The initial [Hg2+] in oligo mineral water are 370 ppb and 330
ppb for static adsorption (A) and microfiltration (dynamic) (B), respectively (Table S7). Removal efficiencies (%)
of 1@Matrimid®, 2@Matrimid® and a pure Matrimid® membrane under the same conditions for static
adsorption (C) and dynamic adsorption (D), respectively (Table S8). (Test with Neat Matrimid® in dynamic mode
was not performed due to the low permeability of membrane and thus low flow, see Table S9. To ensure that the
setup does not absorb mercury, an Hg2+ solution of 2.75 ppm was recirculated through the cell used in the
dynamic adsorption experiment for 80 minutes. Through ICP-MS analysis, it was verified that there was no drop
in concentration at the end of this procedure. Furthermore, each time a separate test with ICP-MS measuring [Hg2+]
of as-prepared solutions before and after recirculating has been performed to confirm that the Hg2+ is not
considerably absorbed by the tubes, pump or other parts of the setup).
11
The loading of the MOFs is already visible to the unaided eye since it changes the membranes
colors, which are yellowish for the neat Matrimid® membranes and green for the MMMs
(Figure 5A). The TGA curves demonstrate that the MOFs in both MMMs do not significantly
influence the pattern of the Matrimid® decomposition and they vary only slightly (Figure S9),
most-likely due to the water contents of the MOF, which is reflected in a 5% weight loss in the
temperature range of 50-400 °C for 1@Matrimid® and 2@Matrimid®. Above 600 °C, there
is ca. 40% weight loss for 1@Matrimid®, 2@Matrimid® and the pure Matrimid® membrane,
which must be due to the partial decomposition of Matrimid® (Figure S9, Supporting
Information).
The MMMs 1@Matrimid® and 2@Matrimid® both have an asymmetric morphology with
large finger-like voids in the center of the membrane and sponge-like layers near the top and
bottom surface (Figure 3 and S3), while the neat Matrimid® membrane has a fully sponge-like
structure (Figure S3). The very fine pore structure in the surface skin layer guarantees the
efficient entrapment of the MOFs inside the membrane and reduces the chances of leaching.
The latter is further improved by the impregnation with dilute Hyflon® AD solution.
The morphological structure of both membranes is characterized by the presence of a sponge-
like layer in the top, which is thicker for 2@Matrimid® (Figure 3E) compared to the
1@Matrimid® (Figure 3F). While both membranes exhibit a region with finger-like
macrovoids under the top layer, due to the high demixing rate during membrane formation via
NIPS, the morphology is suitable for microfiltration. It should be noted that optimum
performance of the MMMs would require a completely sponge-like morphology, such as that
of the neat Matrimid® membrane (Figure S3) (to guarantee a tortuous flow path and intense
contact of the fluid with the MOFs), with a high overall porosity (to allow a high MOF loading)
and small surface pores (to prevent MOF leakage) and high surface porosity (to guarantee high
permeability). This may be achieved by careful optimization of numerous experimental
parameters in the preparation method (type and concentration of polymer, type of solvent,
12
additives in the solvent, type of nonsolvent coagulation bath, solution temperature, casting
temperature, coagulation temperature, additives in the coagulation bath, relative humidity,
MOF size etc). However, even if the present membranes 1@Matrimid® and 2@Matrimid®
have less effective large elongated voids, they are suitable for a convincing proof-of-principle.
The membranes present a density of 0.33 g cm−3 and 0.34 g cm−3 for 1@Matrimid® and
2@Matrimid®, respectively, and both have a similar porosity of about 75%. Both membranes
are stable in water and no MOF leaching is detected after the coating with Hyflon® AD60x,
even after long storage of the produced membranes in water.
The Hg2+ removal efficiency of the two MMMs was studied by both static adsorption in batch
(Figure 5B) and dynamic adsorption during permeation, adapting a microfiltration test cell
(Figure 5C, 2.5 mL min-1, feed pressure = 3 bar; see Table S3 and Figure S10), thus recirculating
the contaminated solutions through the MOF-MMMs via a peristaltic pump. In both methods,
three different solutions were used: deionized water with a high [Hg2+] for benchmark
experiments (Table S4), and oligo mineral water (see composition in Table S5) with two
different [Hg2+] in order to analyze the effect of other ions in solution in a more realistic
situation (Tables S6-S7). The variation of the [Hg2+] was monitored through ICP-MS analysis,
which allowed to follow the absorption in a dynamic process.
In the experiments under static conditions with deionized water, the starting mercury
concentration ([Hg2+] = 2.61 ppm) dropped to 63 and 52 ppb after 48h, and even at 52.8 and
47.5 ppb after 72h with 1@Matrimid® and 2@Matrimid®, respectively (Figure S11A left and
Table S3), which are values about seven times lower than the concentration reached by the neat
Matrimid® membrane (348 ppb). Thus, the presence of both MOFs within the corresponding
MMMs significantly increase the removal efficiency of the pure polymer, which was 98.0% for
1@Matrimid® and 98.2 % for 2@Matrimid® (Figure S11C left and Table S8). Under dynamic
experiments by deionized water ([Hg2+] = 2.21 ppm), the final concentrations after 48h were
13
302 and 275 ppb for 1@Matrimid® and 2@Matrimid®, respectively (Figure S11B right and
Table S4). Thus, the removal efficiency for 1@Matrimid® and 2@Matrimid® are 82.8% and
89.2 %, respectively, demonstrating the suitability of these membranes for Hg2+ removal from
water (Figure S11D right and Table S8). Static experiments using a solution of [Hg2+] = 50 ppm
showed that 1@Matrimid® and 2@Matrimid® exhibit maximum adsorption capacities of 540
and 690 mg g−1, respectively, most-likely due to the higher surface area shown by 2 and
suggesting higher efficiency for the removal of the target Hg2+. Despite the maximum Hg2+
adsorption capacity of these novel MMMs are lower than the reported state-of-the-art thiol-
functionalized COFs[63,64] or other MOFs composites,[23] they are largely enough to prove the
validity of our approach; where carefully designed thioalkyl-based MOFs, derived from amino
acids with high selectivity –at ppb levels– towards Hg2+, are structured and implemented in
MMMs without depletion of their activity and with the associated benefits toward real-world
applications of membranes.
The presence of other ions in oligo mineral water slightly reduces the ability of 1@Matrimid®
to capture Hg2+, whereas 2@Matrimid® retains its performance, thus showing a higher
selectivity for Hg2+ (Table S6 and Figure S12). Most likely the pore’s dimension is at the origin
of this behavior. Thus, despite the chemical similarity of the binding site, ensuring strong S…Hg
interactions in 1 as in 2, the longer aliphatic chain in 1, reduces the size of the virtual diameter
in 1 with respect to 2 (see X-ray structures).
At low [Hg2+] in oligo mineral water, the Hg2+ concentration dropped from 370 ppb to 1.85 ppb
after adsorption by 1@Matrimid®, and below the 1.2 ppb detection limit for 2@Matrimid®
after 72 h under static conditions (Figure 6A and Table S7). Under dynamic conditions (Figure
6B and Table S7), the Hg2+ concentration decreased from the initial 330 ppb to 1.78 ppb with
1@Matrimid® and to 1.26 ppb with 2@Matrimid® after 48 h. These results reveal that both
1@Matrimid® and 2@Matrimid® are capable to remove Hg2+ from water (Figure 6C-D).
14
2@Matrimid® is able to capture up to the 99.9% of the Hg2+ cations from a realistic model
solution (Figure 6D and Table S8), achieving values below the limits for drinking water
established by the WHO[59] or the U.S. Environmental Protection Agency (EPA, < 2ppb for
Hg2+).[65]
For instance, a 30%wt MOF2/Matrimid® (2@matrimid®) membrane can purify 100mL of
water from a dangerous 330 ppb level of [Hg2+] to drinking water (1.26 ppb < 2 ppb EPA) by
recirculating through the same membrane for two days (see also Table S9 for adsorption
measured in 100 mL and expressed as µg/cm2).
Capture experiments described above, demonstrate that the MOFs within the MMMs drastically
improve the removal efficiency of the pure matrimid® polymer. Scanning Electron Microscopy
coupled with Energy Dispersive X–ray (SEM/EDX) measurements of 1@Matrimid® and
2@Matrimid® loaded with HgCl2 show the fundamental role of the MOFs in the Hg adsorption
(Figures S13 and S14). EDX elemental mappings for Cu, Ca, S and Hg elements show a
heterogeneous spatial distribution of Hg atoms through 1@Matrimid® and 2@Matrimid®
with Hg atoms always located next to Cu, Ca and S atoms, which are part of the MOF particles.
This confirms the prominent role of the MOF in the capture properties of the MOF-MMMs and
suggesting that mercury atoms are bound by the thioether groups, as expected. Furthermore,
while neat Matrimid does have a certain affinity for Hg (Figure 6a), in the presence of the MOFs,
the Hg concentration in the polymer drops to very lower values, confirming not only the
sorption capacity of the MOFs, but also their much higher affinity compared to that of the
polymer matrix.
The stability of the membranes for a potential regeneration process was also evaluated. In this
regard, Hg2+ was extracted after suspension of the membranes in a 10% (v/v%) aqueous solution
of 2-mercaptoethanol for 24 hours. The integrity and reusability of 1@Matrimid® and
2@Matrimid® after the extraction process was studied over three subsequent cycles of
adsorption and regeneration, which indicates the stability and still significant efficiency
15
(Supporting Information, Tables S10-S11 and Figures S15-S16) recovering up to 87.7 and
89.1% of Hg2+ of the original adsorption capacities for 1@Matrimid® and 2@Matrimid,
respectively. The reusability of 1@Matrimid® and 2@Matrimid® should be also improved
after optimization of the membrane morphology. Membranes stability and regeneration are key-
factors for their exploitation in industrial applications since they have a direct impact on the
economic feasibility.
In summary, the potential of MOFs and MOF-containing MMMs in water remediation, receives
a very important boost with the reported results. Herein, in a first part, we have described the
preparation and thorough physical characterization by means of synchrotron single-crystal X-
ray crystallography of a novel MOF derived from amino acid S-methyl-L-cysteine (2) and the
resulting adsorbate after mercury chloride removal (HgCl2@2). Then, we showed two
unprecedented MOF-MMMs –1@Matrimid® and 2@Matrimid®, built up with
Matrimid®5218 and two bioinspired MOFs highly performant in the capture of cationic Hg2+–
as a sustainable, effective and cheap solution for mercury removal from water, with the potential
of minimizing the impact of this pollutant on the environment and human bodies. These features
are due to the exquisite control of the thio-alkyl functionalities decorating the MOF pores that
are used, to decontaminate, very efficiently –even from very low concentrations (ppb), which
are usually found in real contaminated waters– aqueous environments to acceptable limits for
potable water. Moreover, dynamic capture experiments have been also performed. Thus, a
novel device –consisting of the recirculation and microfiltration of contaminated solutions
through the MOF-MMMs via the use of a peristaltic pump– is reported here with outstanding
capture results. These last results must be considered as a feasibility study for the proof-of-
principle and open new avenues –after optimizing experimental conditions– for the use of
MOF-MMMs in environmental remediation.
Experimental Section
16
See Supporting Information for a detailed description of MOFs and membranes preparation,
their characterization and capture experiments.
[CCDC Deposition Numbers 2007971-2007972 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.]
Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements
((This work was supported by the Ministero dell’Istruzione, dell’Università e della Ricerca
(Italy) and the Ministerio de Ciencia e Innovación (Spain) (Projects PID2019–104778GB–I00
and Excellence Unit “Maria de Maeztu” CEX2019-000919-M). R. B. thanks the MIUR (Project
PON R&I FSE–FESR 2014–2020) for predoctoral grant. Thanks are also extended to the “2019
Post–doctoral Junior Leader–Retaining Fellowship, la Caixa Foundation (ID100010434 and
fellowship code LCF/BQ/PR19/11700011” (J. F.–S.). M. M. thanks ITQ for the concession of
a contract. D.A. acknowledges the financial support of the Fondazione CARIPLO / “Economia
Circolare: ricerca per un futuro sostenibile” 2019, Project code: 2019–2090, MOCA. E.P.
acknowledges the financial support of the European Research Council under the European
Union's Horizon 2020 research and innovation programme / ERC Grant Agreement No 814804,
MOF–reactors.))
17
References
[1] J. Murria, Nat. Hazards Rev. 2003, 4, 166.
[2] L. Xu, Impact of Climate Change and Human Activity on the Eco-Environment,
Springer Berlin Heidelberg, Berlin, Heidelberg, 2015.
[3] United Nations, “Transforming our world: the 2030 Agenda for Sustainable
Development,” 2015.
[4] P. B. Tchounwou, C. G. Yedjou, A. K. Patlolla, D. J. Sutton, EXS, 2012, pp. 133–164.
[5] S. N. Groudev, S. G. Bratcova, K. Komnitsas, Miner. Eng. 1999, 12, 261.
[6] H. Galal‐Gorchev, Food Addit. Contam. 1993, 10, 115.
[7] D. B. Miklos, C. Remy, M. Jekel, K. G. Linden, J. E. Drewes, U. Hübner, Water Res.
2018, 139, 118.
[8] R. J. Drout, L. Robison, Z. Chen, T. Islamoglu, O. K. Farha, Trends Chem. 2019, 1,
304.
[9] S. Bhattacharya, A. B. Gupta, A. Gupta, A. Pandey, Eds. , Water Remediation,
Springer Singapore, Singapore, 2018.
[10] S. Bolisetty, M. Peydayesh, R. Mezzenga, Chem. Soc. Rev. 2019, 48, 463.
[11] H. Furukawa, K. E. Cordova, M. O’Keeffe, O. M. Yaghi, Science 2013, 341, 974.
[12] Y. Cui, B. Li, H. He, W. Zhou, B. Chen, G. Qian, Acc. Chem. Res. 2016, 49, 483.
[13] G. Maurin, C. Serre, A. Cooper, G. Férey, Chem. Soc. Rev. 2017, 46, 3104.
[14] A. Kirchon, L. Feng, H. F. Drake, E. A. Joseph, H.-C. Zhou, Chem. Soc. Rev. 2018, 47,
8611.
[15] S. K. Ghosh, Ed., Metal-Organic Frameworks (MOFs) for Environmental
Applications, Elsevier, 2019.
[16] M. Feng, P. Zhang, H.-C. Zhou, V. K. Sharma, Chemosphere 2018, 209, 783.
[17] J. Li, X. Wang, G. Zhao, C. Chen, Z. Chai, A. Alsaedi, T. Hayat, X. Wang, Chem. Soc.
Rev. 2018, 47, 2322.
18
[18] R. Ricco, K. Konstas, M. J. Styles, J. J. Richardson, R. Babarao, K. Suzuki, P.
Scopece, P. Falcaro, J. Mater. Chem. A 2015, 3, 19822.
[19] E. Tahmasebi, M. Y. Masoomi, Y. Yamini, A. Morsali, Inorg. Chem. 2015, 54, 425.
[20] H. Saleem, U. Rafique, R. P. Davies, Microporous Mesoporous Mater. 2016, 221, 238.
[21] L. Wang, X. Zhao, J. Zhang, Z. Xiong, Environ. Sci. Pollut. Res. 2017, 24, 14198.
[22] C. Yu, Z. Shao, H. Hou, Chem. Sci. 2017, 8, 7611.
[23] D. T. Sun, L. Peng, W. S. Reeder, S. M. Moosavi, D. Tiana, D. K. Britt, E. Oveisi, W.
L. Queen, ACS Cent. Sci. 2018, 4, 349.
[24] Q.-R. Fang, D.-Q. Yuan, J. Sculley, J.-R. Li, Z.-B. Han, H.-C. Zhou, Inorg. Chem.
2010, 49, 11637.
[25] M. Mon, F. Lloret, J. Ferrando-Soria, C. Martí-Gastaldo, D. Armentano, E. Pardo,
Angew. Chem. Int. Ed. 2016, 55, 11167.
[26] M. Mon, X. Qu, J. Ferrando-Soria, I. Pellicer-Carreño, A. Sepúlveda-Escribano, E. V.
Ramos-Fernandez, J. C. Jansen, D. Armentano, E. Pardo, J. Mater. Chem. A 2017, 5,
20120.
[27] F. Ke, L.-G. Qiu, Y.-P. Yuan, F.-M. Peng, X. Jiang, A.-J. Xie, Y.-H. Shen, J.-F. Zhu, J.
Hazard. Mater. 2011, 196, 36.
[28] J. He, K.-K. Yee, Z. Xu, M. Zeller, A. D. Hunter, S. S.-Y. Chui, C.-M. Che, Chem.
Mater. 2011, 23, 2940.
[29] K.-K. Yee, N. Reimer, J. Liu, S.-Y. Cheng, S.-M. Yiu, J. Weber, N. Stock, Z. Xu, J.
Am. Chem. Soc. 2013, 135, 7795.
[30] T. Liu, J.-X. Che, Y.-Z. Hu, X.-W. Dong, X.-Y. Liu, C.-M. Che, Chem. Eur. J. 2014,
20, 14090.
[31] F. Luo, J. L. Chen, L. L. Dang, W. N. Zhou, H. L. Lin, J. Q. Li, S. J. Liu, M. B. Luo, J.
Mater. Chem. A 2015, 3, 9616.
[32] L. Liang, Q. Chen, F. Jiang, D. Yuan, J. Qian, G. Lv, H. Xue, L. Liu, H.-L. Jiang, M.
19
Hong, J. Mater. Chem. A 2016, 4, 15370.
[33] L. Huang, M. He, B. Chen, B. Hu, J. Mater. Chem. A 2016, 4, 5159.
[34] A. Chakraborty, S. Bhattacharyya, A. Hazra, A. C. Ghosh, T. K. Maji, Chem. Commun.
2016, 52, 2831.
[35] M. Mon, R. Bruno, J. Ferrando-Soria, D. Armentano, E. Pardo, J. Mater. Chem. A
2018, 6, 4912.
[36] M. Urgun-Demirtas, P. L. Benda, P. S. Gillenwater, M. C. Negri, H. Xiong, S. W.
Snyder, J. Hazard. Mater. 2012, 215–216, 98.
[37] S. Weiss, V. Grigoriev, P. Mühl, J. Memb. Sci. 1982, 12, 119.
[38] S. Sangtumrong, P. Ramakul, C. Satayaprasert, U. Pancharoen, A. W. Lothongkum, J.
Ind. Eng. Chem. 2007, 13, 751.
[39] K. Chakrabarty, P. Saha, A. K. Ghoshal, J. Memb. Sci. 2010, 350, 395.
[40] M. Shamsipur, M. H. Mashhadizadeh, G. Azimi, Sep. Purif. Technol. 2002, 27, 155.
[41] C. Fontàs, M. Hidalgo, V. Salvadó, E. Anticó, Anal. Chim. Acta 2005, 547, 255.
[42] A. Oehmen, D. Vergel, J. Fradinho, M. A. M. Reis, J. G. Crespo, S. Velizarov, J.
Hazard. Mater. 2014, 264, 65.
[43] B. Seoane, J. Coronas, I. Gascon, M. E. Benavides, O. Karvan, J. Caro, F. Kapteijn, J.
Gascon, Chem. Soc. Rev. 2015, 44, 2421.
[44] C. Y. Chuah, K. Goh, Y. Yang, H. Gong, W. Li, H. E. Karahan, M. D. Guiver, R.
Wang, T. Bae, Chem. Rev. 2018, 118, 8655.
[45] Y. Liu, Z. Chen, G. Liu, Y. Belmabkhout, K. Adil, M. Eddaoudi, W. Koros, Adv.
Mater. 2019, 31, 1807513.
[46] G. Liu, V. Chernikova, Y. Liu, K. Zhang, Y. Belmabkhout, O. Shekhah, C. Zhang, S.
Yi, M. Eddaoudi, W. J. Koros, Nat. Mater. 2018, 17, 283.
[47] W. J. Koros, C. Zhang, Nat. Mater. 2017, 16, 289.
[48] B. Ghalei, K. Sakurai, Y. Kinoshita, K. Wakimoto, A. P. Isfahani, Q. Song, K.
20
Doitomi, S. Furukawa, H. Hirao, H. Kusuda, S. Kitagawa, E. Sivaniah, Nat. Energy
2017, 2, 17086.
[49] H. B. Park, J. Kamcev, L. M. Robeson, M. Elimelech, B. D. Freeman, Science 2017,
356, eaab0530.
[50] A. Elrasheedy, N. Nady, M. Bassyouni, A. El-Shazly, Membranes (Basel). 2019, 9,
DOI 10.3390/membranes9070088.
[51] M. Kalaj, K. C. Bentz, S. Ayala, J. M. Palomba, K. S. Barcus, Y. Katayama, S. M.
Cohen, Chem. Rev. 2020, acs.chemrev.9b00575.
[52] B.-M. Jun, Y. A. J. Al-Hamadani, A. Son, C. Min Park, M. Jang, A. Jang, N. Chan
Kim, Y. Yoon, Sep. Purif. Technol. 2020, 116947.
[53] M. M. Baneshi, A. M. Ghaedi, A. Vafaei, D. Emadzadeh, W. J. Lau, H. Marioryad, A.
Jamshidi, Environ. Res. 2020, 183, 109278.
[54] P. Wan, M. Yuan, X. Yu, Z. Zhang, B. Deng, Chem. Eng. J. 2020, 382, 122921.
[55] M. U. M. Junaidi, C. P. Leo, S. N. M. Kamal, A. L. Ahmad, Water Sci. Technol. 2013,
67, 2102.
[56] F. Mohammadnezhad, M. Feyzi, S. Zinadini, J. Ind. Eng. Chem. 2019, 71, 99.
[57] A. Karimi, A. Khataee, V. Vatanpour, M. Safarpour, Sep. Purif. Technol. 2019, 229,
115838.
[58] M. Mon, J. Ferrando-Soria, T. Grancha, F. R. Fortea-Pérez, J. Gascon, A. Leyva-Pérez,
D. Armentano, E. Pardo, J. Am. Chem. Soc. 2016, 138, 7864.
[59] World Health Organisation, Guidelines for Drinking-Water Quality, Fourth Edition,
2011.
[60] J. D. Evans, C. J. Sumby, C. J. Doonan, Chem. Soc. Rev. 2014, 43, 5933.
[61] S. M. Cohen, J. Am. Chem. Soc. 2017, 139, 2855.
[62] M. Mon, J. Ferrando-Soria, M. Verdaguer, C. Train, C. Paillard, B. Dkhil, C. Versace,
R. Bruno, D. Armentano, E. Pardo, J. Am. Chem. Soc. 2017, 139, 8098.
21
[63] L. Merí-Bofí, S. Royuela, F. Zamora, M. L. Ruiz-González, J. L. Segura, R. Muñoz-
Olivas, M. J. Mancheño, J. Mater. Chem. A 2017, 5, 17973– 17981.
[64] Q. Sun, B. Aguila, J. Perman, N. Nguyen, S. Q. Ma, J. Am. Chem.
Soc. 2016, 138, 15790– 15796.
[65] U.S. Environmental Protection Agency, “2012 ed. of the Drinking Water Standards and
Health Advisories,” 2012.
22
Supporting Information
Bioinspired Metal-Organic Frameworks in Mixed Matrix Membranes for Efficient
Static/Dynamic Removal of Mercury from Water
Rosaria Bruno, Marta Mon, Paula Escamilla, Jesus Ferrando-Soria,* Elisa Esposito, Alessio
Fuoco, Marcello Monteleone, Johannes C. Jansen,* Rosangela Elliani, Antonio Tagarelli,
Donatella Armentano,* Emilio Pardo*
Experimental Section
Preparations
Chemical: All chemicals were of reagent grade quality. They were purchased from commercial
sources and used as received. The powder of polyimide (PI) Matrimid®5218 was degassed at
453 K overnight under vacuum to remove the adsorbed water. Compound {Cu6Ca[(S,S)-
methox]3(OH)2(H2O)} . 16H2O (1) was prepared as reported earlier (Ref. 57 in the main text).
Preparation of HMeEt-(S,S)-Mecysmox [bis[(S)-methycysteine]oxalyl diamide]: The proligand
was prepared using the following synthetic procedure: First, under a N2 atmosphere, an excess
of thionyl chloride (13.10 mL, 180 mmol) was added dropwise, under stirring at 0 °C on an ice-
bath, to a solution of (S)-methyl-(L)-cysteine amino acid (8.11 g, 60 mmol) in 150 mL of MeOH.
The resulting colorless solution was refluxed for 6 hours. Then, the excess of thionyl chloride
was distilled with MeOH (3 x 150 mL). The reaction mixture was washed with acetone (150
mL) and diethyl ether (100 mL) and further concentrated, under reduced pressure, to afford the
methyl ester derivative of the (S)-methyl-(L)-cysteine amino acid, which was used in the next
step without further purification. Second, the resulting methyl ester derivative of the (S)-methyl-
(L)-cysteine amino acid (8.95 g, 60 mmol) was dissolved in 250 mL of dichloromethane and
charged with triethylamine (8.4 mL, 60 mmol). To the resulting colorless reaction mixture, was
added dropwise another solution containing oxalyl chloride (2.54 mL, 30.0 mmol) in
dichloromethane (150 mL) under vigorous stirring at 0 °C on an ice-bath. The resulting solution
was further stirred during two hours. The small amount of white solid (Et3NHCl) formed was
filtered off and the resulting solution was then concentrated in a rotatory evaporator to a final
volume of 100 mL. The pale-yellow solution was washed three times with water (3x50 mL)
and finally, the solvent was removed in a rotatory evaporator to afford a white solid, which was
collected with water and dried under vacuum. Yield: 9.62 g, 91%; Anal. calcd (%) for
C12H20S2N2O6 (352.4): C 40.98, H 5.72, S 18.20, N 7.95; found: C 40.97, H 5.68, S 18.26, N
7.99; 1H NMR ([D6]DMSO): 2.20 (s, 6H; SCH3), 2.97 (m, 2H; CH2), 3.17 (m, 2H; CH2), 3.62
(s, 6H; OCH3), 4.78 (t, 2H; CH), 9.01 (d, 2H; NH from CONH). IR (KBr): ν = 1763, 1751and
1656 cm–1 (C=O).
Preparation of (Me4N)2{Cu2[(S,S)-Mecysmox](OH)2} . 5H2O: An aqueous suspension (60 mL)
of H2Me2-(S,S)-Mecysmox (10.572 g, 30 mmol) was treated with a 25% methanolic solution
of Me4NOH (36 mL, 125 mmol) until complete dissolution. Another aqueous solution (25 mL)
of CuCl2 (8.07 g, 60 mmol) was then added dropwise while the reaction mixture was stirred.
The resulting deep green solution was concentrated to a volume of ca. 5-10 mL in a rotary
23
evaporator affording a green polycrystalline solid that was gently washed with acetone filtered
off and dried under vacuum. Yield: 14.77 g, 68%; Anal.: calcd for C18H48Cu2S2N4O13 (719.8):
C, 30.03; H, 6.72; S, 8.91; N, 7.78%. Found: C, 30.13; H, 6.63; S, 8.93; N, 7.75%. IR (KBr): ν
= 3621 cm–1 (O-H), 3023, 2964 cm–1 (C-H), 1608 cm–1 (C=O).
Preparation of {CaIICuII6[(S,S)-Mecysmox]3(OH)2(H2O)} . 16H2O (2): (Me4N)2{Cu2[(S,S)-
Mecysmox](OH)2} . 4H2O (4.32 g, 6.0 mmol) was dissolved in 50 mL of water. Then, another
aqueous solution (10 mL) containing CaCl2 (0.22 g, 2.0 mmol) was added dropwise under
stirring. After further stirring for 10 h, at room temperature, a green polycrystalline powder was
obtained and collected via filtration and dried with ethanol, acetone and diethyl ether. Yield:
2.91 g, 83%; Anal.: calcd for C30Cu6CaH72S6N6O37 (1722.7): C, 20.92; H, 4.21; S, 11.17; N,
4.88%. Found: C, 20.91; H, 4.17; S, 11.19; N, 4.93%. IR (KBr): ν = 1602 cm–1 (C=O). Well-
shaped hexagonal prisms of 1 suitable for X-ray structural analysis could be obtained by slow
diffusion, in an H-shaped tube, of H2O/DMF (1:9) solutions containing stoichiometric amounts
of (Me4N)2{Cu2[(S,S)-Mecysmox](OH)2} . 5H2O (0.13 g, 0.18 mmol) in one arm and CaCl2
(0.0067 g, 0.06 mmol) in the other. They were isolated by filtration on paper and air-dried.
Preparation of HgCl2@{CaIICuII6[(S,S)-Mecysmox]3(OH)2(H2O)} . 8H2O (HgCl2@2): Well-
formed hexagonal green prisms of HgCl2@2, which were suitable for X-ray diffraction, were
obtained by soaking crystals of 2 (5.0 mg) in saturated aqueous and H2O/CH3OH (1:1) solutions
of HgCl2 for 48 hours. The crystals were washed with water, isolated by filtration on paper and
air-dried. HgCl2@2: Anal.: calcd for C30Cl2Cu6CaH56HgS6N6O29 (1850.0): C, 19.48; H, 3.05;
S, 10.40; N, 4.54%. Found: C, 19.51; H, 3.09; S, 10.37; N, 4.51%. IR (KBr): ν = 1601 cm–1
(C=O).
Preparation of mixed matrix membranes: A homogeneous 20 wt% dope solution of Matrimid®
was prepared by dissolving 6 g of powder polymer in 24 g of DMA under magnetic stirring at
room temperature for 24 h. The powder of polycrystalline MOFs 1 and 2 were activated at 80 °C
in an oven under vacuum for 5 hours, and then suspended in dimethylacetamide (DMA).
Sonication for 30 minutes yielded homogenous suspensions of both MOFs to which 5 g of the
polymer solution of Matrimid® 5218 was added. The resulting solution/suspension was stirred
for 24 hours. Then the solution was cast on a glass plates by means of a casting knife
(Elcometer®3570) with a gap of 250 m, exposed to the air for 1 minute, and then immersed
in the coagulation bath, consisting of distilled water, to form the final membranes by non-
solvent induced phase separation (NIPS).
The obtained membrane was left in the coagulation bath for 72 hours, then washed in ethanol
and left to dry in air in order to remove residual traces of solvent. The dry membrane was coated
with a protective layer of Hyflon® AD60x by means of a dip-coating procedure using a solution
of Hyflon® AD60x in HFE 7100 (hydrofluoroether) at 0.1 wt%. It is a crucial point the very
low thickness of layer to avoid permeability decrease. Reference membrane of pure
Matrimid®5218 was prepared by the same procedure.
Characterization
Physical Techniques: Elemental (C, H, N), SEM and ICP-MS analyses (of the pure MOFs 1
and 2) and titration experiments were performed at the Microanalytical Service of the
Universitat de València. ICP-MS analyses for the Hg2+ capture experiments on membranes
(vide infra) were performed at the Department of Chemistry of the University of Calabria. FT–
IR spectra were recorded on a Perkin-Elmer 882 spectrophotometer as KBr pellets. The
thermogravimetric analysis was performed on crystalline samples under a dry N2 atmosphere
with a Mettler Toledo TGA/STDA 851e thermobalance operating at a heating rate of 10 ºC min–
1.
24
X-ray Powder Diffraction Measurements: Polycrystalline samples of 2, and HgCl2@2, were
introduced into 0.5 mm borosilicate capillaries prior to being mounted and aligned on a
Empyrean PANalytical powder diffractometer, using Cu Kα radiation (λ = 1.54056 Å). For
each sample, five repeated measurements were collected at room temperature (2θ = 2–60°) and
merged in a single diffractogram.
X-ray crystallographic data collection and structure refinement: Crystals of 2 and HgCl2@2
were selected and mounted on a MITIGEN holder in Paratone oil and very quickly placed in a
nitrogen stream cooled at 100 K to avoid the possible degradation upon desolvation or exposure
to air. Diffraction data were collected using synchrotron radiation at I19 beamline of the
Diamond Light Source at = 0.6889 Å and processed through xia2 software.
1
The structures
were solved with the SHELXS structure solution program, using the Patterson method. The
model was refined with version 2018/3 of SHELXL against F2 on all data by full-matrix least
squares.
2
Bearing in mind that, crystals of HgCl2@2, suitable for X-ray diffraction, were obtained by
soaking crystals of 2 in saturated aqueous and H2O/CH3OH (1:1) solutions of HgCl2, thus after
a crystal-to-crystal transformation, it is reasonable to observe a diffraction pattern sometimes
affected by expected internal imperfections of the crystals that give a quite expected difficulty
to perform a perfect correction of anisotropy (detected as Alerts A in the checkcifs). However,
the solution and refinement parameters are pretty above of the standard MOFs structures
generally reported.
In HgCl2@2 the occupancy factors, of HgCl2 molecules have been defined in agreement with
SEM results performed on loaded MOFs. The use of some C-C and C-S bond lengths restrains
during the refinements or fixed positions of some highly disordered atoms, has been reasonable
imposed and related to extraordinary flexibility of amino acid chains of the Mecysmox ligand
that are dynamic components of the frameworks. In the refinement of both crystal structures
some further restrains, to make the refinement more efficient, have been applied, for instance
in HgCl2@2 ADP components have been restrained to be similar to other related atoms, using
SIMU for disordered sections or EADP for group of atoms of amino acid chains and the guest
molecules, in HgCl2@2, expected to have essentially similar ADPs.
The solvent molecules were highly disordered as expected in porous crystals and for that reason
have not been modeled, the quite large channels featured by this series of MOFs likely account
for that. The hydrogen atoms of the ligand, except for the hydroxo/water oxygen atom O(1H)
(where the OH/H2O statistic distribution is 2:1) both in 2 and HgCl2@2 were set in calculated
positions and refined as riding atoms. The contribution to the diffraction pattern from the highly
disordered and undetected solvent molecules located in the voids was subtracted from the
observed data through the SQUEEZE method, implemented in PLATON.3
1
(a) Evans, P. Scaling and assessment of data quality. Acta Cryst. D 62, 72–82 (2006). (b) Evans, P. R.,
Murshudov, G. N. How good are my data and what is the resolution? Acta Cryst. D 69, 1204–1214 (2013). (c)
Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Cryst. D 67, 235–242 (2011). (d)
Winter, G. xia2: and expert system for macromolecular crystallography data reduction. J. Appl. Cryst. 43, 186–
190 (2010). (e) Winter, G. et al. DIALS: implementation and evaluation of a new integration package Acta Cryst.
2018, D74, 85-97.
2
(a) Sheldrick, G. M. Acta Cryst. A64, 112-122 (2008). (b) Sheldrick, G. M. Acta Cryst. Sect. A Found. Adv. 71,
3–8 (2015).
25
A summary of the crystallographic data and structure refinement for the three compounds is
given in Table S2. The comments for the alerts A and B are described in the CIFs using the
validation reply form (vrf). CCDC Deposition Numbers are 2007971-2007972 for 2 and
HgCl2@2, respectively.
The final geometrical calculations on free voids and the graphical manipulations were carried
out with PLATON
3
implemented in WinGX,
4
and CRYSTAL MAKER
5
programs,
respectively.
Membrane characterization: The surface and the cross-section morphology of the membranes
were characterized by scanning electron microscopy with a HITACHI model S-4800 SEM. The
samples for cross-section SEM characterization were prepared by freeze fracturing in liquid
nitrogen, in order to avoid deformation of the sample during the fracturing procedure. SEM
characterization was performed using the Ultra High-resolution Electron Beam Lithography
and SEM imaging system with accelerating voltage of 5 kV. Only then Neat_Matrimid® sample
was deposited an ultrathin gold layer using the sputter coater, model KW-4a, Chemat
technology for 18 seconds and 500 rpm.
Samples were analyzed without and with a sputter-coating with gold. The images before and
after the coating procedure are shown in Figure S3.
The MOF is homogenously distributed across the membrane. No particles sedimentation or
agglomeration phenomena were observed. The cross section morphology of both membranes
is characterized by the presence of a sponge-like layer in the top with an average thickness
about 5 µm. This is the effective layer of the MMMs that promotes the actual contact between
Hg2+ and sulfur-functional groups of MOF, during the water flow.
Thermogravimetric Analysis. The thermogravimetric analysis was performed on membrane
samples under a dry N2 atmosphere with a Mettler Toledo TGA/STDA 851e thermobalance.
The experiments were carried out within a temperature range from 25 °C up to 800 °C at a
heating rate of 10 K/min. Approximately 20 mg of the membrane was placed in a ceramic pan
for the measurements.
Permeability measurements of the membranes in distilled water and low mineral water. The
study of water permeability transport properties was carried out by using a Millipore UF
Solvent-resistant Stirred Cell 47 mm XFUF04701 cell. The membrane is located in a circular
compartment with an area of 17 cm2, but the effective membrane area that participated to the
separation process is about 13.84 cm2, because the membrane is held in plane and at the same
time protected by an o-ring with a diameter of 4.2 cm. The flow filtration setup is illustrated in
Figure 5. The water solution was fed to upper side of the membrane by means of a circulation
pump and forced to flow through the membrane under a given trans-membrane pressure, which
was set by regulating the recirculation speed of the solution until the desired pressure was
reached. All experiments were carried out at 25 °C. The water flux (Jw) is determined from the
volume of permeate (VP) collected per unit of time (t) through the given membrane surface area
(A) at fixed trans-membrane pressure (ΔP) values:
3
(a) Spek, A. L. Acta Crystal. Sect. D, Biol. Crystal. 65, 148 (2009). b) Spek, A. L. Acta Crystal. Sect. C-Struct.
Chem. 71, 9-18 (2015).
4
Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837.
5
D. Palmer, CRYSTAL MAKER, Cambridge University Technical Services, C. No Title, 1996.
26
𝐽𝑤 =𝑉𝑝
𝑡 × 𝐴
The measurements of the permeation rate were performed at steady-state conditions under
different ΔP. The water permeance is defined as:
𝑃𝑤 = 𝐽𝑤
𝛥𝑃
and is obtained from the slope of the plot of Jw vs. ΔP
By using the cell, it was possible to apply a pressure gradient to a certain volume of distilled
water that would pass through the membrane. In view of the experiment in the low mineral
content matrix for greater accuracy of the data, the flow and permeance values (Table S3 and
Figure S10) for each membrane were recorded, also in this other matrix, in this specific case
commercial water was used (Table S5).
Hg2+ Capture experiments
Prior to the evaluation of the MMMs removal efficiency, the capture behavior of MOF 2 was
first evaluated. 10 mg of a polycrystalline sample of 2 was soaked in an aqueous solution
containing 10 ppm of Hg2+ cations and also other common ions found in drinking water like
Na+, K+, Ca2+, Mg2+, HCO3-, Cl-, NO3- and SO42-. The capture process was monitored, through
ICP-MS analyses and results are shown in Figure S5.
Thereafter, static and dynamic adsorption experiments were performed on circular sections of
flat micro-porous membranes (1@Matrimid®, 2@Matrimid® and pure Matrimid®) with an
area of 13.84 cm2 (Data reported on Tables S3-S7). First of all, the capture of mercury(II) in
distilled water was verified. In addition, in order to mimic the real conditions of polluted
matrices, two other experiments were conducted in an oligomineral matrix contaminated by
different concentrations of mercury. The experiment with higher concentration (oligomineral
matrix I) is aimed to demonstrate the efficiency of the device, while lowering the amount of
contaminant (oligomineral matrix II), it has been shown that due to the strong affinity of MOFs
for mercury, the device is not only effective but also sensitive towards the capture of small
quantities (traces) of mercury(II).
During the dynamic adsorption, the mixed matrix membranes were placed in contact with a
solution containing mercury chloride at a known concentration, recirculating 100 mL of volume
in the Millipore UF Solvent-resistant Stirred Cell 47 mm cell XFUF04701. A Masterflex L / S
Economy Variable-Speed Drive, 20 to 600 rpm, 115 VAC peristaltic pump was used for the
recirculation of the solution.
The analysis on samples were performed for a total number of 72 cycles (a cycle to recirculate
100 mL of the mercury polluted solution; it is performed in 40 minutes). The stabilization of
the mercury in view of the ICP-MS analysis occurred by adding nitric acid. After each cycle, a
volume of 200 µl of solution was picked up and stabilized for ICP-MS analysis of mercury by
addition of 200 µl of nitric acid with a purity of 98 %. The ICP-MS data are reported considering
the dilution factor.
27
Firstly, the effective capture in distilled water was assessed, then, to confirm the effectiveness
of the membranes in operating conditions close to the real matrices, commercial mineral water
with a defined mineral composition was used (reported in Table S5).
In the case of static adsorption, the three different types of membrane were placed in contact
with 100 mL of a solution with a known concentration of mercury(II) chloride. Data reported
in Tables S4, S6-S7. The stabilization of the mercury in solution, in view of the ICP-MS
analyses, was carried out by adding concentrated nitric acid.
Microscopy measurements
Scanning Electron Microscopy (SEM) Imaging of the membrane surface and cross sections
(1@Matrimid® and 2@Matrimid® and pure Matrimid®) was carried out on a LEO 420 SEM.
Samples were freeze-fractured in liquid nitrogen and were not sputter coated with gold before
analysis.
Scanning Electron Microscopy (SEM) elemental analysis was carried out for 1@Matrimid®
and 2@Matrimid® after Hg adsorption, using a HITACHI S-4800 electron microscope coupled
with an Energy Dispersive X–ray (EDX) detector. Data was analyzed with QUANTAX 400.
Both membranes were suspended in a 100 ppm HgCl2 aqueous solution for 24 h. and then
introduced in pure water for 15 minutes to rinse the unbound Hg salt from the pores of the
membrane and for safety reasons. The samples were then dried under atmospheric conditions
for 1 h. and were sputter coated with gold before the analysis.
28
Table S1. Selected dataa from the ICP-MS analysesb
for the aqueous mother solution during the Hg2+
adsorption process by 10 mg of a polycrystalline
sample of MOF 2 (center column) and the previously
reported MOF 1[25] (right column).
Time (min.)
[Hg2+]
[Hg2+]
0
9911
9863
1
5194
6764
5
1741
3454
10
978
1088
15
411
778
30
63.4
166
45
11.2
67.7
60
5.01
47.6
75
5.00
35.5
90
4.92
25.5
120
4.94
12.7
180
4.87
9.22
240
4.70
8.44
300
4.68
8.17
360
4.60
7.96
720
4.61
7.41
1440
4.60
7.43
4320
4.60
-
a Results are given as
g/L. b LOD: 0,012 μg/L
29
Table S2. Summary of Crystallographic Data for 2 and HgCl2@2.a
Compound
2
HgCl2@2
Formula
C30Cu6CaH72S6N6O37
C30Cl2Cu6CaH56HgS6N6O29
M (g mol–1)
1722.61
1849.9701
Å
0.6889
0.6889
Crystal system
Hexagonal
Hexagonal
Space group
P63
P63
a (Å)
17.807
18.3828(13)
c (Å)
12.54520(10)
11.8283(10)
V (Å3)
3445.20(3)
3461.6(6)
Z
2
2
calc (g cm–3)
1.661
1.775
µ (mm–1)
1.912
3.945
T (K)
100
100
range for data
collection (°)
2.028 to 36.067
2.079 to 29.178
Completeness to
=
25.0
100%
100%
Measured reflections
77003
50992
Unique reflections
(Rint)
11411 (0.0632)
6584 (0.0676)
Observed reflections [I
> 2
(I)]
8560
5196
Goof
0.962
1.33
Rb [I > 2
(I)] (all data)
0.0552(0.0676)
0.0770 (0.0920)
wRc [I > 2
(I)] (all
data)
0.1824(0.1931)
0.2179(0.2281)
a Crystallographic Data for 1 can be found at ref.[58]
b R = ∑(|Fo| – |Fc|)/∑|Fo|. c wR = [∑w(|Fo| – |Fc|)2/∑w|Fo|2]1/2.
30
Table S3. Permeability data of distilled water and low mineral water through the tested membranes. The
permeability values are calculated as the slope of the graph given by the measured flow (L m-2 h-1) with respect to
the pressure (bar).
Permeability of deionized water
Permeability of Oligo-mineral water
Membrane
Permeability
(L m-2 h-1 bar-1)
Membrane
Permeability
(L m-2 h-1 bar-1)
1@Matrimid®
37.16
1@Matrimid®
35.72
2@Matrimid®
35.68
2@Matrimid®
35.84
Matrimid®
1.10
Matrimid®
1.08
31
Table S4. Residual Hg2+ concentration in the stock solution analyzed with the ICP-MS under static and dynamic
adsorption processes for 1@Matrimid®, 2@Matrimid® and neat Matrimid®, using deionized water with high
concentrations of pollutant.
Mercury concentration in deionized water (µg/L)
Static adsorption
Dynamic adsorption
time
(min)
Neat_Matrimid®
1@Matrimid®
2@Matrimid®
Number of
cycles
1@Matrimid®
2@Matrimid®
0
2610
2610
2610
0
2210
2210
15
716
403
262
1
1234
1076
30
457
226
225
2
1137
1044
60
430
125
138
3
801
752
90
411
103
114
12
571
550
1440
369
98.0
87.1
36
420
390
2880
352
63.1
52.2
72
377
230
4320
348
52.8
47.5
32
Table S5. Mineral water content concentration expressed in (mg L-1).
Mineral water composition (mg L-1)
Sodium
1.5
Bicarbonate
10
Fluoride
< 0.10
Calcium
2.9
Nitrate
0.81
33
Table S6. Selected data from the ICP-MS analyses for the Hg2+ as static and dynamic adsorption for
1@Matrimid®, 2@Matrimid® and pure Matrimid® in oligo mineral water with added high concentrations of
pollutant.
Mercury concentration in oligo mineral water I (µg/L)
Static adsorption
Dynamic adsorption
time
(min)
Neat_Matrimid®
1@Matrimid®
2@Matrimid®
Number
of cycles
1@Matrimid®
2@Matrimid®
0
2220
2220
2220
0
2090
2090
15
1170
740
520
1
1220
890
30
950
580
228
2
1040
860
60
900
490
120
3
885
645
90
830
405
60.2
12
531
398
1440
662
355
55.9
36
427
310
2880
538
210
50.8
72
380
238
4320
500
160
40.9
34
Table S7. Selected data from the ICP-MS analyses for the Hg2+ as static and dynamic adsorption for
1@Matrimid®, 2@Matrimid® and pure Matrimid® in oligo-mineral water with added low concentrations of
pollutant.
a LOD = 0.1 gL-1
Mercury concentration in oligo mineral water II (µg/L)
Static adsorption
Dynamic adsorption
time
(min)
Neat_Matrimid®
1@Matrimid®
2@Matrimid®
Number
of cycles
1@Matrimid®
2@Matrimid®
0
373
373
373
0
330
330
15
227
227
228
1
236
131
30
219
206
195
2
190
81.5
60
200
170
129
3
133
64.6
90
144
98.9
55.3
12
13.1
6.22
1440
97.8
52.8
32.5
36
2.33
1.89
2880
88.1
21.4
10.1
72
1.78
1.26
4320
76.6
1.85
< LODa
35
Table S8. Hg2+ removal (%) of 1@Matrimid®, 2@Matrimid® and pure Matrimid® for static and dynamic
adsorption.
Static adsorption
time
(min)
Hg2+ solution in Deionized water
Hg2+ solution in Oligo mineral
water I
Hg2+ solution in Oligo mineral
water II
Neat_Matrimid®
1
2
Neat_Matrimid®
1
2
Neat_Matrimid®
1
2
0
0
0
0
0
0
0
0
0
0
15
72.6
84.6
90.0
47.3
66.7
76.6
39.0
39.1
38.7
30
82.5
91.3
91.8
57.2
73.9
89.7
41.3
44.7
47.6
60
83.5
95.2
94.7
59.,5
77.9
94.6
46.4
54.5
65.3
90
84.3
96.1
95.6
62.6
81.8
97.3
61.3
73.5
85.2
1440
85.7
96.2
96.7
70.2
84.0
97.5
73.8
85.8
91.3
2880
86.5
97.6
98.0
75.8
90.5
97.7
76.4
94.3
97.3
4320
86.7
98.0
98.2
77.5
92.8
98.2
79.5
99.5
<LOD
Dynamic adsorption
time
(min)
Number
of cycles
Hg2+ solution in
Deionized water
Hg2+ solution in Oligo
mineral water I
Hg2+ solution in Oligo
mineral water II
1
2
1
2
1
2
0
0
0
0
0
0
0
0
40
1
44.1
51.2
41.6
57.4
28.6
60.3
80
2
48.5
52.7
50.2
58.9
42.4
75.3
120
3
63.7
65.9
57,7
69.1
59.7
80.4
8 hours
12
74.1
75.1
74.6
81.0
96.0
98.1
24 hours
36
81.0
82.3
79.6
85.2
99.3
99.4
48 hours
72
82.9
89.6
81.8
88.6
99.5
99.6
36
Table S9. Adsorption expressed as µg/cm2 for 1@Matrimid®, 2@Matrimid® and pure Matrimid® in static and
dynamic adsorption, obtained for the reported experiments (100 mL of polluted solutions) calculated as following:
a(Ci-Cf)/ surface areab cm2
Dynamic adsorption (µg/cm2)
Membrane
Hg2+ solution in Deionized
water
Hg2+ solution in Oligo
mineral water I
Hg2+ solution in Oligo mineral
water II
1@ Matrimid®
137.9
123.6
23.7
2@ Matrimid®
143.0
133.8
23.7
The green section underlines the reached potable water limits.
a Concentrations expressed in ppb (from Tables S4, S6-S7).
b In the static adsorption, the membrane’s area measured 17.34 cm2 whereas in the dynamic the exposed area was
of 13.84 cm2.
Static adsorption (µg/cm2)
membrane
bHg2+ solution in Deionized
water
Hg2+ solution in Oligo
mineral water I
Hg2+ solution in Oligo mineral
water II
Neat_Matrimid®
130.4
99.2
17.1
1@ Matrimid®
147.5
118.8
26.8
2@ Matrimid®
147.8
125.7
26.9
37
Table S10. Hg2+ removal (g L-1) of 1@Matrimid®, 2@Matrimid® and pure Matrimid® after three regeneration
cycles for static and dynamic adsorption of Hg2+ in oligo mineral water. Values indicate the residual concentration
in the solutions.
Static adsorption
Dynamic adsorption
Time
(min)
Matrimid®
1@Matrimid®
2@Matrimid®
Number of
cycle
1@Matrimid®
2@Matrimid®
0
361
361
361
0
300
300
15
269
231
230
1
248
234
30
240
229
182
2
188
176
60
187
155
135
3
141
84.3
90
151
105
102
12
135
61.7
1440
111
77.2
69.5
36
82.4
54.0
2880
95.8
45.4
43.1
72
40.7
35.9
4320
66.6
37.0
29.7
38
Table S11. Hg2+ removal (%) of 1@Matrimid®, 2@Matrimid® and pure Matrimid® after three regeneration
cycles for static and dynamic adsorption of Hg2+ in oligomineral water.
Static adsorption
Dynamic adsorption
Time
(min)
Matrimid®
1@Matrimid®
2@Matrimid®
Cycle
1@Matrimid®
2@Matrimid®
15
28.0
38.1
38.2
1
24.7
29.0
30
35.7
38.7
51.1
2
43.1
46.6
60
49.8
58.3
63.7
3
57.2
74.5
90
59.5
71.9
72.7
12
59.1
81.3
1440
70.3
79.3
81.4
36
75.0
83.6
2880
74.3
87.8
88.4
72
87.7
89.1
4320
82.1
90.1
92.0
39
Figure S1. Thermo-Gravimetric Analysis (TGA) of 2 (red) and HgCl2@2 (blue) under dry N2 atmosphere.
40
Figure S2. a) Calculated (bold lines) and experimental (solid lines) PXRD pattern profiles of 2 (red) and HgCl2@2
(blue) in the 2θ range 2.0–60.0°.
41
cross-section
Upper layer
1 000 x
10 000 x
50 000 x
Hyflon® coated
Matrimid®
1
2
Uncoated
Matrimid®
1
2
Figure S3. SEM images of mixed matrix membranes 1@Matrimid® and 2@Matrimid® and of the neat
Matrimid® membrane at magnification of 1000 X, 10.000 X and 50. 000 X at an accelerating voltage of 10 Kv.
The Hyflon® coated and uncoated membranes are compared.
0,05
20 µm
5 µm
1 µm
1 µm
1 µm
1 µm
1 µm
1µm
5µm
5 µm
5 µm
5 µm
5 µm
50 µm
50 µm
20 µm
50 µm
50 µm
42
Figure S4. Kinetic profile of the mercury(II) capture by MOF 2 measured as the decrease of the metal
concentration with time after soaking 10 mg of a polycrystalline sample in an aqueous solution containing 10 ppm
of HgCl2 in the 0-72 h. interval (ICP-MS data collected in Table S1). The inset shows the capture in the time
interval of 0-100 min.
43
Figure S5. X-ray crystal structure of 2: A) View along c and B) a crystallographic axis. Atom color code: All
atoms from the coordination network are represented as grey sticks, with the only exception of copper(II) (cyan
spheres), calcium(II) (blue spheres) and sulfur atoms (yellow spheres) from methylcysteine residues residing in
pores and conveying receptor properties to the MOF towards Hg2+ metal ions.
44
Figure S6. Details on X-ray crystal structure of 2: A) Dicopper(II) units, {CuII2[(S,S)-Mecysmox]} acting as
linkers between the CaII ions through the carboxylate groups and perspective views of a single channel along c
crystallographic axis B) underlining the large voids through the representation of a dummy atom (blue sphere in
pore) C); representation of a single channel with space filling model (with Van der Waals radii) D). Atom color
code: All atoms from the coordination network are represented as grey sticks, with the only exception of copper(II)
(cyan spheres), calcium(II) (blue spheres) and sulfur atoms (yellow spheres) from methylcysteine residues residing
in pores. In A and D carbon, nitrogen and oxygen atoms have been depicted in grey, light blue and red color.
45
Figure S7. X-ray crystal structure of HgCl2@2 with HgCl2 molecules linked by S atoms of methylcysteine
residues, densely packed within channels. Atom colour code: All atoms from the coordination network are
represented as grey sticks, with the only exception of copper(II) (cyan spheres), calcium(II) (blue spheres) and
sulfur atoms (yellow spheres) from methylcysteine residues residing in pores together with Hg(II) and chloride
represented by purple and green spheres, respectively.
46
Figure S8. Details of crystal structure of HgCl2@2: A) Dicopper(II) units, {CuII2[(S,S)-Mecysmox]} acting as
linkers via methylcysteine arms towards HgCl2 molecules B) perspective view of a single channel of HgCl2@2
crystal structure along c (B) and a crystallographic axis (C) where the large voids are filled by ‘captured’ pollutant
molecules; D) representation of a single channel with space filling model (with Van der Waals radii. Atom color
code: All atoms from the coordination network are represented as grey sticks, with the only exception of copper(II)
(cyan spheres), calcium(II) (blue spheres) and mercury(II) (purple spheres), Chloride (green spheres) and sulfur
atoms (yellow spheres) from methylcysteine residues residing in pores. In A and D carbon, nitrogen and oxygen
atoms have been depicted in grey, light blue and red color.
47
Figure S9. Thermo Gravimetric Analysis TGA of the 1@Matrimid®, 2@Matrimid® and a pure Matrimid®
shown for comparison.
48
Figure S10. Water flux (J) with distilled water and low mineral water through the tested membranes as a function
of the applied pressure. The permeability is calculated from the slope of the curves and the values are reported in
Table S3.
49
Figure S11. Kinetic profiles of the mercury(II) capture by 1@Matrimid®, 2@Matrimid® and a pure Matrimid®
membrane after soaking a circular flat membrane (surface of 17.34 cm2) in an aqueous solution of HgCl2 in the 0-
72 h interval. The initial [Hg2+] in deionized water are 2.60 ppm and 2.21 ppm for static (A) and dynamic (B)
adsorption, respectively (Table S4). (B) Removal efficiencies (%) of 1@Matrimid®, 2@Matrimid® and a pure
Matrimid® membrane under the same conditions for static (C) and dynamic (D) adsorption (Table S8). 48 h
corresponds to 72 cycles of microfiltration (the duration of a cycle for the recirculation of 100 mL of solution is
40 minutes).
50
Figure S12. Kinetic profiles of the mercury(II) capture by 1@Matrimid®, 2@Matrimid® and a pure Matrimid®
membrane after soaking a circular flat membrane (surface of 17.34 cm2) in an aqueous solution of HgCl2 in the 0-
72 h interval. The initial [Hg2+] in oligo mineral water are 2.22 ppm and 2.09 ppm for static (A) and dynamic (B)
adsorption, respectively (Table S6). (B) Removal efficiencies (%) of 1@Matrimid®, 2@Matrimid® and a pure
Matrimid® membrane under the same conditions for static (C) and dynamic (D) adsorption (Table S8). 48 h
corresponds to 72 cycles of microfiltration (the duration of a cycle for the recirculation of 100 mL of solution is
40 minutes).
A
B
C
D
51
Figure S13. (a) Backscattered SEM image of 1@Matrimid® and the corresponding EDX elemental mapping for
Cu (b), S (c), Hg (d) and Ca (e) elements. (f) Superposition of images a-e. The backscattering detector highlights
the MOF particles as brighter areas due to crystalline MOF structure and to the presence of heavier atoms in the
MOF than in the polymer matrix.
52
Figure S14. (a) Backscattered SEM image of 2@Matrimid® and the corresponding EDX elemental mapping for
Cu (b), S (c), Hg (d) and Ca (e) elements. (f) Superposition of images a-e. The backscattering detector highlights
the MOF particles as brighter areas due to crystalline MOF structure and to the presence of heavier atoms in the
MOF than in the polymer matrix.
53
Figure S15. Adsorption performance (g L-1) of 1@Matrimid®, 2@Matrimid® membranes reused after three
regeneration cycles for static (A) and dynamic (B) adsorption of Hg2+ (data from Table S10).
54
Figure S16. Adsorption performance (%) of 1@Matrimid®, 2@Matrimid® membranes reused after three
regeneration cycles for static (A) and dynamic (B) adsorption of Hg2+ (data from Table S11).
A preview of this full-text is provided by Wiley.
Content available from Advanced Functional Materials
This content is subject to copyright. Terms and conditions apply.
