View Article Online / Journal Homepage / Table of Contents for this issue

APPLICATION www.rsc.org/materials | Journal of Materials Chemistry

Metal–organic frameworks—prospective industrial applications{

Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

U. Mueller,* M. Schubert, F. Teich, H. Puetter, K. Schierle-Arndt and J. Pastré

Received 22nd August 2005, Accepted 26th October 2005
First published as an Advance Article on the web 23rd November 2005
DOI: 10.1039/b511962f

The generation of metal–organic framework (MOF) coordination polymers enables the tailoring
of novel solids with regular porosity from the micro to nanopore scale. Since the discovery of
this new family of nanoporous materials and the concept of so called ‘reticular design’, nowadays
several hundred different types of MOF are known. The self assembly of metal ions, which act
as coordination centres, linked together with a variety of polyatomic organic bridging ligands,
results in tailorable nanoporous host materials as robust solids with high thermal and
mechanical stability.
Describing examples of different zinc-containing structures, e.g. MOF-2, MOF-5 and IRMOF-
8 verified synthesis methods will be given, as well as a totally novel electrochemical approach for
transition metal based MOFs will be presented for the first time.
With sufficient amounts of sample now being available, the testing of metal–organic
frameworks in fields of catalysis and gas processing is exemplified. Report is given on the catalytic
activation of alkynes (formation of methoxypropene from propyne, vinylester synthesis from
acetylene). Removal of impurities in natural gas (traces of tetrahydrothiophene in methane),
pressure swing separation of rare gases (krypton and xenon) and storage of hydrogen (3.3 wt% at
2.0 MPa/77 K on Cu-BTC-MOF) will underline the prospective future industrial use of metal–
organic frameworks in gas processing. Whenever possible, comparison is made to state-of-art
applications in order to outline possibilities which might be superior by using MOFs.

1. Introduction based on metals like zinc, nickel, iron, aluminium (but also on
thorium and uranium) employing bi- to tetravalent aromatic
As early as 1965 a first publication by Tomic1 on novel solids carboxylic acids are described . Interesting features of these
was introduced which, nowadays, would be categorized and compounds such as high thermal stability and high metal
addressed as metal–organic frameworks, coordination poly- content were already investigated.
mers or supramolecular structures. Already in the aforemen- However, decades later interest in the field was stimulated
tioned contribution simple syntheses of coordination polymers by the group of O. M. Yaghi, which published the structure of
MOF-5 in late 1999,2 and the concept of reticular design, with
BASF Aktiengesellschaft, Chemicals Research & Engineering, D-67056, totally different carboxylate linkers, in 2002.3–5 Meanwhile,
Ludwigshafen, Germany. E-mail: [email protected]
numerous reviews have addressed this fast growing research
{ Presented at Symposium T: Porous materials for emerging applica-
tions, International Conference on Materials for Advanced efforts, the most comprehensive ones given by Kitagawa6 and
Technologies (ICMAT 2005), Singapore, 3–8 July 2005. Yaghi.7 Structures, properties and possible applications as

Ulrich Mueller, born 1957 in purification, piloting of pro-

Katzenelnbogen, Germany. pylene epoxidation catalysts.
1977: studied chemistry in 1999: Synthesis, scale-up,
Mainz (thesis on the synthesis modification and testing of
of large zeolite crystals and various metal–organic frame-
sorption properties) and work compositions. 2005:
recieved his PhD in the group BASF Research Director.
of Prof. K.K. Unger; research
activities at CNRS Markus M. Schubert, born
‘Tian&Calvet’, Marseille, ILL 1971 in Munich, Germany.
Grenoble, and with G.T. 1991: study of chemistry in
Kokotailo, Univ. Ulm and PhD in group of
Pennsylvania. 1989: Ammonia Prof. Behm on catalysis and
Laboratory BASF: zeolite surface chemistry. 2000:
synthesis and application in Postdoc at ETH Zürich with
catalysis and adsorption. 1999: Senior Scientist, zeolite catalysis: Prof. Baiker. 2001: Ammonia Laboratory BASF: catalysts
CFC-free polyurethane foams, catalysts for crop protection carriers, acid–base catalysis. 2004: Scale-up and piloting of
agents, chemical intermediates, sorptive olefin feedstream metal–organic frameworks for gas processing.

626 | J. Mater. Chem., 2006, 16, 626–636 This journal is ß The Royal Society of Chemistry 2006
 View Article Online

storage media were again studied by Rowsell, from Yaghi’s velocity of molecular traffic through these open structures
group.7,8 Comparisons with oxides, molecular sieves, porous are closely related to the regularity of pores in nanometer size
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

carbon and heteropolyanion salts has been filed by Barton as well.

and coauthors.9 Nowadays several hundred different MOFs Thus the combination of so far unreached porosity, surface
are known. The self assembly of metal ions, which act as area, pore size and wide chemical inorganic–organic composi-
coordination centres, linked together with a variety of tion recently brought these materials to the attention of many
polyatomic organic bridging ligands, results in tailored researchers in both academia and industry, with about 1000
nanoporous host materials as robust solids with high thermal publications on ‘coordination polymers’ per annum.6
and mechanical stability (Fig. 1). Interestingly, unlike other This paper, however, aims to describe how MOF-materials
solid matter, e.g. zeolites, carbons and oxides, a number of can be synthesized using verified synthetic methods as well as
coordination compounds are known to exhibit high frame- by a totally novel electrochemical approach.10
work flexibility and shrinkage/expansion due to interaction With a large range of samples now available, the testing of
with guest molecules.6 The most striking difference to state-of- metal–organic frameworks in fields of catalysis and gas
art materials is probably the total lack of non-accessible bulk processing is enabled. A report is given on the catalytic
volume in metal–organic framework structures. Although high activation of alkynes (the formation of methoxypropene
surface areas are already known from activated carbons and from propyne, vinylester synthesis from acetylene).11 Further
zeolites as well, it is the absence of any dead volume in MOFs examples like olefin polymerization, Diels–Alder reaction,
which principally gives them (on a weight-specific basis) the transesterification6 or cyanosilylation12 are referenced in the
highest porosities and world record surface areas (Fig. 2), literature.
especially with MOF-177, for which values of 4500 m2 g21 are The removal of impurities in natural gas (i.e. traces of
reported.5 Of course, properties like the drastically increased tetrahydrothiophene in methane), pressure swing separation of

Friedhelm Teich, born 1955 in Ludwigshafen, development of

Trier, Germany. 1973: study direct and indirect organic
of Chemical Engineering electrosyntheses, commerciali-
at Karlsruhe (PhD on gas sation of the first electro-
processing). 1986: Dyestuff & dialysis processes in our
Pigments laboratory BASF. company. 1985–1992: plant
2004: BASF Chemicals manager of a chloralkali plant
Research & Engineering, in Ludwigshafen. 1990: estab-
20 years experience in design- lishment of the first chlorine
ing and developing processes based ‘‘chemistree’’ of
for the production of pigments Germany for VCI. During this
and fine chemicals. 2004: time: papers, public discussions
process simulation for applica- and lectures on chlorine
tion of metal–organic frame- chemistry. Since 1993:
work materials. Manager of R&D activities on
all electrochemical processes of our company. Since 1994:
Hermann Pütter, born 1944 in Duesseldorf, Germany. 1951– lectures on environmental and sustainability aspects of chemistry.
1964: school in Duesseldorf. 1964–1972: study of chemistry in 1999: BASF innovation award for the first technical paired
Wuerzburg (thesis on the preparation and elucidation of the electrosynthesis, a process with high atom efficiency that avoids
electrochemical properties of squaric acid derivatives). 1969: emissions and halves the energy demand of an electrosynthesis.
summer/autumn: electroanalytical studies at the Heyrovsky 2001: BASF Research Fellow. 2003, Synthesis of metal–organic
Institute in Prague. 1973–1985: Main Laboratory BASF, frameworks using electrochemistry.

Kerstin Schierle-Arndt, born Joerg Pastré, born 1970, in

1971 in Köln, Germany. 1990: Groß-Gerau, Germany. 1977:
study of chemistry in Bonn; study of chemistry in
PhD in organic electro- Darmstadt. 2000: PhD on
chemistry. 1998: Ammonia Chemical Engineering at ETH
Laboratory BASF: electro- Z ü r i c h . 2 0 0 0 : A m m o n i a
chemical research. 2003: laboratory BASF: process
Chemicals Research & development department.
Engineering, Head of 2005: New business develop-
Controlling. 2005: New ment at BASF Inorganic
Business development at Specialties.
BASF’s Inorganic Specialties.

This journal is ß The Royal Society of Chemistry 2006 J. Mater. Chem., 2006, 16, 626–636 | 627
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM. View Article Online

Fig. 3 Principle flowsheet scheme of industrial MOF synthesis

procedure. Cost efficiency and sustainability requires solvent recycling
and zinc oxide rather than zinc nitrate.
Fig. 1 Illustration of metal–carboxylate building units of MOFs
(upper left: MOF-5 with a Zn4O-cluster linked to the terephthalic acid
repeatedly checked and scaled-up to the order of kg. A typical
molecules, upper right: IRMOF-8 with a Zn4O-cluster attached to the
2,6-naphthalene dicarboxylic acid, lower left: Cu-BTC with a dimeric scheme of a semi-technical process is given in Fig. 3, indicating
Cu-cluster terminated by 1,3,5-benzenetricarboxylic acid, lower right: the different steps of preparation, as well as recycling of
MOF-2 showing the paddle-wheel of a dimeric Zn-cluster linked to the the solvent and further processing of the dried powders into
terephthalic acid units). shaped particles.

2.1 Verified lab-scale synthesis recipes of MOF samples

MOF-2. A glass reactor equipped with a reflux condenser

and a teflon-lined stirrer was filled with 24.9 g of terephthalic
acid (BDC) and 52.2 g of zinc nitrate tetrahydrate (Merck)
were dissolved in a mixture of 43.6 g of N-methyl-2-
pyrrolidone, 8.6 g of chlorobenzene and 24.9 g of dimethyl-
formamide (Merck) and heated up to 70 uC for a total of 3 h.
After about 60 min, 30 g triethylamine was added. The white
precipitate formed was filtered off, dried at room temperature
and finally heated at 200 uC for 8 h. The molar yield based on
zinc amounted to 87%.

MOF-5. Uniformly large crystals of MOF-5 were synthe-

sized by using an optimized procedure starting from
terephthalic acid, zinc nitrate and diethylformamide as organic
solvent.
In a glass reactor equipped with a reflux condenser and a
teflon-lined stirrer, 41 g of terephthalic acid (BDC) and 193 g
of zinc nitrate tetrahydrate (Merck) were dissolved in 5650 g of
diethylformamide (BASF AG; ,100 ppm water) and heated
Fig. 2 Framework of MOF-5 displaying free access to nanosized
voids and the absence of non-accessible bulk-volume. The picture
up to 130 uC for 4 h. After about 45 min, crystallization started
shows a MOF-5 particle of about 500 nanocells with a cube edge and the formerly clear solution turned slightly opaque. After a
length of about 100 nm. total of 4 h, the reaction product was cooled down to room
temperature. The solid was filtered off, washed three times
rare gases (krypton and xenon) and storage of hydrogen with 1 L of dry acetone and dried under a stream of flowing
(3.3 wt% at 2.5 MPa/77 K on Cu-BTC-MOF) will underline nitrogen. Finally the product was activated at 60 uC for at least
the prospective industrial use of metal–organic frameworks. 3 h under a reduced pressure of ,0.2 mbar.
Whenever possible, comparison is made to state-of-art applica- The wet chemical analysis of the thus obtained solid
tions in order to outline possibilities of processes which might yielded 33 wt% Zn, equivalent to 91% molar yield of MOF-5
be beneficially run by using MOFs. calculated as Zn4O(BDC)3. The concentration from residual
nitrate amounted to 0.05 wt% N. Cubic shaped crystals in
between 50–150 mm size are to be observed by scanning
2. Experimentals
micrographs (Fig. 4). The PXRD pattern is shown in Fig. 5.
In order to enable readers to look into the properties of MOFs, Specific surface area measurements with nitrogen at 77 K were
some easy-to-follow recipes are listed hereafter. They were all determined as 3400 m2 g21.

628 | J. Mater. Chem., 2006, 16, 626–636 This journal is ß The Royal Society of Chemistry 2006
 View Article Online

zinc nitrate tetrahydrate, diethylformamide and N-methyl-2-

pyrollidone as organic solvent.
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

In a glass reactor equipped with a reflux condenser and a

teflon-lined stirrer 7.5 g of naphthalenedicarboxylic acid
(NDC) and 67.4 g of zinc nitrate tetrahydrate (Merck) were
dissolved in 883 g of diethylformamide (BASF AG) and heated
up to 130 uC for 4 h. After about 75 min crystallization started
and the formerly clear solution became turbid. After the reac-
tion the product was cooled down to room temperature and
the precipitate was filtered off, washed three times with 1 L of
dry chloroform and dried under a stream of flowing nitrogen.
Finally the product was activated at 60 uC for 3 h under
reduced pressure of ,0.2 mbar giving 9.5 g of the final product.
Wet chemical analysis of the solid yielded 27.3 wt% Zn
equivalent to 15% molar yield of IRMOF-8 calculated on zinc
and 92% on NDC (calculated as Zn4O(NDC)3). Concentration
from residual nitrate amounted to 0.78 wt% nitrogen. The
Fig. 4 SEM-picture of large MOF-5 crystals from laboratory Langmuir specific surface area reached 1750 m2 g21. All
preparation (scale bar: 1 mm). crystals of IRMOF-8 had a cubic shape of about 100 mm size,
however, the scaly morphology indicated a high degree of
Using argon adsorption at liquid argon temperature (87 K) intergrowth.
the adsorption properties were compared to state-of-art
materials like zeolite X and activated carbon (Ceca, carbon Cu-MOF. For the first time, to our knowledge, MOFs are
AC40). As depicted in Fig. 6 the outstanding uptake behaviour synthesized using an electrochemical route.
of argon on MOF-5 is obvious and clearly exceeds zeolites and Bulk copper plates, thickness 5 mm, are arranged as the
carbon components. anodes in an electrochemical cell with the carboxylate linker,
viz. 1,3,5-benzenetricarboxylic acid, dissolved in methanol as
IRMOF-8. Large crystals of IRMOF-8 were grown from a solvent and a copper cathode. Details are to be found in.10
starting mixture containing 2,6-naphthalenedicarboxylic acid, During a period of 150 min at a voltage of 12–19 V and a

Fig. 5 PXRD of large MOF-5 crystals from laboratory preparation indicating superior crystallinity.

This journal is ß The Royal Society of Chemistry 2006 J. Mater. Chem., 2006, 16, 626–636 | 629
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM. View Article Online

Fig. 6 Adsorption isotherms of argon at 87 K on MOF-5 compared

to activated carbon and zeolite X.

Fig. 8 Synthesis cell for electrochemical preparation of MOFs with

Cu-plates as electrode material (laboratory scale).

Selected samples of electrochemically prepared Cu-MOF

were compared to conventionally synthesized materials by
EXAFS and XANES. Spectra were collected on beamline E4
at HASYLAB in the DESY synchrotron, Hamburg. Single
crystal data of monoclinic tenorite mineral (Cc-symmetry)
served as model basis. Although the X-ray powder diffraction
pattern of both Cu-BTC-MOF conventionally synthesized and
Fig. 7 SEM-picture of Cu-MOF crystals from novel electrochemical
preparation (scale bar: 1 mm).
electrochemically prepared are very close to each other, the
samples offer clear differences with regard to the local fine-
structure of the copper. The electrochemically prepared
currency of 1.3 A, a greenish blue precipitate was formed.
Cu-MOF gives Cu–K edge spectra which nicely agree with
After separation by filtration and drying at 120 uC overnight
the spectra of the mineral tenorite having as model a fourfold
pure Cu-MOF was obtained. The surface area determination
coordination of Cu, whereas the Cu-BTC-MOF following
yielded 1649 m2 g21 and after activation at 250 uC finally
literature recipes12–14 is indicative of having a disturbed
1820 m2 g21 giving a dark blue coloured solid of octahedral
Cu-environment and an additional peak at 8.98 keV photon
crystals from 0.5 to 5 mm size (Fig. 7)
energy. The latter one is presumably due to the existence of
The experimental setup of the electrochemical cell in a
occluded nitrate moieties close to the open metal copper
laboratory glass reactor being under operation is depicted in
ligand site, which could also explain the lower surface areas
Fig. 8.
of only 917 m2 g21.14 Beneficially, this does not occur on the
electrochemically prepared material.10 The latter is a very
2.2 Analysis and characterization
useful adsorbent with a strong possibility to attract electron-
Adsorption measurements were performed with commercially rich molecules on the open copper-sites, as is illustrated in our
available equipment (Autosorb 6, Quantachrome Corp.) using gas purification example in section 2.5, below.
nitrogen sorption at 77 K. Prior to measurement, samples In order to measure intracrystalline self-diffusion, pulsed
were activated down to 1024 mbar, first at 60 uC and finally field gradient NMR measurements were performed on the
at 120 uC, until a constant vacuum was achieved for 14 h. NMR spectrometer FEGRIS 400 NT using the 13-interval
Surface area values were calculated according to the Langmuir stimulated spin echo pulse sequence with two pairs of
equation. alternating pulsed magnetic field gradients at the group of
X-Ray powder diffraction was monitored on sealed samples Kärger.15 The results were compared to literature data
stored under dry nitrogen using Cu Ka radiation (D 5000, obtained on NaX zeolite crystals and will be discussed in
Siemens). Scanning electron micrographs were taken at a 10 kV detail elsewhere.15
electron beam using field emission cathode arrangement (JSM Nevertheless, for ethane and benzene, in summary from
6400F; Jeol). Fig. 9, it can be seen that diffusion in MOF-5 is clearly orders

630 | J. Mater. Chem., 2006, 16, 626–636 This journal is ß The Royal Society of Chemistry 2006
 View Article Online

Table 1 Vinyl group esterification on MOFs

Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

Surface Conversion Selectivity

Catalyst Area/m2 g21 (mol% acid) (mol% ester)

MOF-2 330 94 83
MOF-5 3400 91 56

Interestingly, although both MOF catalyst materials reach

at almost the same conversion of tert-butylbenzoic acid, the
selectivity towards the vinyl ester is expressed far more by
the zinc paddle-wheel containing MOF-2, with about a
tenfold lower surface area.16 MOF-5 with fully saturated
zinc-coordination obviously is quite poor in performing in the
Fig. 9 Comparison of diffusivity of ethane and benzene molecules in
selective esterification.
MOF-5 over zeolite X experimentally determined from PFG-NMR
Although a clear understanding of the underlying
measurements.15
mechanism is not yet available, there is without any doubt
proof that MOFs can act as catalysts. Successful catalysis on
of magnitude faster than in zeolite NaX. This is considered to
zinc-containing MOFs in the activation of alkoxides and
be a consequence of the difference between the diameters of the
carbon dioxide into polypropylene carbonate has already
two large nanoporous cavities in the MOF-5 (1.1–1.3 nm2)
been reported.17,18 Even enantioselective conversions with
over the smaller NaX supercage (1.2 nm) with an even lower
an enatiomeric excess of 8% on homochiral metal–organic
entrance window size (0.7 nm). The effective self-diffusion
POST-1 have been addressed by Kim and for heterogeneous
coefficient of benzene in the MOF-5 structure is only slightly
asymmetric catalysis by Lin19,20 towards chiral secondary
smaller than the self-diffusion coefficient in the neat liquids at
alcohols.
the same temperature (C6H6: 2.5 61029 m2 s21). For both
Further catalytic reactions from other research groups on
catalysis and gas processing, this is an important observation
and promises fast molecular transport and low mass transfer MOFs were recently reviewed and collected by Kitagawa.6
resistance as additional benefit when using MOF materials Ziegler–Natta-type polymerization, Diels–Alder-reaction,
rather than zeolites. In industrial applications of MOFs this transesterification, cyanosilylation of aldehydes, hydrogena-
might contribute to permanent benefits in variable energy cost tion and isomerization reactions were reported.
over state-of-art solids. Future research is now dedicated to find out if the metal
centers, the ligands or functionalized ligands, or even metal–
2.3 Catalysis ligand interactions or differences in particle size, may cause
unusual catalytic properties and if MOFs can compete with
For testing the catalytic activity in methoxypropene formation, well-known heterogeneous industrial catalysts. Beyond activity
a differentially-running reactor (volume 200 ml) filled with and selectivity, also the questions of time-on-stream behaviour
55 g of MOF-2 tablet material, obtained according to the and leaching stability will be crucial.
preparation, using pyrrolidone as solvent, were packed in a
fixed bed basket-type arrangement. The reactor was heated to 2.4 Gas purification
250 uC and fed with a mixture of methanol–cyclohexane at a
10 : 1 ratio and at a rate of 1.5 g h21. Propyne was added at a Cu-BTC-MOF, prepared as described above, was used in
flux of 6 g h21. Product analysis was done on line using gas demonstrating the removal of sulfur odorant components from
chromatography. natural gas. In a fixed bed reactor vessel (inner diameter of
When the system had reached steady state conditions the 10 mm) about 10 g of granular Cu-MOF of a particle size
conversion of propyne was calculated to be 30% with a fraction of 1–2 mm were thoroughly packed. At a temperature
selectivity of 80% into 2-methoxypropene.11 of 25 uC a gas stream of methane odorized with 13 ppm of
Using zinc silicate as catalyst in a comparative example the tetrahydrothiophene (THT) was fed over the packing and
selectivity towards 2-methoxypropene was found to be 77%, analyzed in the reactor effluent by means of gas chromato-
however, full conversion at temperatures of only 180–200 uC graphy until breakthrough occurred (Fig. 10).
was obtained.21 The used catalyst was removed from the reactor and
In a different reaction, the esterification of 4-tert-butylben- analyzed for residual sulfur content using bulk wet chemical
zoic acid was performed in a batch type test configuration. An analysis. For comparison, the trials were repeated using
autoclave was filled with 2.5 g of MOF-2 material suspended two commercially available activated carbon materials as
in 100 g of N-methyl-2-pyrrolidone and 40 g of 4-tert- adsorbents, viz. Norit (type RB4) and CarboTech (type C38/4).
butylbenzoic acid. After inflating to 0.5 MPa of nitrogen, the Taking the breakthrough curve of tetrahydrothiophene as
reactor vessel was heated to 180 uC and 2 MPa of acetylene depicted in Fig. 10, it can be clearly seen that the sulfur
were introduced. The acetylene pressure was kept constant for odorant is completely omitted from the natural gas in the
24 h. The final liquid product after cooling down was analyzed effluent and even detection by smelling is not any longer
by means of gas chromatography.11 In a repeated trial MOF-5 indicative. The analysis of the used Cu-MOF after the test
was used instead of MOF-2. Table 1 collects the results. cycle reveals an volume specific uptake of 70 g THT L21 MOF

This journal is ß The Royal Society of Chemistry 2006 J. Mater. Chem., 2006, 16, 626–636 | 631
 View Article Online

a thermostatted Dewar vessel. This experimental setup

guaranteed a sufficiently high resolution of storage capacity
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

at a sensitivity of about 104. The reproducibility usually was

better than ¡2%.
Balance B monitored the weight of a high pressure hydrogen
feed flask B (e.g. 20 MPa). By controlled stepwise opening of
valves in the connecting piping steel system between containers
A and B, there was a redundant checking of weight and
pressure increase in the MOF-containing vessel, whereas
weight and pressure drop was indicated in the hydrogen
feeding container B. Of course uptake on one side and release
data on the other side had to fit to each other in terms of
proving a leak-free mass balance. Blank runs without MOF-
sample were repeated prior to each measurement. In order to
Fig. 10 Breakthrough-curve of continuous tetrahydrothiophene- obtain weight-specific uptake the data was corrected for bulk
removal from natural gas using Cu-BTC-MOF out of electrochemical volume using the helium density of the MOF-sample.
synthesis.
As for the MOF-5 sample adsorption isotherms with either
argon, krypton or xenon at room temperature were registered
which is considerably higher than 0.5 g THT L21 for Norit- and compared to their conventional compression curve up to
carbon or 6.5 g THT L21 for the CarboTech sample. about 5 MPa.
Interestingly, the colour of the Cu-MOF had changed from Obviously, as shown in Fig. 11, for all rare gases under
a deep blue into a light greenish one. investigation (Ar, Kr, Xe) the volume specific uptake is
Obviously, Cu-MOF is a powerful material for the separa- higher in case of a gas cylinder being filled with MOF-5. This
tion of polar components from nonpolar gases. Looking at effect is becoming even more expressed from argon over
the structure, the special arrangement of channels in krypton to xenon. In the latter case, at about 10 bar pressure,
Cu-BTC-MOF together with open metal ligand sites offers a the amount of xenon stored in a MOF-filled container is
dual type sorption behaviour.12–14 Bulk adsorption, on one about fourfold higher (if one compares the curves vertically).
hand, exemplified e.g. during nitrogen loading at 77 K, aiming Of course such an enhanced uptake of one gas over another
at pore filling to deduce pore volumes, and secondly, the can be exploited to separate gas mixtures of e.g. krypton and
predominant existence of open ligands at the copper xenon. Details on the performance of such a process are given
paddle wheel-cluster (cf. Fig. 1, lower left structure unit), give in section 2.6.
rise to possible chemisorptive sites, the latter being nicely Looking at a storage level in the xenon curve of about
kept free and unaffected during BASF’s electrochemical 100 g Xe L21 in Fig. 11, it becomes clear from a horizontal
synthesis starting from bulk metal. The anion-free preparation comparison of the isotherms at a constant uptake value that
omits the unwanted occlusion of nitrates into the metal with a MOF-filled container the pressure which is needed
organic framework that is usually applied in state-of-the-art for holding a gas in a given volume is reduced from about 17
preparations.13,14,22,23 to 2 bar.
Further electron-rich molecules that are successfully Thus, storage of a gas in MOF-filled canisters can be used
removed by Cu-MOF in a similar manner are amines and either way to enhance capacity in a given volume or to
ammonia, water traces, alcohols and oxygenates. transport an equivalent amount of gas at a far lower pressure.
In addition, in all cases a dominant color change readily
allows visible detection of breakthrough and contaminant
saturation on the MOF. During removal of the contaminant
by vacuum treatment or heating, the original color reappears,
indicating a possible regeneration of the adsorbent. Depending
on time and temperature applied, chemical analysis of the
sulfur content as well as the final weight of the used Cu-MOF
versus the fresh sample helps to monitor the degree of
regeneration.

2.5 Gas storage

Gas storage both at room temperature and 77 K up to 10 MPa
was measured on a homebuilt gravimetric prototype equip-
ment comprising two highly sensitive balances A and B
(Mettler, PG 5002-S, ¡0.01 g displayed). Balance A carried a
stainless steel (material type 1.4541) fabricated container A of Fig. 11 Compression curves of rare gases Ar, Kr, Xe and comparison
about 460 ml content which was filled up to the neck with over inflation curves into MOF-5-filled gas containers (room
about 150 g of ‘in-situ’ activated MOF sample sitting in temperature, up to 60 bar; lecture bottles).

632 | J. Mater. Chem., 2006, 16, 626–636 This journal is ß The Royal Society of Chemistry 2006
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM. View Article Online

Fig. 12 Compression of propane into gas container with and without Fig. 13 Volume-specific storage curves of hydrogen on different
MOF-filling (MOF-5 tablets in lecture bottles, room temperature). MOF-materials in comparison to compression curve of hydrogen into
empty gas containers (77 K, 40 bar; measurement from BASF
prototype setup as described in section 2.5).
Similarly, other gases like methane and hydrocarbons can be
held at a denser state in the same manner.22,24
The same finding holds for propane as well. In Fig. 12 are the empty container and with steepest incline below 10 bar. At
depicted the curves for shaped MOF-5 pellets in gas containers 40 bar the PVT-relation of hydrogen in an empty canister is
over non-filled containers. Again the inflation curves are registered as 12.8 g H2 L21, whereas Cu-BTC-MOF filled into
markedly different. The non-MOF-filled container represents containers reach a plus 44% capacity of up to 18.5 g H2 L21.
the almost linear uptake behaviour, whereas the MOF-filled For comparison, the volume-specific density of liquid hydro-
can is far higher and with a steeper increase at the beginning. gen at its boiling point (20 K) is 70 g H2 L21.
Taking the pressure at about 10 bar it becomes obvious that Above 10 bar, the curves run mostly parallel to the
one MOF-filled container substitutes the amount of three conventional H2-pressure–volume relation. On a per weight-
state-of-art pressurized cylinders. calculation it becomes clear that the saturation of MOFs with
Already today much interest is attributed to the storage of hydrogen is already achieved at pressures of less than 15 bar
hydrogen for mobile and portable fuel-cell application by (Fig. 14). For electrochemically-prepared Cu-BTC-MOF this
using hydrogen as a carrier for electric power supply.25 Due to attributes to about 3.3 wt% H2-uptake.
the need for alternative fuel sources and energy carriers, the This data from our large scale prototype compares reason-
target values of US Department of Energy were recently ably well with the literature.8
reviewed,8 compiling storage data of some 0.2–3.8 wt% of It should be mentioned here, that for many volume-limited
hydrogen on MOFs, the maximum on a weight-specific
fuel-cell applications, i.e. the mobile and portable cases, it
calculation being found by Férey on MIL-537 derived from
will be industrially much more relevant to compare storage
aluminium salts and BDC. Pillaring with secondary amine
data on a volume-specific rather than on a weight-specific
(triethylenediamine) linkers as strategy was employed by
storage capacity basis. Typically the packing densities of
Kim27 and Seki22 and MOF-505 again by Yaghi-group28
MOF powders are around 0.2 to 0.4 g cm23 increasing to
with Cu paddle-wheels connected by 3,39,5,59-biphenyltetra-
0.5–0.8 g cm23 when shaped into tablets or extrudates. So far,
carboxylic acid ranged up to more than 2 wt%. Doubly
this material density is so low that weight limitation in an
interpenetrated nets of zinc frameworks built by NTB-linkers
application is non-existent, in contrast of course to the
(4,49,40-nitrilotrisbenzoic acid) by Suh29 were reported to reach
alternative use of metal hydride as storage media. Much
1.9 wt%. of hydrogen uptake at 77 K. However, it is not
more importantly, however, there is only a strictly confined
yet possible to foresee if large surface area materials like
volume available in both mobile automotive as well as in
MOF-177,29 MIL-10032 or MOF-5 and isoreticular mem-
bers,2–4 or materials with an average surface of between 1000
to 1500 m2 g2126,29 or even small pore MOFs like30,33 will be
the most promising storage media. Neither can it be concluded
if divalent or trivalent metal-clusters34 are the most favourable
ones. Simulations obviously support metal–organic frame-
works as being favorable over zeolites.31
From results of our prototype equipment (77 K tempera-
ture, up to 40 bar) it can be seen, how different MOF-materials
contribute differently to volume-specific hydrogen storage
(cf. Fig. 13).
Comparing to the pressurizing of an empty container with
hydrogen MOF-5, IRMOF-8 and Cu-BTC-MOF increasingly
take up higher amounts of hydrogen, all of them exceeding the Fig. 14 Volume-specific versus weight-specific storage curve of
standard pressure–volume–temperature (PVT) uptake curve of hydrogen on Cu-BTC-MOF from electrochemical preparation.

This journal is ß The Royal Society of Chemistry 2006 J. Mater. Chem., 2006, 16, 626–636 | 633
 View Article Online

portable electronic device applications. Hence, storage data

communicated as weight-specific rather than volume-specific
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

numbers can be misleading by merely focusing on only one of

them. Depending on the intended application and the given
boundary conditions, it might be necessary in future to
indicate both types of values.
Anyhow, the most important issue is the amount of
hydrogen which, in a reasonable time-scale, can be discharged
from storage media. Here, MOFs really do have a fully
reversible uptake and release behaviour. As the storage
mechanism is predominately physisorption, there are no huge
activation energy barriers to be overcome as compared to e.g.
metal hydrides when trying to liberate the stored hydrogen.
Simple pressure reduction by controlled valve opening is Fig. 15 Gas separation of Kr–Xe mixture by continuous adsorption
sufficient to draw off hydrogen from MOFs within a few on electrochemically produced Cu-BTC-MOF.
seconds.
Energy density values of 1.1 kW h L21, as requested in the
1.6 cm diameter) filled with 193 g Cu-MOF at a given
European Hydrogen and Fuel Cell Strategic Research Agenda
temperature of 55 uC and 40 bar (Fig. 15). The flow rate
and Deployment Strategy,25 are equivalent to a volumetric
(60 L h21) and reactor dimensions were chosen appropriately
hydrogen storage capacity of about 33 g H2 L21. As
in order to prevent complete saturation of the MOF material
demonstrated, almost two thirds of this value can be reached
during the time-span of the measurement. The pressure
by storing hydrogen in MOFs at 77 K and a moderate pressure
was adjusted manually by a needle valve at the reactor outlet.
of 40–50 bar. Increasing this volumetric capacity by shaping
The gas composition leaving the adsorber was detected by
MOF-powders into tablets and extrudates is proved to be
an on-line mass spectrometer (Pfeiffer Vacuum OmniStar2
feasible. Presently, a high amount of storage at room
QMS-200).
temperature has not been attained.
In the gas stream leaving the adsorber, Xe was reduced to a
Further research concepts are dedicated to increase the
level of ca. 50 ppm. After more than 100 min on-stream the
storage capacity and to shift it to higher temperatures. Once it
MOF became saturated with Xe and a rapid breakthrough was
is clarified where the hydrogen molecules are favourably
observed. The calculated capacity of the Cu-MOF for Xe was
adsorbed and attached in the different MOFs, it will be easy to
more than 60 wt%. This is almost twice as much Xe as a high
browse through the huge variety of several hundred (and
constantly still-increasing) numbers of structures and to surface active carbon (Ceca, AC 40, ca. 2000 m2 g21) could
identify the most promising examples. In this respect, take up under identical conditions.
molecular modelling tools might become as important as Due to the fact that MOFs exhibit a gas molecule mobility
elaborate experimental synthesis efforts.31 It should be kept in of about two to three orders higher than state-of-art molecular
mind that depending on the temperature required for a sieves or active carbons, a faster swing operation period
possible application, highly porous MOFs might be favourable between adsorption and desorption cycle seems possible. In
for low temperatures, whereas rather small pore materials,30,33 terms of economic consideration, this contributes to fairly
or highly attractive and flexible ones,26,35 could be favourites reduced purge time, energy consumption and overall variable
for room temperature storage. New mechanisms bridging costs thus rendering MOFs beneficial over established
chemisorption (as in hydrides) and physisorption (as in metal– technology. Again novel material properties of metal–organic
organic frameworks) might be also requested to meet future frameworks, viz. their porosity with nanometer size pores in
challenges. regular arrays and the absence of blocked bulk volume
contribute to principal differences in performance.
2.6 Gas separation Looking into the literature where some zeolite-like (MTN
topology) metal–organic frameworks have already been
Already during the monitoring of the diverse uptake behaviour observed, it will be interesting to watch if, with the small pore
of rare gases on MOF-5 (Fig. 11) it became obvious that materials, separations based on molecular sieving might
this property could possibly be used to separate mixtures become possible.30,33,36 Given this possibility, MOFs would
of rare gases by adsorption on MOFs. Such mixtures are grow to be competitors to zeolitic molecular sieves, which
usually to be found technically in cryogenic air separation currently have a market volume of some hundred thousand
units and once the gases are separated, xenon and krypton tons capacity per year.
can be marketed separately, e.g. xenon as narcotic medical
gas and krypton as filler for lamp industry. Unlike cryogenic
3. Conclusion and prospects
distillation, the following experiments indicate a far simpler
process of pressure–swing adsorption to separate rare gases In this article, we have shown that metal–organic frameworks
mixtures. or coordination polymers are not merely a new class of porous
A mixture of Kr (ca. 94% mol) and Xe (ca. 6% mol) was fed materials for combining inorganic and organic chemistry
continuously to an isothermal tubular reactor (2 m length, classifications. From an industrial point of view, they offer,

634 | J. Mater. Chem., 2006, 16, 626–636 This journal is ß The Royal Society of Chemistry 2006
 View Article Online

in principle, many interesting and promising features over Acknowledgements

prior art, viz.
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

N world records in surface area Technical assistance of Dr O. Metelkina-Schubert, Dr Cox,

N ultimate porosity with absence of blocked volume in solid S. Lutter, W. Kippenberger, U. Diehlmann, R. Hess, R. Ruetz,
matter I. Schwabauer, R. Senk, and H. Sichler is gratefully acknowl-
edged. U.M. thanks O.M. Yaghi and S. Kitagawa for many
N combined flexible and robust frameworks
stimulating discussions on metal–organic frameworks (MOFs)
N full exposure of metal sites
and coordination polymers (CPLs).
N high mobility of guest species in regular framework
nanopores
N fast growing number of novel inorganic–organic chemical References
compositions. 1 E. A. Tomic, J. Appl. Polym. Sci., 1965, 9, 3745–3752.
Obviously, many applications might (and surely will) be 2 H. Li, M. Eddaoudi, M. O’Keeffe and O. M. Yaghi, Nature, 1999,
tested once verified synthesis recipes of MOFs are available. 402, 276–279.
3 M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. O’Keefe
The recipes given in the Experimental section will allow the and O. M. Yaghi, Science, 2002, 295, 469–472.
reader to prepare these new compounds in laboratory-scale 4 O. M. Yaghi, M. Eddaoudi, H. Li, J. Kim and N. Rosi, WO 2002/
amounts. However, industrial synthesis at BASF is understood 088148, 2002, University of Michigan.
5 H. K. Chae, D. Y. Siberio-Pérez, J. Kim, Y. B. Go, M. Eddaoudi,
to be far more advanced, already into barrel-size pilot scale, A. J. Matzger, M. O’Keeffe and O. M. Yaghi, Nature, 2004, 427,
and additional issues need to be taken into account during the 523–527.
manufacturing procedures, which of course are beyond the 6 S. Kitagawa, R. Kitaura and S. Noro, Angew. Chem., Int. Ed.,
scope of this paper. 2004, 43, 2334–2375.
7 J. L. C. Rowsell and O.M. Yaghi, Microporous Mesoporous
Unlike many other novel materials, e.g. carbon polymorphs, Mater., 2004, 73, 3–14.
fullerenes, bucky-balls, CNT, the metal–organic framework 8 J. L. C. Rowsell and O. M. Yaghi, Angew. Chem., 2005, 117,
materials’ preparation and fabrication does not necessarily 4748–4758.
need additional capital investment into a totally new synthesis 9 T. J. Barton, L. M. Bull, W. G. Klemperer, D. A. Loy,
B. McEnaney, M. Misono, P. A. Monson, G. Pez, G. W. Scherer,
technology. Simply adaptation of conventionally available J. C. Vartuli and O. M. Yaghi, Chem. Mater., 1999, 11, 2633–2656.
precipitation and crystallisation manufacturing methods needs 10 U. Mueller, H. Puetter, M. Hesse and H. Wessel, WO 2005/049892,
to be done. Shaping of metal–organic framework powders 2005, BASF Aktiengesellschaft.
11 U. Mueller, M. Hesse, L. Lobree, M. Hoelzle, J. D. Arndt and
into industrially widespread geometries of mm-sized tablets, P. Rudolf, WO 2002/070526, 2002, BASF Aktiengesellschaft.
extrudates, honeycombs, etc. can be performed on MOFs as 12 K. Schlichte, T. Kratzke and S. Kaskel, Microporous Mesoporous
well without any major obstacle. Mater., 2004, 73, 81–88.
The examples which we gave for catalysis as well as for gas 13 Q. M. Wang, D. Shen, M. Buelow, M. L. Lau, F. R. Fitch and
S. Deng, US Pat. 6 491 740, 2002, The BOC Group, Inc.
processing and storage already indicate that there is much 14 S. S.-Y. Chui, S. M.-F. Lo, J. P. H. Charmant, A. G. Orpen and
room left for many future research efforts (viz. storage of I. D. Williams, Science, 1999, 283, 1148–1150.
alternative energy carriers like small hydrocarbons, odour 15 F. Stallmach, S. Groeger, V. Kuenzel, J. Kaerger, O. M. Yaghi,
M. Hesse and U. Mueller, Angew. Chem., Int. Ed. (submitted).
removal in both stationary—e.g. household—as well as
16 M. Eddaoudi, H. Li and O.M. Yaghi, J. Am. Chem. Soc., 2000,
mobile (bus, train, subway, ship) environments or the adverse 122, 1391–1397.
odorization in carrying perfumes, etc. Pick-up of liquids 17 U. Mueller, G. Luinstra and O. M. Yaghi, US Pat. 6 617 467,
without swelling of solids could be of interest as well, e.g. for 2004, BASF Aktiengesellschaft.
18 R. Eberhardt, M. Allmendiger, M. Zintl, C. Troll, G. A. Luinstra
food packaging or removal of hazard liquids like organic and B. Rieger, Macromol. Chem. Phys., 2004, 205, 42–47.
solvents, oils, brake fluids and the like). 19 Chuan-De Wu, A. Hu, L. Zhang and W. Lin, J. Am. Chem. Soc.,
It is worthwhile to mention that, unlike state-of-art 2005, 127, 8940.
heterogeneous catalysts, the metal sites in MOFs are usually 20 J. S. Seo, D. Whang, H. Lee, S. I. Jun, J. Oh, Y. J. Jeon and
K. Kim, Nature, 2000, 404, 982–986.
fully exposed, therefore giving an ultimately high degree of 21 Ch. Miller, P. Rudolf and H. J. Teles, WO 2004/009523, 2004,
metal-dispersion. From supplementary work we already BASF Aktiengesellschaft.
know that these metal sites usually behave differently from 22 K. Seki and W. Mori, J. Phys. Chem. B, 2002, 106, 1380–1385.
23 O. M. Yaghi, US Pat. 5 648 508, 1997, Nalco Chemical Company.
bulk metals. Compared to zeolites the amount of metals in
24 U. Mueller, K. Harth, M. Hoelzle, M. Hesse, L. Lobree, W. Harder
MOFs are by almost a factor of ten higher and many of the and O. M. Yaghi, WO 2003/064030, 07.08.2003, BASF
metal species belong for chemists to the interesting class of Aktiengesellschaft.
transition metals. 25 S. Barrett, Fuel Cells Bull., 2005, 12–19.
26 G. Férey, M. Latroche, C. Serre, F. Millange, T. Loiseau and
In summary and perspective, all this might lead to a fast A. Percheron-Guégan, Chem. Commun., 2003, 2976–2977.
growing, prosperous and widespread innovation in materials 27 D. N. Dybtsev, H. Chun and K. Kim, Angew. Chem., 2004, 116,
science, both in academia and industry. However, one 5143–5146.
certainly has to keep attention to find superior performance 28 B. Chen, N. W. Ockwig, A. R. Millward, D. S. Contreras and
O. M. Yaghi, Angew. Chem., 2005, 117, 4823–4827.
by applying MOFs over state-of-art technologies. It will 29 E. Y. Lee, S. Y. Jang and M. P. Suh, J. Am. Chem. Soc., 2005, 127,
never be sufficient just to find a ‘me-too’ solution instead 6374–6381.
of looking for considerable improvement of the best existing 30 D. N. Dybtsev, H. Chun, S. H. Yoon, D. Kim and K. Kim, J. Am.
Chem. Soc., 2004, 126, 32–33.
one. Only the latter approach will finally contribute to true
31 Q. Yang and Ch. Zhong, J. Phys. Chem. B, 2005, 109, 24, 11862–4.
innovation and value-added growth of industrial companies 32 G. Férey, C. Serre, C. Mellot-Draznieks, F. Millange, S. Surblé,
and society. J. Dutour and I. Margiolaki, Angew. Chem., 2004, 116, 6456–6461.

This journal is ß The Royal Society of Chemistry 2006 J. Mater. Chem., 2006, 16, 626–636 | 635
 View Article Online

33 R. Matsuda, R. Kitaura, S. Kitagawa, Y. Kubota, 35 E. Choi, K. Park, C. Yang, H. Kim, J. Son, S. W. Lee,
R. V. Belosludov, T. C. Kobayashi, H. Sakamoto, T. Chiba, Y. H. Lee, D. Min and Y. Kwon, Chem.—Eur. J., 2004, 10,
Published on 23 November 2005. Downloaded by Pakistan Institute of Engineering and Applied Sciences on 7/5/2022 8:06:06 AM.

M. Takata, Y. Kawazoe and Y. Mita, Nature, 2005, 436, 238–241. 5535–5540.

34 C. Serre, F. Millange, S. Surblé and G. Férey, Angew. Chem., 2004, 36 Q. Fang, G. Zhu, M. Xue, J. Sun, Y. Wei, S. Qiu and R. Xu,
116, 6446–6449. Angew. Chem., 2005, 117, 3913–3916.

636 | J. Mater. Chem., 2006, 16, 626–636 This journal is ß The Royal Society of Chemistry 2006