Assessing the adsorption selectivity of linker functionalized, moisture-stable metal-organic framework thin films by means of an environment-controlled quartz crystal microbalance

Article abstract of DOI:10.1039/c2cc34608g

The stepwise thin film deposition of the robust, hydrophobic [Zn ₄O(dmcapz)₃]_n (dmcapz = 3,5-dimethyl-4-carboxy- pyrazolato) is reported. The adsorption of small organic probe molecules, including alkanols, toluene, aniline and xylenes, was monitored by an environment-controlled quartz crystal microbalance setup. The adsorption selectivity was tuned by introducing alkyl side chains in the dmcapz linker.

Full text of DOI:10.1039/c2cc34608g

View Online / Journal Homepage
Dynamic Article Links
ChemComm
Cite this: DOI: 10.1039/c2cc34608g
www.rsc.org/chemcomm
COMMUNICATION
Assessing the adsorption selectivity of linker functionalized,
moisture-stable metal–organic framework thin films by means
of an environment-controlled quartz crystal microbalancewz
´
´
Angelique Betard, Suttipong Wannapaiboon and Roland A. Fischer*
Received 27th June 2012, Accepted 5th September 2012
DOI: 10.1039/c2cc34608g
The stepwise thin film deposition of the robust, hydrophobic
[Zn4O(dmcapz)3]n (dmcapz = 3,5-dimethyl-4-carboxy-pyrazolato)
is reported. The adsorption of small organic probe molecules,
including alkanols, toluene, aniline and xylenes, was monitored
by an environment-controlled quartz crystal microbalance setup.
The adsorption selectivity was tuned by introducing alkyl side
chains in the dmcapz linker.
(SBU) and the same topology as MOF-5.10 That is, the
structure is cubic with the Zn4O tetrahedron sitting at the
vertices and linked by six molecules of dmcapz to the next
vertex. Unlike MOF-5, [Zn4O(dmcapz)3]n is highly water and
air stable as a result of the presence of both the pyrazole group
at one side of the linker and the two methyl groups which
point toward the carboxylic acid group, protecting the more
labile Zn–O bonds from nucleophilic attacks. [Zn4O(dmcapz)3]n
was reported to capture toxic gases even in the presence of
moisture.9 The high stability and cubic structure make it an
ideal candidate for applications in filters or membranes,
chromatography and pervaporation. Its close structural
relationship to MOF-5 strongly suggests its compatibility with
the stepwise film deposition method using the Controlled SBU
Approach (CSA).4,11
The processing of metal–organic frameworks (MOFs) as thin films
and coatings is relevant for a wide range of applications, such as
separation by membranes, gas chromatography, catalysis and
sensing.1 For integration into practical devices, good air, water
and thermal stability as well as an easy and quick thin film
fabrication method at low temperature, compatible with the sub-
strate limitations, is required. The stepwise deposition is one of
these methods, in which the metal and the linker components are
kept separate from each other and are successively put into contact
with the growing film on the substrate.2 It has been used for the
coating of textile fibers,3 the coating of gas chromatography
columns,4 the fabrication of membranes for gas separation5 and
of sensing devices.6 However, until now only few MOFs have been
grown as dense coatings in a stepwise fashion, namely: HKUST-1
[Cu3(btc)2]n2 (btc = 1,3,5-benzenetricarboxylate), pillared-layered
[M2L2P]n structures where M = Cu2+ or Zn2+, L = dicarboxylate
linker, P = dinitrogen pillar ligand,7,8 and MOF-5 [Zn4O(bdc)3]n4
(bdc = 1,4-benzenedicarboxylate). The drawbacks associated with
these MOFs are either their limited functionalization possibilities,
i.e. for HKUST-1, or their limited resistance to moisture; MOF-5 is
even only attainable when working in a dry environment.
For device applications more robust and easily tailored MOFs
are desirable. The stepwise film growth of such a candidate is
reported below.
We first synthesized polycrystalline powder samples of [Zn4O-
(dmcapz)3]n as reference by employing the CSA. A solution of
basic zinc acetate [Zn4O(OAc)6] was mixed with a solution of
H2dmcapz in the presence of ethanol and water (for details see
ESIz). Phase pure [Zn4O(dmcapz)3]n was formed as proven by
powder X-ray diffraction (XRD) (Fig. 1b).
Interestingly, the same experiment without water led to the
formation of another and yet unknown phase, whose powder
XRD pattern could be indexed in an orthorhombic space group
(Fig. S2, ESIz). Following these results, stepwise deposition of
[Zn4O(dmcapz)3]n was performed. A gold coated substrate suited
for quartz crystal microbalance (QCM) measurements was func-
tionalized with a COOH-terminated SAM (self-assembled mono-
layer) and was then exposed alternately to a solution of basic zinc
acetate [Zn4O(OAc)6] in ethanol and to a solution of H2dmcapz in
an ethanol–water mixture, using a computer controlled automatic
deposition setup described previously.12 These growth steps were
separated by rinsing steps with ethanol only. Characterization by
XRD and scanning electron microscopy (SEM) (Fig. 1, top)
revealed the deposition of phase pure material with preferred
orientation along the [100] direction. This corresponds to cubic
crystallites preferably attached to the surface by their bottom
facet. A dense coating of intergrown crystallites is achieved after
approximately 30 cycles (Fig. 1c), leading to an overall film
thickness in the range of 0.5–1 mm. The deposition method is
compatible with a wide range of substrates: gold surfaces
functionalized with SAMs of COOH, OH or pyridyl termination,
as well as unmodified flat or porous silica and alumina substrates
Barea, Navarro and co-workers9 recently reported the synthesis
of a fairly robust MOF with interesting adsorption properties:
[Zn4O(dmcapz)3]n (dmcapz = 3,5-dimethyl-4-carboxypyrazo-
lato). This MOF exhibits the same secondary building unit
Inorganic Chemistry II—Organometallics and Materials Chemistry,
Ruhr-University Bochum, D-44780 Bochum, Germany.
E-mail: roland.fischer@rub.de
w This article is part of the ChemComm ‘Metal–organic frameworks’
web themed issue.
z Electronic supplementary information (ESI) available: Methods and
materials. Fig. S2–S12. See DOI: 10.1039/c2cc34608g
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun.

View Online
Fig. 2 Time dependent mass uptake of a [Zn4O(dmcapz)3]n film for
alkanols of increasing size at a rel. humidity of 85% at 20 1C (1 atm).
The uptake curves are presented in Fig. 2. A striking difference
can be seen between methanol and ethanol, for which 4 to 5
molecules per cavity are absorbed within a few seconds, compared
with isopropanol, for which the kinetic uptake is much slower; tert-
butanol is not adsorbed. From these measurements the effective size
of the pore window can be estimated to be around 5.3 A, which is
somewhat smaller than the value of 5.6 A estimated from crystal-
lographic data.9 We then studied in some detail the adsorption of
molecules whose sizes are close to this characteristic pore opening.
Toluene and aniline were chosen on one hand, and the xylene
series on the other. A very good selectivity of toluene over aniline
(4.9 at 90 min), despite their very close sizes (about 5.3 A), is
evident from the recorded mass gain over time (Fig. 3, left). The
observed differences in the adsorption rate and saturation level
can be explained by the different polarities of the analytes:
[Zn4O(dmcapz)3]n is highly hydrophobic and therefore shows
a higher affinity for non-polar molecules over polar ones.
A significant selectivity of p-xylene over m-/o-xylene (5.3 at
90 min) was also observed, which is certainly based on the
molecular size: in this case p-xylene is slightly smaller than the
pore opening, whereas the other isomers are slightly larger.
In order to vary the pore size and to modulate the adsorp-
tion response of the films, a range of derivatives of the parent
linker were chosen bearing various substituents other than just
methyl on the pyrazole ring, i.e. ethyl (Et), n-propyl (nPr),
isopropyl (iPr), cyclopropyl (cPr). The respectively function-
alized 4-carboxy-pyrazoles H2L were synthesized via a facile
and efficient two-step route starting from acetoacetate and the
respective derivatives. This allows us to vary independently
the substituents on the two sides of the linker, giving access
to non-symmetrically substituted linkers, thus affecting the
Fig. 1 (top) SEM micrographs of a 30-cycle film grown on COOH-
terminated gold (bottom). Comparison of the X-ray diffraction pat-
terns (a) simulated from published data, (b) of a powder obtained by
the CSA and (c) of a film grown on COOH-functionalized gold.
(Fig. S4, ESIz). Note that the growth of [Zn4O(dmcapz)3]n was
carried out with low-cost technical ethanol, making the film
deposition much less expensive as compared with other so far
reported stepwise depositions of MOFs on SAMs as surface
templates, which all required very pure and dry ethanol for
achieving crystalline, densely packed, phase-pure and oriented
(SUR)MOFs.2,7,8
The methanol adsorption isotherm (0–95% rel. humidity)
of a typical film of [Zn4O(dmcapz)3]n grown on the above-
mentioned substrates was obtained by using an environmental
controlled QCM setup as described in the ESIz. The isotherm has
the expected type I shape (Fig. S5, ESIz) and nicely demon-
strates the microporosity of the MOF film. The amount
of methanol adsorbed at 20 1C and relative humidity of
0.95 p/p0 was 7.6 mmol gꢀ1. This represents a pore volume of
0.36 cm3 per cm3 of MOF (calculated from the density of liquid
methanol) whereas a value of 0.45 cm3 per cm3 of MOF has been
measured on bulk samples with standard N2 adsorption.9 The
difference can be attributed to the size of the methanol molecules
which cannot fill every empty space in the structure and to the
activation process, carried out for the thin film samples in situ
(see ESIz) and which might be incomplete to some extent.
As a model for pervaporation or chromatographic separa-
tion, we studied the adsorption selectivity of our films for a
choice of volatile organic compounds using the environment
controlled QCM equipment mentioned above. At t = 0 min, a
helium flow with a partial vapor pressure p/p0 = 0.85 of the
desired analyte reached the QCM adsorption cell and the mass
change over time was monitored. We first investigated the
adsorption of a series of simple alkanol molecules of increasing
sizes: methanol, ethanol, isopropanol and tert-butanol.
Fig. 3 Time dependent mass uptake of a [Zn4O(dmcapz)3]n film for
toluene vs. aniline (left) and for p-xylene vs. o-,m-xylene (right) at a rel.
humidity of 85% at 20 1C (1 atm).
c
Chem. Commun.
This journal is The Royal Society of Chemistry 2012

View Online
possibility of fine tuning the adsorption selectivity for quite
similar molecules by the choice of substituents at the linker L.
In addition, adsorption isotherms of water were measured on
various films (one day at 20 1C, 0–95% rel. humidity, Fig. S12,
ESIz). No or little differences in crystallinity and adsorption
properties after the experiments were observed. This is an encoura-
ging indication of good stability of these films toward moisture.
To sum up, the MOF thin film stepwise deposition concept was
applied to a chemically very robust, MOF-5-analogue family. The
obtained coatings on COOH-functionalized QCM substrates
allowed the rapid screening and quantitative assessment of the
adsorption properties of the various thin film materials. The
selective adsorption of probe molecules depending on their size
and/or hydrophobicity as a function of the linker functionalization
of the MOF is demonstrated. From our data we conclude that
[Zn4O(dmcapz)3]n and its derivatives could be used as an active
separating layer in applications such as gas chromatography or
pervaporation.y The well-controlled and facile stepwise deposition
method, the robustness and compatibility of this type of MOF
with a wide range of substrates open the way for integration into
many functional devices.
Fig. 4 Amount of analyte adsorbed at saturation (or after 90 min) of
linker functionalized films [Zn4O(L)3]n at rel. humidity of 85% (20 1C,
1 atm). The analytes are close in size but different in polarities.
multivariate complexity of the structures. Surprisingly, how-
ever, we failed to prepare most of the target MOFs
[Zn4O(L)3]n as phase-pure polycrystalline powder materials.
The orthorhombic phase of so far unknown crystal structure,
already mentioned in the case of parent [Zn4O(dmcapz)3]n,
was nearly always present as a (minor) contaminant which
could not be avoided or separated.
Support within the Priority Program 1362 ‘‘Metal–Organic
Frameworks’’ of the German Research Foundation (Fi-502/
26-1) is acknowledged. A.B. is grateful for a PhD fellowship
by the German Academic Exchange service (DAAD).
Nevertheless, phase-pure films were finally obtained by
using the seeding layers concept (details in the ESIz and
Fig. S7–S9).8 The seeding approach consisted in depositing some
parent [Zn4O(dmcapz)3]n material (just one deposition cycle)
on the substrate surface prior to the deposition of the function-
alized [Zn4O(L)3]n, in order to direct the nucleation and
growth of the desired phase. The respective methanol adsorp-
tion isotherms (0–95% rel. humidity) of the deposited films
were then recorded at 20 1C (Fig. S10 and S11, ESIz). They all
represented type I isotherms and from this the accessible pore
volumes were derived (Table S2, ESIz). As expected, a constant
decrease of the pore volume Vp is observed upon the addition of
increasingly more bulky alkyl substituents. Moreover, a striking
Notes and references
y Preliminary experiments (Fig. S4) showed that the growth is also
possible on porous substrates. A detailed study will be published later.
1 A. Betard and R. A. Fischer, Chem. Rev., 2012, 112, 1055;
´
O. Shekhah, J. Liu, R. A. Fischer and C. Woll, Chem. Soc. Rev.,
¨
2011, 40, 1081.
2 O. Shekhah, H. Wang, S. Kowarik, F. Schreiber, M. Paulus,
M. Tolan, C. Sternemann, F. Evers, D. Zacher, R. A. Fischer
and C. Woll, J. Am. Chem. Soc., 2007, 129, 15118.
¨
3 M. Meilikhov, K. Yusenko, E. Schollmeyer, C. Mayer,
H.-J. Buschmann and R. A. Fischer, Dalton Trans., 2011, 40, 4838.
n
i
difference appears between Pr (Vp = 0.15 cm3 gꢀ1) and Pr
4 A. S. Munch, J. Seidel, A. Obst, E. Weber and F. O. R. L. Mertens,
¨
Chem.–Eur. J., 2011, 17, 10958.
´
5 A. Betard, H. Bux, S. Henke, D. Zacher, J. Caro and
R. A. Fischer, Microporous Mesoporous Mater., 2012, 150, 76.
6 M. D. Allendorf, R. J. T. Houk, L. Andruszkiewicz, A. A. Talin,
J. Pikarsky, A. Choudhury, K. A. Gall and P. J. Hesketh, J. Am.
Chem. Soc., 2008, 130, 14404; L. E. Kreno, J. T. Hupp and
R. P. Van Duyne, Anal. Chem., 2010, 82, 8042.
(Vp = 0.052 cm3 gꢀ1) substitution. This is certainly due to the
i
high rigidity of Pr, which therefore blocks the pore entrance
more efficiently than the conformationally flexible nPr. This
high rigidity has an effect on the accessibility of the MOF bulk
structure. Indeed, we failed to obtain the MOF-5-analogue
phase when the linker was symmetrically functionalized with
two iPr groups, whereas a symmetric functionalization with two
nPr groups led to the formation of the desired material. (The
characterization and structural elucidation of the obtained
phases are out of scope of this communication.)
¨
7 B. Liu, O. Shekhah, H. K. Arslan, J. Liu, C. Woll and R. A. Fischer,
Angew. Chem., Int. Ed., 2012, 51, 807; O. Shekhah, K. Hirai, H. Wang,
H. Uehara, M. Kondo, S. Diring, D. Zacher, R. A. Fischer, O. Sakata,
S. Kitagawa, S. Furukawa and C. Woll, Dalton Trans., 2011, 40, 4954;
¨
O. Shekhah, H. Wang, M. Paradinas, C. Ocal, B. Schupbach,
¨
A. Terfort, D. Zacher, R. A. Fischer and C. Woll, Nat. Mater., 2009,
¨
We then estimated the pore size opening of the functiona-
lized derivatives [Zn4O(L)3]n by studying the uptake of the
deposited films at the saturation limit (for 85% rel. humidity)
after 90 min for some analytes of similar size. The results are
grouped in Fig. 4. Isopropanol is somewhat smaller (5.12 A)
than toluene and aniline (5.3 A), and is therefore more readily
adsorbed, in general, than the other analytes. However, with
increasing bulkiness of one substituent, the uptake of toluene
and aniline decreases dramatically due to a smaller pore size of
8, 481.
8 D. Zacher, K. Yusenko, A. Be
T. Ladnorg, O. Shekhah, B. Schupbach, T. D. L. Arcos,
¨
M. Krasnopolski, M. Meilikhov, J. Winter, A. Terfort, C. Woll
¨
and R. A. Fischer, Chem.–Eur. J., 2011, 17, 1448.
9 C. Montoro, F. t. Linares, E. Quartapelle Procopio, I. Senkovska,
S. Kaskel, S. Galli, N. Masciocchi, E. Barea and J. A. R. Navarro,
J. Am. Chem. Soc., 2011, 133, 11888.
10 H. Li, M. Eddaoudi, M. O’Keeffe and O. M. Yaghi, Nature, 1999,
402, 276.
´
tard, S. Henke, M. Molon,
11 C. Serre, F. Millange, S. Surble
Ed., 2004, 43, 6285.
12 K. Yusenko, M. Meilikhov, D. Zacher, F. Wieland, C. Sternemann,
and G. Ferey, Angew. Chem., Int.
´ ´
i
the functionalized MOFs. The (mono) Pr functionalized film
does not adsorb any isopropanol, suggesting that its pore size
is smaller than 5.1 A. These data clearly demonstrate the
X. Stammer, T. Ladnorg, C. Woll and R. A. Fischer, CrystEng-
Comm, 2010, 12, 2086.
¨
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun.