Supercritical processing as a route to high internal surface areas and permanent microporosity in metal#organic framework materials

Full text of DOI:10.1021/ja808853q

Published on Web 12/24/2008
Supercritical Processing as a Route to High Internal Surface Areas and
Permanent Microporosity in Metal-Organic Framework Materials
Andrew P. Nelson,†,§ Omar K. Farha,† Karen L. Mulfort,†,‡ and Joseph T. Hupp*,†
Department of Chemistry and International Institute for Nanotechnology, Northwestern UniVersity, 2145 Sheridan
Road, EVanston, Illinois 60208, and DiVision of Chemical Sciences and Engineering, Argonne National Laboratory,
Argonne, Illinois 60439
Received November 11, 2008; E-mail: j-hupp@northwestern.edu
Crystalline metal-organic frameworks (MOFs)1 have emerged
as attractive materials for a remarkably wide range of potential
applications, including chemical separations,2 size-selective mo-
lecular catalysis,3 and chemical fuel storage and release.4 Among
the materials properties favoring these particular applications are
permanent microporosity and large internal surface areas. Unfor-
tunately, the surface areas attained experimentally are often less
than anticipated from computational studies or single-crystal X-ray
structural studies.5 Furthermore, they can differ substantially from
laboratory to laboratory. The dispartities and discrepancies have
most often been attributed to channel collapse upon solvent removal
or channel blockage due to solvent retention.6 In many instances,
porosity can be recovered and surface areas can be increased by
exchanging the MOF-incorporated solvent remaining from synthesis
for a lower boiling point solvent and then removing the solvent
under relatively mild conditions.7 Nevertheless, in our experience,
cates, as expected, that the channels of “as-synthesized” 5 are fully
occupied, presumably by disordered DEF.
and presumably that of others, the liquid solvent exchange strategy
still occasionally fails to elicit MOF microporosity or, more
commonly, succeeds in enabling access to the internal surface area
of a given MOF, but to a lesser extent than anticipated from
computations.
A similar solvothermal synthesis in DMF, but with ligand 2 rather
than 1, yielded a red solid, 6. While we were unable to obtain a
suitable single crystal for X-ray structural determination, powder
X-ray diffraction (PXRD) measurements revealed that the material
Here we report on a very promising alternative approach entailing
processing of solvent-containing MOF materials with liquid and
supercritical carbon dioxide. To illustrate, we report the behavior
of two new MOFs as well as two previously described materials.
In all four cases the materials comprise dicarboxylated organic
ligands as struts and Zn(II)-containing clusters as nodes. As shown
below, the new approach is capable of very substantially enhancing
access to MOF internal surface area, relative to treatments entailing
either (a) thermally assisted evacuation of the solvent used for
synthesis (DMF or DEF) (“conventional activation”) or (b) liquid
solvent exchange (e.g., DMF S CHCl3; DEF S THF) followed
by pore evacuation at moderate temperature (DMF ) dimethyl-
formamide, DEF ) diethylformamide, THF ) tetrahydrofuran).
As detailed in the Supporting Information (SI), the first of the
four investigated materials was obtained by solvothermal synthesis
from a DEF solution of ZnNO3 ·6H2O and a napthalenediimide-
containing ligand, 1. Figure 1 shows several views of the resulting
channel-containing material, 5.8 The nodes of 5 consist of Zn4O
clusters, coordinated by carboxylates in both a mono- and bidentate
fashion. Each node is also coordinated by one water and two DEF
molecules. In contrast to the cubic frameworks typically obtained
with these clusters, the unusual coordination produces a canted
rhombohedral framework. The single-crystal X-ray structure indi-
1
is crystalline (see SI). H NMR measurements of a sample of 6
that had been thermally evacuated and then dissolved in D2SO4
showed the presence only of 2 and a trace amount of DMF.
Thermal gravimetric analysis (TGA) of 5 indicated a ca. 35%
mass decrease (solvent loss) upon heating to 250 °C (see SI), a
Figure 1. Crystallographically derived: (A) structure of 5, (B) corner
cluster, showing coordination modes of 1 and two DEF molecules, (C) ac-
plane of 5, looking down b-channels, and (D) ab-plane of 5, looking down
c-channels. For clarity, disordered DEF molecules are omitted from images
A, C, and D. For image A, coordinated DEF molecules are also omitted.
† Northwestern University.
‡ Argonne National Laboratory.
§ Present address: Synthesis and Formulations Branch, Naval Air Warfare Center,
Weapons Division, China Lake, California 93555.
9
458 J. AM. CHEM. SOC. 2009, 131, 458–460
10.1021/ja808853q CCC: $40.75
2009 American Chemical Society

C O M M U N I C A T I O N S
Figure 2. (I) 77K N2 adsorption isotherms of 5 following ScD activation
(top), exchange with THF and evacuation at 25 °C (middle), or conventional
activation at 110 °C (bottom). (II) 77K N2 isotherms of 6 following ScD
activation (top), exchange with chloroform, followed by evacuation at 25
°C (bottom, red), or traditional activation 110 °C (bottom, blue (overlap)).
Figure 3. ( I) 77 K N2 isotherms of IRMOF-3 following ScD activation
(top), exchange with CHCl3 followed by evacuation at 25 °C (middle), or
conventional activation at 100 °C (bottom). (II) 77 K N2 isotherms of
IRMOF-16 following ScD activation (top) or exchange with CHCl3 followed
by activation at 25 °C (bottom).
Table 1. Nitrogen-Accessible Surface Areas (m2/g) of Various
then with liquid CO2. The sample was then taken above the carbon
dioxide critical point (T ) 31 °C; P ) 73 atm), held for 30 min,
and finally, slowly vented (18 h; further details in SI). As illustrated
in Figure 2, the results are striking: the N2-accessible surface area
increases to 400 m2/g for 5 (3-fold increase) and 430 m2/g for 6
(12-fold increase).
Framework Materials as a Function of Method of Pore Evacuation
conventional
thermal
evacuation
CHCl3 or
THF
exchange
material
ScD
IRMOF-3
IRMOF-16
5
6
10
na
5
1800
470
135
36
2850
1910
400
With these results in hand, we extended the investigation to
noncatenated, cubic, “isoreticular” MOFs constructed from ligands
4 (IRMOF-3) and 3 (IRMOF-16).7,10 IRMOF-3 has attracted
considerable recent attention because of its susceptibility to
postsynthetic covalent modification via the available amine group.11
For IRMOF-3, activation via solvent exchange (DMF S CHCl3)
yielded a surface area of 1800 m2/g. ScD activation increases the
N2 accessible surface area to 2850 m2/g (Figure 3). For IRMOF-
16 there are, to our knowledge, no previous reports of surface area.
From Figure 3, we observe that activation of IRMOF-16 via solvent
exchange yields a N2-accessible surface area of 470 m2/g but that
the surface area is quadrupled via ScD activation (1910 m2/g).
We hypothesize that the key feature of supercritical activation
is the elimination of solvent (CO2) surface tension at temperatures
and pressures above the critical point. For aerogel fabrication, the
elimination of surface tension, and therefore capillary forces,
prevents the pore collapse that would otherwise occur upon removal
of solvent.9 In contrast, because of their tiny size and the relative
strength of the zinc-oxygen bonds that define them, the crystalline
micropores that permeate the four MOFs should be resistant to
collapse via capillary effects, even with solvents such as DEF and
DMF. The persistence of crystallinity in solvent-exchanged 6,
despite its inability to adsorb nitrogen, supports this contention.
Similarly, the ability to resolvate 6 with DMF after nitrogen
5
430
result consistent with the porosity implied by the X-ray structure
and suggesting substantial internal surface area. TGA measurements
of 6 showed a 17% mass loss upon heating to 300 °C (see SI; the
framework itself is stable to ca. 400 °C). The diminished solvent
capacity despite the larger strut size is suggestive of network
catenation, the formation of a 2D rather than 3D structure, or both.
In contrast to the implications from X-ray and/or TGA measure-
ments, N2 adsorption studies (77 K) indicate negligible accessible
surface areas for either 5 or 6, following conventional thermal
activation (Figure 2 and Table 1). Solvent exchange with 6 (CHCl3)
yields no meaningful improvement, while solvent exchange with
5 (THF) yields a modest Brunauer-Emmet-Teller (BET) surface
area of 135 m2/g.
Confronted with these disappointing results, we turned to
supercritical drying (ScD) as a potential activation method. ScD
has been used previously in polymer synthesis and preparation of
organosilicates and high-area silica aerogels.9 Additionally, the low
cost of CO2(g) and the dryer make this a practical method. Briefly,
the solvent remaining from MOF synthesis was first exchanged with
EtOH (miscible with CO2 and compatible with our instrument) and
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 459

C O M M U N I C A T I O N S
Scheme 1a
Postdoctoral Fellowship. K.L.M. thanks Argonne National Labora-
tory for a Laboratory-Grad Fellowship.
Supporting Information Available: Experimental procedures for
the synthesis of 1, 2, 3, 5, and 6. X-ray crystallographic files for 5 in
1
CIF format. PXRD, TGA, and sorption isotherm data. H NMR data
on residual solvent content of 3 as a function of activation method.
This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Recent reviews: (a) Collins, D. J.; Zhou, H.-C. J. Mater. Chem. 2007, 17,
3154–3160. (b) Fe´rey, G. Chem. Soc. ReV. 2008, 37, 191–214.
(2) For example, see: (a) Bastin, L.; Barcia, P. S.; Hurtado, E. J.; Silva, J. A. C.;
Rodrigues, A. E.; Chen, B. J. Phys. Chem. C 2008, 112, 1575–1581. (b)
Lee, E. Y.; Jang, S. Y.; Suh, M. P. J. Am. Chem. Soc. 2005, 127, 6374–
6381. (c) Dinca, M.; Long, J. R. J. Am. Chem. Soc. 2005, 127, 9376–
9377. (d) Snurr, R. Q.; Hupp, J. T.; Nguyen, S. T. AIChE 2004, 50, 1090–
1095. (e) Bae, Y. S.; Mulfort, K. L.; Frost, H.; Ryan, P.; Punnathanam, S.;
Broadbelt, L. J.; Hupp, J. T.; Snurr, R. Q. Langmuir 2008, 24, 8592–8598.
(f) Bae, Y. S.; Farha, O. K.; Spokoyny, A. M.; Mirkin, C. A.; Hupp, J. T.;
Snurr, R. Q. Chem. Commun. 2008, 4135–4137.
a (A) Supercritical CO2 activation of MOF microcrystals, (B) adsorption
of gas molecules after supercritical CO2 activation, (C) traditional activation
of MOF crystals, (D) adsorption of gas molecules after traditional activation,
and (E) resolvation.
(3) (a) Wu, C.-D.; Hu, A.; Zhang, L.; Lin, W. J. Am. Chem. Soc. 2005, 127,
8940–8941. (b) Horike, S.; Dinca, M.; Tamaki, K.; Long, J. R. J. Am. Chem.
Soc. 2008, 130, 5854–5855. (c) Cho, S.-H.; Ma, B.; Nguyen, S. T.; Hupp,
J. T.; Albrecht-Schmitt, T. E. Chem. Commun. 2006, 256, 3–2565.
(4) For example, see: (a) Collins, D. J.; Zhou, H.-C. J. Mater. Chem. 2007,
17, 3154–3160. (b) Fe´rey, G. Chem. Soc. ReV. 2008, 37, 191–214. (c) Nouar,
F.; Eubank, J. F.; Bousquet, T.; Wojtas, L.; Zaworotko, M. J.; Eddaoudi,
M. J. Am. Chem. Soc. 2008, 130, 1833–1835. (d) Chen, B.; Ockwig, N. W.;
Millard, A. R.; Contreras, D. S.; Yaghi, O. M. Angew. Chem., Int. Ed.
2005, 44, 4745–4749. (e) Dinca, M.; Dailly, A.; Liu, Y.; Brown, C. M.;
Neumann, D. A.; Long, J. R. J. Am. Chem. Soc. 2006, 128, 16876–16883.
(f) Latroche, M.; Surble´, S.; Serre, C.; Mellot-Draznieks, C.; Llewellyn,
P. L.; Lee, H.; Chang, J.; Jhung, S. H.; Fe´rey, G. Angew. Chem., Int. Ed.
2006, 45, 8227–8231. (g) Mulfort, K. L.; Hupp, J. T. J. Am. Chem. Soc.
2007, 129, 9604–9605. (h) Farha, O. K.; Spokoyny, A. M.; Mulfort, K. L.;
Hawthorne, M. F.; Mirkin, C. A.; Hupp, J. T. J. Am. Chem. Soc. 2007,
129, 12680–12680.
adsorption isotherms (TGA experiments; see SI), indicates pres-
ervation of micropores in nitrogen-impermeable samples. Similar
behavior is seen for other MOFs. Thus, collapse of crystallographi-
cally defined channels appears not to be the primary cause of
surface-area loss in the four MOFs examined (although it may be
a secondary contributor).
As illustrated in Scheme 1, we hypothesize that thermal
evacuation of solvent instead causes the collapse of interparticle
mesopores. Misalignment of micropores at particle/particle bound-
aries then inhibits access by gas molecules to internal (microporous)
surfaces. At the same time, re-exposure to liquid solvent should
enable particles to separate and to take up comparatively large
solvent molecules (as is indeed observed). In contrast, ScD
activation prevents mesopore collapse, leaving micropores acces-
sible to gas molecules. Consistent with this description, we find
that thermally evacuated samples of 5, exhibiting negligible surface
area, can be rendered microporous by resolvating the material and
then subjecting it to ScD (steps E f A f B in Scheme 1; final
surface area ) 370 m2/g)).
(5) Duren, T.; Milliange, F.; Ferey, G.; Walton, K. S.; Snurr, R. Q. J. Phy.
Chem. C 2007, 111, 15350–15356.
(6) (a) Kepert, C. J.; Rosseinsky, M. J. Chem. Commun. 1999, 375–376. (b)
Yaghi, O. M.; Li, H.; Davis, C.; Richardson, D.; Groy, T. L. Acc. Chem.
Res. 1998, 31, 474–484.
(7) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’Keeffe, M.;
Yaghi, O. M. Science 2002, 295, 469–472.
(8) Inclusion of pairs of methyl groups on the phenyl carboxylate units proved
essential for formation of a crystalline coordination polymer. Attempts to
prepare MOFs using analogues of 1 lacking one or both methyl groups
yielded only amorphous solids. As shown in the crystal structure of 5, the
methyl groups render the napthelendiimide geometrically orthogonal to the
phenyl carboxylate.
(9) (a) Reichenauer, G.; Scherer, G. W. Colloids Surf., A 2001, 187-188, 41–
50. (b) Reichenauer, G.; Scherer, G. W. J. Non-Cryst. Solids 2001, 285,
167–174. (c) Cooper, A. I. J. Mater. Chem. 2000, 10, 207–234. (d)
Lubguban, J. A.; Gangopadhyay, S.; Lahlouh, B.; Rajagopalan, T.; Biswas,
N.; Sun, J.; Huang, D. H.; Simon, S. L.; Mallikarjunan, A.; Kim, H.-C.;
Hedstrom, J.; Volksen, W.; Miller, R. D.; Toney, M. F. J. Mater. Res.
2004, 19, 3224–3233.
To summarize, careful processing of four representative MOFs
with liquid and supercritical carbon dioxide leads to substantial, or
in some cases spectacular (up to 1200%), increases in gas-accessible
surface area. Maximization of surface area is key to the optimization
of MOFs for many potential applications. Preliminary evidence
points to inhibition of mesopore collapse, and therefore micropore
accessibility, as the basis for the extraordinarily efficacious outcome
of ScD-based activation.
(10) IRMOF-16 was prepared by a method similar to that in the literature,7 but
from DMF rather than DEF as solvent. Powder X-ray diffraction (PXRD)
measurements confirm that IRMOF-16 is formed but do not rule out
contamination with IRMOF-15 (the catenated analogue of IRMOF-16), as
IRMOF-15 displays no unique PXRD peaks. IRMOF-15 contamination
would be expected to produce falsely low surface areas for the IRMOF-16
sample.
Acknowledgment. We gratefully acknowledge the U.S. Dept.
of Energy, Office of Science, Basic Energy Science Program (Grant
no. DE-FG02-08-ER15967) and the Northwestern Nanoscale Sci-
ence and Engineering Center for financial support. A.P.N. thanks
the ACS Petroleum Research Fund for an Alternative Energy
(11) (a) Wang, Z.; Cohen, S. M. J. Am. Chem. Soc. 2007, 129, 12368–12369.
(b) Wang, Z.; Cohen, S. M. Angew. Chem., Int. Ed. 2008, 47, 4699–4702.
(c) Tanabe, K. K.; Wang, Z.; Cohen, S. M J. Am. Chem. Soc. 2008, 130,
8508–8517.
JA808853Q
9
460 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009