Control of morphology and catalytic activity of a cationic layered material

Article abstract of DOI:10.1021/cg200826d

We have discovered three new synthetic pathways to our two-dimensional cationic inorganic material, antimony oxide hydroxide ethanedisulfonate, [Sb₄O₄(OH)₂](O₃SCH ₂CH₂SO₃)?H₂O. These new methods include room temperature conditions and allow control of crystal size and morphology for improved catalytic performance of the material. As a benchmark reaction, we studied ketalization of 2-butanone with 1,2-ethanediol, a common step for carbonyl group protection. Our material is easily recovered for reuse and gives a higher yield and turnover rates than the typical iodine or toluenesulfonate homogeneous catalyst. The yield ranged between 79 and 92%, with the crystals from reflux conditions giving the highest yield when used as-synthesized. The morphology and properties for each crystal growth method are described in detail. A repeated catalysis study with powdered samples increases the yield to ca. 97%.

Full text of DOI:10.1021/cg200826d

Article
pubs.acs.org/crystal
Control of Morphology and Catalytic Activity of a Cationic Layered
Material
Claudia H. Swanson, Harris A. Shaikh, Yashar Abdollahian, Jesse L. Hauser, Mariko Ikehata,
and Scott R. J. Oliver*
University of California, Santa Cruz, Department of Chemistry and Biochemistry, 1156 High Street, Santa Cruz, California 95064,
United States
ABSTRACT: We have discovered three new synthetic
pathways to our two-dimensional cationic inorganic material,
antimony oxide hydroxide ethanedisulfonate, [Sb4O4(OH)2]-
(O3SCH2CH2SO3)·H2O. These new methods include room
temperature conditions and allow control of crystal size and
morphology for improved catalytic performance of the
material. As a benchmark reaction, we studied ketalization of
2-butanone with 1,2-ethanediol, a common step for carbonyl
group protection. Our material is easily recovered for reuse and gives a higher yield and turnover rates than the typical iodine or
toluenesulfonate homogeneous catalyst. The yield ranged between 79 and 92%, with the crystals from reflux conditions giving
the highest yield when used as-synthesized. The morphology and properties for each crystal growth method are described in
detail. A repeated catalysis study with powdered samples increases the yield to ca. 97%.
if there is water present in the reaction. Several traditional
Lewis acids such as AlCl3 and FeCl3 also have inherent
disadvantages, including reactivity with moisture as well as the
necessity of stoichiometric equivalents in order for catalysis to
occur. Other catalysts require inert conditions and/or
calcination.3,16
1. INTRODUCTION
Layered inorganic materials have broad application in ion-
exchange,1 exfoliation,2 drug delivery systems, and catalysis.3,4
New synthetic methods such as self-assembly, modular
nanobuilding blocks, and polymer intercalation have given
rise to a wide range of inorganic−organic hybrids and metal−
organic frameworks (MOFs). These materials display potential
applications such as improvement in strength and toughness,
electrical resistance, and lower thermal expansion.5 Naturally
occurring clays,3 particularly montmorillonite, kaolinite, and
hydrotalcites (also known as layered double hydroxides),
continue to be extensively explored as heterogeneous catalysts.
Montmorillonite, (Na,Ca)0:33(Al,Mg)2(Si4O10)(OH)2·nH2O, is
a layered aluminosilicate anionic clay with intercalated sodium
and calcium cations as well as water molecules. It has been
proven to be an effective Lewis as well as Brønsted acid.6 The
cations provide Lewis acidity, while the hydroxy groups of the
aluminosilicate layers provide Brønsted acidity.7 In addition to
clays,3 rare-earth metal triflates8,9 have been extensively studied
for their homogeneous Lewis acid catalytic properties.
Heterogeneous solid catalysts, however, can be easily separated
from reaction products, also allowing for possible reuse and
may operate without the need for a solvent.10,11 In general,
Lewis acids catalyze a wide variety of significant reactions
including Friedel−Crafts,12 Diels−Alder,13 C−C bond for-
mation, esterification, and ketalization.14,15
Recently, we reported the hydrothermal synthesis of a
cationic layered antimony oxide hydroxide ethanedisulfonate,
[Sb4O4(OH)2](O3SCH2CH2SO3)·H2O (denoted SLUG-5 for
University of California, Santa Cruz, structure No. 5).17,18 We
provided initial examples of its catalytic activity as a
heterogeneous Lewis acid. The layered structure of SLUG-5
is defined by antimony oxide templated by ethanedisulfonate
anions (Figure 1). The bonding between the layers is
electrostatic, with only a long contact between one sulfonate
oxygen on each end and the antimony centers. Water molecules
residing between the ethanedisulfonate anions increase the
stability of the structure via strong hydrogen bonding. The
anionic moieties necessitate a positive charge on the material
for electroneutrality. The material is thermally stable up to ca.
350 °C under atmospheric conditions, stable in aqueous
medium and can be used as prepared.
Here, we report the controlled synthesis of SLUG-5 by
several new synthetic methods, allowing us to tune the
morphology and average particle size, ranging from 0.1 to 0.7
mm. The relationship between size and morphology of the
crystals and their reactivity toward ketalization catalysis is
Though excellent catalysts, clays have two major drawbacks:
(i) they have to undergo postsynthetic calcination and in many
cases cation exchange in order to be catalytically active;6 (ii) the
layers are known to swell when exposed to water.7 This
expansion is advantageous if larger cations need to be
intercalated, but the catalytic performance significantly declines
Received: June 30, 2011
Revised: November 28, 2011
Published: November 29, 2011
© 2011 American Chemical Society
93
dx.doi.org/10.1021/cg200826d | Cryst. Growth Des. 2012, 12, 93−98

Crystal Growth & Design
Article
ically for 3 days at 150 °C. Colorless crystals were obtained after
filtering and rinsing with ethanol [yield based on antimony(III)
acetate: 97.4%]. Denoted as SLUG-5 (I).
2.1.2. Stirring-Induced Crystallization of SLUG-5 (II). Antimony-
(III) acetate, 1,2-ethanedisulfonic acid, and water were added into a
vial in a molar ratio of 1:5:100. The mixture was sealed and
mechanically stirred for 4 days at ambient temperature. The product
was recovered by filtration and rinsed with ethanol [yield based on
antimony(III) acetate: 89.5%]. Denoted as SLUG-5 (II).
2.1.3. Reflux-Induced Crystallization of SLUG-5 (III). Antimony-
(III) acetate, 1,2-ethanedisulfonic acid, and water were added into a
round-bottom flask in a molar ratio of 1:5:100. The mixture was
mechanically stirred and refluxed for 4 days. The product was filtered
and rinsed with ethanol [yield based on antimony(III) acetate: 81.6%].
Denoted as SLUG-5 (III).
2.1.4. Evaporation-Induced Crystallization of SLUG-5 (IV).
Antimony(III) acetate, 1,2-ethanedisulfonic acid, and water were
added into a 10 mL beaker. The beaker was covered with a watch glass
and allowed to evaporate at room temperature and pressure for 10
weeks. The product was filtered and rinsed with ethanol [yield based
on antimony(III) acetate: 64.0%]. Denoted as SLUG-5 (IV).
2.2. Catalytic Reactions. The following general procedure of
ketal formation was used to characterize the performance of SLUG-5
as a Lewis acid catalyst: first, 3.10 μmol (2.5 mg) of catalyst was added
to a round-bottomed glass reactor (capacity 50 mL). A 70 mmol
portion of 2-butanone (methyl ethyl ketone, CH3COCH2CH3, Acros
Organics, 99%) and 70 mmol of ethylene glycol (HOC2H4OH, Acros
Organics, 99%) were then added. Toluene (C6H5CH3, EM Science,
99.5%) was used as solvent. The reactants were used as received, with
no further treatment. The reaction mixture was refluxed for various
time intervals under Dean−Stark conditions, which continuously
removed evolved water. The catalyst was recovered by filtration after
the desired reaction time and left overnight to air-dry. The product
was separated from ethylene glycol using a separatory funnel, and the
final product was obtained by fractional distillation. Structure and yield
were determined by integrated 1H NMR peak areas of the ketal
product (2-ethyl-2-methyl-[1,3]-dioxolane, δ 3.93−4.00, m, 4H)
versus 2-butanone and ethylene glycol (δ 3.73−3.78, m, 4H).
Figure 1. Space-filling view of SLUG-5 emphasizing the surface of the
antimonate layers. The cylindrical ethanedisulfonate (bottom)
templates the antimony oxide hydroxide into a 2D layer (top). The
template is negatively charged and the surface has a positive charge,
held together by electrostatic forces. Color coding: antimony (blue),
sulfur (yellow), oxygen (red), carbon (gray), and hydrogen (white).
discussed. The thicker, more defined morphology from reflux
conditions gave the strongest performance.
2. EXPERIMENTAL SECTION
2.1. Crystal Synthesis Methods. 2.1.1. Hydrothermal Syn-
thesis of SLUG-5 (I). Starting reagents antimony(III) acetate [Sb-
(OOCCH3)3, Alfa Aesar, 97%] and 1,2-ethanedisulfonic acid
(HO3SC2H4SO3H·H2O, TCI Inc., 95%) were used as received,
without further purification. Crystals of [Sb4O4(OH)2]-
(O3SCH2CH2SO3)·H2O were hydrothermally synthesized as reported
previously,17 with a molar ratio of 100:1:5 for H2O:Sb-
(OOCCH3)3:HO3SCH2CH2SO3H. The autoclave was heated stat-
Figure 2. PXRD confirms that all synthesis methods gave rise to the SLUG-5 material: (I) hydrothermal synthesis, light green; (II) stirring-induced
crystallization, dark green; (III) reflux-induced crystallization, blue; (IV) evaporation, red. Inset: an expansion of the low-angle majority peak.
94
dx.doi.org/10.1021/cg200826d | Cryst. Growth Des. 2012, 12, 93−98

Crystal Growth & Design
Article
2.3. Instrumental. All materials were first characterized by
powder X-ray diffraction (PXRD) with a Rigaku Americas Miniflex
Plus powder diffractometer equipped with Cu−Kα radiation.
Diffraction patterns were recorded from 2 to 60° 2θ, with a step
size of 0.04° at 1°/min. Thermogravimetric analysis (TGA) was
measured on a TA Instruments 2050 TGA with nitrogen purge, heated
from 30 to 800 °C at 10 °C/min and coupled to a mass spectrometer
(Pfeiffer Vacuum, ThermoStar, GSD 301 T3) with a 70 eV ionizing
ascribed to S−O stretching in the ethanedisulfonate. Peaks
between 1139 and 1129 cm−1 represent sulfonate peaks19 and
occur for all four morphologies, indicating a similar degree of
interaction with the layers. The strong band between
approximately 3200 and 3600 cm−1, for each morphology, is
due to hydrogen bonding.
Thermogravimetric analysis shows that morphologies II, III,
and IV have an almost identical thermal stability (Figure 4).
Weight loss starting at ca. 350 °C corresponds to the
1
potential. H NMR spectroscopy was carried out with a 600 MHz 5
mm Indirect Detection Probe with PFG (Pulse Field Gradients) at
298 K. Scanning electron microscope (SEM) images were collected on
an SEM Hitachi S-2700. Fourier transform infrared spectroscopy (FT-
IR) spectra were recorded using a Perkin-Elmer Spectrum One FT-IR
spectrophotometer. The scan range was from 400 to 4000 cm−1 at a
resolution of 1 cm−1.
3. RESULTS AND DISCUSSION
3.1. Characterization of Morphologies. The PXRD
pattern for all samples I−IV are almost identical, confirming
that the same product and crystallinity was obtained for all
synthetic methods (Figure 2). The structure of SLUG-5 was
previously determined by single crystal X-ray dffraction.17 The
characteristic low angle peak at 8.37 2θ (d ∼ 10.50 Å, which is
the strongest reflection) indicates the layer-to-layer distance.
The slight shift in the higher angle peaks and a change in
intensity are likely due to preferred orientation, a consequence
of the different morphologies (vide infra). The crystal growth is
affected by several factors such as crystallization period
(evaporation rate, cooling temperature) and stirring speed,
with varying degree of preferred orientation for each
morphology. For example, the (h00) peaks for SLUG-5 (II)
are lowest intensity of the four methods and shifted to slightly
lower d-spacing (Figure 2, inset). SLUG-5 (IV) displays the
strongest intensity (200) peak. During the SLUG-5 (II)
crystallization, the agitation by the stir bar likely interferes
with and limits the degree of crystallization.
Figure 4. Thermogravimetric analysis shows similar traces for all four
synthesis methods: (I) hydrothermal, (II) stirring-induced, (III) reflux-
induced, and (IV) evaporation.
decomposition of the structure and accounts for the loss of
the ethanedisulfonate template (experimental: 18.0%; theoreti-
cal: 17.9%) and crystalline water (experimental: 2.0%;
theoretical: 2.3%). The hydrothermal product (I) has a slightly
higher thermal stability, presumably due to a more ordered
lattice.17 In all cases, PXRD after heating to 600 °C indicates
collapse to valentinite, Sb2O3 (ICDD #11-689). SLUG-5 (II,
stirring) and (IV, evaporation) show a slight mass loss of ∼2%
around 280 °C, likely due to additional interlayer chemisorbed
water.
FT-IR spectra were also obtained to confirm the presence of
the ethanedisulfonate template between the layers (Figure 3).
Sharp bands observed between 1020 and 1013 cm−1 are
After it was confirmed that the three new morphologies are
the same as our crystal structure reported from hydrothermal
synthesis, the size and morphology were investigated by SEM.
Although the powder patterns are similar, the particle size and
morphology greatly depend on the synthetic method. Hydro-
thermal crystals (I) are up to several millimeters in size with an
uneven surface (Figure 5-I). Stirring-induced crystals (II) are
very thin platelets, 1−5 μm in width (Figure 5-II). Reflux-
induced crystals (III) are significantly larger and thicker, with a
much greater size distribution of ∼1 to 20 μm (Figure 5-III).
Evaporation-induced crystals (IV) display the most homoge-
neous morphology and smallest particle size (Figure 5-IV). As
observed for other materials such as zirconium phosphates,20
reflux conditions often give rise to smaller particles than
hydrothermal conditions. The static conditions of hydrothermal
(I) and evaporation (IV) led to a narrower particle size
distribution.
Figure 3. FT-IR spectra of the different morphologies confirm the
presence of interlayer EDS anions: (I) hydrothermal synthesis, (II)
stirring-induced crystallization, (III) reflux-induced crystallization, and
(IV) evaporation.
Clearly, the size of the crystals can be effectively controlled
by the synthetic technique. A given method affects the
solubility of the precursors and thus supersaturation level and
95
dx.doi.org/10.1021/cg200826d | Cryst. Growth Des. 2012, 12, 93−98

Crystal Growth & Design
Article
maximum yield of 92%, followed by (I), (II), and (IV) (80−
84%, Table 1). The yield profile over 4 h is plotted in Figure 6.
No reaction occurred in the absence of the catalyst or with
EDSA only. It is unlikely any species including EDS are
Table 1. Ketal Formation Yield for the Four Morphologies
after 4 h Reaction Time
crystallization method
abbreviation
catalytic yield [%]
hydrothermal
stirring-induced
reflux-induced
evaporation
I
83.9
82.5
92.3
79.6
II
III
IV
Figure 5. SEM of SLUG-5 synthesized by each method: (I)
hydrothermal yields a rough surface and large particle size; (II)
stirring-induced crystallization leads to thin platelets; (III) reflux-
induced crystallization gave thicker crystals ranging from 1 to 20 μm in
size. The crystal edges are much sharper than for the other methods;
(IV) evaporation-induced crystallization results in evenly shaped
nanoplatelets, the smallest particles of the four methods.
Figure 6. Percent conversion of methyl ethyl ketone to ketal with
different morphologies of SLUG-5 [hydrothermal (I), stirring-induced
(II), reflux-induced (III), evaporation (IV)], as well as antimony oxide
(V, square symbol at 4 h) and with no solid catalyst present (VI,
triangle symbol at 4 h).
rate of crystal growth. Hydrothermal conditions (higher
temperature, autogenous pressure created in the autoclave)
generally leads to larger crystals than reflux conditions due to
the elevated pressure and reduced supersaturation, as well as
the ripening effect due to greater temperature fluctuation and
thermal gradients. Reflux gives rise to a higher level of
supersaturation, for rapid growth of smaller crystals, with the
agitation caused by stirring and boiling further limiting crystal
size. Stirring and evaporation occurred under ambient
conditions and hence formed smaller particles. While particle
size is inversely proportional to specific surface area, the
crystallinity and morphology are known be equally important in
determining catalytic properties.
catalyzing homogeneously. We have published several related
cationic inorganic materials that contain interlamellar EDS
anions.21−23 In all cases, we removed the solid catalyst after
several hours and allowed the reaction to continue. No
1
additional ketal formed, while H NMR showed that no EDS
was present in any of the reaction mixtures. Antimony(III)
oxide as catalyst resulted in a finite but very low yield of 3.0%,
underscoring the activity of SLUG-5.
One possible explanation of the enhanced catalytic properties
of the reflux-induced SLUG-5 (III) is the well-defined surfaces
of the polygonal crystals. The crystals are larger than the
rounded disk-like morphologies from stirring-induced (II) and
evaporation (IV). Another possible explanation for the
enhanced performance of (III) is the sharp crystal edges of
the material. Such areas are known to have more defects as well
as adsorb reactants onto the catalyst surface.24,25 This
explanation would account for the lower performance by the
3.2. Catalytic Activity. As a benchmark reaction to
compare the four morphologies, ketal formation from 2-
butanone (methyl ethyl ketone, MEK) using 1,2-ethanediol
(ethylene glycol) was studied. Generally, this step serves as
protection of a carbonyl group from undesired side reaction
(Scheme 1). The reaction was performed for 4 h (Experimental
Section 2.2).
The morphology has a significant impact on the performance
of the catalyst. The most efficient was catalyst (III), with a
a
Scheme 1. Ketal Formation from 2-Butanone and 1,2-Ethanediol Results in a Protective Group for the Carbonyl Group
a
Water is formed as a side product and removed during the reaction in order to shift the equilibrium to the right.
96
dx.doi.org/10.1021/cg200826d | Cryst. Growth Des. 2012, 12, 93−98

Crystal Growth & Design
Article
Figure 7. Percent ketal conversion by powdered (A) and preheated-powdered (B) SLUG-5 for the different synthesis methods [hydrothermal (I),
stirring-induced (II), reflux-induced (III), and evaporation (IV)].
rounded morphology of evaporation (IV). Indeed, the rough
edges of hydrothermal (I) and reflux (III) likely give higher
ketal formation than the much smaller crystals of stirring (II)
and evaporation (IV). The lack of porosity of SLUG-5 implies
that the reactions are occurring on the surface of the particles
and thus are sensitive to the surface morphology.
that of hydrothermal (I) but with a far simpler, cost-effective
synthetic method. Furthermore, the catalytic yield for all four
forms of SLUG-5 increased to 97% when the samples were
powdered. We are also investigating other reactions for SLUG-
5 including Diels−Alder, esterification, and Friedel−Crafts
alkylation. The cationic charge of the layers coupled with
morphological control provides SLUG-5 a significant advantage
over traditional catalysts. The material can be easily recovered
by filtration for reuse, making it an attractive choice over
homogeneous toluenesulfonate, the standard ketalization
catalyst.
Another possible reason for the varying yields, however, is
the structurally contained water. Water reduces the speed of the
reaction by shifting the equilibrium toward the reactants and
needs to be constantly removed from the system. Indeed, TGA
below 250 °C (Figure 4) indicates a greater amount of physi-
and chemisorbed water in SLUG-5 (I, II, and III). We therefore
further examined all four morphologies by repeating the
catalysis in both powdered and preheated-powdered forms. To
remove extra water from the surface of the latter, the samples
were kept at 125 °C for 30 min. The conversion yield for the
catalysis using all four SLUG-5 morphologies is ca. 97% after 4
h (Figure 7). Clearly, powdering increased specific surface area
and in turn conversion yield. The powdered-preheated samples
showed no increase in final yield as compared to the powdered
SLUG-5 morphologies. Table 2 compares the conversion yields
of as-synthesized, powdered and powdered-preheated SLUG-5.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: soliver@chemistry.ucsc.edu.
REFERENCES
■
(1) Tran, D. T.; Zavalij, P. Y.; Oliver, S. R. J. J. Am. Chem. Soc. 2002,
124, 3966−3969.
(2) Ma, R.; Liu, Z.; Takada, K.; Iyi, N.; Bando, Y.; Sasaki, T. J. Am.
Chem. Soc. 2007, 129, 5257−5263.
(3) Centi, G.; Perathoner, S. Microporous Mesoporous Mater. 2008,
107, 3−15.
(4) Tagusagawa, C.; Takagaki, A.; Hayashi, S.; Domen, K. J. Am.
Chem. Soc. 2008, 130, 7230−7231.
Table 2. Comparison of Ketal Yield for the Four
Morphologies after 4 h Reaction Time
(5) Sanchez, C. A.; Crump, K. S.; Krieger, R. I.; Khandaker, N. R.;
Gibbs, J. P. Environ. Sci. Technol. 2005, 39, 9391−9397.
(6) Hawxwell, S. M.; Adams, H.; Brammer, L. Acta Crystallogr. Sect. B
2006, 62, 808−814.
catalytic yield [%]
crystallization method
as-synthesized
powdered
preheated
(7) Vaccari, A. Appl. Clay Sci. 1999, 14, 161−198.
(8) Kobayashi, S.; Sugiura, M.; Kitagawa, H.; Lam, W. W. L. Chem.
Rev. 2002, 102, 2227−2302.
hydrothermal
stirring-induced
reflux-induced
evaporation
I
83.9
82.5
92.3
79.6
96.3
97.3
96.6
96.9
98.4
95.4
97.7
96.1
II
(9) Longbottom, D. Synlett 1999, 2023.
III
IV
(10) Forster, P. M.; Cheetham, A. K. Top. Catal. 2003, 24, 79−86.
(11) Corma, A.; Garcia, H. Chem. Rev. 2003, 103, 4307−4366.
(12) Tsuchimoto, T.; Tobita, K.; Hiyama, T.; Fukuzawa, S.-i. J. Org.
Chem. 1997, 62, 6997−7005.
4. CONCLUSIONS
(13) Kobayashi, S.; Hachiya, I.; Araki, M.; Ishitani, H. Tetrahedron
Lett. 1993, 34, 3755−3758.
Several new synthetic pathways to SLUG-5 requiring only
ambient conditions have led to increased catalytic performance.
The crystalline structure of SLUG-5 via hydrothermal (I),
stirring (II), reflux (III), and evaporation (IV) are identical, yet
their morphology influences the catalytic behavior. Reflux (III)
gave the strongest catalytic yield with reactivity greater than
(14) Patel, S. M.; Chudasama, U. V.; Ganeshpure, P. A. J. Mol. Catal
A: Chem. 2003, 194, 267−271.
(15) Banik, B. K.; Chapa, M.; Marquez, J.; Cardona, M. Tetrahedron
Lett. 2005, 46, 2341−2343.
(16) Kobayashi, S.; Manabe, K. Pure Appl. Chem. 2000, 72, 1373−
1380.
97
dx.doi.org/10.1021/cg200826d | Cryst. Growth Des. 2012, 12, 93−98

Crystal Growth & Design
Article
(17) Swanson, C. H.; Shaikh, H. A.; Rogow, D. L.; Oliver, A. G.;
Campana, C. F.; Oliver, S. R. J. J. Am. Chem. Soc. 2008, 130, 11737−
11741.
(18) Oliver, S. R. J. Chem. Soc. Rev. 2009, 38, 1868−1881.
(19) Fadini, A.; Schnepel, F. M. Schwingungsspektroskopie: Methoden
und Anwendung; Thieme: Stuttgart, 1985.
(20) Kaschak, D. M.; Johnson, S. A.; Hooks, D. E.; Kim, H. N.; Ward,
M. D.; Mallouk, T. E. J. Am. Chem. Soc. 1998, 120, 10887−10894.
(21) Rogow, D. L.; Fei, H. H.; Brennan, D. P.; Ikehata, M.; Zavalij, P.
Y.; Oliver, A. G.; Oliver, S. R. J. Inorg. Chem. 2010, 49, 5619−5624.
(22) Fei, H. H.; Paw, L.; Rogow, D. L.; Bresler, M. R.; Abdollahian, Y.
A.; Oliver, S. R. J. Chem. Mater. 2010, 22, 2027−2032.
(23) Fei, H. H.; Rogow, D. L.; Oliver, S. R. J. J. Am. Chem. Soc. 2010,
132, 7202−7209.
(24) Marra, G. L.; Tozzola, G.; Leofanti, G.; Padovan, M.; Petrini, G.;
Genoni, F.; Venturelli, B.; Zecchina, A.; Bordiga, S.; Ricchiardi, G.
Stud. Surf. Sci. Catal. 1994, 84, 559−566.
(25) Roeffaers, M. B. J.; Sels, B. F.; Uji-i, H.; De Schryver, F. C.;
Jacobs, P. A.; De Vos, D. E.; Hofkens, J. Nature 2006, 439, 572−575.
98
dx.doi.org/10.1021/cg200826d | Cryst. Growth Des. 2012, 12, 93−98