Supramolecular Assembly of Tris(4-carboxyphenyl)arenes: Relationship between Molecular Structure and Solid-State Catenation Motifs

Article abstract of DOI:10.1021/acs.cgd.5b01416

The crystal structures of seven 1,3,5-tris(4-carboxyphenyl)arenes with functionalized central arene rings are reported. The formation of (6,3) hcb hexagonal sheets as a result of carboxylic acid dimer formation was observed in most of the crystal structures, with the exception of two compounds with functional groups capable of forming hydrogen bonds, namely, 2,4,6-tris(4-carboxyphenyl)-1,3-diaminobenzene and 2,4,6-tris(4-carboxyphenyl)-3-methylaniline. These structures were found to incorporate THF solvent molecules in their hydrogen-bonding motif, giving rise to distorted pseudohexagonal arrays. Functional groups on the central ring were found to influence stacking distances, stacking offsets, inclination angles, degree of catenation, and dimensions of solvent-occupied channels. To better understand and appreciate these complicated crystal structures, they were categorized into four distinct stacking/catenation families: simple stacking, single-layer offset catenation, double-layer offset catenation, and rotated-layer catenation. The unique structure of the unfunctionalized parent compound 1,3,5-tris(4-carboxyphenyl)benzene is rationalized in light of the structural behavior of its derivatives.

Full text of DOI:10.1021/acs.cgd.5b01416

Article
pubs.acs.org/crystal
Supramolecular Assembly of Tris(4-carboxyphenyl)arenes:
Relationship between Molecular Structure and Solid-State
Catenation Motifs
Holden W. H. Lai,*,†,# Ren A. Wiscons,† Cassandra A. Zentner,† Matthias Zeller,‡
and Jesse L. C. Rowsell†,§
†Department of Chemistry and Biochemistry, Oberlin College, 119 Woodland Street, Oberlin, Ohio 44074, United States
‡Department of Chemistry, Youngstown State University, 1 University Plaza, Youngstown, Ohio 44555, United States
S
* Supporting Information
ABSTRACT: The crystal structures of seven 1,3,5-tris(4-
carboxyphenyl)arenes with functionalized central arene rings
are reported. The formation of (6,3) hcb hexagonal sheets as a
result of carboxylic acid dimer formation was observed in most
of the crystal structures, with the exception of two compounds
with functional groups capable of forming hydrogen bonds,
namely, 2,4,6-tris(4-carboxyphenyl)-1,3-diaminobenzene and
2,4,6-tris(4-carboxyphenyl)-3-methylaniline. These structures
were found to incorporate THF solvent molecules in their
hydrogen-bonding motif, giving rise to distorted pseudohex-
agonal arrays. Functional groups on the central ring were found to influence stacking distances, stacking offsets, inclination
angles, degree of catenation, and dimensions of solvent-occupied channels. To better understand and appreciate these
complicated crystal structures, they were categorized into four distinct stacking/catenation families: simple stacking, single-layer
offset catenation, double-layer offset catenation, and rotated-layer catenation. The unique structure of the unfunctionalized parent
compound 1,3,5-tris(4-carboxyphenyl)benzene is rationalized in light of the structural behavior of its derivatives.
Individual molecules of 1 have three carboxylic acid
functional groups and molecular 3-fold rotational symmetry.
Carboxylic acids are known to form R22(8) dimers (Figure 1),4 a
robust synthon present in both the solid and liquid states.3 A
recent survey of the Cambridge Structural Database (CSD)
found that over a third of all carboxylic acids form R22(8)
dimers.5 If carboxylic acid dimers were formed with no
interruption in the self-assembly of 1, (6,3) hcb hexagonal
sheets with >2 nm pore diameter would result. Silly,6 Ruben
and co-workers,7 as well as Lackinger and his group8,9
employed scanning tunneling microscopy to study the 2-D
self-assembly of 1 at the liquid−solid interface. Consistent with
the carboxylic acid dimer synthon, it was determined that (6,3)
hexagonal sheets were overall the most prevalent assembly.
However, 2-D phases with alternative hydrogen-bonded
networks were also observed.
INTRODUCTION
■
The understanding of extended lattice motifs of crystalline
molecular compounds has greatly expanded within the past
decades. Despite these advances, it is not yet possible to reliably
predict a crystalline phase from solely its molecular
constituents, making crystal structure prediction a challenge
even for simple systems. Crystal engineering seeks to solve this
predicament, aiming to rationalize and reliably predict crystal
structures based on understanding and exploitation of
intermolecular interactions, thus enabling the design and
synthesis of molecular solid-state structures.1,2 Central to
crystal engineering is the concept of supramolecular synthons,
structural units formed by predictable intermolecular inter-
actions.3 While supramolecular synthons are helpful for
predicting local intermolecular interactions, they often provide
little insight into the global topology of a crystal structure.
Carboxylic acids are a popular choice for supramolecular
synthons in the rational design of molecular crystals because of
the directionality and predictability of their interactions. In
recent years, our group and others have attempted to
understand the self-assembly of 1,3,5-tris(4-carboxyphenyl)-
benzene (tcpb, 1, Table 1) and its derivatives. Among this
compound and its chemical derivatives, the carboxylic acids are
expected to provide the driving force that guides the overall
topology of tcpb derivatives’ structures.
The R22(8) carboxylic acid dimer synthon was also found in
the solid-state 3D assembly of 1. Hexagonal sheets formed in
the crystal structure of 1,10 reported by our group, and that of
one of its derivatives, 2,4,6-tris(4-carboxyphenyl)mesitylene (9,
Table 1), described by Moorthy and co-workers.11 However,
these two crystal structures have distinct packing motifs that
Received: October 3, 2015
Revised: November 26, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
compare the motifs in the solid-state assembly of tcpb
derivatives.
Table 1. Molecular Structures of 1,3,5-Tris(4-
carboxyphenyl)benzene Derivatives
The present paper follows up on our recent report on the
structure and properties of parent compound 1 and systemati-
cally explores the effect of central ring chemical functionaliza-
tion on its supramolecular assembly. The effect of functional
groups on the molecular conformation of tcpb derivatives will
be discussed, followed by the effect of molecular conformation
on the formation, stacking, and catenation of the hexagonal and
pseudohexagonal sheets. Finally, the solvent-occupied channels
and the overall stacking and catenation motifs of the structures
are compared, and the findings are used to rationalize the
unique structure of the unfunctionalized parent compound 1.
source
code
R1
R2
R3
ref 10
1
2
3
4
5
6
7
8
9
H
H
H
RESULTS AND DISCUSSION
■
this work
this work
this work
this work
this work
this work
this work
ref 11
H
H
CH3
Synthesis and Crystallization of 2−8. The synthesis and
crystallization of 1 and 9 have been reported elsewhere.10,11
Compounds 2−7 were synthesized by Suzuki-Miyaura coupling
of three equivalents of 4-methylcarboxyphenylboronic acid with
the corresponding functionalized tribromobenzenes, followed
by hydrolysis of the methyl esters to yield carboxylic acids
(Table 2). Although the hydrolyzed homocoupling product,
4,4′-biphenyldicarboxylic acid, was observed in the coupling
reactions, this side product could be minimized by exclusion of
oxygen and water during synthesis. The small traces of the
homocoupling products that still formed were readily removed
in the crystallization process. Because of deactivation of
nitrobenzene toward bromination,14 an alternative approach
avoiding the use of tribromonitrobenzene was developed for
the preparation of 8. Compound 8 was synthesized by the
cyclocondensation of 2 equiv of 4-methylacetophenone and 1
equiv of 4-methylbenzaldehyde to form 2,4,6-tris(4-
methylphenyl)pyrylium tetrafluoroborate. The pyrylium core
was then reacted with nitromethane to yield 2,4,6-tris(4-
methylphenyl)nitrobenzene, which was oxidized with potas-
sium permanganate to yield the triacid (Scheme 1).
H
H
OCH3
OCH3
NH2
NH2
NH2
NO2
CH3
H
CH3
H
H
H
CH3
NH2
H
H
H
CH3
CH3
Figure 1. R22(8) carboxylic acid dimer, an eight-membered ring
consisting of two hydrogen bond donors and two hydrogen bond
acceptors.
cannot be rationalized by the carboxylic acid dimer synthon
alone. Crystal structures of 1 and 9 both exhibit inclined
polycatenation (ICAT) but differ in their degrees of catenation
(DOC),12 whereas the structure of 1 has a DOC of (7/7), the
structure of 9 has a DOC of (5/5) (see Note 1 in Supporting
Information (SI)). A DOC of (7/7) indicates that there are two
motifs, with seven external rings from one motif catenating a
single ring of the other motif, and vice versa (Figure 2).13 To
better understand these distinct structural features of 1 and its
derivatives, a systematic study was conducted to analyze and
Single crystals of 3−8 were grown by vapor diffusion using
either THF or 1,4-dioxane as the solvent, and either acetonitrile
or propionitrile as the antisolvent. Single crystals of 2 were
grown by slow cooling from a heated solvent mixture of 1:1
(v:v) THF/water.
Single-Crystal X-ray Diffraction (SCXRD) of 2−8.
Structure determinations of the crystals for these compounds
were particularly challenging because of their solvent-occupied
channels and tendency to quickly desolvate upon removal from
the mother liquor. Back Fourier transform methods
(SQUEEZE, as implemented in the program Platon15) were
applied to account for electron density of unresolved or
extensively disordered solvent molecules in the channels of the
frameworks. For several tcpb derivatives, disorder of framework
and/or solvent molecules was too pronounced, resulting in
diffraction to only very low resolution. Compounds for which
no data with sufficiently high resolution could be obtained
include 7 and 2,4,6-tris(4-carboxyphenyl)-3,5-dimethylaniline,
grown by vapor diffusion using the solvent/antisolvent pair 1,4-
dioxane/acetonitrile. Some of the remaining structures, such as
6 grown by vapor diffusion using the solvent/antisolvent pair
1,4-dioxane/acetonitrile, had sufficient diffraction intensity but
suffered from non-Bragg behavior such as streaking of diffracted
intensity between Bragg peaks (see Figure S12 for examples of
representative diffraction images). Possible causes for this
behavior include stacking faults of hexagonal sheets or other
dislocations. However, due to the labile nature of the materials
Figure 2. DOC of (7/7) found in the structure of 1. Seven layers of
hexagonal sheets (three dark green and four blue) catenate one
hexagon (shown in light green).
B
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
a
Table 2. Synthesis of 2−7
code
R1
R2
R3
overall yield (%)
2
3
4
H
H
H
H
H
H
H
CH3
93
62
75
94
98
78
H
OCH3
OCH3
NH2
CH3
H
b
5
b
6
b
CH3
NH2
NH2
7
NH2
a
b
The first step was performed under nitrogen atmosphere for all compounds. The second step, hydrolysis of the methyl esters, was performed
under nitrogen atmosphere.
the study. Of the structures included, that of 3 is the only
structure that is noncatenated; i.e., layers of hexagonal sheets
are not anchored against shifting relative to one another, which
would make this material more likely to suffer from dislocations
and stacking faults and also explains its particular structural
instability toward desolvation. On the basis of this behavior, it is
reasonable to speculate that crystals with even more
pronounced non-Bragg behavior might also have noncaten-
tated, layered structures. Crystal structure data of compounds
2−8 are summarized in Table 3. Additional structure
determination details and ORTEP style plots for all compounds
are given in the SI.
Scheme 1. Synthesis of 8
Powder X-ray Diffraction (PXRD) and Thermal
Gravimetric Analysis (TGA) of 2−8. Crystals of 2−8 were
filtered, dried over a glass frit, and analyzed by PXRD and TGA.
Upon filtration, all crystals, with the exception of 5, slowly
became opaque as solvent evaporated from within the crystals;
macroscopic fractures and cracks became evident for larger
crystals. Crystals of 2, 4, 6, and 7 had PXRD diffractograms that
did not match the calculated diffractograms from SCXRD data,
likely due to structural transformations as a result of solvent
evaporation. Crystals of 3 and 8 lost crystallinity completely
upon filtration. Compound 5 retained the core features of its
calculated diffractogram from SCXRD data. The peaks are
shifted to lower diffraction angles, indicating an expansion of its
framework between the single crystal structure measured at 100
K and the experimental powder diffractogram taken at room
temperature (see Figures S13−19 for comparisons of the
measured PXRD diffractograms and diffractograms calculated
from SCXRD data). The observed disagreement between
experimental and calculated PXRD diffractograms is consistent
with a general trend for porous molecular crystals as a result of
framework collapse upon evacuation of solvent molecules in
voids or channels.16 Tcpb 1 is a notable exception, which was
found to be unusually stable even after complete solvent
evacuation.10 Despite the drastic changes in the crystal
structures of 2−4 and 6−8 upon filtration and air-drying,
TGA, in conjunction with NMR analyses of the residues
revealed that solvent molecules remained in the solids of 3−8
and their rapid loss of crystallinity upon removal from mother
liquor, it was not possible to use electron microscopy or
diffraction techniques to further investigate the origin and exact
nature of the non-Bragg behavior. For some of the materials,
the degree of non-Bragg diffraction was too pronounced to
allow for a meaningful refinement of the overall structures.
These structures were therefore excluded from this report.
Crystals falling into this category include 8 and 2,4,6-tris(4-
carboxyphenyl)-meta-xylene grown by vapor diffusion using the
solvent/antisolvent pair tetrahydrofuran/acetonitrile. The
reported structures of 6, 7, and 8 were determined from
crystals grown from different solvents (see the crystallization
conditions in the Experimental Section).
Of the reported structures, compounds 8 and especially 3 did
show some non-Bragg behavior, resulting in increased R values
and low structure quality indicators. The nature of the
frameworks and structural motifs obtained for compounds 3
and 8 are, however, unambiguous, and they are thus included in
C
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
a
Table 3. Crystal Structure Data of 1−9
c
compound
1
2
3
4
5
molecular formula
formula weight
crystal system
space group
a (Å)
C27H18O6
438.44
C28H20O6
452.46
C28H20O7
468.46
C29H22O7
482.49
C27H19NO6
453.43
monoclinic
I2
orthorhombic
Cmca
monoclinic
C2/c
orthorhombic
Pca21
orthorhombic
Cmca
31.419(6)
30.116(6)
45.320(9)
90.412(2)
42880(14)
37.1
30.4631(10)
23.4366(8)
8.4589(3)
90
53.575(3)
7.2941(5)
33.145(2)
33.380(2)
90
32.015(3)
24.655(2)
23.540(2)
90
b (Å)
31.8450(18)
25.6252(14)
118.323(3)
38485(4)
68.6
c (Å)
β (deg)
volume (Å3)
6039.2(4)
31.9
8070.1(9)
44.7
18581(3)
34.1
b
% void volume
Z, Z′
56, 14
8, 0.5
24, 3
8, 2
24, 1.5
final R indices
R1 = 0.060, wR2 = 0.138 R1 = 0.0334,
wR2 = 0.0957
R1 = 0.1479,
wR2 = 0.5277
R1 = 0.0938,
wR2 = 0.2687
R1 = 0.0672,
wR2 = 0.2851
(I > 2σ(I))
compound
6-THF
7-THF
C31H28N2O7
540.57
8
9−2DME11
molecular formula
formula weight
crystal system
space group
a (Å)
C32H29NO7
539.58
C27H17NO8
C30H24O6
483.42
660.75
orthorhombic
Pbca
orthorhombic
Pbca
orthorhombic
Iba2
orthorhombic
Pbcn
26.571(3)
7.2638(8)
32.672(4)
6305.9(13)
16.2
26.523(4)
7.2629(6)
32.349(3)
6231.5(12)
16.8
14.692(9)
29.009(18)
31.930(19)
13609(14)
36.7
27.090(3)
9.0376(9)
31.579(3)
7731.3(14)
41.6
b (Å)
c (Å)
volume (Å3)
b
% void volume
Z, Z′
8, 1
8, 1
16, 2
8, 1
final R indices (I > 2σ(I))
R1 = 0.0610, wR2 = 0.1664
R1 = 0.0658, wR2 = 0.1519
R1 = 0.1095, wR2 = 0.2755
R1 = 0.1222, wR2 = 0.3324
a
The SQUEEZE routine was used to correct for the influence of disordered solvent molecules in all structures, with the exception of 9, in which the
b
solvent molecules in the channels were modeled to be dimethoxyethane (DME). The void volume was determined after removal of all
nonframework molecules using CCDC Mercury computer software, calculated using contact surface, with a probe radius of 1.2 Å and approximate
c
grid spacing of 0.7 Å. Data of the I2 polymorph of 1, from ref 10.
up to temperatures between 100 and 200 °C (see the SI for
NMR spectra and TGA traces).
Molecular Conformations of 1−9. Functional groups on
the central arene ring affect the molecular conformations of
tcpb derivatives because of the steric repulsion between the
functional groups and the hydrogen atoms at the ortho
positions of the outer arene rings. This repulsion leads to
increases in torsion angles between the central and outer arene
rings of tcpb derivative molecules. To probe this effect, the
average torsion angles between the central and outer arene
rings were plotted against the average steric A values17 (see
Note 2 in SI) of the three substituents on the central arene ring
in compounds 1−9 (Figure 3). As expected, torsion angles
increase with the steric A values of the substituents on the
central arene ring. Compound 9, which has three methyl
groups on the central arene ring and the highest average steric
A value (1.7 kcal/mol for methyl), has three outer arene rings
that are nearly orthogonal to the central ring (average torsion
angle = 87.66°).
Figure 3. Correlation between the average torsion angle (deg)
between central and outer arene rings and the average steric A value of
substituents on the central ring for the crystal structures of 1−9. The A
value15 (kcal/mol) of hydrogen is zero (see SI for tables of all torsion
angles and steric A values).
An understanding of the ways in which functional groups on
the central arene ring of the tcpb derivatives affect their
molecular conformations aids in the rationalization of their
solid-state assemblies; molecules become less planar the larger
the torsion angles are between the central and substituent arene
rings, which affects their ability to π-stack in the solid state. This
phenomenon will be discussed in more detail in the following
sections.
the crystal structures of 1−5 and 8−9. The hexagonal sheets
formed were found to be especially distorted among tcpb
derivatives that have large torsion angles between central and
outer arene rings, such as 4 and 9 (Figure 4d,h). The structures
of 6 and 7 also have large torsion angles and distorted
pseudohexagonal sheets, but the distortion is primarily due to
interruption of the carboxylic acid dimer pattern by solvent
molecules. Two parallel sides of the hexagons are shortened by
hydrogen bonding of the carboxylic acid toward a THF
molecule and a methyl or amine group (Figure 4f). The
hydrogen-bonding motif here forms an eight-membered ring
Formation of (6,3) hcb Hexagonal and Pseudohex-
agonal Sheets. The formation and stacking of (6,3) hcb
sheets held together entirely by R22(8) dimers are conserved in
D
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
Figure 5. (a) One third of the carboxylic acid groups of 6 form R23(8)
connections that consist of COOH and NH2 groups from two
molecules of 6 and two THF molecules, forming one side of the
hexagon. The gray, white, blue, and red atoms are carbon, hydrogen,
nitrogen, and oxygen atoms, respectively; (b) this R23(8) hydrogen-
bonding motif consists of two donors, three acceptors, and forms an
eight-membered ring. Crystal structure refinement determined that in
the structure of 6 the methyl group can take the place of the amino
group and form the same hydrogen-bonding motif. Red dashed lines
indicate hydrogen bonds.
carboxylate, creating coulombic repulsion that would prevent
the formation of R22(8) dimers and hydrogen bonding between
the carboxylic acid group and THF molecules as observed in
5−7.
Stacking of (6,3) Hexagonal and Pseudohexagonal
Sheets. Stacking of the (6,3) hexagonal sheets is conserved in
the crystal structures of 1−5, 8, and 9. Analogous stacking of
pseudohexagonal sheets is observed in 6 and 7. To better
understand and quantify the stacking distances between
hexagonal sheets, the distance S (Figure 6) was measured
Figure 4. Complete hexagons from the (6,3) hcb sheets in the crystal
structure of (a) 1, (b) 2, (c) 3, (d) 4, (e) 5, (g) 8, and (h) 9, and
pseudohexagonal sheet in the structures of (f) 6 and 7. The hexagonal
sheets are held together by R22(8) carboxylic acid hydrogen bonds (for
6 and 7, R22(8) and R23(8)). The gray, white, blue, and red atoms are
carbon, hydrogen, nitrogen, and oxygen atoms, respectively. Lengths
of the diagonals of the hexagons were determined in units of Å by
measuring the interatomic distance between the atoms directly
attached to the central arene rings that are located at the opposite
vertices of the hexagons, and subtracting the van der Waals radii18 of
the atoms. The directions of the black arrows also indicate the
direction along which catenating hexagonal or pseudohexagonal sheets
stack, except in the case of 3 (c), which does not exhibit
polycatenation.
Figure 6. Distance S (in Å) used to quantify the stacking distances
between hexagonal sheets. The red spheres are the centroids of the
central arene rings of the tcpb derivative molecules.
between the central arene rings of the stacked tcpb derivative
molecules. Generally, tcpb derivative molecules with greater
torsion angles between the central and outer arene rings were
found to have larger stacking distances S (Figure 7). This
correlation is expected because orthogonal outer rings push
stacked molecules further apart.
Inclined Polycatenation. Inclined polycatenation (ICAT)
is observed in all of the tcpb derivative crystal structures with
the exception of 3. The DOC, however, are different among
these structures (Figure 8). DOC and Inclination angles
(Figure 9), the acute angles formed by two catenating hcb
sheets in the ICAT structures, were determined using
TOPOS19 and are tabulated in Table 4.
Among the polycatenated crystal structures, functional
groups on the central ring affect the degree of catenation by
changing two structural features: the dimensions of the
hexagonal channels and the stacking distances between
hexagonal sheets. As depicted in Figure 4, all of the catenating
hexagonal sheets stack along a diagonal of the hexagonal
channel, indicated by the black arrows. Crystal structures of 2−
9 all have shorter diagonals relative to 1 because of the
functional groups that are present at the vertices of the
hexagonal channels. As a result, fewer catenating hexagonal
and consists of three donors and two acceptors (Figure 5). It is
therefore categorized as an R23(8) motif.4 While the amino
functional group, which is weakly basic and capable of both
accepting and donating hydrogen bonds, did form hydrogen
bonds with solvent molecules to interrupt the hcb sheet
formations in 6 and 7, these interactions were not observed in
5, which has a single amino group. There was no evidence of
proton transfer between the amine and carboxylic acid groups
in the crystal structures of 5−7. Hydrogen atoms at both
carboxylic acid as well as amine groups were resolved in
difference density electron maps for fully occupied and major
disordered moieties. Any significant proton transfer from
carboxylic acid to amine would create a negatively charged
E
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
Figure 9. Inclination angle, θ, from the crystal structure of 1.
Table 4. Inclination Angles (Figure 8) and DOC of the
ICAT Structures of 1, 2, 4−9
Figure 7. Correlation between the average stacking distances S (Figure
6) and average torsion torsion angle (deg) between central and outer
arene for the crystal structures of 1−9.
crystal structure
inclination angle (deg)
DOC
1
2
4
5
6
7
8
9
67.2
49.6
84.5
55.3
84.9
84.8
66.7
61.9
(7/7)
(6/6)
(5/5)
(6/6)
(4/4)
(4/4)
(6/6)
(5/5)
sheets will fit along the diagonal, leading to lower DOC. This
explains why crystal structures of tcpb derivatives with one
functional group on the central ring (2, 5, 8) have a DOC of
(6/6), lower than the DOC of (7/7) in the crystal structure of
nonfunctionalized 1. With multiple functional groups on the
central ring, as is the case with both 4 and 9, the torsion angles
between the central and outer arene rings increases due to
steric repulsion. This leads to increased stacking distances
(Figure 7). With shorter diagonal lengths in the hexagonal
channels and increased stacking distances between hexagonal
sheets, both 4 and 9 have a DOC of (5/5), lower than what is
observed in the structure of the monofunctionalized tcpb
derivatives other than 3. Crystal structures of 6 and 7 have
pseudohexagonal channels with the shortest diagonals as a
result of the THF-amine/methyl hydrogen-bonding motif
(Figure 5). As a result, these structures have the lowest DOC
of (4/4) among the polycatenated crystal structures.
Solvent-Filled Channels. Polycatenation influences the
dimensions of the solvent-occupied channels in the structures
of 1−9 (Figure 10). As expected, the noncatenated crystal
structure of 3 has the largest % void volume reported here, at
69%, and a channel diameter of greater than 2 nm. The
catenated structures with hexagonal sheets connected entirely
through carboxylic acid dimers have solvent-occupied volumes
within a range of 30−50% and channel diameters of less than 1
nm. Structures of 6 and 7 on the other hand, which have
smaller and distorted pseudohexagonal sheets due to hydrogen
bonding with THF molecules, have reduced solvent-occupied
volumes of less than 20%. For the calculation of % void
volumes in 6 and 7, the tightly bound THF molecules are
considered an intrinsic part of the hydrogen-bonded framework
for 6 and 7, not as disordered guest molecules in the voids.
Stacking and Catenation Motifs. As discussed in the
previous sections, supramolecular synthons and functional
groups on the central arene ring can help rationalize structural
motifs in the solid-state assembly of tcpb derivatives. While
hydrogen bonding and the R22(8) carboxylic acid dimer synthon
reliably lead to the formation of (6,3) hcb hexagonal or
pseudohexagonal sheets, steric demand of the functional groups
on the central arene ring affects the molecular conformations of
Figure 8. DOC in crystal structures of 1, 2, and 4−9. (a) 1 DOC (7/7), (b) 2 DOC (6/6), (c) 8 DOC (6/6), (d) 5 DOC (6/6), (e) 4 DOC (5/5),
(f) 9 DOC (5/5), and (g) 6 and 7 DOC (4/4), which are isomorphous. Green molecules form catenating hexagonal or pseudohexagonal sheets that
are analogous to the purple ones. The green molecules are shown in alternating light and dark for clarity. The gray, white, and red atoms in (g) are
respectively carbon, hydrogen, and oxygen atoms that belong to the ordered THF molecules that are part of the hcb sheets (Figure 5).
F
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
Figure 10. Depiction of solvent-occupied channels of the 1 × 1 × 1 unit cells of the crystal structures of 1 (a, view along the c-axis), 2 (b, view along
the c-axis), 3 (c, view along the c-axis), 4 (d, view along the a-axis), 5 (e, view along the c-axis), 6 and 7 (f, view along the b-axis), 8 (g, view along the
a-axis), and 9 (h, view along the b-axis). Channels are displayed in purple using the CCDC Mercury computer software, calculated using contact
surface, with a probe radius of 1.2 Å and approximate grid spacing of 0.7 Å.
tcpb derivatives, which has implications for the stacking
distances between hexagonal sheets and the DOC. Though
these structural motifs are essential for the most basic analysis
of tcpb derivatives’ crystal structures, they are inadequate for
rationalizing the subtle differences between the structures’
stacking and catenation motifs. To better understand and
appreciate the complicated structures of tcpb derivatives, crystal
structures of 1−9 were categorized into four distinct stacking/
catenation families: simple stacking, single-layer offset catenation,
Figure 11. (a) Stacking of (6,3) hexagonal sheets in the crystal
double-layer offset catenation, and rotated-layer catenation.
structure of 3; (b) centroid-to-centroid distances (Å) between the
The hexagonal sheets in the crystal structure of 3 exhibit
central arene rings of the three stacking crystallographically
independent molecules (centroids shown as small orange spheres).
Molecules are colored by symmetry equivalence.
simple stacking because the layers are tightly stacked with no
offsets or gaps of sufficient size to allow interweaving with
inclined hexagonal sheets. The structure consists of three
crystallographically independent molecules, as depicted in
Figure 11. The centroid-to-centroid distance between the
central arene rings of molecules depicted in violet and green in
Figure 11b is at the upper limit of what is typically considered
parallel displaced π−π stacking (3.3−3.8 Å).20 Centroid-to-
centroid distances between the green and blue molecules and
between the violet and blue molecules are out of that range.
However, close packing facilitated through edge-to-face CH-π
interactions between the outer arene rings (Figure 11b) leaves
little to no space between hexagonal sheets. Polycatenation,
which requires space between stacked hcb sheets for catenating
layers to interweave, is therefore not possible and not observed
in the crystal structure of 3. Molecules are stacked in one
direction only, with continuous solvent-occupied channels
propagating the entire length of the crystal, but being isolated
from each other. This results in a structure with an unusually
large void volume of ca. 69%, double that of its polycatenated
analogues.
Single-layer offset catenation, the most common catenation
among the structures reported here, also features the simplest
form of catenation. Crystal structures belonging to this family
(2, 4, 6, 7, 9) are made up of a single crystallographically
independent layer of a (6,3) hexagonal or pseudohexagonal
sheet. Parallel layers are slipped along each other, having just
enough offset to create a gap that can accommodate a single
layer of another catenating hexagonal or pseudohexagonal sheet
(Figure 12). The way this catenation motif is realized differs
substantially between structures 2, 4, 6, 7, and 9. All structures
belonging to this catenation family, with the exception of the
isomorphous structures of 6 and 7, feature different space
groups, inclination angles, DOC, as well as distinct π-stacking
G
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
Figure 12. (a) Stacking of (6,3) hexagonal sheets in the crystal structure of 2. Offset stacking leads to a gap (indicated by black arrows) between the
hexagonal sheets; (b) the gap is ideally sized to accommodate a single layer of a catenating hexagonal sheet (shown in blue). A similar topology is
also found in the crystal structures of 4, 6, 7, and 9.
Figure 13. (a) Stacking of (6,3) hexagonal sheets in the crystal structure of 8. Offset stacking leads to a gap (indicated by black arrows) between
every other hexagonal sheet; (b) the gap is ideally sized to accommodate two layers of catenating hexagonal sheets (shown in dark green and dark
blue).
Figure 14. (a−c) Stacking (6,3) hexagonal sheets in the crystal structure of 1. Rotated stacking leads to openings between the hexagonal sheets. (d)
The gap indicated by yellow and red arrows are ideally sized to accommodate four and three layers of catenating hexagonal sheets (shown in light
blue and light green), respectively. A similar topology is also found in the crystal structure of 5, which has 3 + 3 rotated stacking as opposed to 3 + 4.
H
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
distances. Despite these differences, there seems to be a
preference among tcpb derivatives to form crystals with this
catenation motif, pointing toward kinetic and/or thermody-
namic factors favoring this arrangement.
of nearly 200 °C, and differentiates both 1 and 5 drastically
from compounds 2−4 and 6−8, which all undergo substantial
changes in their structures, or lose crystallinity altogether, upon
removal from mother liquor. The combination of catenation
with extensive closest π-stacking and rotation of stacks seems to
give rise to a unique stability to tcpb derivatives not observed
for compounds that are π-stacked but not catenated (such as
3), or compounds that are only catenated, but without
extensive closest π-stacking. Further experiments, including
variable-temperature PXRD and gas adsorption studies, will be
necessary to demonstrate that 5 can retain its original crystal
structure and porosity after complete solvent evacuation. If 5
proves to be permanently porous and similarly stable as 1, it
would be exciting to test for its potential applications in
molecular separations, as there has been growing interests in
developing hydrogen-bonded organic frameworks (HOF) for
separation and catalysis.21−23 These experiments, however, are
outside the scope of the present paper, which focuses on
supramolecular assembly.
The unique rotated-layer catenation observed for 1 and 5
also helps to rationalize the unusually large number of
crystallographically independent molecules observed in the
two polymorphs of 1, with Z′ numbers of 14 for the I2
polymorph, and 56 for the P1 polymorph, the largest number
of crystallographically unique but chemically identical entities
within one crystal lattice reported to date.10,24 In the structure
of 5, there are stacks of three symmetry-independent molecules,
which are π-stacked with another three-molecule thick stack of
the same type. The three molecules on top are related to those
at the bottom by an inversion center in the middle of the six-
molecule stack. Inversion centers are also located between
neighboring stacks, at the center of the carboxylate dimers of
the second molecule of the three-molecule stacks. The overall
number of crystallographically independent molecules in 5 is
thus three, the thickness of half of the six-molecule stack. In the
structure of 1, no symmetry operation is conceivable that would
be able to convert a three-molecule stack into a four molecule
stack, either within one seven-molecule stack or between
molecules in parallel stacks. The smallest symmetry independ-
ent unit in 1 has to consist of two parallel stacks of each seven
molecules as there cannot be any nontranslational symmetry
either within one stack, or between directly neighboring stacks.
Fourteen independent molecules are indeed observed in the
simpler of the two polymorphs, and in the less symmetric case
it is four times that value, 56.
Less prevalent is multilayer offset catenation, with only one
case of double-layer offset catenation observed among the tcpb
derivatives. Multilayer offset catenation is similar to single-layer
offset catenation, but with stacking of hexagonal sheets
alternating between closest stacking and offset stacking such
that after a certain number of tightly stacked layers, there is an
offset ideally sized to accommodate the same number of
catenating hexagonal sheets (Figure 13). The crystal structure
of 8, the only structure belonging to this catenation family,
consists of one unique double-layer of (6,3) hexagonal sheets.
This motif emulates single-layer polycatenation, with a double
layer in 8 taking the role of single layers in 2, 4, 6, 7, and 9.
Both single and double layer offset catenation seem to be
associated with medium-to-large torsion angles between the
central and peripheral arene rings, and all structures with
torsion angles above 40° belong to one of these families (Figure
3).
An entirely different catenation motif is realized in the crystal
structures of 1 and 5, both of which can be categorized as
rotated-layer catenated. In the crystal structure of 1, closest
stacking occurs within stacks of three or four hexagonal sheets
(green and blue molecules, Figure 14a). Within these stacks, 1
stacks the same as 3, leaving no space or gap between layers for
other molecules to interweave. Immediately above a closest-
packed three-molecule stack follows a four-molecule stack of
closest-stacked hexagonal sheets. The four-molecule stack
(blue, Figure 14a) is rotated by ca. 60° with respect to the
three-molecule stack (green, Figure 14a). These seven-
molecule thick stacks fit perfectly within one opening formed
by a catenating hexagonal sheet (Figure 14b). The same kind of
rotated 3 + 4 stacking continues above the blue molecules in
Figure 13b, but the stack above is shifted by one molecule (dark
green molecules, Figure 14c). As a result of this rotated
stacking, undulating channels are created through which
interweaving hexagonal sheets can fit (Figure 14d,e). An
analogous rotated-layer catenation is observed in the structure
of 5, which has 3 + 3 rotated stacking as opposed to 3 + 4 in 1.
The DOC is (6/6) for the 3 + 3 stacks of 5 as opposed to (7/
7) for the 3 + 4 arrangement in 1.
There is a correlation between the extent of π-stacking and
the torsion angles between central and outer arene rings.
Among all investigated structures, 1, 3, and 5 have the smallest
average torsion angles between the central and outer arene
rings, with 33.7°, 36.4°, and 36.4°, respectively (Figure 3).
Small torsion angles appear to be a prerequisite for the
formation of closest, as opposed to offset, π-stacks made up of
three or more molecules as observed in 1, 3, and 5. All other
structures investigated have arene−arene torsion angles above
40° and exhibit single or double layer offset catenation.
The similarities between the structures of 1 and 5 extend
beyond their solid-state arrangements, differentiating them-
selves from the other examples by an increase in framework
stability. Among the examples investigated, 1 and 5 are the only
two structures that survive removal from the mother liquor and
air-drying, as evidenced by the agreement of the powder
diffractogram of 5 measured after filtration and air-drying with
that of its SCXRD structure. These data suggest that the
structure of 5 may have stability similar to that of 1, which was
shown to retain its single crystal structure up to a temperature
CONCLUSIONS
■
A series of 1,3,5-tris(4-carboxyphenyl)arenes with function-
alized central arene rings were synthesized, and their crystal
structures were determined. The two important supramolecular
synthons found in the nonfunctionalized 1,3,5-tris(4-
carboxyphenyl)benzenecarboxylic acid dimers leading to
(6,3) hcb sheets and π-stacking of the (6,3) sheetsare
conserved in all of the structures reported, with the exception
of 6 and 7, which incorporate THF molecules in their
pseudohexagonal sheets. The different structures demonstrate
that functionalization at the central arene ring affects both
molecular conformations and solid-state self-assembly. Func-
tionalization at the central ring leads to steric repulsion between
the functional groups and the hydrogen atoms at the ortho
position of the outer arene rings, leading to larger torsion
angles between the central and outer arene rings. Molecules
with larger torsion angles were generally found to have larger
I
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
by direct methods using the SHELXTL suite of programs26,27 and
stacking distances between hexagonal sheets, leading to lower
DOC, and to single or double layer offset catenation. Small
torsion angles, below 40°, allow for extensive closest π-stacking.
In the case of compound 3 this led to infinite closest stacking
with complete elimination of catenation, resulting in open
hexagonal channels with 69% of solvent-occupied volume in its
structure. The other structure type found for molecules with
small torsion angles and extensive π-stacking is rotated-layer
catenation, as observed for structures of 1 and 5. This latter
solid-state arrangement is the only one among tcpb derivatives
yet observed that survives removal from mother liquor and air
drying without collapse of the hcb framework, making these
two rotated-layer catenation structures uniquely stable among
this family of closely related structure types.
Although the carboxylic acid dimer synthon did not itself
provide rationalization of the global topologies of the crystal
structures reported, it reliably predicted the 2-D (6,3)
hexagonal sheet assembly, even though pseudohexagonal sheets
found in 6 and 7 were interrupted by solvent molecules. This
prediction simplified the analysis of the crystal structures and
focused the analysis on the stacking and catenation of (6,3)
hexagonal or pseudohexagonal sheets. The tcpb derivative
crystal structures reported demonstrate the significant crystal
structural changes that are possible with subtle molecular
modifications, but more examples are necessary to fully
understand the relationship between solid-state assembly and
molecular structure. As evident from crystal structures of 6 and
7, which incorporated solvent molecules into their frameworks,
crystallization conditions can affect the resulting crystal
structures. Future work on this system should therefore not
only focus on obtaining structures of new tcpb derivatives, but
also consider the possibility of polymorphism and/or different
solvates for a given molecule.
refined by full matrix least-squares against F2 with all reflections using
Shelxl2013 or Shelxl201428,29 and the graphical interface Shelxle.30
H
atoms attached to carbon, nitrogen, and oxygen atoms were positioned
geometrically and constrained to ride on their parent atoms, with
carbon hydrogen bond distances of 0.95 Å for aromatic C−H, 0.99 and
0.98 Å for aliphatic CH2 and CH3, 0.88 Å for NH2 moieties and 0.84 Å
for carboxylic acid moieties, respectively. Methyl and hydroxyl H
atoms were allowed to rotate but not to tip to best fit the experimental
electron density. Uiso(H) values were set to a multiple of Ueq(C/O/N)
with 1.5 for CH3 and OH, and 1.2 for C−H, CH2, and N−H units,
respectively. SCXRD details are reported in Table 3 (compounds 2−
8) or in Table S1 of the SI (homo coupling product dimethyl
biphenyl-4,4′-dicarboxylate and synthetic intermediates 2,4,6-tris(4-
carboxyphenyl)anisole trimethyl ester, 2,4,6-tris(4-carboxyphenyl)-
aniline trimethyl ester and 2,4,6-tritolylnitrobenzene). Additional
details for each structure are given in the SI, including disorder,
pseudosymmetry handling, restraints and constraints used, and details
on solvent-filled void space and numbers of electrons corrected for.
Complete crystallographic data, in CIF format, have been deposited
with the Cambridge Crystallographic Data Centre. CCDC 1422491−
1422501 contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/
cif.
Synthesis of 2,4,6-tris(4-methylphenyl)pyrylium Tetrafluor-
oborate. The procedure described by Kotra and co-workers31 was
modified as follows: A 100 mL 3-neck round-bottom flask, equipped
with two septa and a condenser topped with a gas inlet adapter, was
purged with nitrogen. Twenty milliliters of anhydrous toluene was
added using a gastight syringe, followed by the addition of 4-
methylacetophenone (6.5 mL, 0.049 mol) and 4-methylbenzaldehyde
(3.0 mL, 0.025 mol). The reaction mixture was heated to 80 °C. Boron
trifluoride diethyl etherate (6.5 mL) was then added via syringe, after
which the color of the reaction mixture turned from yellow to deep
red. The reaction mixture was then heated to reflux for 2 h under
nitrogen. After the mixture was allowed to cool to room temperature,
10 mL of acetone were added, followed by 250 mL of diethyl ether.
The resulting precipitate was collected by suction filtration and dried
over a glass frit. Yield 3.73 g (8.5 mmol, 35%); yellow solid: 1H NMR
(400 MHz, DMSO-d6) δ 9.03 (s, 2H), 8.54 (d, 2H, J = 8.4 Hz), 8.48
(d, 4H, J = 8.4 Hz), 7.62−7.60 (m, 6H), 3.343 (s, 9H); 13C NMR
(100 MHz, DMSO-d6) δ 169.88, 164.41, 147.07, 146.49, 130.91,
130.45, 130.07, 129.07, 126.96, 113.88, 21.91, 21.86. APCI-MS+ m/z
EXPERIMENTAL SECTION
■
All reagents were purchased from Sigma-Aldrich and used without
further purification, with the following exceptions: 4-(methylcarboxy)-
phenylboronic acid (OxChem), 2,4,6-tribromotoluene and tripotas-
sium phosphate (Acros Organics), 2,4,6-tribromoaniline (TCI
America), 2,4,6-tribromo-3-methylaniline (Alfa Aesar), concentrated
HCl and potassium carbonate (Fisher Scientific), ethyl acetate,
ethanol, and methanol (Pharmco-Aaper). A Varian 400 MHz NMR
spectrometer was used. All NMR spectra were taken at room
temperature with spinning at 20 Hz. Shimming and locking were
performed using the Varian VnmrJ software. An Agilent 1100 Series
LC/MSD system was used for atmospheric pressure chemical
ionization mass spectrometry. Samples were dissolved in HPLC
grade methanol (Pharmco-Aaper) and carried into the mass
spectrometer by a mobile phase consisting of pure methanol at a
flow rate of 0.500 mL/min. High-resolution mass spectrometry for
crystals of 2-6 and 8 were performed at the University of Akron Mass
Spectrometry Facility. A Rigaku Ultima IV diffractometer (λ = 1.5418
Å) with Bragg−Brentano geometry was used for room temperature
PXRD experiments. A scan speed of 4 deg/min was used.
−
Calcd for C26H23O+ [M − BF4 ] 351.17, found 351.2.
Synthesis of 2,4,6-tris(4-methylphenyl)nitrobenzene. The
procedure described by Dimroth and co-workers32 for the synthesis
of 2,4,6-triphenylnitrobenzene was modified as follows: In a 100 mL
round-bottom flask, 2,4,6-tris(4-methylphenyl)pyrylium tetrafluorobo-
rate (3.39 g, 7.7 mmol), nitromethane (0.6 mL, 11.2 mmol), and 34
mL of absolute ethanol were stirred at room temperature and purged
with nitrogen. Triethylamine (1.8 mL, 12.9 mmol) was added rapidly
using a gastight syringe. The reaction mixture was then heated at reflux
for 3 h, during which time the mixture turned from a yellow slurry to a
red/brown solution. The reaction mixture was rotary evaporated to
reduce the solvent volume by half and the residue was allowed to cool
in a refrigerator overnight. The resulting precipitate was collected by
suction filtration and recrystallized from acetic acid. Yield 1.35 g (3.4
1
Single Crystal X-ray Diffraction. Instruments used include a
Bruker SMART APEX CCD diffractometer with Mo K-α radiation, a
Bruker AXS Prospector CCD diffractometer with Cu K-α radiation,
and a Bruker AXS D8 Quest CMOS diffractometer with Mo K-α
radiation. Data for the crystal structure of 5 were collected at the
Advanced Photon Source at Argonne National Lab using source
Beamline 15-ID B ChemMatCARS and a Bruker AXS APEXII CCD
diffractometer as the measurement device. The wavelength of the
synchrotron radiation was 0.41328 Å. The Apex225 suite of programs
was used for collection of diffraction data, unit cell determinations,
data integration and correction for absorption and other systematic
errors. The space groups were assigned and the structures were solved
mmol, 4%): light yellow needles: H NMR (400 MHz, DMSO-d6) δ
7.74 (d, 2H, J = 8.0 Hz), 7.71 (s, 2H), 7.35 (d, 4H, J = 8.0 Hz), 7.32−
7.28 (m, 6H); 13C NMR (100 MHz, DMSO-d6) δ 148.24, 142.83,
138.68, 138.64, 135.55, 134.74, 133.65, 130.11, 129.87, 128.34, 128.26,
127.71, 21.19, 21.16. SCXRD data are given in the SI.
Synthesis and Crystal Growth of 2,4,6-Tris(4-
carboxypheny)nitrobenzene (8). 2,4,6-Tris(4-methylphenyl)-
nitrobenzene (1.0 g, 2.5 mmol) was dissolved in 15 mL of pyridine
in a 100 mL round-bottom flask, and 15 mL of deionized (DI) water
and potassium permanganate (2.04 g, 12.9 mmol) were added. The
resulting purple slurry was then heated to reflux. When the color of the
reaction mixture turned completely from purple to brown/black, more
J
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
potassium permanganate (2.95 g, 18.7 mmol) was added directly to
the reaction mixture. Heating at reflux was then continued until the
reaction mixture once again turned from purple to brown/black. The
reaction mixture was then filtered to remove manganese(IV) oxide
(deep brown solid). More potassium permanganate (3.1 g, 19.6
mmol) was then added to the filtrate, and the resulting purple solution
was heated at reflux. When the reaction mixture turned from purple to
brown/black, more potassium permanganate was added (2.1 g, 13.2
mmol). The mixture was then heated at reflux until all of the purple
color disappeared. Manganese(IV) oxide was removed from the
mixture by suction filtration, and the filtrate was washed with ethyl
acetate (3 × 20 mL) and acidified with 1 M HCl. The resulting
precipitate was collected by suction filtration and dried over a glass frit.
Yield 778 mg (1.6 mmol, 64.3%): light yellow powder: 1H NMR (400
MHz, DMSO-d6) 13.14 (s, 3H), 8.04−8.02 (m, 8H), 7.97 (s, 2H),
7.60 (d, 4H, J = 8.4 Hz); 13C NMR (100 MHz, DMSO-d6) 167.39,
167.25, 148.38, 142.21, 142.04 140.41, 134.34, 131.57, 131.29, 130.36,
130.21, 129.77, 128.81, 128.27; APCI-MS− m/z Calcd for
J = 8.0 Hz), 7.61 (m, 6H), 2.01 (s, 3H). 13C NMR (100 MHz, DMSO-
d6) 167.58, 167.54, 146.13, 143.85, 142.82, 137.04, 132.89, 130.38,
130.11, 130.05, 129.73, 127.93, 127.33, 18.85. APCI-MS+ m/z Calcd
for C28H19O6 [M − H+] 451.1182, found 451.1204. H NMR data
are consistent with those previously reported in the patent literature.35
Single crystals for SCXRD analysis were grown by slow cooling.
Crude product (400 mg) was mixed with 40 mL of 1:1 (v:v)
tetrahydrofuran/water and sonicated in a 50 mL media bottle. The
resulting slurry was tightly capped and heated at 70 °C in an oven,
during which the media bottle was shaken every 15 min. After 2 h of
heating, the mixture phase-separated. The media bottle was then
removed from the oven and allowed to cool to room temperature
overnight. The two phases were then combined by gently swirling the
mixture. The resulting solution was refrigerated, and needle-shaped
crystals grew after 2 days.
−
1
2,4,6-Tris(4-carboxyphenyl)anisole (3). White solid. Yield: 270
1
mg (0.58 mmol, 62% yield): H NMR (400 MHz, DMSO-d6) 12.99
(s, 3H), 8.04 (d, 4H, J = 8.0 Hz), 8.00 (d, 2H, J = 8.0 Hz), 7.92 (d, 2H,
J = 8.0 Hz), 7.80 (d, 4H, J = 8.0 Hz), 7.75 (s, 2H), 3.12 (s, 3H). 13C
NMR (100 MHz, DMSO-d6) 167.60, 167.56, 155.18, 143.67, 142.55,
135.90, 135.33, 130.34, 130.21, 130.08, 129.83, 129.77, 127.45 61.05.
−
C27H16NO8 [M − H+] 482.0876, found 482.0829.
Single crystals for SC-XRD analysis were grown by vapor diffusion
experiments on the benchtop at 20 °C. The crude product (33.3 mg)
was dissolved in 1 mL of 1,4-dioxane. The solution was filtered by
syringe filtration, followed by vapor diffusion with propionitrile. Plates
suitable for SCXRD developed within 3 days.
−
APCI-MS+ m/z Calcd for C28H19O6 [M − H+] 467.1131, found
467.1085.
Single crystals for SCXRD analysis were grown by vapor diffusion
with the solvent/antisolvent pair 1,4-dioxane/acetonitrile: the crude
product (150 mg) was mixed with 2.8 mL of 1,4-dioxane and
sonicated. The mixture was then filtered by syringe filtration, separated
into 2 aliquots, followed by vapor diffusion with acetonitrile. Block-
shaped crystals grew in 2 days.
Synthesis of 1,3-Diamino-2,4,6-tribromobenzene. In a 100
mL round-bottom flask, a mixture of m-phenylenediamine (1.78 g,
16.5 mmol) and 50 mL of acetic acid was stirred at room temperature
until all solids were completely dissolved. Stirring was then continued
with the round-bottom flask immersed in an ice water bath. Bromine
(3 mL, 60 mmol) was added dropwise using a Pasteur pipet. After the
bromine was added, the reaction mixture was removed from the ice
water bath and allowed to stir at room temperature for 2 h. The black
mixture was suction filtered, and the black solid was allowed to air-dry
over a glass frit. The black solid was then extracted with chloroform,
filtered, and the filtrate was completely evaporated to leave behind a
2,4,6-Tris(4-carboxyphenyl)-3-methylanisole (4). White solid,
contains 25 mol % pyridine. Yield: 334 mg (0.70 mmol, 75% yield):
1H NMR (400 MHz, DMSO-d6) 12.97 (s, 3H), 8.03 (d, 2H, J = 8.0
Hz), 7.99 (m, 4H), 7.73 (d, 2H, J = 8.0 Hz), 7.55 (d, 2H, J = 8.0 Hz),
7.48 (d, 2H, J = 8.0 Hz), 7.31 (s, 1H), 3.02 (s, 3H), 1.91 (s, 3H). 13C
NMR (100 MHz, DMSO-d6) 167.64, 167.58, 167.56, 154.55, 145.69,
142.59, 142.44, 137.68, 136.63, 135.00, 131.63, 131.37, 130.51, 130.07,
130.07, 129.96, 129.88, 129.86, 129.70, 129.60, 129.43, 60.76, 19.08.
1
brown solid. Yield: 1.8 g (5.2 mmol, 31.5%): H NMR (400 MHz,
CDCl3) 7.44 (s, 1H), 4.51 (s, 4H). 13C NMR (100 MHz, CDCl3)
141.65, 133.23, 96.10, 95.28. NMR data are consistent with those
previously reported.33
−
APCI-MS+ m/z Calcd for C29H21O7 [M − H+] 481.1287, found
481.1317.
General Procedure for Suzuki Coupling and Hydrolysis. In a
nitrogen glovebox, substituted tribromobenzene (0.93 mmol), 4-
(methylcarboxy)phenylboronic acid (500 mg, 2.8 mmol), tetrakis-
(triphenylphosphine)palladium(0) (40 mg, 0.043 mmol), and
tripotassium phosphate (620 mg, 2.9 mmol) were mixed with 3 mL
of commercial anhydrous methanol and 10 mL of commercial
anhydrous tetrahydrofuran in a 20 mL scintillation vial charged with a
stir bar. The vial was tightly capped, removed from the glovebox, and
stirred at 90 °C in an aluminum heating block for 3 days. After the
reaction mixture cooled to room temperature, the solvent was
completely evaporated under a stream of air. The homocoupling
side product, dimethyl biphenyl-4,4′-dicarboxylate, was observed in
varying amounts. The identity of this side product was confirmed by
SCXRD. Structure details are given in the SI and agree with data
reported previously.34 The residue was extracted with dichloro-
methane and filtered through a glass frit. The solvent was completely
evaporated from the filtrate to obtain the trimethyl esters (SCXRD
details for the trimethyl esters of 2,4,6-tris(4-carboxyphenyl)anisole
and 2,4,6-tris(4-carboxyphenyl)aniline are given in the SI). The crude
trimethyl esters were dissolved in 15 mL of pyridine. Fifteen milliliters
of 2.5 M aqueous NaOH were added to the pyridine solution, and the
resulting biphasic mixture was heating at reflux for 20 min. The
mixture was allowed to cool to room temperature, and enough 1 M
aqueous HCl was added such that the two phases combined. The
solution was then washed twice with 15 mL of ethyl acetate.
Concentrated HCl was then added to the aqueous layer. The resulting
precipitate was collected by suction filtration and allowed to air-dry
over a glass frit.
Single crystals for SCXRD analysis were grown by vapor diffusion
with the solvent/antisolvent pair 1,4-dioxane/acetonitrile: the crude
product (65 mg) was mixed with 3.0 mL of 1,4-dioxane and sonicated.
The mixture was then filtered by syringe filtration, separated into three
aliquots, followed by vapor diffusion with acetonitrile. Block-shaped
crystals grew in 1 week.
2,4,6-Tris(4-carboxyphenyl)aniline (5). Hydrolysis of the
methyl esters was performed under nitrogen atmosphere. Green
solid, contains 23 mol % ethyl acetate. Yield: 338 mg (0.87 mmol, 94%
1
yield): H NMR (400 MHz, DMSO-d6) 12.93 (s, 3H), 8.05 (d, 4H, J
= 8.0 Hz), 7.93 (d, 2H, J = 8.0 Hz), 7.79 (d, 2H, J = 8.0 Hz), 7.68 (d,
4H, J = 8.0 Hz), 7.47 (s, 2H), 4.65 (s). 13C NMR (100 MHz, DMSO-
d6) 167.66, 167.58, 144.38, 144.09, 142.33, 130.46, 130.36, 130.34,
130.03, 129.81, 128.78, 128.44, 127.23, 126.13. APCI-MS+ m/z Calcd
for C27H18NO6− [M − H+] 452.1134, found 452.1091. NMR data are
consistent with those previously reported.36
Single crystals for SCXRD analysis were grown by vapor diffusion
with the solvent/antisolvent pair tetrahydrofuran/acetonitrile: the
crude product was mixed with tetrahydrofuran and sonicated. The
mixture was then filtered by syringe filtration, followed by vapor
diffusion with acetonitrile. Plate-shaped crystals grew in a few weeks.
2,4,6-Tris(4-carboxyphenyl)-3-methylaniline (6). Hydrolysis of
the methyl ester was performed under nitrogen atmosphere. Yellow
solid, contains 5 mol % biphenyl-4,4′dicarboxylic acid (homocoupling
product). Yield: 424 mg (0.90 mmol, 98% yield): 1H NMR (400 MHz,
DMSO-d6) 12.94 (s, 3H), 8.08 (d, 2H, J = 8.0 Hz), 7.99 (d, 2H, J = 8.0
Hz), 7.94 (d, 2H, J = 8.0 Hz), 7.62 (d, 2H, J = 8.0 Hz), 7.48 (d, 2H, J
= 8.0 Hz), 7.44 (d, 2H, J = 8.0 Hz), 7.00 (s, 1H), 1.83 (s, 3H). 13C
NMR (100 MHz, DMSO-d6) 167.65, 167.60, 167.55, 146.56, 144.08,
143.48, 133.42, 131.18, 130.81, 130.77, 130.46, 130.36, 130.29, 130.07,
129.63, 129.59, 129.57, 129.03, 127.87, 127.61, 124.09, 19.20. APCI-
2,4,6-Tris(4-carboxyphenyl)toluene (2). White solid. Yield: 400
1
mg (0.87 mmol, 93% yield): H NMR (400 MHz, DMSO-d6) 12.99
(s, 3H), 8.03 (d, 4H, J = 8.0 Hz), 7.98 (d, 2H, J = 8.0 Hz), 7.88 (d, 2H,
K
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
−
MS+ m/z Calcd for C27H19N2O6 [M − H+] 466.1291, found
Crystallography at the Advanced Photon Source, SCrAPS,
program supported by Indiana University. The authors thank
Drs. Josh Chen and Maren Pink for data collection.
ChemMatCARS Sector 15 is principally supported by the
Divisions of Chemistry (CHE) and Materials Research (DMR),
National Science Foundation, under grant number NSF/CHE-
1346572. Use of the Advanced Photon Source, an Office of
Science User Facility operated for the U.S. Department of
Energy (DOE) Office of Science by Argonne National
Laboratory, was supported by the U.S. DOE under Contract
No. DE-AC02-06CH11357. The authors thank Professor
Davide M. Proserpio for help with using the TOPOS program.
The authors also thank Selim Gerislioglu and Chrys
Wesdemiotis at the University of Akron Mass Spectrometry
Facility for collecting high-resolution mass spectrometry data.
466.1327.
Single crystals for SCXRD analysis were grown by vapor diffusion
with the solvent/antisolvent pair tetrahydrofuran/acetonitrile: the
crude product was mixed with tetrahydrofuran and sonicated. The
mixture was then filtered by syringe filtration, followed by vapor
diffusion with acetonitrile. Needle-shaped crystals grew in a few weeks.
2,4,6-Tris(4-carboxyphenyl)-1,3-diaminobenzene (7). Hy-
drolysis of the methyl ester was performed under nitrogen atmosphere.
Brown solid, contains 10 mol % pyridine and 14 mol % biphenyl-4−
4′dicarboxylic acid (homocoupling product). Yield: 338 mg (0.72
1
mmol, 78% yield): H NMR (400 MHz, DMSO-d6) 12.89 (s, 3H),
8.10 (d, 2H, J = 8.0 Hz), 7.97 (d, 4H, J = 8.0 Hz), 7.64 (d, 4H, J = 8.0
Hz), 7.57 (d, 2H, J = 8.0 Hz), 7.01 (s, 1H). 13C NMR (100 MHz,
DMSO-d6) 167.61, 167.58, 146.85, 143.66, 143.10, 139.38, 132.39,
131.80, 131.12, 131.09, 130.22, 129.60, 129.45, 127.78. APCI-MS+ m/
z Calcd for C27H19N2O6 [M − H+] 467.12, found 467.2.
−
Single crystals for SCXRD analysis were grown by vapor diffusion
with the solvent/antisolvent pair tetrahydrofuran/acetonitrile: the
crude product was mixed with tetrahydrofuran and sonicated. The
mixture was then filtered by syringe filtration, followed by vapor
diffusion with acetonitrile. Needle-shaped crystals grew in a few days.
REFERENCES
■
(1) Desiraju, G. R.; Gavezzotti, A. Acta Crystallogr., Sect. B: Struct. Sci.
1989, 45, 473−482.
(2) Price, S. L. Acc. Chem. Res. 2009, 42, 117−126.
(3) Desiraju, G. R. J. Am. Chem. Soc. 2013, 135, 9952−9967.
(4) Etter, M. C. Acc. Chem. Res. 1990, 23, 120−126.
(5) Sanphui, P.; Bolla, G.; Das, U.; Mukherjee, A. K.; Nangia, A.
CrystEngComm 2013, 15, 34−38.
(6) Silly, F. J. Phys. Chem. C 2012, 116, 10029−10032.
(7) Ruben, M.; Payer, D.; Landa, A.; Comisso, A.; Gattinoni, C.; Lin,
N.; Collin, J.-P.; Sauvage, J.-P.; De Vita, A.; Kern, K. J. Am. Chem. Soc.
2006, 128, 15644−15651.
(8) Kampschulte, L.; Lackinger, M.; Maier, A.-K.; Kishore, R. S. K.;
Griessl, S.; Schmittel, M.; Heckl, W. M. J. Phys. Chem. B 2006, 110,
10829−10836.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.cgd.5b01416.
■
S
Details of single crystal structure determinations, Ortep
style plots for all new structures reported, and
experimental details for structures of side products and
intermediates. Selected diffraction images, comparisons
of measured PXRD diffractograms and diffractograms
simulated from SCXRD experiments, TGA/DSC traces,
1H and 13C NMR spectra, tables of torsion angles, as well
as CIFs, are available free of charge (PDF)
(9) Gutzler, R.; Sirtl, T.; Dienstmaier, J. F.; Mahata, K.; Heckl, W. M.;
Schmittel, M.; Lackinger, M. J. Am. Chem. Soc. 2010, 132, 5084−5090.
(10) Zentner, C. A.; Lai, H. W. H; Greenfield, J. T.; Wiscons, R. A.;
Zeller, M.; Campana, C. F.; Talu, O.; FitzGerald, S. A.; Rowsell, J. L.
C. Chem. Commun. 2015, 51, 11642−11645.
Accession Codes
CCDC 1422491−1422501 contains the supplementary crys-
tallographic data for this paper. These data can be obtained free
of charge via www.ccdc.cam.ac.uk/data_request/cif, or by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
Cambridge Crystallographic Data Centre, 12, Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
(11) Bajpai, A.; Venugopalan, P.; Moorthy, J. N. Cryst. Growth Des.
2013, 13, 4721−4729.
(12) Carlucci, L.; Ciani, G.; Proserpio, D. M.; Mitina, T. G.; Blatov,
V. A. Chem. Rev. 2014, 114, 7557−7580.
(13) Carlucci, L.; Ciani, G.; Proserpio, D. M. Coord. Chem. Rev. 2003,
246, 247−289.
(14) Sobolev, V. I.; Radchenko, V. B.; Ostvald, R. V.; Filimonov, V.
D.; Zherin, I. I. Procedia Chem. 2014, 11, 88−91.
(15) van der Sluis, P.; Spek, A. L. Acta Crystallogr., Sect. A: Found.
Crystallogr. 1990, 46, 194−201.
AUTHOR INFORMATION
Corresponding Author
*E-mail: hwhlai@stanford.edu.
■
(16) Tian, J.; Thallapally, P. K.; McGrail, B. P. CrystEngComm 2012,
14, 1909−1919.
Present Address
#Department of Chemistry, Stanford University, Stanford, CA
94305, United States.
(17) Winstein, S.; Holness, N. J. J. Am. Chem. Soc. 1955, 77, 5562−
5578.
(18) Bondi, A. J. Phys. Chem. 1964, 68, 441−451.
(19) Blatov, V. A.; Shevchenko, A. P.; Proserpio, D. M. Cryst. Growth
Des. 2014, 14, 3576−3586.
(20) Janiak, C. J. Chem. Soc., Dalton Trans. 2000, 3885−3896.
(21) Li, P.; He, Y.; Guang, J.; Weng, L.; Zhao, J.C.-G.; Xiang, S.;
Chen, B. J. Am. Chem. Soc. 2014, 136, 547−549.
(22) Wang, H.; Li, B.; Wu, H.; Hu, T.-L.; Yao, Z.; Zhou, W.; Xiang,
S.; Chen, B. J. Am. Chem. Soc. 2015, 137, 9963−9970.
(23) Yang, W.; Li, B.; Wang, H.; Alduhaish, O.; Alfooty, K.; Zayed,
M. A.; Li, P.; Arman, H. D.; Chen, B. Cryst. Growth Des. 2015, 15,
2000−2004.
Notes
The authors declare no competing financial interest.
§J.L.C.R., who passed away on January 30, 2015, was the
principal investigator who designed this research. The authors
are grateful for his unsurpassed mentorship and friendship.
ACKNOWLEDGMENTS
■
The X-ray diffractometers at Youngstown State University were
funded by NSF Grants DMR-1337296, CHE-0087210, by Ohio
Board of Regents Grant CAP-491, and by Youngstown State
University. The powder X-ray diffractometer at Oberlin College
was purchased with support of NSF grant DMR-0922588. Data
for 5 were collected at Beamline 15-ID B ChemMatCARS at
the Advanced Photon Source through the Synchrotron
(24) Steed, K. M.; Steed, J. W. Chem. Rev. 2015, 115, 2895−2933.
(25) Apex2, Saint, Sadabs, Bruker Advanced X-ray Solutions; Bruker
AXS Inc.; Madison, Wisconsin, USA, 2014.
(26) SHELXTL, Bruker Advanced X-ray Solutions, Version 6.14;
Bruker AXS Inc.: Madison, Wisconsin: USA, 2000−2003.
L
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design
Article
(27) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
2008, 64, 112−122.
(28) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Adv. 2015, 71,
3−8.
(29) Sheldrick, G. M. University of Gottingen, Germany, 2013/2014.
̈
(30) Hubschle, C. B.; Sheldrick, G. M.; Dittrich, B. J. Appl.
Crystallogr. 2011, 44, 1281−1284.
̈
(31) Bello, A. M.; Kotra, L. P. Tetrahedron Lett. 2003, 44, 9271−
9274.
(32) Dimroth, K.; Reichardt, C.; Vogel, K. Org. Synth. 1973, 5, 1135−
1137.
(33) Giannicchi, I.; Jouvelet, B.; Isare, B.; Linares, M.; Dalla Cort, A.
D.; Bouteiller, L. Chem. Commun. 2014, 50, 611−613.
(34) Ritzerfeld, V.; Pyrlik, A.; Wang, Y.; Englert, U. Acta Crystallogr.,
Sect. E: Struct. Rep. Online 2009, 65, o1677.
(35) Zhou, H.-C.; Dawei, F.; Kecheng, W. Eur. Pat. Appl. 2015, EP
2876112A1 20150527.
(36) Dutta, A.; Koh, K.; Wong-Foy, A. G.; Matzger, A. J. Angew.
Chem., Int. Ed. 2015, 54, 3983−3987.
M
DOI: 10.1021/acs.cgd.5b01416
Cryst. Growth Des. XXXX, XXX, XXX−XXX