Key words
organic cage compounds - imines - pyrrols - dynamic covalent chemistry
Introduction
Organic cage compounds have gained a lot of interest in the past decades[1]–[5] due to emerging applications in materials science. They found their use in sensor
films,[6] as coatings for quartz crystal microbalances,[7],[8] for selective ion recognition,[9] to make porous membranes,[10],[11] for selective gas sorption[12]–[15] and separation,[16]–[20] catalysis,[21] or as stationary phases for gas chromatography.[22],[23] To create cages for specific applications it is important to control parameters
such as size,[24] shape,[25] geometry[3],[4] as well as number and positions of functional groups.[9],[26] Most cage compounds are synthesized by dynamic covalent chemistry (DCC).[27] Among DCC reactions, the reactions forming boronic acid esters or imines are the
most common ones and for the latter a huge variety of geometries such as (truncated)[28],[29] tetrahedra,[30]–[33] trigonal prisms,[34]–[36] cubes,[16],[37],[38] rhombicuboctahedra,[39]–[41] adamantoids,[35],[42],[43] tetrapods,[44] cucurbitimines[45] or more complex structures such as catenanes[46],[47] have been realized and some of these structures have even been post-stabilized to
enhance their chemical robustness.[6],[48]–[50] Cages with nitrogen-containing pyridine[9],[51] or pyrrol-units[52] showed superior behavior in, e.g., ion recognition and could act as ligands for
several transition metal ions as shown for numerous metal–organic cage compounds or
complexes.[53],[54] To extend the portfolio of organic cage compounds, we envisioned the synthesis of
tetrahedral [4 + 4] imine cages, which are based on a trispyrrol aldehyde, a unit used by Beer and co-workers for flexible [2 + 3] cages,[55],[56] yet this time with substituted 1,3,5-trimethylamino benzenes[9],[28] to retain a higher degree of shape persistency.
Results and Discussion
The desired cage compounds were synthesized by condensation reactions of methyl-trispyrrolyl-aldehyde 1
[56] with equimolar amounts of the differently substituted triamines 2-Me, 2-Et, 2-OPr and 2-Br in chloroform at room temperature or 60 °C ([Scheme 1]). Due to their solubility in common organic solvents, cage compounds 3-Me, 3-Et and 3-OPr were successfully isolated by recycling gel permeation chromatography (rec-GPC) in
THF after 6 (3-Me and 3-Et) or 15 (3-OPr) recycling cycles in yields between 10 and 18% (for chromatograms, see the Supporting
Information, SI).
Scheme 1 Synthesis of the cage compounds. a) CHCl3, TFA (3 mol-%), rt or 60 °C, 2 d.
Due to its low solubility, the purification of 3-Br by rec-GPC was not possible and attempts to remove impurities by recrystallization
failed, so that minor impurities remained. The successful synthesis of the cages was
verified by 1H NMR spectroscopy by the disappearance of the signals of the aldehyde units of 1 (δ = 9.13 ppm) as well as of the amine units of 2-Me, 2-Et, 2-OPr and 2-Br (δ = 1.30 – 1.63 ppm). Instead, signals of the imine protons of cages 3-Me, 3-Et, 3-OPr and 3-Br at δ = 8.15 – 8.24 ppm were found (labeled as protons 7 in [Figure 1a – d]). Furthermore, MALDI-TOF mass spectrometry shows signals at m/z 1850.0210 (3-Me; calc: 1850.0201), m/z 2018.2058 (3-Et; calc: 2018.2079), m/z 2378.3341 (3-OPr; calc: 2378.3346) and m/z 2618.7553 (3-Br; 2618.7568) representing the corresponding [4 + 4] condensates ([Figure 1e – h]).
Figure 1 a – d) 1H NMR spectra and of cage 3-Me (a) (600 MHz), 3-Et (b) (600 MHz), 3-OPr (c) (700 MHz) and 3-Br (d) (700 MHz) in CDCl3. # = BHT, * = THF, $ = pentane, & = methanol. e – h) MALDI-TOF mass spectra (matrix:
DCTB) of 3-Me (a), 3-Et (b), 3-OPr (c) and 3-Br (d). Simulated (top) and experimental (bottom) isotopic patterns are shown as insets.
The low isolated yields of cages 3-Me, 3-Et and 3-OPr after rec-GPC purification can be rationalized by the formation of several other
species as obvious by multiple peaks in the GPC-chromatograms (see the SI). MALDI-TOF
MS analyses of the isolated fractions showed the 4 + 4 condensates as main components
alongside minor impurities for each fraction. Nevertheless, 1H NMR analyses revealed more complex patterns leading to the hypothesis of isomeric
cage structures. To get further insight into possible cage isomers, we conducted quantum
chemical calculations at the B3LYP/6 – 31 G(d) level of theory and calculated the
interconversion energy of a pro-chiral trispyrrol unit to elucidate whether these units lead to a potential mixture of diastereoisomeric
cages ([Figure 2]). Even considering the highest energy barrier between intermediate II and transition
state III of 23.7 kJ/mol would result in a rapid interconversion rate of 4.35·108 · s−1 and a half-life time of 1.59 ns, suggesting a fast interconversion in solution. This
is further underlined by the comparably low relative energies of MMMM, MMMP, MMPP, MPPP and PPPP cages of 6.7 to 11.1 kJ/mol independent of the side chains (see SI), indicating again
a fast interconversion excluding diastereomeric cages causing the several fractions
obtained by rec-GPC. Another possibility is the formation of in–out-isomers referring to the central methyl-units of the trispyrrols. The relative energies of the possible in–out-combinations show that the all-in isomers should represent > 98.5% of the cages in a dynamic equilibrium in solution.
Thus, it can be concluded that the different fractions obtained by rec-GPC contain
different in–out isomers of the cages that are kinetically trapped.
Figure 2 DFT-calculated (B3LYP/6 – 31 G(d)) interconversion of P- to M-trispyrrol.
For all isolated all-in cages, suitable crystals for single-crystal X-ray diffraction (SCXRD) analyses have
been obtained, unambiguously proving the molecular structures of the [4 + 4] cages
([Figure 3]). The geometrical shapes can best be described as truncated cubes with tetrahedral
symmetry. In all studied cases, the central methyl group of the trispyrrols is pointing into the cage cavities resulting in distances to adjacent phenyl
rings of d
Me-Ph = 1.2 – 1.3 nm and outer diameters of d = 1.5 – 1.6 nm. As previously observed, for 3-Me all ethyl substituents are conformationally stable and point outwards, whereas propoxy
groups in 3-OPr point either out- or inwards.
Figure 3 Single-crystal X-ray structures of the cages 3-Me, 3-Et, 3-OPr and 3-Br as capped-stick models. The distances of a phenyl unit to the opposing internal methyl
group and to the outer carbon atoms of the trispyrrol moiety are given for 3-Me exemplarily. Colors: carbon, gray; nitrogen, blue; oxygen, red; bromine, brown; hydrogen,
white.
Besides the solvate 3-Etα
crystallized from dichloromethane and hexane (cubic, P
a3̄) and cage 3-OPr (orthorhombic, Aba
2), all obtained structures crystallized in the triclinic space group P1̄. Due to steric repulsion, each trispyrrol unit adopts a chiral C
3-symmetrical conformation (for the sake of simplicity, we name them here M or P orientation), which can interconvert quickly by rotation around the C–C single bonds
to the central carbon atom. Thus, each trispyrrol unit is prochiral leading to overall five possible cage isomers due to their
tetrahedral symmetry. The solid-state structures of 3-Me, 3-Etβ, 3-OPr and 3-Br adopt single enantiomeric MMMM or PPPP orientation, an effect comparable to the assembly of face-orientated polyhedra derived
from prochiral building blocks as reported by Caoʼs group ([Table 1] and [Figure 5]).[57],[58] Since four out of eight “faces” of the synthesized truncated cubes possess helical
chirality, the cages can be understood as semi-face-orientated polyhedra.
Table 1 Selected crystallographic data of cages 3-Me, 3-Et, 3-OPr and 3-Br.
|
Compound
|
Solvents
|
Crystal system
|
Space group
|
Z
|
Cell volume
|
Isomers observed
|
Packing
|
|
3-Me
|
CDCl3/MeOH
|
Triclinic
|
P1̄
|
2
|
7578.4 Å3
|
MMMM/PPPP
|
Enantiopure layers
|
|
3-Etα
|
CH2Cl2/C6H14
|
Cubic
|
P
a3̄
|
8
|
31 222.2 Å3
|
MMMP/PPPM
|
3D-alternating
|
|
3-Etβ
|
CH2Cl2/MeOH
|
Triclinic
|
P1̄
|
2
|
7646.5 Å3
|
MMMM/PPPP
|
Enantiopure layers
|
|
3-Etγ
|
CDCl3
|
Triclinic
|
P1̄
|
2
|
8142.8 Å3
|
MMMP/PPPM
|
Enantiopure layers
|
|
3-OPr
|
CH2Cl2/MeOH
|
Orthorhombic
|
Aba
2̄
|
4
|
18 567.7 Å3
|
MMMM/PPPP
|
Enantiopure layers
|
|
3-Br
|
CH2Cl2/MeOH
|
Triclinic
|
P1̄
|
2
|
7399.5 Å3
|
MMMM/PPPP
|
Enantiopure layers
|
Cages 3-Me and 3-Br both pack in the triclinic space group P1̄ in an isomorphic fashion, which is obvious for example from the comparable unit
cell volumes of 7578.4 Å3 (3-Me) and 7399.5 Å3 (3-Br) ([Table 1]). In both cases face-to-face π–π-stacking motifs are found on one face of each cage
with distances of d
π–π = 3.43 Å (3-Br) or 3.48 Å (3-Me) ([Figure 4]). The residual faces interact, e.g., via C – H⋯π interactions in the case of 3-Me with d
C – H⋯π = 2.99 Å ([Figure 4c]). This interaction is mimicked in 3-Br by C–Br⋯π interactions of d
C–Br⋯π = 3.26 Å ([Figure 4f]), leading in both cases to enantiopure layers within the crystallographic ab-planes ([Figure 5f]).
Figure 4 Intermolecular interactions found in the single-crystal X-ray structures of 3-Me (a – c) and 3-Br (d – f).
Changing the solvent diffused into saturated solutions of 3-Et from methanol (leading to 3-Etβ
with enantiopure cages) to hexane or if the crystals were grown by slow evaporation
of a CDCl3 solution of 3-Et, solvates 3-Etα
and 3-Etγ
are obtained showing the MMMP and PPPM isomers exclusively ([Figure 5]).
In the case of 3-OPr, packing in the orthorhombic space group Aba 2, similar layers as in 3-Me and 3-Br are found in the crystallographic bc-plane with one-dimensional channels along the c-axis filled with methanol and water molecules (see [Table 1], [Figure 5], [Figure 6] and discussion below). For 3-Et, the homochiral solvate 3-Etβ, as well heterochiral solvate 3-Etγ
, form layered structures as well, while 3-Etα
crystallizes in the cubic space group P
a3̄ and the enantiomeric cages
MMMP-
3-Et and
PPPM-
3-Et can be found in an alternating order in all three dimensions ([Figure 5f]).
Figure 5 a) Schematical depiction of the P- and M-alignment of the prochiral trispyrrol units. Note: The methyl group pointing out of the paper plane here is pointing inside the cage
voids, which has to be taken into account in the following figures. b) Schematic representations
of the homochiral PPPP (shown in blue for 3-Me) and MMMM (shown in red for 3-Me) as well as heterochiral PPPM (shown in green for 3-Etα
) and MMMP (shown in orange for 3-Etα
) cages. d – h) Packing motifs of the obtained solvates with different enantiomers
colored according to [Figure 3b] as capped-stick models. One cage each is shown as space-fill model in element colors.
Figure 6 Residual solvent analysis of single-crystal X-ray structures of cages 3-Me and 3-OPr. a) Cage 3-Me (shown in black as capped-stick model) with 12 molecules of methanol and one molecule
of CHCl3 (shown as space-fill models). b) Top-view and c) side-view of the hydrogen bonding
pattern in detail. d) Crystal packing of cage 3-OPr (shown as capped-stick models in white with one molecule shown as space-fill model
in element colors) with ten methanol molecules (shown as capped-stick model in element
colors) and ten water molecules (shown as a space-fill model in red without hydrogen
atoms).
The nitrogen-rich inside of the cagesʼ cavities should be a perfect environment for
polar guests. Crystallizing 3-Me using CHCl3 and methanol led to 12 enclathrate-ordered methanol molecules in defined H-bonded
arrays interacting with the pyrrol-imine subunits via synergetic H-bonds of d
N–H⋯O = 2.00 Å and d
O–H⋯N = 2.11 Å ([Figure 6a – c]). This alignment generates a “shell” of methanol molecules by the interaction of
the highly polar OH-groups and the inner surface of the cage leaving the methyl groups
pointing to the inside of the cage creating a more nonpolar environment. Inside this
nonpolar pocket, exactly one molecule of CHCl3 can be found to be interacting via van-der-Waals interactions ([Figure 6a]).
While Russian-doll alignments with three spheres have been realized on a molecular
level;[59]–[62] to the best of our knowledge, the structure discussed here is the first organic
cage compound with two of these spheres being two different solvents. Furthermore,
seven more methanol and one water molecule fill the inter-cage voids in the crystalline
packing. For 3-Etβ
, 10 methanol molecules are found in similar alignments as discussed for 3-Me ([Figure 6]), yet due to disorder the residual solvent molecules had to be SQUEEZED. Ten methanol
molecules are also found in the cavity in 3-OPr ([Figure 6d]). This time the residual voids are filled with ten water molecules which align in
a chain-like fashion along the crystallographic a-axis ([Figure 6d]). Due to fast exchange, the H-atoms of the water molecules were not crystallographically
found and a detailed discussion of the hydrogen bonding pattern is not possible.
Conclusions
To summarize, we have constructed four cage molecules based on a trispyrroltrialdehyde. While 3-Me, 3-Et and 3-OPr have been successfully isolated and fully characterized, 3-Br was mainly investigated by SCXRD. A detailed study of six single-crystal X-ray structures
in total revealed that the prochiral trispyrroltrialdehyde orient in homoenantiomeric PPPP and MMMM cages in four cases and only cage 3-Et can be found in the heteroenantiomeric MMMP and PPPM form in two solvates in the solid state. Furthermore, the nitrogen-rich interior
of the cages is able to interact with polar molecules such as methanol or water and
defined alignments like methanol shells or hydrogen-bonded chain-like structures were
obtained. These findings can be beneficial for applications of the cages in the selective
binding of polar guests or proton conductivity.[63]–[65] Furthermore, the nitrogen-rich cavity can be used as a cage ligand for transition
metal ions or clusters.[66]–[69]
Funding Information
We thank the European Research Council (ERC) in the frame of the consolidator grant
CaTs n DOCs (grant no. 725 765) and Deutsche Forschungsgemeinschaft (DFG) under Germanyʼs
Excellence Strategy Cluster 3D Matter Made to Order (EXC-2082/1 – 390 761 711) for
funding this project. The authors acknowledge support by the state of Baden-Württemberg
through bwHPC and the German Research Foundation (DFG) through grant no. INST 40/575 – 1
FUGG (JUSTUS 2 cluster).
