Article pubs.acs.org/JACS
Viologen-Based Conjugated Covalent Organic Networks via Zincke Reaction Gobinda Das,† Tina Skorjanc,† Sudhir Kumar Sharma,‡ Felipe Gándara,§ Matteo Lusi,∥ D. S. Shankar Rao,⊥ Sridurai Vimala,⊥ Subbarao Krishna Prasad,⊥ Jesus Raya,# Dong Suk Han,∇ Ramesh Jagannathan,‡ John-Carl Olsen,⊗ and Ali Trabolsi*,† †
Chemistry Program, New York University Abu Dhabi, Experimental Research Building (C1), Saadiyat Island, United Arab Emirates Engineering Division, New York University Abu Dhabi, Experimental Research Building (C1), Saadiyat Island, United Arab Emirates § The Materials Science Factory, Instituto de Ciencia de Materiales de Madrid−CSIC, Sor Juana Inés de la Cruz 3, 28049 Madrid, Spain ∥ Department of Chemical and Environmental Science, University of Limerick, Limerick V94 T9PX, Republic of Ireland ⊥ Centre for Nano and Soft Matter Sciences, Jalahalli, Bangalore 560013, India # CNRS/Université de Strasbourg, 1, Rue Blaise Pascal, Strasbourg 67000, France ∇ Chemical Engineering Program, Texas A&M University at Qatar, Education City, Doha, Qatar ⊗ Department of Chemistry, University of Rochester, RC Box 270216, Rochester, New York 14627, United States ‡
S Supporting Information *
ABSTRACT: Morphology influences the functionality of covalent organic networks and determines potential applications. Here, we report for the first time the use of Zincke reaction to fabricate, under either solvothermal or microwave conditions, a viologen-linked covalent organic network in the form of hollow particles or nanosheets. The synthesized materials are stable in acidic, neutral, and basic aqueous solutions. Under basic conditions, the neutral network assumes radical cationic character without decomposing or changing structure. Solvent polarity and heating method determine product morphology. Depending upon solvent polarity, the resulting polymeric network forms either uniform selftemplated hollow spheres (HS) or hollow tubes (HT). The spheres develop via an inside-out Ostwald ripening mechanism. Interestingly, microwave conditions and certain solvent polarities result in the formation of a robust covalent organic gel framework (COGF) that is organized in nanosheets stacked several layers thick. In the gel phase, the nanosheets are crystalline and form honeycomb lattices. The use of the Zincke reaction has previously been limited to the synthesis of small viologen molecules and conjugated viologen oligomers. Its application here expands the repertoire of tools for the fabrication of covalent organic networks (which are usually prepared by dynamic covalent chemistry) and for the synthesis of viologen-based materials. All three materialsHT, HS, and COGFserve as efficient adsorbents of iodine due to the presence of the cationic viologen linker and, in the cases of HT and HS, permanent porosity.
■
been shown to be beneficial for drug delivery,17 catalysis,18 sensing,19 and storage.20 The majority of reported covalent organic networks have been constructed from neutral building blocks, which limits the range of non-covalent interactions the materials can have with guest molecules. Charged building blocks would, alternatively, allow for the introduction of stronger electrostatic forces for host−guest recognition, while redox-active sites would open the door to tunable properties and functions.21 Viologens, i.e., 4,4′bipyridinium ions, undergo reversible one- and two-electron
INTRODUCTION Extended covalent organic networks, in the form of either amorphous covalent organic polymers (COPs) or crystalline covalent organic frameworks (COFs), exhibit a variety of features that are conducive to practical application.1−4 They are often porous and robust, and their high physicochemical stabilities make possible their application at high temperatures and under humid conditions.5 These characteristics and their favorable chemical properties make them potentially useful for gas storage 6−8 and separation, 9 chemical sensing, 10−12 catalysis,13,14 drug delivery,15 and photovoltaic light harvesting.16 Furthermore, their morphologies have a strong influence on their functionalities. For example, hollow structures have © 2017 American Chemical Society
Received: March 24, 2017 Published: May 16, 2017 9558
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society
Figure 1. Synthesis of the viologen-based self-templated materials via Zincke reaction under solvothermal conditions (ST) or microwave (MW) irradiation: (a) amorphous hollow spheres (HS) and tubes (HT) and (b) a crystalline covalent organic gel framework (COGF).
strategies are used to generate covalently linked networks that have (i) specific morphologies, (ii) order that ranges from amorphous to crystalline, and (iii) mechanical properties that vary from those of sol−gels to those of rigid solids.28 Self-templation and hollow structure are two important aspects of covalent organic assembly.31,32 Although not very common, self-templated (template-free) synthesis is more costeffective than, and is considered superior to, templatedependent synthesis.20 In an effort to build highly stable π-conjugated viologenbased covalent materials, we have investigated the use of the Zincke reaction (Figure 1) between 1,1′-bis(2,4-dinitrophenyl)[4,4′-bipyridine]-1,1′-diium dichloride (BDB) and an aromatic amine, TAPB, in different solvent systems, and using either solvothermal (ST) or microwave (MW) heating modes. Interestingly, we have observed that the morphology of the prepared viologen-based material is solvent-dependent and varies from self-templated hollow spheres (HS) in ethanol:water solvent (4:1, v:v) to self-templated hollow tubes (HT) in dioxane (Figure 1a). Both products are amorphous and can be prepared under ST conditions in 3 days, or under MW irradiation in 2 h. A detailed study revealed that HS formation is the result of an inside-out Ostwald ripening process.20 More interestingly, when ethanol:water (1:1, v:v) was used as solvent under MW irradiation, a self-standing crystalline covalent organic gel framework (COGF) was obtained (Figure 1b, SI video). The formation of crystalline covalent organic gels by the Zincke reaction is unprecedented. We speculate that, under the gel-forming conditions, either reaction reversibility or increased yield contributes to the observed crystallinity. The mechanism of the Zincke reaction involves nucleophilic attack by a primary amine onto an N-(2,4-dinitrophenyl)pyridinium salt. Subsequent electrocyclic ring opening gives a linear
reductions and are promising components for the construction of charged covalent organic networks that have tunable properties.22−26 The Menshutkin reaction,22,26 Sonogashira− Hagihara coupling,23,27 and the Zincke reaction22 are several methods that have been used for incorporating viologens into such networks. Li et al. reported the synthesis of a viologenbased COF using a Schiff base reaction between 1,3,5 tris(4aminophenyl)benzene (TAPB) and 1,1-bis(4-formylphenyl)4,4′-bipyridinium dichloride.25 D’Alessandro et al. synthesized a porous COP by Sonogashira−Hagihara palladium-catalyzed cross coupling and showed that its rate of gas uptake depended on the redox state of its viologen subunits.23 Upon reduction, the selectivity for CO2 over N2 was enhanced. Recently, our group reported the synthesis of a viologen-based COP by alkylation and demonstrated that its affinity for toxic dyes could be adjusted by changing the redox state of its viologen linkers. When the linkers were doubly cationic, the material displayed high fluorescein uptake, whereas when its linkers were in their fully reduced neutral form, the polymer had a greater affinity for Nile red.24 Because of their cationic character, viologen-based materials have also shown great practical potential, especially for iodine capture and the removal of toxic anions from water.22,25 Control of the morphology and specific surface area of covalent organic materials is generally achieved by three different approaches.28,29 The first involves the use of kinetically controlled reactions to covalently and irreversibly connect sterically demanding building blocks.30 These reactions tend to be high-yielding and to generate networks with large free volumes. The second method relies on reversible covalent reactions that allow constituent building blocks to self-assemble into crystalline networks. The third method uses reactions between presynthesized templates with predefined geometries that determine the connectivity of the structure. These 9559
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society
Figure 2. Microscopic characterization [SEM (top row), HRTEM (middle row), AFM (bottom row)] of HS, HT, and COGF: (a−c) HS obtained under solvothermal conditions; (d−f) HS obtained with microwave heating; (g−i) HT obtained under solvothermal conditions; (j−l) HT obtained with microwave heating; and (m−o) COGF obtained with microwave heating.
disappearance of the characteristic nitro peaks at ∼1300 cm−1 in the FTIR spectra of the final product (Figure S1) indicates loss of 2,4-dinitroaniline and provides evidence for the formation of the expected product. CP-MAS 13C NMR spectra of HS showed broad peaks between 127 and 147 ppm corresponding to the aromatic carbon atoms of the network (Figure S2). As synthesized, HS and HT exhibit an electron paramagnetic resonance (EPR) signal (Figures S3 and S4). The radical character of the as-synthesized networks is the result of a charge transfer from the aromatic core to the viologen unit and is an indication of the highly conjugated structures23 obtained from the Zincke reaction. SEM micrographs of HS prepared under ST conditions in ethanol:water (4:1, v:v) showed uniform hollow capsules with a raspberry-like surface and an average diameter of 1.8 μm (Figures 2a,b). The spherical microcapsules obtained under MW irradiation in the same solvent mixture had rougher surfaces compared to those obtained under ST conditions (Figures S5 and S6). The size of the capsules synthesized under MW conditions was comparable to that obtained under ST conditions. Hollow interiors can be clearly observed through partially broken capsule walls (Figure S6). The appearance of a light regime at the center of the capsules and a well-defined dark area at their outer surface indicate hollow structure (Figure 2b,e).20 The outer diameter of the capsules was found to be ∼2.6 μm and the cavity diameter was ∼2.0 μm for HS prepared by either method. In ST reactions, smaller capsules with an outer diameter of 457 nm and a cavity diameter of ∼290 nm were also observed (Figures S6 and S7). After heating under MW irradiation, relatively uniform capsules were formed. In both cases, the shell thickness varied between 85 and 300 nm. AFM reveals that the surfaces of these spheres are grainy and rough (Figure 2c,f). Their average cross-sectional diameter and height were found to be approximately 1.5 μm and ∼400−
structure known as a König salt. Ring closing followed by elimination of 2,4-dinitroaniline yields a pyridinium ion product. Ring opening is thought to be the rate-determining step, though thorough studies of its reversibility/irreversibility under different conditions have not, to our knowledge, been reported.33,34 Reversibility would allow for error correction and a greater degree of polymerization and would explain the formation of more extensive (and more thermodynamically stable) sheet-like structures that crystallize by π-stacking. On the other hand, the reaction may remain under kinetic control but simply proceed with increased yield that, in itself, leads to larger sheets that crystallize.35 We also found that MW irradiation accelerates reaction rates and seems to enhance the crystallinity of the products.36 The gel did not form in ethanol:water (1:1, v:v) under ST conditions. Instead, a powdered product composed of HS particles was produced.
■
RESULTS AND DISCUSSION HS and HT Synthesis and Characterization. Both hollow viologen-based materials discussed here, HS and HT, were synthesized using the Zincke reaction. Each was prepared by either solvothermal (ST) or microwave (MW) heating (Figure 1). In two typical reactions, 0.175 g (0.5 mmol) of TAPB and 0.280 g (0.5 mmol) of BDB were mixed and heated for 3 days under ST conditions or for 2 h under MW irradiation. A 4:1 EtOH:H2O mixture was shown by Rivard et al.35 to be an optimal solvent for Zincke reactions. Water was found to play an important role in minimizing degradation of the Zincke salt. A ratio of 1:1 EtOH:H2O was used to check the effect of polarity on the morphology of the product. The successful coupling of the triamine and the Zincke salt under both conditions was confirmed by Fourier transform infrared spectroscopy (FTIR) and by solid-state cross-polarization magic angle spinning (CP-MAS) 13C NMR analysis. The 9560
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society
Figure 3. Sequential micrographs depicting (from left to right) the inside-out Ostwald ripening mechanism of HS formation.
chloride counterions. Pore size distributions were calculated on the basis of nonlocal density functional theory (NLDFT), but because of the blocked pores, no sharp pore distribution could be estimated. Nevertheless, we conclude that the porosities of the materials are in the range of micro- to mesoporous. HS and HT were thermally stable at temperatures up to 400 °C (Figure S25). A small weight loss attributable to solvent evaporation occurred during the initial heating. The additional weight loss that occurred above 600 °C was consistent with framework degradation. The chemical stabilities of HS and HT in water were investigated by immersing 50 mg of each in 25 mL of boiling water for 3 days. Characterization by FTIR (Figure S13) and SEM (Figures S14 and S15) shows that both materials have retain their structural integrity. In addition, both HS and HT were found to be stable in 6 M aqueous HCl at room temperature for 3 days. FTIR and SEM characterizations of acid-treated HS and HT were identical to those of the assynthesized networks. When HS and HT were treated with 1 M NaOH, they were instantly reduced to their radical cationic forms as a result of electron transfer from the base to the viologens,39 as confirmed by strong, sharp signals in solid state EPR spectra. Such signals are a strong indication of paramagnetic character and the absence of pimerization.40,41 These results contrast with the EPR spectra of the assynthesized materials, which, in their neutral forms, showed very weak signals (Figure S16). Despite reduction, the surface morphologies were unchanged relative to those of the assynthesized products, a result indicating that HS and HT are remarkably chemically stable not only in neutral water and aqueous acid but also in aqueous base (Figures S13−S15). COGF Synthesis, Characterization, and Rheological Studies. When the ethanol:water ratio of the solvent mixture of the Zincke reaction was increased to 1:1 (v:v), a strong organogel was obtained under MW conditions. In contrast, gel synthesis in the same solvent mixture was unsuccessful under conventional ST conditions, which instead yielded an amorphous polymer. With the use of the right solvent combination and under MW irradiation, the presence of an intrinsic dipole moment in the gel network and an enhanced dipole moment generated by dielectric heating promoted gel formation.42 As synthesized, the gel exhibited a strong EPR
500 nm, respectively (Figures S6 and S7). The soft organic surfaces of the capsules appear flattened because spheres are hollow and have partially collapsed under their own weight (Figures S9 and S10). The hydrodynamic diameter of the hollow capsules was measured by dynamic light scattering (DLS) and in both cases was found to be ∼1.5−1.8 μm with a polydispersity of 0.06 (Figure S11). HS characterization data were consistent in SEM, TEM, and AFM investigations. Further studies to evaluate the effect of counterions on the morphology of the synthesized materials are currently underway in our laboratory. To understand the mechanism of HS formation under ST conditions, we sampled a HS formation reaction at 12, 24, 36, and 72 h. All samples were characterized by SEM and HRTEM. SEM and TEM images for the 12 h sample (Figure 3) exhibit a flower-like aggregation of small nonspherical filaments with sizes ranging from ∼700 nm to 1.5 μm. It was found that after longer reaction times (24 and 36 h) these particles selfassembled and displayed a strong tendency to form hollow spherical structures. After 72 h, the interiors of the smaller structures that had formed by nucleation and growth had become smaller and the material residing in the central part of the spheres had slowly diffused via inside-out Ostwald ripening to produce central cavities, i.e., hollow structures.20 The formation of covalent network materials by this mechanism is uncommon and has rarely been documented.37,38 When polymer synthesis was carried out in 1,4-dioxane under ST or MW conditions, the morphology of the resultant particles observed by SEM was tube-like (HT) and hollow, with tubes having an average diameter of ∼260 nm. The hollow nature of these nanotubes was confirmed by HRTEM, which reveals rigid walls surrounding internal cylindrical cavities (Figures 2h,k and S8). Again, the lack of the characteristic nitro peaks at ∼1300 cm−1 in the FTIR spectra of the final product (Figure S1) indicates loss of 2,4-dinitroaniline and formation of the expected product. Nitrogen adsorption isotherms at 77 K revealed that HS and HT exhibited low permanent porosity. Both materials showed Type-II reversible adsorption isotherms. Their Brunauer− Emmet−Teller (BET) surface areas were 12 and 35 m2/g, respectively (Figure S12). These relatively low values may be the result of the pores of the materials being occupied by 9561
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society
showed that, for an applied strain of 0.25% and the frequency range of 0.1−100 rad s−1, the storage modulus G′ was significantly greater than the respective loss modulus G″ over the entire investigated angular frequency range, which is typical of a viscoelastic organogel. Steady shear flow curves (Figure 4b) show that the apparent viscosity of the gel matrix decreases gradually with increasing shear rate, indicating that the gel network behaves as a pseudoplastic system. With decreased shear, the process exhibited the reversibility and hysteresis characteristic of a thixotropic material. These results indicate the formation of a strong stable gel network. To further probe the mechanical properties of the gel, we performed a strain (γ)-sweep experiment at a fixed angular frequency of 1 rad s−1 (Figure 4c). G′ was initially constant for γ < 0.004, but over the window 0.02 < γ < 0.1, it fell off sharply following a power law of G′(γ) ∼ γ−0.75±0.01. Concomitantly, G″ passed through a maximum at a critical strain (γm) of 0.02 and then diminished according to G″(γ) ∼ γ−0.46±0.01 for higher strains. This G′(γ)-dominant linear viscoelasticity with an associated clear, albeit with a weak maximum in G″(γ) before the full strain softening sets in, is often seen in soft glassy systems.43−50 Toward the upper limit of the strain applied in this study, the storage and the loss moduli intersected (G″= G′), marking a breakdown of the network that resulted in fluid-like behavior. A parameter that differentiates between a highly viscous liquid and a gel is the difference between the complex viscosity determined from the small oscillation measurements,
signal which confirmed its radical nature (Figure S17) and may be the result of its donor−acceptor properties as well as electron transfer from its aromatic cores to the viologen units. As with HS and HT, the sharp EPR signal is an indication of the paramagnetic character of COGF. In an inverted sample tube, the thixotropic gel remained suspended at the upper end of the container. Even with vigorous shaking, the gel did not break or flow, which suggests high material strength. This property was further investigated by rheological studies. Rheological measurements were carried out with a stresscontrolled rheometer (model ARG2, Waters) and a parallel plate (20 mm diameter) geometry. A 1 mm gap between the plates defined the sample’s thickness. The experiments were performed at ambient temperature with a solvent trap containing 3:2 EtOH:H2O to maintain a solvent-rich environment. The oscillatory frequency-sweep experiments (Figure 4a)
(G ′2 + G ″2 )1/2
η* = , and η, the steady-state viscosity.50 ω Maxwellian fluids follow the Cox−Merz rule with identical η and η*. This rule is disobeyed by weak gels as well as soft glassy systems. In addition, the frequency dependence of η* follows the power law, η*∝ η−p, with p ≈ 0 indicating liquid-like behavior, and p ≈ 1 showing a solid-like response. The sheardependent η and the frequency-dependent η* are shown in Figure 4d, which highlights the disobedience of the Cox−Merz rule. Fitting the η* data to η−p yields p = 0.933 ± 0.001, a value expected for a solid-like response of the system and characteristic of well-formed gels.
Figure 4. Rheological measurements of COGF. (a) Frequency-sweep measurements of the gel; (b) Graphical representation of the relationship between the viscosity and shearing rates of the gel; (c) Strain-sweep measurements of the gel; (d) Plot of shear-dependent viscosity η and frequency-dependent η* of the gel.
Figure 5. PXRD pattern and simulated crystal structure of COGF. (a) Experimental PXRD pattern of the as-synthesized product (red line), simulated XRD pattern using staggered stacking (black line), and side view of a space-filling representation of three stacked layers of COGF (inset). (b) Top view of a space-filling representation of two stacked layers of the COGF crystal structure, with counterions omitted for clarity. The simulated 2D unit cell parameters are a = b = 35.08 Å and c = 4.1 Å. 9562
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society The morphology of the gel was studied by SEM, TEM, and AFM. As shown in Figures 2m−o and S18, the synthesized material has a 2D sheet-like morphology. AFM (Figures 2o and S9c,d) shows an average profile height of approximately 2 nm. While PXRD patterns of HS and HT display only a broad featureless peak indicating the amorphous nature of the materials (Figure S19), the PXRD of the COGF reveals a strong peak at 2θ = 2.91° (30.4 Å) and three low-intensity peaks at 2θ = 5.04° (17.5 Å), 2θ = 5.83° (15.2 Å), and 2θ = 7.73° (11.4 Å) (Figure 5a). These reflections are attributed to (100), (110), (200), and (210) planes of a hexagonal lattice, respectively. The ratio of the spacing with respect to d100 is 1:1/ √3:1/√4:1/√7, which is typical for a hexagonal lattice formed in a covalent network.49 The absence of sharp diffraction peaks corresponding to hkl planes with the l component and the presence of only a broad peak centered at 2θ ≈ 25° suggest that there is no periodicity along the c axis. Crystal models were generated based on the geometry of the building blocks forming the honeycomb layers (Table S1). Various stacking sequences were modeled, including those corresponding to bnn (Figure S20) and gra (Figure S22) topologies. The models were geometrically optimized with force field and DFT-based energy minimizations. A comparison of the simulated PXRD pattern with the experimental one shows the correspondence of the strong diffraction peaks observed at 2θ = 2.9°, 5.0°, and 5.8° to the (100), (110), and (200) hkl planes, respectively. The unit cell parameters of the gra model are in good agreement with the experimental pattern (Table S1) and result in a good profile fitting (Rwp = 1.95%). According to the optimized model, the pyridinium rings are not coplanar, but are rotated with respect to each other and with respect to the triphenylbenzene moieties. This conformation is reminiscent of the one observed in the single crystal structure determined for the TBAP monomer as described below. Although the exact stacking sequence of the layers, if any, cannot be unambiguously determined from the PXRD pattern, the lack of mesoporosity expected for a bnn structure (Figure S20) possibly rules out this stacking. In addition, the AFM study indicates that only a few layers stack in each particle, which is also consistent with the broad scattering in the 20°− 30° 2θ region. Large-scale synthesis of nanosheets that does not rely on exfoliation or mechanical force is a challenging task. TEM images of the synthesized polymer also depict sheet-like morphology with well-defined lattice fringes (Figure 2n, inset, and Figure S18), which further confirm the presence of an ordered layered structure. The nanosheets became welldispersed in water and exhibited a clear Tyndall effect (Figure S23), which is typical for colloidal suspensions. The sheet-like structure of the gel could also be observed from its staggered geometry in our computational models. Layer-flapping and the staggered orientation of the layers prevented N2 from accessing the gel’s internal pores and made the material completely nonporous. Crystal Structure of the Trisviologen (TV) Monomer. To better understand the nature of materials synthesized under different conditions, we crystallized the TV monomer with PF6− counteranions (Figure 6). This salt packs in a triclinic P1̅ unit cell with one molecule per asymmetric unit. The crystal structure reveals (Figure 6c) that the three central phenyl rings of the monomer do not lie in the same plane. Tribranched organic cations, with their branches rotated ∼60° out of register, are paired to form bivalve-like structures. π−π interactions exist between the central rings, which are spaced
Figure 6. Graphical representations of the TV monomer. (a) Chemical structure of the PF6− salt of TV. Single crystals were obtained by slow evaporation of diisopropyl ether in an acetonitrile solution of the monomer. (b) Top view of a ball-and-stick representation of the crystal structure. (c) Top view of a space-filing representation of two staggered monomers. (d) Side view of a space-filling representation of two staggered monomers.In b−d, PF6− counterions have been omitted for clarity.
3.6 Å apart. The cationic dimers are surrounded by six independent PF6− anions that balance charge and stabilize the structure. Lack of planarity and imperfect π−π stacking, similar to what is seen in the monomer, likely contribute to the formation of hollow spheres and nanotubular structures during polymer synthesis. The solvent dependence of the polymer morphologies can be explained as follows. When the reaction is carried out in polar solvents, such as ethanol:water mixtures, there is a favorable interaction between the polar surfaces of incipient particles and the solvent molecules. These surfaces form hydrogen bonds with the solvent, which delays nucleation and leads to the formation of hollow capsules. In 1,4-dioxane, on the contrary, polymer growth rate is faster because no hydrogen bonds are formed between the growing polymer and the solvent. Iodine Capture. The porosity and cationic nature of our materials, together with our previous positive results involving the use of viologen-based polymers as adsorbents,22 prompted us to investigate the abilities of HS, HT, and COGF to capture iodine from solution and vapor phases. For solution measurements, 15.0 mg of each polymer was added to a 1 mM solution of I2 in cyclohexane. The process of I2 uptake was then monitored by UV−vis spectrophotometry (Figure S24). Upon polymer addition, the purple color of the I2 solution began to fade and, in the presence of the most adsorbent materials, became completely colorless within 1 h. HT was the most efficient at removing I2 from solution, as ∼88% was removed in the first 2 h (Figure 7a). The fast uptake of iodine from solution by HT can be attributed to this material’s having a higher surface area (35 m2/g) than HS (12 m2/g) and the nonporous gel. Experiments involving vapor phase I 2 capture were performed using an excess amount of solid I2, which was heated to 40 °C and maintained at ambient pressure. COGF proved to be the most efficient in the long run as its mass increased by ∼140% within 22 h. However, the HT morphology proved to be by far the most efficient adsorbent 9563
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society
obtained free of charge via www.ccdc.cam.ac.uk/data_request/ cif. FTIR and TGA of synthesized materials; CP/MAS 13C NMR of HS; EPR spectra of HS, HT, and COGF; SEM, TEM, and AFM images of HS; DLS of HS; porosity measurements for HS, HT, and COGF; SEM and FTIR analysis of HS and HT after exposure to acid, hot water, and base; EPR analysis of HS after base treatment; HRTEM images of COGF; Tyndall effect; PXRD of HS and HT; simulations on COGF in different conformations; iodine adsorption raw data; iodine release experiment; and XPS analysis of iodine-loaded HS, including Figures S1−S27 (PDF) Video showing the self-standing crystalline covalent organic gel (AVI) X-ray crystallographic data for trisviologen (CIF) X-ray crystallographic data for the gra topology (CIF)
Figure 7. Iodine adsorption measurements. (a) Graph of percent removal of I2 from a 1 mM solution in cyclohexane over time by HT (black squares), HS (red circles), or COGF (blue triangles). (b) Graph of the % mass increase of HT, HS, or COGF samples due to the adsorption of iodine vapor over time.
in the first 7 h. To assess the reusability of these materials, we performed an I2 release study in ethanol (Figure S26). We observed that the rate of release varied exponentially, with a sharp increase in the first 20 min, which subsequently plateaued after about 60 min. This suggests that our synthesized materials can be regenerated and reused for I2 capture. X-ray photoelectron spectroscopy (XPS) allowed us to investigate the interactions between iodine and the viologen-based materials (Figure S27). The XPS study revealed the presence of both neutral I2 and triiodide ions (I3−). In a TGA experiment, we observed a significant iodine weight loss from 90 to 300 °C (Figure S25).
■
*
[email protected] ORCID
Felipe Gándara: 0000-0002-1671-6260 Matteo Lusi: 0000-0002-9067-7802 Subbarao Krishna Prasad: 0000-0002-9367-1369 Dong Suk Han: 0000-0002-4804-5369 Ali Trabolsi: 0000-0001-7494-7887
■
Notes
CONCLUSION Application of the Zincke reaction has previously been limited to the synthesis of small viologen-containing molecules and oligomers. Here, we have presented the reaction’s efficient use for the preparation of viologen-based networks. The morphologies, crystallinities, and mechanical properties of the networks depended on reaction solvent and the heating mode. The networks were produced by the reaction of an aromatic amine, 1,3,5-tris(4-aminophenyl)benzene, and 1,1′-bis(2,4dinitrophenyl)-[4,4′-bipyridine]-1,1′-diium dichlodichloride. Hollow spherical particles were obtained in ethanol:water (4:1, v:v) with either solvothermal or microwave heating. Hollow tube-like structures were isolated in dioxane with either type of heating. Interestingly, a robust and highly ordered selfstanding gel formed under microwave irradiation when the reaction was carried out in ethanol:water (1:1, v:v). In the gel phase, the framework self-organized into nanosheets that stacked several layers thick. All of the viologen-based materials synthesized were resistant to moisture and highly stable in acidic or basic aqueous solutions. We took advantage of their remarkable stability and cationic nature by using them as efficient adsorbers of iodine in solution and vapor phases. We are currently using the Zincke reaction to synthesize viologenbased covalent organic frameworks that contain different core moieties.
■
AUTHOR INFORMATION
Corresponding Author
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The research described here was sponsored by New York University Abu Dhabi (NYUAD), UAE. G.D., T.S., S.K.S., R.J., and A.T. thank NYUAD for its generous support of the research program at NYUAD. The research was carried out by using the Core Technology Platform resources at NYUAD. The authors are grateful to F. Ravaux for technical assistance and M. Jouiad for providing access to the state-of-the-art characterization facility available at Masdar Institute, Abu Dhabi, UAE.
■
REFERENCES
(1) Cote, A. P.; Benin, A. I.; Ockwig, N. W.; O’Keeffe, M.; Matzger, A. J.; Yaghi, O. M. Science 2005, 310, 1166−1170. (2) Feng, X.; Ding, X.; Jiang, D. Chem. Soc. Rev. 2012, 41, 6010− 6022. (3) Liu, W.; Luo, X.; Bao, Y.; Liu, Y. P.; Ning, G.-H.; Abdelwahab, I.; Li, L.; Nai, C. T.; Hu, Z. G.; Zhao, D.; Liu, B.; Quek, S. Y.; Loh, K. P. Nat. Chem. 2017, DOI: 10.1038/nchem.2696. (4) Perepichka, D. F.; Rosei, F. Science 2009, 323, 216−217. (5) Kandambeth, S.; Mallick, A.; Lukose, B.; Mane, M. V.; Heine, T.; Banerjee, R. J. Am. Chem. Soc. 2012, 134, 19524−19527. (6) Li, Z.; Feng, X.; Zou, Y.; Zhang, Y.; Xia, H.; Liu, X.; Mu, Y. Chem. Commun. 2014, 50, 13825−13828. (7) Rabbani, M. G.; Sekizkardes, A. K.; Kahveci, Z.; Reich, T. E.; Ding, R.; El-Kaderi, H. M. Chem. - Eur. J. 2013, 19, 3324−3328. (8) Li, Z.; Zhi, Y.; Feng, X.; Ding, X.; Zou, Y.; Liu, X.; Mu, Y. Chem. Eur. J. 2015, 21, 12079−12084. (9) Tong, M.; Yang, Q.; Ma, Q.; Liu, D.; Zhong, C. J. Mater. Chem. A 2016, 4, 124−131. (10) Dalapati, S.; Jin, S.; Gao, J.; Xu, Y.; Nagai, A.; Jiang, D. J. Am. Chem. Soc. 2013, 135, 17310−17313.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b02836. Crystallographic details for the PF6− salt of the trisviologen monomer can be found in the Cambridge Crystallographic Database, CCDC deposition no. 1535240. These data can be 9564
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565
Article
Journal of the American Chemical Society (11) Ding, S.-Y.; Dong, M.; Wang, Y.-W.; Chen, Y.-T.; Wang, H.-Z.; Su, C.-Y.; Wang, W. J. Am. Chem. Soc. 2016, 138, 3031−3037. (12) Das, G.; Biswal, B. P.; Kandambeth, S.; Venkatesh, V.; Kaur, G.; Addicoat, M.; Heine, T.; Verma, S.; Banerjee, R. Chem. Sci. 2015, 6, 3931−3939. (13) Wang, X.; Han, X.; Zhang, J.; Wu, X.; Liu, Y.; Cui, Y. J. Am. Chem. Soc. 2016, 138, 12332−12335. (14) Xu, H.; Gao, J.; Jiang, D. Nat. Chem. 2015, 7, 905−912. (15) Fang, Q.; Wang, J.; Gu, S.; Kaspar, R. B.; Zhuang, Z.; Zheng, J.; Guo, H.; Qiu, S.; Yan, Y. J. Am. Chem. Soc. 2015, 137, 8352−8355. (16) Dogru, M.; Handloser, M.; Auras, F.; Kunz, T.; Medina, D.; Hartschuh, A.; Knochel, P.; Bein, T. Angew. Chem. 2013, 125, 2992− 2996. (17) Sun, L.; Xiong, X.; Zou, Q.; Ouyang, P.; Burkhardt, C.; Krastev, R. J. Appl. Polym. Sci. 2017, 134, 4425-1−4425-10. (18) Arnal, P. M.; Comotti, M.; Schüth, F. Angew. Chem. 2006, 118, 8404−8407. (19) Pan, L.; Chortos, A.; Yu, G.; Wang, Y.; Isaacson, S.; Allen, R.; Shi, Y.; Dauskardt, R.; Bao, Z. Nat. Commun. 2014, 5, 3002-1−3002-8. (20) Kandambeth, S.; Venkatesh, V.; Shinde, D. B.; Kumari, S.; Halder, A.; Verma, S.; Banerjee, R. Nat. Commun. 2015, 6, 6786-1− 6786-10. (21) Ma, H.; Liu, B.; Li, B.; Zhang, L.; Li, Y.-G.; Tan, H.-Q.; Zang, H.-Y.; Zhu, G. J. Am. Chem. Soc. 2016, 138, 5897−5903. (22) Das, G.; Prakasam, T.; Nuryyeva, S.; Han, D. S.; Abdel-Wahab, A.; Olsen, J.-C.; Polychronopoulou, K.; Platas-Iglesias, C.; Ravaux, F.; Jouiad, M.; Trabolsi, A. J. Mater. Chem. A 2016, 4, 15361−15369. (23) Hua, C.; Chan, B.; Rawal, A.; Tuna, F.; Collison, D.; Hook, J. M.; D’Alessandro, D. M. J. Mater. Chem. C 2016, 4, 2535−2544. (24) Das, G.; Skorjanc, T.; Prakasam, T.; Nuryyeva, S.; Olsen, J.-C.; Trabolsi, A. RSC Adv. 2017, 7, 3594−3598. (25) Yu, S.-B.; Lyu, H.; Tian, J.; Wang, H.; Zhang, D.-W.; Liu, Y.; Li, Z.-T. Polym. Chem. 2016, 7, 3392−3397. (26) Chen, G.; Zhou, Y.; Wang, X.; Li, J.; Xue, S.; Liu, Y.; Wang, Q.; Wang, J. Sci. Rep. 2015, 5, 11236-1−11236-14. (27) Buyukcakir, O.; Je, S. H.; Choi, D. S.; Talapaneni, S. N.; Seo, Y.; Jung, Y.; Polychronopoulou, K.; Coskun, A. Chem. Commun. 2016, 52, 934−937. (28) Addicoat, M. A.; Tsotsalas, M. Front. Mater. 2015, 2, 1−5. (29) Segura, J. L.; Mancheño, M. J.; Zamora, F. Chem. Soc. Rev. 2016, 45, 5635−5671. (30) Zeng, Y.; Zou, R.; Luo, Z.; Zhang, H.; Yao, X.; Ma, X.; Zou, R.; Zhao, Y. J. Am. Chem. Soc. 2015, 137, 1020−1023. (31) Hota, R.; Baek, K.; Yun, G.; Kim, Y.; Jung, H.; Park, K. M.; Yoon, E.; Joo, T.; Kang, J.; Park, C. G.; et al. Chem. Sci. 2013, 4, 339− 344. (32) Kim, D.; Kim, E.; Lee, J.; Hong, S.; Sung, W.; Lim, N.; Park, C. G.; Kim, K. J. Am. Chem. Soc. 2010, 132, 9908−9919. (33) Marvell, E. N.; Caple, G.; Shahidi, I. J. Am. Chem. Soc. 1970, 92, 5641−5645. (34) Van Dunghen, E. D.; Nasielski, J.; Van Laer, P. Bull. Soc. Chim. Belg. 1957, 66, 661−663. (35) Zeghbib, N.; Thelliere; Rivard, M.; Martens, T. J. Org. Chem. 2016, 81, 3256−3262. (36) Calatayud, D. G.; Jardiel, T.; Peiteado, M.; Caballero, A. C.; Fernández-Hevia, D. Nanoscale Res. Lett. 2014, 9, 273−277. (37) Zhou, Q.; Xiang, H.; Fan, H.; Yang, X.; Zhao, N.; Xu, J. J. Mater. Chem. 2011, 21, 13056−13061. (38) Levin, A.; Mason, T. O.; Adler-Abramovich, L.; Buell, A. K.; Meisl, G.; Galvagnion, C.; Bram, Y.; Stratford, S. A.; Dobson, C. M.; Knowles, T. P. J.; Gazit, E. Nat. Commun. 2014, 5, 5219-1−5219-8. (39) Shi, W.; Xing, F.; Bai, Y.-L.; Hu, M.; Zhao, Y.; Li, M.-X.; Zhu, S. ACS Appl. Mater. Interfaces 2015, 7, 14493−14500. (40) Trabolsi, A.; Khashab, N.; Fahrenbach, A. C.; Friedman, D. C.; Colvin, M. T.; Cotí, K. K.; Benítez, D.; Tkatchouk, E.; Olsen, J.-C.; Belowich, M. E.; et al. Nat. Chem. 2010, 2, 42−49. (41) Juetten, M. J.; Buck, A. T.; Winter, A. H. Chem. Commun. 2015, 51, 5516−5519.
(42) Issadore, D.; Humphry, K. J.; Brown, K. A.; Sandberg, L.; Weitz, D. A.; Westervelt, R. M. Lab Chip 2009, 9, 1701−1706. (43) Agarwal, P.; Archer, L. A. Phys. Rev. E 2011, 83, 041402-1− 041402-5. (44) Sollich, P. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 1998, 58, 738−759. (45) Wyss, H. M.; Miyazaki, K.; Mattsson, J.; Hu, Z.; Reichman, D. R.; Weitz, D. A. Phys. Rev. Lett. 2007, 98, 238303-1−238303-4. (46) Robertson, C. G.; Wang, X. Phys. Rev. Lett. 2005, 95, 075703-1− 075703-4. (47) Miyazaki, K.; Wyss, H. M.; Weitz, D. A.; Reichman, D. R. EPL Europhysics Lett. 2006, 75, 915−921. (48) Nair, G. G.; Prasad, S. K.; Bhargavi, R.; Jayalakshmi, V.; Shanker, G.; Yelamaggad, C. V. J. Phys. Chem. B 2010, 114, 697−704. (49) Li, S.; Zhao, G.; Chen, H. J. Dispersion Sci. Technol. 2005, 26, 415−419. (50) Ç elik, Ö .; Dag, Ö . Angew. Chem. 2001, 113, 3915−3919.
9565
DOI: 10.1021/jacs.7b02836 J. Am. Chem. Soc. 2017, 139, 9558−9565