Functional supramolecular systems: design and applications

Tetrapyrroles are directly involved in the supramolecular self-assembly through aggregation induced by intermolecular π – π interactions, which first of all decrease the solubility of the compounds. In terms of practical applications, aggregation of tetrapyrroles has long been viewed as an adverse phenomenon that deteriorates the functional characteristics of materials based on them, such as quantum yields of luminescence and singlet oxygen generation. Thus, the use of porphyrins, phthalocyanines and related compounds as agents for photodynamic therapy and diagnosis of cancer, and for photoinactivation of pathogenic microorganisms ¹⁷ is largely limited by the ability of these molecules to aggregate in aqueous media, which affects their photophysical properties. In this regard, numerous efforts of researchers were aimed at obtaining water-soluble compounds that do not tend to aggregate. ¹⁸

Nevertheless, it was found that phthalocyanine-based aggregates stable in aqueous solutions could be used for biomedical purposes by implementing other mechanisms of relaxation of excited states. For example, photochemical properties of amphiphilic zinc phthalocyanine complex 1 containing one tris(dimethylaminomethyl)phenoxy group have been studied (Scheme 1). ¹⁹ It was shown that this complex, when transferred from DMF to water or phosphate buffer, forms stable colloidal solutions. According to dynamic light scattering and transmission electron microscopy (TEM) data, molecules in these solutions form spherical aggregates, nanodots nano-1 with a narrow size distribution (55 ± 20 nm at a concentration of 20 μmol L⁻¹). The surface of the nanodots bears hydrophilic tris(dimethylaminomethyl)phenoxy groups, which ensure stability in aqueous solutions, while their interior consists of hydrophobic phthalocyanine moieties bound by stacking interactions. This architecture of nanodots causes fluorescence quenching, restricts the excited state vibrational relaxation and completely inhibits singlet oxygen generation. However, these aggregates acquire the ability to generate another reactive oxygen species, the superoxide radical anion (). The formation of this species was confirmed using hydroethidine, a blue fluorescent dye, as a chemical probe. A study of the possibility of using nano-1 particles for photoinactivation of pathogenic microorganisms showed their high efficiency: even at a concentration of 50 nmolL⁻¹, they suppressed the growth of both Gram-positive (Staphylococcus aureus) and Gram-negative (Escherichia coli) antibiotic-resistant microorganisms.

Scheme 1 

Zoom In Zoom Out Reset image size

Controlled aggregation of tetrapyrrole compounds can also be induced by using electrostatic interactions. Thus, the interaction of anionic and cationic water-soluble phthalocyanines 2 and 3 gave nano-aggregates nano-[2·3] of 30–100 nm in size, which, unlike the initial phthalocyanines, possess neither fluorescence nor singlet oxygen generation capability. ²⁰ However, these aggregates exhibited a photothermal effect: their colloidal aqueous solution with a concentration of 5.6 μmol L⁻¹ warmed up from 22 to 43 °C under laser irradiation (with a wavelength of 655 nm and a power density of 0.6 W cm⁻²) for 5 min, whereas solutions of the initial phthalocyanines warmed up to no more than 35 °C under the same conditions. Based on this observation, the possibility of using nano-[2·3] assemblies for photothermal therapy was demonstrated in vivo with a 4T1 murine mammary carcinoma. In addition, these structures exhibited a pronounced optoacoustic effect under laser irradiation (Fig. 1), which allowed their accumulation and distribution in the tissues to be monitored by optoacoustic microscopy.

Structures 2, 3 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Electrostatic interactions between differently charged porphyrins have been used to create binary ionic materials. ^{21, 22} Thus, the interaction of cationic tin porphyrin complexes 4a,b with anionic porphyrin 5 produces either crystalline plate aggregates nano-[4a · 5] or X-ray amorphous clover-like structures nano-[4b · 5] (Fig. 2). The deposition of such nanoparticles on the platinum surface produced hybrid nanomaterials, which were used for the photodegradation of water in the presence of methyl viologen and triethanolamine. ²² The efficiency of these photocatalysts was determined by the structure of the aggregates: a significantly higher rate of hydrogen evolution was observed in the case of a photocatalyst based on nano-[4b · 5] (Fig. 3). This effect was explained on the basis of X-ray crystallography data, indicating the absence of intermolecular interactions in the structure of nano-[4a · 5], which are necessary for the appearance of electronic conductivity. At the same time, the conclusion about the presence of such interactions in nano-[4b · 5] was made on the basis of electron absorption spectroscopy data.

Structures 4, 5 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

A number of studies are devoted to self-assembly of tetrapyrrole systems via electrostatic interactions. For example, the self-assembly mechanisms of protonated tetrakis(4-sulfonatophenyl)porphyrin (H₄TSPP^2–, 6) to form porphyrin nanotubes due to electrostatic interaction of zwitter ions were discussed in a review of Sheinin et al. ²³ and a similar controlled formation of J- and H-aggregates was also studied ²⁴ for its analogue 6' with an inverted porphyrin macrocycle (Fig. 4). ²⁵

Zoom In Zoom Out Reset image size

The presence of an extended aromatic system allows tetrapyrrole macrocycles to form supramolecular assemblies with other polyunsaturated compounds, including carbon nanomaterials, such as fullerenes, graphene, nanotubes, etc., through intermolecular stacking interactions. In these hybrid systems, tetrapyrroles act as electron donors, while carbon components act as acceptors. ²⁶ As a result, under the action of light, they efficiently generate charge separated states and, hence, they are considered as promising components of photovoltaic materials. A significant advantage of supramolecular assembly of these systems compared to covalent bonding is that the conjugation chain of the carbon components is not broken and their unique photophysical properties are preserved.

Structures 6, 6' 

Zoom In Zoom Out Reset image size

The introduction of additional receptor groups to the periphery of tetrapyrrole compounds increases the efficiency of their binding to carbon nanomaterials. Thus, M(7) complexes containing eight conformationally flexible O-benzyldiethylene glycol residues are capable of forming supramolecular complexes with fullerenes C₆₀ and C₇₀, and the receptor molecules behave like octopuses capturing spherical guest molecules by tentacles. ²⁷ It was shown that the interaction of phthalocyanine derivatives of magnesium Mg(7) and zinc Zn(7) is accompanied by quenching of fluorescence of phthalocyanines due to the formation of charge transfer complexes in the excited state.

Analysis of the fluorescence quenching curves using the Stern – Volmer equation made it possible to determine the stability constants of these assemblies. ²⁷ It turned out that the stability of the supramolecular complex with C₇₀ is 1.5 – 2.5 times higher than that for the complex formed with fullerene C₆₀. This selectivity to C₇₀ is also consistent with the results of quantum chemical calculations using the PM6-DH2 Hamiltonian parameterized for the study of non-covalent systems (Fig. 5). Based on the EPR data, it was concluded that cobalt(II) complex Co(7) binds to C₇₀ via dipole – dipole interactions without changes in the spin density distribution. ²⁷

Zoom In Zoom Out Reset image size

Structures M(7) 

Zoom In Zoom Out Reset image size

An increase in the selectivity of binding of various fullerenes can be used to develop methods for the separation of fullerene mixtures. ²⁹ For example, palladium acetate-induced assembly of zinc meso-tetrakis(4-carboxyphenyl)-porphyrin 8 with macrocyclic ligand 9 in acetonitrile yielded macrocyclic capsule [(8)₂ · (9)₄ · (PdOAc)₈]⁸⁺. This capsule was capable of binding fullerenes C₆₀ and C₇₀ in a homogeneous solution; after being dissolved in acetonitrile, it can extract solid fullerenes (Fig. 6). ³⁰ X-Ray diffraction data indicate that, upon binding of fullerenes to this receptor, the size of the cavity decreases to better adjust to the guest molecule. The binding constants (log K_a ) of this capsule with C₆₀ and C₇₀ found by spectrofluorimetric titration were 7.47 ± 0.03 and 8.60 ± 0.30 L mol⁻¹, respectively, which is consistent with a significant increase in selectivity of binding of fullerene C₇₀. Such a high difference between the log K_a values made it possible to develop a chromatographic method for isolation of C₆₀ from a mixture of fullerenes containing also C₇₀ and C₈₂.

Zoom In Zoom Out Reset image size

One of the most successful solutions to the problem of controlling the assembly of tetrapyrrole compounds was the synthesis of their crown-substituted analogues, from which supramolecular assemblies of a given architecture can be obtained using the cation-induced assembly method. In addition, binding of alkali metal cations to crown-substituted tetrapyrroles directly affects the physicochemical properties of these heterotopic receptors, which makes them applicable for the fabrication of materials with controlled properties.

Thus, porphyrazines (Pz) modified with azacrown groups were proposed as fluorescent sensors for alkali metal cations (Fig. 7). ³¹ These sensors have no intrinsic fluorescence because of the nonradiative deactivation of their excited states via charge transfer from the azacrown ether nitrogen atom to the tetrapyrrole macrocycle. However, binding to alkali metal cations blocks the conjugation of the nitrogen lone pair with the porphyrazine ring, thereby preventing the photoinduced charge transfer, which results in the observed enhancement of the receptor fluorescence. Thus, when receptor 10 interacts with potassium and sodium cations, the fluorescence intensity increases by a factor of 10.5 and 11.5, respectively; however, the presence of ions of other alkali, alkaline earth and heavy metals does not increase fluorescence.

Zoom In Zoom Out Reset image size

To exhibit a selective fluorescence response to potassium or sodium cations, pincer receptors 11a,b containing a pair of azacrown ether groups were synthesized. ³² In this case, the method of binding the cations depended on which crown ether groups were introduced into the receptor molecule and gave rise to selectivity.

The fluorescence intensity of receptor 11a based on 12-azacrown-4 in the presence of sodium cations was selectively enhanced 291-fold, whereas receptor 11b, containing 15-azacrown-5 ether moieties exhibited selectivity to potassium cations, enhancing fluorescence 22-fold (Fig. 8). Importantly, the introduction of macrocycles 11a,b into SiO₂ nanoparticles made it possible to obtain hybrid materials in which the receptors retained a selective fluorescence response toward sodium and potassium cations.

Structures 11 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The above example emphasizes the importance of structural compliance and cooperative interaction between the receptor (crown ether) and the substrate (alkali metal cation) for the formation of supramolecular assemblies of a given architecture and obtaining the required analytical response.

Using these ideas, a highly sensitive sensor system for detecting and determining the concentration of beryllium salts was fabricated on the basis of zinc tetra(9-crown-3)-substituted phthalocyanine complex Zn[(9C3)₄Pc] (where 9C3 is 9-crown-3 ether, and Pc is phthalocyanine) (Fig. 9). ³³ It was shown that in the presence of Be²⁺ cations, this complex forms J-aggregates (Scheme 2), which leads to significant changes in the absorption and fluorescence spectra even at an analyte concentration of 10⁻⁹ mol L⁻¹. Other cations (Li⁺, Na⁺, K⁺, Cs⁺, Mg²⁺, Ca²⁺, Ba²⁺, Al³⁺, Co²⁺, Hg²⁺, Ni²⁺, Pb²⁺, Zn²⁺) cause no changes in the absorption/fluorescence spectra of the sensor.

Zoom In Zoom Out Reset image size

Scheme 2 

Zoom In Zoom Out Reset image size

On going to 15-crown-5-substituted (15C5) complexes, it was found that aggregation can be caused by cations with large ionic radii, K⁺, Rb⁺ and Cs⁺. In contrast to Zn[(9C3)₄Pc], the M[(15C5)₄Pc] complexes in the presence of these cations form cofacial supramolecular dimers with effective intramolecular exciton interaction between the aromatic systems. ³⁴ The formation of these dimers in solutions was confirmed by various physicochemical methods not only for phthalocyanines, but also for their analogues with extended aromatic systems such as naphthalocyanines (Nc) ³⁵ and oxanthrenocyanines (Oc) ^{36, 37} (Scheme 3).

Scheme 3 

Zoom In Zoom Out Reset image size

Only in 2018, the structure of supramolecular assemblies formed by crown-substituted phthalocyanines in the presence of alkali metal cations was established using X-ray diffraction analysis in relation to the dimer {[Rb₄(NicAl(15C5)₄Pc)₂(μ-O)]²⁺(Nic⁻)₂}, where Nic is a nicotinate (Fig. 10). ³⁸ This dimer was obtained by the reaction of a hydroxy aluminium tetra(15-crown-5)-substituted phthalocyanine complex [(15C5)₄Pc]Al(OH) with rubidium nicotinate. It was shown that the supramolecular assembly is accompanied by the formation of an μ-oxobridge between aluminium atoms. The Al–O–Al moiety formed in this case has a linear structure and is located perpendicular to the plane of phthalocyanine macrocycles, with the Al–Al distance being 3.762 Å.

Zoom In Zoom Out Reset image size

The analogy between the architecture of these supramolecular dimers and the structure of dimers formed by chlorophyll a molecules in plant photosynthetic systems made it possible to fabricate a donor – acceptor biomimetic system, which formed a long-lived charge-separated state upon photoexcitation. For this, an electron-donor zinc crown-substituted phthalocyanine complex Zn[(15C5)₄Pc] was used as a donor, while 2-(4-pyridyl)-5-(5-ammoniopentyl)pyrrolidino[3.4:1.2][60]fullerene served as an acceptor (Fig. 11). ³⁹ The lifetime of charge-separated state of ZnPc^+·–C₆₀ ^−· was 4.8 μs, and it increased to 6.7 μs on going from the dyad with a monomeric donor to the tetrad in which a supramolecular dimer acted as the donor.

Zoom In Zoom Out Reset image size

Indium(III) tetra(15-crown-5)-substituted phthalocyanine complex ([(15C5)₄Pc]In(OH)) was used to demonstrate the possibility of targeted and reversible cation-induced control of photophysical properties of compounds such as luminescence, singlet oxygen generation and photostability. ⁴⁰ In the monomeric state, this complex had low photostability due to its ability to generate singlet oxygen, whereas supramolecular dimerization resulted in inhibition of¹O₂ generation and increased the photostability.

Reversible cation-induced supramolecular assembly was also used to control the aggregation and photophysical properties of crown-substituted macrocycles exhibiting an increased propensity for intermolecular π – π interactions, tetra(15-crown-5)naphthalocyanines M[(15C5)₄Nc], where M=2 H, Mg, Zn, ³⁵ and tetra(15-crown-5)oxanthrenocyanines M[(15C5)₄Oc], where M=Mg, Zn. ³⁶

Using UV-vis spectroscopy, it was shown that naphthalocyanines M[(15C5)₄Nc] exist in an aggregated state in a CHCl₃ + MeOH mixture. ³⁵ However, under the action of potassium acetate, these aggregates are converted to soluble supramolecular dimers 2 M[(15C5)₄Nc] · 4 K⁺. The removal of potassium cations from the dimers in the presence of [2.2.2]cryptand leads to the formation of monomeric forms of complexes, which do not re-aggregate. It is important to note that solutions containing supramolecular dimers and aggregated forms of M[(15C5)₄Nc] remained photostable upon irradiation with white light. At the same time, solutions of monomeric naphthalocyanines underwent rapid photodegradation due to the generation of singlet oxygen, which causes the destruction of tetrapyrrole macrocycles. The photostability of crown naphthalocyanines M[(15C5)₄Nc] decreased in the M series as follows: 2 H > Mg > Zn, which is consistent with the increase in the quantum yields of singlet oxygen generation on going from the naphthalocyanine ligand to metal complexes.

Tetra(15-crown-5)oxanthrenocyanines M[(15C5)₄Oc], structurally related to crown naphthalocyanines, ³⁶ also formed soluble supramolecular dimers 2 M[(15C5)₄Oc] · 4 K⁺ in the presence of potassium cations. However, in contrast to naphthalocyanine analogues, the destruction of these dimers in the presence of [2.2.2]crypt- and did not lead to the formation of monomeric forms of complexes. Instead, repeated aggregation of the complexes was observed with the formation of nanoparticles characterized by panchromatic absorption in the 250 – 900 nm range. ⁴¹

Structure 12 

Zoom In Zoom Out Reset image size

A hybrid material with graphite-like carbon nitride modified with platinum nanoparticles was obtained using an asymmetric photosensitizer, zinc 4-carboxybenzoannulated tri(15-crown-5)phthalocyanine complex 12 (Pt/g-C₃N₄), (Fig. 12). ⁴²

Zoom In Zoom Out Reset image size

This material showed the ability to photocatalytically form hydrogen from an aqueous solution of triethanolamine with a turnover number (TON) of 3105 h⁻¹ due to photoinduced electron transfer from phthalocyanine molecules 12 to an inorganic matrix with subsequent reduction of protons to molecular hydrogen. It was shown that binding of alkali metal cations to crown ether substituents has different effects on the quantum yield, the rate of hydrogen evolution, and the TON of the catalytic system. These characteristics virtually did not depend on binding of lithium cations, while in the presence of sodium and especially potassium cations, a significant increase in the rate of hydrogen evolution was observed (Fig. 13). In the latter case, the TON value reached 9067 h⁻¹. The authors used a wide range of methods, including time-resolved fluorescence spectroscopy and quantum chemical calculations, to study the processes of electron transfer from a photosensitizer to Pt/g-C₃N₄.

Zoom In Zoom Out Reset image size

In relation to the heteroleptic double-decker terbium(III) phthalocyanine complex [(15C5)₄Pc]Tb(Pc), which behaves as single-molecule magnet (SMM), it was shown that the formation of a supramolecular dimer (Fig. 14 a ) leads to a 1000-fold decrease in the relaxation rate of magnetization at 2 K, and the temperature range in which the hysteresis of magnetization is observed expands from 7 to 15 K. ⁴³ Such an improvement in the magnetic properties is primarily due to ferromagnetic interaction between the terbium ions. The distance between the Tb³⁺ cations (6.2 Å), estimated from NMR spectroscopy data, is in line with the value of 6.249 Å found by X-ray diffraction for a structurally similar six-decker dimer obtained by supramolecular dimerization of the complex [(15C5)₄Pc]Y(Pc)Y(Pc) in the presence of KBPh₄ (see Fig. 14 b ). ⁴⁴ This structure is the first direct evidence for the architecture of supramolecular dimers based on sandwich crown-substituted phthalocyanine complexes of rare earth elements.

Zoom In Zoom Out Reset image size

It should be noted that the nature of the solvent and heat treatment of crown phthalocyanines can also lead to the formation of assemblies with new physicochemical properties due to numerous non-covalent interactions involving crown ether groups, aromatic moieties and solvent molecules. Thus, it was shown that the supramolecular self-assembly of the [(15C5)₄Pc]Ru(Pyz)₂ (13) (Pyz is pyrazine) complex in a mixture of tetrachloroethane and polymer composites with polyvinylcarbazole induced by repeated heat treatment with heating the solution to 70 °C and subsequent cooling to 5 °C gives rise to a nonlinear optical response of the material, which is absent in the case of the monomeric complex (Fig. 15). ⁴⁵

Structure 13 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Hydrogen bonds (H-bonds) play a crucial role in the formation of supramolecular structures based on natural tetrapyrrole macrocycles, providing their high activity and exceptional selectivity in various biochemical processes in life (chlorophyll in photosynthesis, haem proteins in oxygen transport and in enzymatic processes). These structures can arise via the formation of intermolecular hydrogen bonds between appropriate functional groups at the periphery of the tetrapyrrole complex or via axial coordination of proton-donating small molecules to the complex-forming metal, or via a combination of both types of binding. ^{46, 47}

Chlorophyll structures 

Zoom In Zoom Out Reset image size

Fig. Figure 16 shows the hydrogen bonding between bacteriochlorophyll a molecules and tryptophan (Trp) or tyrosine (Tyr) residues in light-harvesting complex II from Gram-negative purple non-sulfur bacteria Rhodoblastus Acidophilus. ⁴⁶ Recently, model studies ⁴⁷ showed that H-bond formation involving the keto group of chlorophyll a (see 16) as a part of supramolecular complex with protein in photosystem II of higher plants allows smooth changes of the energy of electronic transitions in a supramolecular assembly, providing high efficiency of cascade energy transfer by a light harvesting antenna.

Zoom In Zoom Out Reset image size

Hydrogen-bonded supramolecular structures also play a key role in haem protein structure and function. Participation of a distal water molecule coordinated to the haem in the H-bonding to the tyrosine residue Tyr107 (Fig. 17 a )was shown to regulate H₂O₂ binding and activation in model myoglobins, ⁴⁸ while H-bonding to histidine His29 (see Fig. 17 b ) determines their hydrolytic activity. ⁴⁹

Zoom In Zoom Out Reset image size

Similar principles of the assembly of supramolecular systems via hydrogen bonding are used in the case of synthetic porphyrins and their meso-aza-substituted analogues (porphyrazines and phthalocyanines). Over the last decade, this approach has been actively used for the development of both biomimetics and new functional materials. These issues are addressed using the same types of interactions that occur in nature, and the porphyrin macrocycle can serve as both a donor and an acceptor of H-bonds, thus forming dimeric, oligomeric or polymeric framework structures, which have been named hydrogen-bonded organic frameworks (HOFs).

Free porphyrin bases generally have a planar structure, their pyrrole NH groups form intramolecular H-bonds with pyrrolenine nitrogen atoms and are incapable of intermolecular interactions. However, protonation of these molecules usually disrupts the planarity of the macrocycle and changes the H-bonding capacity. For example, due to the saddle-shaped distortion of the macrocycle in the diprotonated forms of porphyrins, especially in the case of dodeca substituted derivatives, NH groups become good H-bond donors and can form supramolecular structures via intermolecular hydrogen bonding to various anions. Thus, various photoactive supramolecular systems with carboxyl-containing redox-active guest molecules capable of photo-induced electron transfer (H₄DPP²⁺, 14a) were derived from dodecaphenyl-substituted porphyrin dication (Fig. 18). ⁵⁰ Using benzyl viologen derivative 15a as an electron acceptor, the H₄DPP²⁺ porphyrin mediated the photoinduced electron transfer from decamethylferrocene (indicated by the blue circle with the letter D in Fig. 18 a ) to the acceptor molecule. In the case of H₄DPP²⁺ binding to the ruthenium(II) polypyridyl complex 15b, intramolecular photoinduced electron transfer was observed. This is the first example of application of protonated porphyrins as photoactive platform for the creation of charge transfer systems via hydrogen bonding.

Haem structure 

Zoom In Zoom Out Reset image size

Structure 14a 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Extended two- and three-dimensional supramolecular structures based on tetrapyrrole macrocycles formed via H-bonds have recently attracted particular interest. Thus, supramolecular assemblies of various dimensionalities were obtained from dodecaaryl-substituted porphyrins 14b – d. ⁵¹

Structures 14b – d 

Zoom In Zoom Out Reset image size

A self-assembly process involving porphyrin 14b containing two p-carboxyphenyl groups in the opposite meso-positions resulted in one-dimensional chains (Fig. 19 a ). Upon the introduction of four p-carboxyphenyl groups, porphyrin 14c forms supramolecular 2D networks (see Fig. 19 c ). The presence of 3,5-dicarboxyphenyl groups in the meso-positions of porphyrin macrocycle 14d results in the formation of H-bonded ladder-type structures (see Fig. 19 d ). An interesting feature of these supramolecular structures is the presence of nanochannels containing anions (see Fig. 19 b ).

Zoom In Zoom Out Reset image size

It has recently been shown that the self-assembly of meso-tetrakis(4-carboxyphenyl)porphyrin-based HOF (H₂TCPP) can be finely tuned by kinetically and thermodynamically controlled blocking of a portion of the carboxyl groups upon interaction with DMF at various temperatures. ⁵² Thus, HOFs with porphyrin – DMF stoichiometric ratios of 1 : 2, 1 : 4 and 1 : 6 were obtained at 120, 130 and 140 °C, respectively. X-Ray diffraction analysis was used to elucidate the nature of the non-covalent interactions responsible for the formation of each type of HOF (Fig. 20). The resulting supramolecular structures were used for the photocatalytic decomposition of 9,10-diphenylanthracene, with the 1 : 6 assembly being the most active. The same assembly showed the highest intrinsic photostability.

Zoom In Zoom Out Reset image size

Three-dimensional HOF-type supramolecular structures can also be obtained by simultaneous introduction of two types of functional groups, e.g., pyridyl and carboxy groups, into the porphyrin molecule. For example, in the Mn ^II and Zn ^II complexes with 5,15-di(4-pyridyl)-10,20-bis(4-carboxyphenyl)porphyrin (16), HOF structures are formed through interaction of carboxyl groups among themselves and with the pyridine moieties not involved in coordination with metal cations (Fig. 21). ⁵³

Zoom In Zoom Out Reset image size

The formation of supramolecular structures through hydrogen bonding is also possible upon the interaction of tetrapyrrole macrocycles of various types. Thus, heteroleptic sandwich complexes of Eu ^III (17a,b) containing octylamine-substituted tetraphenylporphyrin and phthalocyanine ligands were used to form nanoaggregates of various architectures via the formation of meso-N...NHC₈H₁₇-n hydrogen bonds.

Complex 17a containing one octylaminophenyl group forms nanorods organized into J-type aggregates (Fig. 22 a ), while its analogue 17b with four aminoalkyl groups forms extended nanolayers which assemble into H-aggregates (see Fig. 22 b ). It was shown that H-aggregates based on complex 17b have higher conductivity as compared to that of J-aggregates of complex 17a. ⁵⁴

Structures 17 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Porphyrin 18 containing 2,4-diaminotriazine rings in meso-positions was employed to obtain a highly porous supramolecular HOF structure due to several types of noncovalent interactions. ⁵⁵ First of all, molecules 18 form dimers via hydrogen bonding between axially coordinated water molecules and the nitrogen atoms of triazine heterocycles (Fig. 23 a ). Further, these dimers form intermolecular hydrogen bonds involving diaminotriazine moieties (shown in relation to 2,6-diamino-4-phenyltriazine; see Fig. 23 b ), which gives rise to a porous structure (see Fig. 23 c ). Due to specific structural features, this HOF system is capable of selective sorption of CO₂.

Structure 18 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Intermolecular hydrogen bonds play a paramount role in the formation of supramolecular structures of tetrapyrrole metal complexes with an axially coordinated water molecule. The formation of H-bonded assemblies with participation of water molecules is particularly characteristic of Mg ^II and Zn ^II complexes, which efficiently adsorb traces of water from solvents. For example, a recent study ⁵⁶ demonstrated the formation of supramolecular dimers based on magnesium(II) and zinc(II) aqua complexes with unsubstituted phthalocyanine via hydrogen bonding between the coordinated water molecule and meso-nitrogen atoms of the adjacent phthalocyanine molecule; simultaneously, an H-bond was formed with the 4-methylmorpholine solvation molecules (Fig. 24).

Zoom In Zoom Out Reset image size

In the case of meso-aza-substituted porphyrins, in particular porphyrazines 19a – c, meso-nitrogen atoms can also participate in the formation of intermolecular H-bonds. Thus, for porphyrazines with annulated 1,4-diazepine moieties, a unique complementary H-bonds were found to arise between the meso-nitrogen atoms of one molecule and the hydrogen atom of the methylene group having enhanced acidity in the diazepine ring of another molecule, which gives rise to stable dimers (Fig. 25). ^{57, 58}

Structures 19 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

It was found that the tendency of tetrakis(1,4-diazepino)porphyrazines to form dimers is influenced by the nature of both the substituent in the diazepine rings and the solvent. ^{57 – 59} Thus, complex 19a with a 4-tert-butylphenyl group in aprotic solvents (DMF, DMSO) exists in the monomeric form, and upon the addition of water, dimers are formed, which is accompanied by characteristic changes in the absorption spectra. ⁵⁷ Additional stabilization of the dimeric form can occur due to the π – π interactions of the peripheral styryl substituents; therefore, the styryl-substituted complex 19b also exists as dimers in DMSO. ^{58, 59} However, the addition of fluoride anions as strong H-bond acceptors to the solution leads to the transformation of the dimer into a monomer. Since the dimeric form, unlike the monomeric form, does not show fluorescence, diazepinoporphyrazines can be used as responsive fluorescent sensors, for example, to detect water or fluoride anions in aprotic solvents (Fig. 26).

Zoom In Zoom Out Reset image size

Interestingly, the nature of the complexing metal also plays a key role in the formation of H-bonded supramolecular assemblies involving water molecules. Thus, for magnesium(II) and zinc(II) complexes with oxanthrenocyanine M[(α-BuⁿO)₈(15C5)₄Oc], containing n-butoxy groups in the α-position (see Scheme 3), single crystals were obtained. X-Ray diffraction analysis showed that the magnesium complex additionally contains two axially coordinated water molecules, while the zinc complex contains only one such molecule (Fig. 27). ³⁷ The oxanthrene moieties in the complexes have an almost flat structure, with the crown ether moieties participating in the formation of the crystal packing via hydrogen bonding with the water molecules axially coordinated to the metal cations. In the case of the zinc complex, 1D coordination polymer is formed (see Fig. 27 a ), while for the magnesium complex, 2D network arises (see Fig. 27 b ).

Zoom In Zoom Out Reset image size

It is noteworthy that the number of axially coordinated water molecules determines the type of supramolecular assembly and leads to different participation of these crown-containing molecules in the formation of cation-induced supramolecular assemblies.

The ability of crown ether macrocycles to form hydrogen bonds with ammonium ions enables the development of functional multicomponent assemblies of various architectures. The deposition of a solution of tetrakis(15-crown-5)phthalocyanine (H₂[(15C5)₄Pc]) on the surface of an aqueous solution of poly(L-lysine) and subsequent transfer of the assemblies to a quartz surface gave rise to fluorescent sensors capable of detecting chiral substrates, sugars and amino acids (Fig. 28). ⁶⁰ The analytical signals used to distinguish between enantio- and diastereomers were either changes in fluorescence intensity (enhancement, quenching) for amino acids containing additional carboxy groups (Glu, Asp) or changes in fluorescence quenching rate for aromatic amino acids (Phe, His) or D-monosaccharides (glucose, galactose, mannose). The authors explained the observed patterns by changes in the degree of aggregation of the H₂[(15C5)₄Pc] fluorophore applied on the chirality-inducing poly(L-lysine) upon interaction with different stereoisomers. Thus, the degree of aggregation increased upon interaction with D-enantiomer of glutamic acid, leading to fluorescence quenching, while in the case of L-enantiomer, a decrease in the degree of aggregation was observed and the fluorescence intensity increased. This conclusion was based, in particular, on atomic force microscopy data.

Zoom In Zoom Out Reset image size

Disubstituted ammonium cations are able to pass through fairly large crown ether macrocycles to form assemblies with the rotaxane topology. This approach was used to prepare rotaxanes based on tetrakis(benzo-24-crown-8) phthalocyanine M(20) and porphyrins M(21), M(22) complexes containing four dialkylammonium substituents with terminal azide groups (Fig. 29 a ). For the introduction of bulky stoppers preventing the dissociation of rotaxanes, click reactions of azide groups with triethyl phosphite ⁶¹ or with a cyclic strained alkyne were used; ⁶² this gave supramolecular structures [M(20) · M(21)] or [M(20) · M(22)].

Zoom In Zoom Out Reset image size

It was shown that the reversible deprotonation of ammonium and amidodiester groups in rotaxane [Cu(20) · Cu(21)] makes it possible to control the intramolecular distance between the porphyrin and phthalocyanine components, which affects the photophysical and magnetic properties of the material. ⁶¹ According to EPR and magnetochemical studies performed for a complex containing two copper(II) ions, the amplitude of motion was about 5 Å (see Fig. 29 b ).

The presence of positively charged ammonium groups in rotaxane leads to interaction with anionic porphyrins to form heteronuclear supramolecular assemblies with a strictly defined arrangement of metal centres. ^{63, 64} The architectural control achieved in this way made it possible to develop a catalyst with tunable properties based on rotaxane [Fe(20) · (μ-N) · Fe(22)] containing a (μ-N)Fe₂ catalytic site for the ethane oxidation reaction (Fig. 30 a ). ⁶⁵ The original rotaxane showed TON of 195 for ethane oxidation by hydrogen peroxide (60 °C, 24 h), whereas supramolecular assemblies Cu(23) and Ni(23) formed by rotaxane with copper and nickel meso-tetrakis(4-sulfophenyl)porphyrin complexes (MTSPP) had TONs of 376 and 394 respectively (see Fig. 30 a ,b). These values surpass those found for iron μ-nitridobis(tetra-tert-butylphthalocyanine) (TON = 289), which reported earlier to be the most efficient catalyst for this reaction. ⁶⁶ An electrochemical study showed that the interaction of rotaxane with porphyrin M(23) leads to a decrease in the oxidation potential of the complex. This indicates the electron-donating nature of MTSPP with respect to rotaxane, and the enhanced catalytic activity is presumably due to facilitation of the heterolytic cleavage of the O–O bond in hydroperoxy complex A to form the highly reactive oxo complex B (see Fig. 30 c ).

Zoom In Zoom Out Reset image size

Thus, the presented results demonstrate a great potential of tetrapyrrole compounds as components of supramolecular materials. Prospects for the development of this research area include the molecular design of novel macrocycles and their complexes, development of synthetic approaches related to the preparation of compounds with desired properties, and investigation of supramolecular assembly processes that determine the architecture and properties of the resulting assemblies. Each of these stages contributes to the development of all fields of modern chemistry, and the functional diversity of the resulting materials includes both biomimetic systems (catalysts and photosensitizers) and assemblies that exhibit properties not found in nature, such as molecular magnetism and optical confinement. Finally, the molecular recognition underlying the supramolecular assembly of tetrapyrroles enables the development of highly selective sensor systems and smart materials.

It is known that the properties of any material are determined by the nature of its constituent components and by the spatial arrangement of these components relative to each other within the material. In this regard, the methods of supramolecular chemistry as a science that studies the processes of self-organization of molecules at the nanoscale represent an effective modern tools for solving various applied problems, in particular, owing to the development of sequential self-assembly of molecules according to the bottom-up approach with a built-in recognition programme. This gives rise to smart materials with unique properties due to control over the spatial organization of molecules.

Several basic approaches to the targeted design of homoleptic (consisting of ligands of the same type) and heteroleptic (consisting of ligands of different types) MOPs have been developed as a result of studying the self-organization of supramolecular systems. ⁹⁰ The formation of structures with different topologies depends on the arrangement of binding centres in both organic (ligand) and inorganic (metal cation, metal complex) complementary molecular building blocks.

Various methods for the synthesis of functional MOPs based on different types of molecular recognition between certain functional groups and metal centres have been described in the literature. ⁹¹ There are known MOPs that have been obtained by combining one metal cation with a rigid coordination sphere and complex metal ligands containing another type of metal. This approach was utilized to prepare novel heterometallic coordination cages. ⁹² The use of ligands and metal complexes based on cavitands such as calix[4]arenes, ⁹³ calix[4]resorcinarenes ^{94 – 97} and pyrogallol[4]arenes ^{98, 99} displaying a pre-organized macrocyclic structure, served for studying the regularities of formation of macrocyclic coordination cages and coordination polymers based on them. ^{100 – 105}

The data available from the literature on the formation of the most common motifs of homometallic coordination cages (and their complexes) based on spatially pre-organized acyclic structurally rigid (aromatic) polytopic ligands are analyzed below. Such ligands contain several (two to four) homogeneous coordination sites that participate in the molecular recognition with metal cations due to their mutual complementarity in terms of spatial and electronic factors. The selected models of molecular self-assembly of coordination cages are classified according to increasing coordination capacity of ligands containing monodentate or bidentate (chelating) coordination sites and metal centres (metal connecting nodes).

One of the first types of molecular recognition forming the basis for a large number of coordination cages was the interaction of cations (metal complexes) that have two (angular, or V-shaped metal connecting node) or four (tetragonal metal connecting node) vacant sites for coordination to ligands containing monodentate coordination sites (Table 1) in their coordination sphere. The V-shaped metal connecting nodes usually include cations with a square or octahedral coordination sphere (Pd²⁺, Pt²⁺, Ru²⁺, Ir²⁺, Rh²⁺). In this sphere, two or four sites are occupied by bidentate chelating co-ligands coordinated in the cis-position (ethylene-1,2-diamine, ethylene-1,2-diphosphine, cyclohexane-1,2-diamine, 2,2'-dipyridine, etc.), while the remaining two (or four) sites in the coordination sphere are available for coordination of other ligands (Fig. 31). By combining V-type metal connecting nodes 24 – 28 with three or four monodentate ligands, usually having a rigid structure (angular, trigonal or tetragonal geometry, respectively), and pyridyl- or imidazolyl-containing ligands, one can generate coordination cages of different geometries.

Table 1. Geometry of the coordination cage based on V-shaped (angular) metal connecting nodes and polytopic ligands with monodentate binding sites. ^106-113

Note. Here and in Tables 2 — 5, the yellow and red triangles denote the complementary binding sites of the ligand and the metal connecting node.Tildes in structures 24 — 27 show the attachment site of the binding polytopic ligands.

Zoom In Zoom Out Reset image size

One of the first coordination cages was obtained using intermolecular recognition of this type: interaction of tri(4-pyridyl)triazine 29 with metal connecting node 24 afforded cationic octahedral structure [(24)₆-(29)₄]¹²⁺ (type A¹, see Table 1) (Fig. 32). This electron-deficient structure was subsequently used to adsorb water molecules, nitrate ions and perfluorinated hydrocarbon molecules. ¹⁰⁶ The size of the inner cavity can be characterized by the distance between the palladium atoms at the opposite vertices of the octahedron: d_Pd–Pd = 18.8 Å.

Zoom In Zoom Out Reset image size

Other ways of obtaining similar cages based on ligand 29 with various metal connecting nodes to fabricate new functional materials have also been described in the literature. In particular, mention should be made of redox-active cage [(27)₆-(29)₄]¹²⁺ (type A¹, see Table 1) with a potential between 0 and –2 V, which was obtained from complex 27 as the V-shaped metal connecting node. This system is of interest as a promising electron-switchable supramolecular container in an organic medium for binding – release processes of various substrates. ¹⁰⁷

A similar supramolecular structure was obtained from 1,3,5tris(4-pyridylethynyl)trifluorobenzene 30 and complex 25. Compared with cage [(24)₆-(29)₄]¹²⁺, the analogue [(25)₆-(30)₄]¹²⁺ (type A¹, see Table 1) (Fig. 33), as expected, had a larger inner cavity (d_Pd–Pd = 24.1 Å). This fact demonstrates the possibility of controlling the size of the inner cavity of the cage by varying the ligand used to bind certain guest molecules. ¹⁰⁸

Zoom In Zoom Out Reset image size

Using 4,4'-bipyridine (bpy) as a co-ligand, cationic cage [(26)₆-(29)₂(bpy)₃]¹²⁺ (Fig. 34) was obtained as a trigonal prism (type B, see Table 1). In this structure, two parallel faces, represented by molecules 29 coordinated to the metal centre of complex 26, were spatially separated by three 4,4'-bipyridine molecules due to interaction at the second vacant site in the metal coordination sphere. This cage was successfully used for the selective recognition of mononucleotides in an aqueous medium and for the synthesis of dinucleotides. ¹⁰⁹ In addition, a similar cage based on the tetramethyl derivative of 4,4'-bipyridine, ligand 29 and complex 1,2-ethylenediamine-coordinated Pd ^II cations was used for the formation of an inclusion complex of naphthalene diimide with corannulene, which was accompanied by flattening of the bowl-like structure of the latter. ¹¹⁰

Zoom In Zoom Out Reset image size

Coordination cages are easily formed when the pyridine ligands are replaced with molecules containing imidazole binding groups. Thus, an octahedral supramolecular structure of type A² (see Table 1) was obtained via molecular recognition of the tris(imidazole) derivative 31 and complex 24. The cage [(24)₆-(31)₄]¹²⁺ (Fig. 35) is more flexible than [(24)₆-(29)₄]¹²⁺, which makes it adaptable to binding of particular substrates. For example, it was shown that this cage is capable of accomodating molecules of photosensitive compounds such as azobenzene and a spiropyran derivative, which can be reversibly uptaken and released from the cage surface under the action of light; this was used for the design of data storage media. ^{111, 112}

Zoom In Zoom Out Reset image size

Using this method of molecular recognition, it s possible to obtain, apart from octahedral structures, neutral coordination cage [(28)₄-(32)₂] (Fig. 36) with a cubic geometry (type C, see Table 1). This was demonstrated for metal complex 28 and tetragonal planar tetrasubstituted monodentate organic linker 32 containing four pyridine moieties. It should be noted that complex 28, used as a metallic bisconnecting node, was prepared in advance by the reaction of [(cymene)RuCl₂]₂ (cymene is 4-isopropyltoluene) with 3,4-dimethoxyfuran-2,5-dicarboxylate. Crystallization experiments carried out in various solvents showed that cage [(28)₄-(32)₂] is converted to dimer [(28)₂-(32)] upon recrystallization from a solution. ¹¹³

Zoom In Zoom Out Reset image size

One more method of obtaining cages based on polytopic monodentate ligands is based on molecular recognition between cage-forming monodentate ligands and metal cations with a characteristic square coordination sphere (Pd ^II , Pt ^II ) in which all four sites are free for coordination or octahedral cations in which two coordination sites in the trans position are occupied by solvent molecules or anions (Table 2). By varying the geometry of the monodentate binding sites of the ligand (Fig. 37) and the type of tetragonal planar metal connecting node, it is possible to obtain dimeric, octahedral and cuboctahedral coordination cages.

Table 2. Geometry of coordination cage based on tetragonal metal connecting nodes and polytopic ligands with monodentate binding sites ^114–121

Zoom In Zoom Out Reset image size

The coordination of Pd ^II to bis(3-pyridylethynyl) ligand 33, which is capable of reversible photoisomerization under the action of light owing to the bis(thienyl)ethene moiety, leads to the cationic cage (Fig. 38) of a capsular dimeric structure (type A, see Table 2). This cage can be used for controlled binding and release of guest molecules. This was demonstrated, in particular, in relation to photocontrolled adsorption – desorption of the spherical [B₁₂F₁₂]^2– anion. ^{114, 115}

Zoom In Zoom Out Reset image size

Molecular recognition between molecules containing polyaromatic substituents and pyridyl binding sites and palladium(II) cations afforded coordination cages. The cages were used for the adsorption of aromatic pigments and dyes from water, which is a relevant problem related to environmental protection. ⁸¹ Cages with anthracene moieties, for example (type A, see Table 2) (Fig. 39) are capable of luminescent response depending on the guest adsorbed in their cavity. The presence of coumarin molecules leads to increase in the fluorescence of this system. ^{116, 117} In addition, this system was used as a matrix for MALDI-TOF mass spectrometry experiments for the detection of single sulfur clusters [S₆], [S₈] and [S₁₂]. ¹¹⁸

Zoom In Zoom Out Reset image size

When naphthalene diimide derivative 35 acts as a V-shaped cage-forming dipyridyl ligand, supramolecular cage is obtained, which has a similar structuralmotif (type A, see Table 2). When dissolved in DMSO, it forms a gel and exhibits selective sensing activity towards pyrene and many anionic dyes (methylene blue and methyl orange). ¹¹⁹

The reaction of Pd ^II and Pt ^II cations with tripyridyl derivative 36 with trigonal binding centres produced coordination cages (Fig. 40) of octahedral structure (type B, see Table 2). These cages were used for binding to neutral aromatic molecules such as anthracene, pyrene and 1,8-naphthalimide. They are also able to bind selectively to the anionic guest, p-toluenesulfonate, due to spatial complementarity. ¹²⁰

Zoom In Zoom Out Reset image size

Neutral coordination cage [(Cu ^II (OH)Br)₆-(37)₈] (Fig. 41) of octahedral shape (type B, see Table 2), in which the octahedral Cu ^II cation acts as a square metal connecting node, was obtained using a relatively flexible tridentate ligand based on tripyridyl derivative 37. The resulting cage was found to form smart multifunctional vesicles and to participate in controlled delivery of anticancer drugs, such as doxorubicin (DOX), to living cells. Control of the drug release from the vesicle can be achieved by switching the pH of the medium. In addition, these vesicles have demonstrated high stability in biological media, allowing them to be used as biomembrane models. ¹²¹

Zoom In Zoom Out Reset image size

Coordination cages obtained via complementary intermolecular interaction of this type can be divided into two groups, depending on the number of metal cations in the metal centres: mononuclear and polynuclear ones.

Cages with mononuclear metal connecting nodes provide recognition between ligands having two, three or four terminally located chelating moieties and transition metal cations, usually Fe ^II/III ,Co ^II ,Zn ^II ,Ga ^IV ,Sn ^IV ,Ti ^IV , with an octahedral coordination sphere. This results in the formation of either a trigonal metalbinding centre (upon reaction with bidentate functional groups) or a V-shaped metal connecting node (when a tridentate chelating moiety is used by the ligand to bind to a metal) (Table 3). Ligands with N-donor centres, such as Schiff bases, bipyridyl or similar bidentate moieties, can then act as chelates (Fig. 42). Among chelating ligands containing O-donor cavities, catechols are most encountered. Mixed-type chelating cavities containing both O- and N-donor atoms have also been described in the literature. ¹²² Thus, depending on the number and geometry (linear, V-shaped, trigonal or tetragonal) of the chelating cavities they have, ligand molecules can form a dimeric, tetrahedral or cubic coordination cage upon interaction with a trigonal metal connecting node. It should be noted that due to the chelating type of ligand binding to the metal cation possessing octahedral geometry, both D and L enantiomers can be formed (Fig. 43). ^{123, 124}

Table 3. Geometry of the coordination cage formed by trigonal and linear metal connecting nodes and polytopic ligands with bidentate and tridentate chelating binding sites. ^122-138

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

For ligands 38 and 40, containing two and three catechol chelating moiety, respectively, anionic tetrahedral coordination cages (Fig. 44) and were obtained (Fig. 45), in which hard octahedral cations M ^IV = Ga, Ti, Sn were used as the metal coordination centres. Whereas in the cages formed by dicatechol derivative 38, organic ligands are located along the edges of tetrahedra (type A, see Table 3), ^{125, 126} in the supramolecular structure based on ligand 40 with a rigid trigonal arrangement of catechol binding sites, organic moieties of the cage occupy the positions of faces of the supramolecular tetrahedron (type C, see Table 3). ¹²⁷ It was shown that such hollow structures can act as catalysts for organic reactions, ⁷³ in particular, the cage accelerates the formation of a C–C single bond in the reductive elimination (cross-coupling reaction) taking place in the coordination sphere of gold cations. ¹²⁵

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The coordination cages of the cationic type were obtained using ligands containing N-donor chelating moieties. The most common examples of coordination cages reported in the literature are systems constructed using molecular recognition between ligands with pyridine pyrazolyl coordination sites. For example, ligand 39a in combination with Co ^II cations forms cage (Fig. 46) of cubic geometry (type B, see Table 3), which showed catalytic properties in the Kemp elimination reaction (synthesis of cyanophenolates from benzoxazoles). ¹²⁸ It should be noted that the hydrophobic cavity of the cubic cage demonstrated selective adsorption properties towards carbon dioxide in the presence of nitrogen, ¹²⁹ as well as towards alkylphosphonates ¹³⁰ and isoquinoline N-oxide. ¹³¹

Zoom In Zoom Out Reset image size

At the same time, Piper et al. ¹³² found that a similar cage is able to provide photoelectron charge transfer from naphthalene substituents to the guests adsorbed inside the cavity, which can be used in photocatalysis. In addition, the possibility of obtaining isostructural cages (type B, see Table 3) with different metal cations (M = Ru, Os), meant for enhancing the photoadsorption properties and consequently improving the photoelectron transfer function, was shown.

Upon the interaction of ligand 41, which has a trigonal arrangement of dipyridyl coordinating moieties, with Fe ^II cations, coordination cage of tetrahedral geometry (type C, Table 3) was obtained. It exhibits distinct electron-deficient properties, which was utilized for selective binding of the triflate anion (triflate is trifluoromethanesulfonate). In addition, such systems are capable of adsorbing polyaromatic contaminants from aqueous media. ⁸¹ Importantly, many Fe ^II -based cages demonstrate spin-crossover properties. ¹³⁴

Another effective and common recognition model used to produce coordination cages is the interaction of octahedral cations with bidentate N-donor groups (analogues of dipyridyl and pyridylpyrazolyl moieties) that contain iminopyridyl binding sites (pyridine Schiff bases). The advantages of this approach include the ease of formation of such functional groups in situ and the ability to design supramolecular cages by varying both the amine and aldehyde components.

Compound 42, containing three iminopyridyl groups and an electron-deficient boron atom, was shown to be an excellent cage-forming ligand. It reacts with Fe ^II cations to give cage (Fig. 47) with tetrahedral structure (type C, see Table 3) capable of selective binding of fluoride anions. ¹³⁵

Zoom In Zoom Out Reset image size

The use of ligand 43, which has a rigid structure with a tetragonal arrangement of iminopyridylcoordinating moieties, and Zn ^II cations yielded supramolecular cage (Fig. 48), which has a cubic geometry (type D, see Table 3). Owing to CH – π and hydrophobic interactions, this structure is able to effectively encapsulate anthranilamides and aromatic aldehydes, as well as to accelerate successive condensation – amino addition reactions of guests adsorbed in the cavity of the cage with participation of 2,3-dihydroquinazolinone. ¹³⁶

Zoom In Zoom Out Reset image size

In addition, similar cubic cages were synthesized by using porphyrin 44 and its complexes with Ni ^II and Zn ^II as building blocks. In this case, Fe ^II cations behaved as metal connecting nodes. The hollow positively charged structures , enclosed by electron-rich porphyrin moieties, demonstrated high selective adsorption capacity towards polyaromatic guests, such as coronene and higher fullerenes (C₇₀–C₈₄). ¹³⁸

An excellent result was obtained in the formation of coordination cages using tridentate iminophenanthrolyl binding sites located at the terminal positions of meso-substituents in porphyrin 45. The interaction of compound 45 with zinc(II) cations gives a pseudo-cuboctahedral coordination cage (Fig. 49) (type E, see Table 3). In this cage, metal atoms act as V-shaped (angular) metal connecting nodes, which are coordinated to two tridentate moieties of iminophenanthrolyl substituents incorporated in two adjacent porphyrin molecules 45. A study of the receptor properties of the resulting cage demonstrated its affinity for fullerene molecules with efficient formation of stable 1 : 2 inclusion complexes. This cage was also found to bind various spherical anions , and . ¹³⁷

Zoom In Zoom Out Reset image size

Polynuclear species can be considered as one more example of trigonal metal connecting nodes that have been used to prepare MOPs with polytopic ligands containing chelating groups. In particular, it was shown that the complex Cp₂Zr ^IV Cl₂ (Cp is cyclopentadienyl) is hydrolyzed in situ in the presence of carboxylates to give triuclear cluster (denoted by [Zr₃]³⁺), in which three carboxy groups point outwards, thus forming a conical trigonal metal connecting node (Fig. 50). ¹³⁹ As the organic counterparts, dicarboxyl ligands are often used (Fig. 51), leading to the formation of capsular dimeric and tetrahedral coordination cages (Table 4).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 4. Geometry of the coordination cage based on [Zr3]³⁺ trigonal metal connecting nodes and linear ligands containing carboxyl binding sites. ^139–144

One of the first coordination structures obtained by molecular recognition of this type was the neutral coordination cage [(Zr₃)₄-(46)₆] (Fig. 52) with a tetrahedral geometry (type A, see Table 4). ¹³⁹ This structure was subsequently modified by introducing amino groups into the edge positions of the cage involving 2-aminoterephthalic acid (47), resulting in the cage [(Zr₃)₄-(47)₆] of a similar structure (type A, see Table 4). This cage was found to exhibit pronounced photocatalytic properties in the hydrogen evolution reactions. ¹⁴⁰ Besides, the possibility of postsynthetic modification was demonstrated: two cages were cross-linked by the acylation reaction with a dicarboxylic acid chloride containing a flexible hydrocarbon spacer. ¹⁴¹ This gave a new polymeric material with a microporous structure. Thus, by covalent cross-linking of coordination cages, it is possible to design new materials with controlled porosity, and also to obtain new adsorbents with tunable properties by varying the building blocks (coordination cages and cross-linking reagents). Moreover, this strategy allowed the synthesis of a microporous polyfunctional hybrid membrane based on imine binding groups, which had water permeability and antibacterial properties. ¹⁴²

Zoom In Zoom Out Reset image size

Another method of functionalization of cages of this type consists in the introduction of additional binding sites into the structure of the organic cage-forming ligand. Coordination cage [(Zr₃)₄-(48)₆] of tetrahedral geometry (type A, see Table 4) has been obtained; the structure was based on dipyridine chelating moieties capable of coordination to d-element cations. The use of Re(CO)₃Cl⁺ complex cations gave rise to nanoreactors, which were successfully used for photocatalytic reduction of carbon dioxide. ¹⁴⁴

Coordination capsules [(Zr₃)₂-(49)₃] (Fig. 53) (type B, see Table 4) showed an interesting result in terms of selective adsorption and also proton conductivity. It was found that the above complexes are capable of selective binding of carbon dioxide in the presence of nitrogen and oxygen molecules; owing to the presence of sulfonate groups, they can adsorb a large number of water molecules (via hydrogen bonding) promoting proton conductivity. ¹⁴³

Zoom In Zoom Out Reset image size

It was shown that in some cases, the interaction of d-element cations, such as Cr ^II , Co ^II , Fe ^II , Cu ^II , Mo ^II , Rh ^II , with carboxyl ligands under solvothermal conditions leads to bidentate tetragonal paddle wheel-like metal connecting nodes . They consist of two metal cations that are in a pentagonal pyramidal coordination sphere and are linked together at the base of the pyramids by four carboxylate groups of adjacent organic linkers (Fig. 54). A number of dimeric, octahedral and cuboctahedral coordination cages have been synthesized based on molecular recognition of this type (Table 5). Li et al. ¹⁴⁵ analyzed the effect of the angle between the coordination sites of the V-shaped organic linker and its size on the volume and supramolecular motif of the resulting coordination cage.

Zoom In Zoom Out Reset image size

Table 5. Geometry of the coordination cage based on tetragonal binuclear metal connecting nodes and angular and trigonal ligands containing carboxyl binding sites. ^145-161

It has been established that dimeric capsules (type A, see Table 5) were formed most often when the carboxylate sites of the ligands were located parallel to each other in the molecule. In particular, the reaction of ligand 50 with Cu ^II cations gave neutral coordination capsule (Fig. 55), capable of displaying the so-called gate opening effect. It consists in the proneness of a crystalline material to exhibit a strong cooperative increase in the adsorption capacity, accompanied by a reversible solid-phase transition in which the crystallinity is preserved. It was found that this dimeric coordination cage is indifferent with respect to adsorption of nitrogen molecules, but shows a gate opening effect in the case of CO₂, allowing the adsorption of 12 CO₂ molecules per molecule of the metal-organic complex in the crystalline phase. ¹⁴⁶

Zoom In Zoom Out Reset image size

While coordination cages based on metal connecting nodes consisting of coordinatively unsaturated Cu ^II cations exhibit adsorption properties towards carbon dioxide molecules, for octahedral (M=Cu, Cr, Mo; type B, see Table 5 and Fig. 56) ¹⁴⁷ and cuboctahedral systems (M=Cr, Mo, Ru; type C, see Table 5 and Fig. 57), ¹⁴⁸ a pronounced affinity towards methane molecules was found. Moreover, the adsorption properties of the material were improved by binding the cuboctahedralcages (M=Fe, Co) via additional coordination of metal connecting nodes to 1,4-diazabicyclo[2.2.2]octane (DABCO), which resulted in coordinated cross-linked molecular cages. ¹⁴⁸

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The cages of similar cuboctahedral structure (type C, see Table 5) containing two types of d-elements (M₁ = Cu, Ni, Zn; M₂=Pd) in the metal connecting node behave as efficient adsorbents for binding hydrogen molecules due to the presence of coordinatively unsaturated metal cations in the desolvated (activated) coordination cage structure. ¹⁵⁰ The use of ligand 52c also resulted in a similar cuboctahedral structure (type C, see Table 5), capable of absorbing organic dye molecules and iodine molecules from aqueous and ethanol solutions. ¹⁵¹

Examples of coordination cage-based systems obtained by post-synthetic modification to enhance their adsorption properties and develop new polymeric functional materials with controlled porosity are often encountered in the literature. Two strategies are employed for this purpose. The first one is to use the possibilities associated with the additional coordination to coordinatively unsaturated metal cations; the second one is based on covalent bonding of cages to form a stable membrane with controlled porosity. In both cases, the coordination cages play the role of secondary building blocks. The coordination cage (type C, see Table 5) was shown to form coordination-bonded aggregates with constant porosity upon the reaction with a bis(imidazolyl) derivative containing a phenyl spacer. ¹⁵² The adsorption properties of this solid phase material towards methyl orange dye were superior to those of the initial coordination cage monomer. ¹⁵³

It should be noted that the indicated molecular cage was used to fabricate ultra-thin membranes (2.5 nm) for carbon dioxide extraction from gas mixtures by the Langmuir – Blodgett method. ¹⁵⁴ Binding by relatively labile sites in the coordination sphere of metal cations in the metal connecting node of (type C, see Tabel 5) was applied to separate thiophene and thiophane molecules in heterogeneous liquid systems (water – organic solvent). ¹⁵⁵

The possibility of depositing cages of this type on the surface of graphene and a single-walled carbon nanotube has been demonstrated. It was found that the thus obtained composite cage-based material (type C, see Table 5) shows catalytic properties in the synthesis of propylene carbonate from propylene oxide and CO₂. ¹⁵⁶

As noted above, covalent binding of coordination cages to one another was used to create new polymeric membranes. Cages , where M = Cu, Rh, and (type C, see Table 5) were copolymerized with each other and with n-butyl methacrylate ¹⁵⁷ and styrene. ¹⁵⁸ The resulting microporous materials had enhanced adsorption properties towards carbon dioxide and methylene blue dye compared to the initial monomer cages. This is probably due to the increased structural stability of the cage pores in the covalently bonded polymer matrix and to the appearance of new pores as a result of copolymerization.

An important achievement in the development of systems simulating physiological processes was the synthesis of the cage (type C, see Table 5). It showed high efficiency as a catalyst for hydrogen peroxide decomposition at a pH of the medium close to that of a living cell. ¹⁵⁹

It is worth noting that varying the nature of the substituent in the isophthalic acid makes it possible to obtain coordination cages with photo-switchable adsorption properties. Coordination cage (type C, see Table 5) based on a ligand containing a photoisomerizable azophenyl moiety is capable of controlled adsorption of methylene blue dye (in solid form) from solutions and release of the dye upon irradiation with UV light. ¹⁶⁰ In addition, deposition of this cage on a mesoporous silica gel surface enhances its photo-switchable adsorption properties towards propylene molecules and brilliant blue dye. This is due to the fact that aggregation, which decreases the adsorption capacity, is less pronounced for the cages deposited on silica gel. ¹⁶¹

In addition to the dicarboxylate ligands, ligand 53, which has a triangular geometry of the carboxyl binding sites, was also used to produce a cage-based molecular recognition of the indicated type. This resulted in the octahedral supramolecular structure (type D, see Table 5 and Fig. 58). The study of its adsorption properties in the crystalline phase showed that this structure efficiently adsorbs carbon dioxide molecules under ambient conditions. ¹⁴⁹

Zoom In Zoom Out Reset image size

Thus, in this Section, we summarized the main types of molecular recognition using different building blocks to produce supramolecular functional coordination cages. It was shown that by varying the geometry and nature of the binding sites of both the ligand and the metal connecting node, their structure and properties can be specifically tuned up. The use of molecular recognition principles makes it possible to control the spontaneous molecular self-assembly of MOPs and produce new materials with specific adsorption, catalytic, sensing and biological properties.

In recent years, the annual number of publications focusing on metal-organic cages has been increasing. The authors pay special attention to the description of possible ways to stabilize these nanoscale structures in order to find the optimal balance between their kinetic stability and the potential for use as materials for various applications with a controlled porosity. In this regard, approaches to the search for new models of molecular recognition are actively developing, new building blocks are involved in the process, for example, calixarenes ^{162 – 165} and resorcinarenes, ^{166, 167} which expands the possibilities of the design of discrete nanoporous structures in order to endow them with the desired functional properties combined with increased stability. One of the possible options for increasing the porosity of such materials is fabrication discrete multicage structures based on a combination of the presented types of molecular recognition. ¹⁶⁸

Covalently bonded polymeric functional membranes, micelles, gels and nanoscale films have been obtained using modern methods of post-synthetic modification of MOPs. ¹⁶⁹ In the future, coordination cages could become parts of many new smart materials based on molecular host – guest recognition and constructed using the bottom-up approach, namely, membrane reactors, efficient molecular transport systems for drug delivery to the cells, film molecular separators, selective adsorbents, etc.

Mixtures of xylene isomers (molecular formula C₈H₁₀) are formed in catalytic reforming and toluene disproportionation processes; xylenes are also present in light petroleum coke products and pyrolysis gasoline. The world production output of xylene isomers is ∼40 million tons per year. ¹⁸⁴ The separation of these isomers is considered to be highly important for modern chemical industry. ¹⁸⁵ Xylenes are separated from other accompanying aromatic hydrocarbons (mainly benzene and toluene) by fractional distillation ¹⁸⁶ and then single isomers are isolated, of which p- and o-xylenes are most valuable for industry. p-Xylene is then oxidized to terephthalic acid or dimethyl terephthalate, which are used as monomers to produce polyesters, first of all, poly(ethylene terephthalate) (PETP). o-Xylene is oxidized to phthalic anhydride, which serves for the production of phthalates, used as plasticizers for polymers. m-Xylene is converted to isophthalic acid — a monomer for the synthesis of polyesters — and can also be isomerized to p- and o-xylenes. Xylene isomers have similar physicochemical properties (e.g., the boiling points of o-, m- and p-xylenes are 144.4, 139.1 and 138.4 °C, respectively); ¹⁸⁷ therefore, separation of these isomers is a challenge. The difference of 5 °C between the boiling points of o-xylene and the other two isomers allows separation of o-xylene by distillation on high-performance columns with high reflux ratios. The similarity of the boiling points of m- and p-xylenes makes their separation by distillation virtually impossible; therefore, pure para-isomer is isolated by crystallization (the melting points of ortho-, meta- and para-isomers amount to –25.2, –47.9 and 13.3 °C, respectively) ¹⁸⁷ or by adsorption. The natural mineral Faujasite ^{188, 189} and synthetic zeolites ZSM-5, ¹⁹⁰ SMZ-100 ¹⁹¹ and HZSM-5 ¹⁹² were used as sorbents. High industrial significance of p-xylene stimulates further search for new methods for separating this isomer from the mixture; therefore, extractive distillation ¹⁹³ and simulated moving bed chromatography ^{194, 195} procedures were recently proposed.

The possibility of wide variation of the pore size and topology of metal-organic frameworks and the nature of functional groups on their surface can be used to tune MOFs for supramolecular recognition of a desired molecule, which makes them fairly promising sorbents for the separation of xylene isomers. The results of the first studies of inclusion of p-xylene molecules into MOF based on zinc ions and napthalene-2,6-dicarboxylic acid (H₂ndc) were reported in 2004. ¹⁹⁶ The authors demonstrated selective inclusion of the para-isomer in the framework pores in the presence of an equivalent amount of ortho- or meta-isomer and determined the crystal structure of the p-xylene inclusion compound described by the formula {[Zn₂(ndc)₂(DMF)₂] · C₈H₁₀}_n . Quantum chemical calculations showed that inclusion of p-xylene into the framework causes the least structural rearrangement of the unit cell, which accounts for the selectivity to this particular isomer.

The inclusion compounds of o-, m- and p-xylenes into the complex {[Cd(ClO₄)₂(L¹)₂] · H₂O}_n (L¹ is 4-amino-3,5-bis[3-(4-pyridyl)phenyl]-1,2,4-triazole) were studied by X-ray crystallography. ¹⁹⁷ It was found that the guest molecules interact with the framework via the C–H...π-contacts with interatomic distances in the range of 2.9–3.2 Å. A study of the selectivity of formation of inclusion compounds with three xylene isomers being present simultaneously demonstrated that o- and m-xylenes are adsorbed in approximately equal amounts, whereas p-xylene does not enter the framework pores under these conditions.

The complex [VO(bdc)]_n (commercial code MIL-47, bdc is the terephthalate anion) was studied most comprehensively as applied to the liquid-phase separation of xylenes. ¹⁹⁸ This framework demonstrates selective adsorption of p-xylene from mixtures with other isomers and with ethylbenzene with a selectivity coefficient of ∼3. A powder X-ray diffraction study of inclusion compounds formed by xylenes and MIL-47 with Rietveld refinement showed that encapsulation of the guest (xylene) into framework pores almost does not change the unit cell parameters, although each unit cell accommodates two xylene molecules The p-xylene molecules are located in parallel planes, with methyl groups being shifted relative to each other, which ensures most efficient π – π interactions. Somewhat weaker interactions are found for o-xylene, and the interactions involving m-xylene are the weakest because of steric hindrance, which prevents molecules from being arranged in parallel planes. ¹⁹⁸ Finsy et al. ¹⁹⁹ studied the adsorption of xylenes with the MIL-47 framework from the gas phase and demonstrated that virtually no selectivity is observed at low partial pressures, but with increasing degree of filling of the sorbent channels with xylene molecules, the selectivity increases. As a result, it was found that the adsorption selectivity can be controlled by varying pressure or temperature. The adjustment of sorbent activation conditions makes it possible to attain the selectivity coefficient of 14 for a p-xylene–m-xylene binary mixture. ²⁰⁰

Chromium terephthalate [Cr₃O(OH)(H₂O)₂(bdc)₃]_n (MIL-101) was used as the stationary phase for the analytical separation of xylenes by gas chromatography ^{201, 202} and high-performance liquid chromatography. ²⁰³ Analysis of adsorption isotherms showed that aromatic hydrocarbons are more strongly adsorbed by the MIL-101 framework than aliphatic hydrocarbons, with the highest Henry constant being inherent in o-xylene. ²⁰² This framework also demonstrated record-high uptake capacity of xylenes equal to ∼120 mass %. ²⁰²

Huang et al. ²⁰⁴ investigated sorption of xylene from an equimolar liquid mixture of the three isomers by zinc biphenyl-3,5-dicarboxylate [Zn(μ₄-L²)]_n , which had extended one-dimensional channels with a diameter of ∼1.7 nm. A predominant inclusion of p-xylene into this framework was observed.

Causes for the selective adsorption of p-xylene by the boron imidazolate framework [Zn₂(BH(mim)₃)₂(obb)]_n , where mim is 2-methylimidazolate, obb is 4,4'- oxybis(benzoate), (BIF-20), were addressed by Lyndon et al. ²⁰⁵ Despite high and nearly equal uptake capacities of o- and p-xylene (∼30 mass %), the p-xylene adsorption had a selectivity coefficient of ∼3 due to the smaller size of its molecule compared with the ortho-isomer, which was characterized by faster diffusion. The crystal structures of p-and o-xylene inclusion compounds in the cavities of this framework were determined. According to the results, the ortho-isomer molecules are disordered to a much higher degree, which indicates that they weakly interact with the framework. The molecules of p-xylene are located near the imidazole borate linkers and are involved in π – π interactions with the imidazole rings and boron atoms (Fig. 59). It is noteworthy that the aperture of framework pores (∼3 Å) is markedly smaller than the kinetic diameter of xylene molecules (∼6 Å). Nevertheless, xylene molecules can enter the pores. This is attributable to structural mobility of the framework, enabling conformational changes of the imidazole rings, which depart from the equilibrium conformation, and hence xylene molecules can slip into the pores. ²⁰⁵

Zoom In Zoom Out Reset image size

The [Zn(mpba)]_n framework (MAF-X8), where mpba is the 4-(1H-3,5-dimethylpyrazol-4-yl)benzoic acid dianion, showed a selectivity towards p-xylene, which was only slightly inferior to the selectivity of BaX zeolite currently used in industry, while the uptake capacity of MAF-X8 was markedly higher, which increased the operating life of the sorbent between regenerations. ²⁰⁶ Molecular modelling demonstrated that the channels of MAF-X8 match p-xylene molecules almost perfectly.

Experimental and theoretical study of selective sorption of p-xylene from an isomer mixture with the [Cu(ndc)(DABCO)_0.5]_n framework [DUT-8(Cu)] was reported by Kim et al. ²⁰⁷ DUT-8(Cu) is a 3D metal-organic framework with 1D channels with a square cross-section of 9.6 Å side. In the adsorption of xylenes from a three-component mixture, the selectivity coefficient for the p-isomer was ∼2. Molecular dynamics simulation showed that p-xylene molecules are packed in the channels in an ordered manner and mainly in the vicinity of DABCO linkers, whereas the other two isomers fill the channels in a random manner.

An interesting 'molecular imprinting' effect in the synthesis of MOFs was demonstrated by Yang et al. ²⁰⁸ The reaction of 4-(1H-pyrazol-4-yl)pyridine (Hpypz) with copper(II) ions in the presence of p-xylene results in the formation of the layered metal-organic framework [Cu₂(pypz)₂] · 0.5 pX, where pX is p-xylene, (MAF-36) in which zigzag-like channels are filled with the para-isomer molecules. The removal of p-xylene molecules affords a metal-organic framework, which retains high selectivity for the p-xylene over other xylene isomers, despite the structural rearrangement taking place upon removal of the guests. The adsorption of an equimolar liquid mixture with MAF-36 demonstrated the record-high selectivity for p-xylene, amounting to 51, which was several times as high as that exhibited by other MOFs (7 to 10) and by the commercially used BaX zeolite (∼7).

Selective sorption of p-xylene, which has the smallest kinetic diameter of the molecule among xylenes, can be attained by using flexible ligands for MOF fabrication. Mukherjee et al. ¹⁸⁴ described the layered metal-organic framework [Zn₄O(L³)₃(DMF)₂]_n · x G (L³ is 4,4'-[(4-tert-butyl-1,2-phenylene)bis(oxy)]dibenzoate with two flexible ester groups, G is the guest molecule). The removal of guests from the MOF pores leads to a considerable contraction of the framework: the fraction of free space in the unit cell decreases from 34 to 12%. The repeated immersion of the MOF into the medium of a guest with the appropriate size of molecules, in particular p-xylene, leads to restoration of the crystal structure and uptake of a considerable amount of guest molecules into the framework (up to 4 p-xylene molecules per formula unit). Meanwhile, in the case of o-and m-xylenes, only 0.08 and 0.34 molecules per formula unit, respectively, are adsorbed. ¹⁸⁴

The conformational mobility of the ligand with amide groups played an important role in the selective encapsulation of xylene into the linear MOF {[Cu(L⁴)₂(H₂O)](ClO₄)₂}_n , [L⁴ is N¹,N⁴-bis(4-pyridylmethyl)fumaramide]. ²⁰⁹ Upon crystallization, polymer chains are assembled into layers through hydrogen bonding, while the layers are stacked to form 19.3×11.3 Å channels. The synthesis of such a metal-organic framework in the presence of binary or ternary mixtures of xylene isomers leads to competitive inclusion of only one of them, particularly, only o-xylene is adsorbed from a ternary mixture and only the m-xylene is adsorbed when a mixture of m- and p-xylenes is used. It is worth noting that, depending on the adsorbed xylene isomer, the dimensionality of the resulting MOF bearing the guest molecules changes. Particularly, the presence of o- and p-xylenes induced crystallization of 2D MOF (Fig. 60), while the meta-isomer caused crystallization of 1D MOF.

Zoom In Zoom Out Reset image size

There are also MOFs that demonstrate adsorption selectivity not only for p-xylene, but also for other isomers. Nicolau et al. ²¹⁰ investigated the adsorption of xylene and ethylbenzene vapours for binary and quaternary mixtures. The authors showed that the [Zn(bdc)(DABCO)_0.5]_n framework exhibits selectivity for o-xylene with a selectivity coefficient of 1.88 due to stronger interaction of this isomer with the framework surface.

The layered metal-organic framework [Co(bpy)₂(NCS)₂]_n showed high uptake capacities towards o- and p-xylene (85 mass %), which are inferior only to those of MIL-101. ²¹¹ Powder X-ray diffraction of the inclusion compounds [Co(bpy)₂(NCS)₂]_n · x C₈H₁₀ demonstrated that xylene vapour fills the space between MOF layers (Fig. 61), the uptake being four xylene molecules per formula unit. This is in line with the observed adsorption capacity. The nature of the intermolecular interactions between the guests and the framework is somewhat different over the series of isomeric xylenes. The shortest distance (4.003 Å) between the xylene molecule and the pyridine rings of the linker is characteristic of the ortho-isomer. In addition, the hydrogen atoms of the aromatic ring of o-xylene are involved in the C–H...π interactions, whereas the meta- and para-isomers form contacts only by their methyl groups. Finally, this MOF shows high selectivity (∼10) for o-xylene over p-xylene. ²¹¹

Zoom In Zoom Out Reset image size

A very high selectivity for o-xylene was found for chromium terephthalate [Cr(OH)(bdc)]_n [MIL-53(Cr)]. The o-xylene – p-xylene selectivity for sorption from a liquid isomer mixture (13.75) makes this material a promising alternative to zeolites in the industrial sorption processes using a simulated moving bed. ²¹²

Among the factors affecting the selectivity of adsorption of a particular xylene isomer is the presence of coordinatively unsaturated sites in the MOF structure. Gonzalez et al. ²¹³ carried out a structural study of the compounds formed upon encapsulation of xylenes into the [Co₂(dobdc)]_n framework (dobdc is 2,5-dihydroxybenzene-1,4-dicarboxylate). The framework contains hexagonal channels, the inner surface of which has coordinatively unsaturated cobalt(II) sites. It was found that encapsulation of o-xylene induces a substantial distortion of three out of each four framework pores and that each distorted pore accommodates four o-xylene molecules, while the undistorted pores contain three such molecules (Fig. 62). In the case of para- and meta-isomers, characterized by lower adsorption capacity, no framework distortions were detected. Three cobalt(II) cations can be identified as binding sites for o-xylene molecules, whereas only two such sites are present for m-xylene and only one site is found for p-xylene. As a result, the o-xylene – p-xylene selectivity is 3.9, while the volumetric capacity of this framework (3.8 – 4.2 mmol cm⁻³) exceeds those of zeolites that are currently used in industrial processes of xylene separation. ²¹³

Zoom In Zoom Out Reset image size

Cyclohexane is a large-scale product of chemical industry, used as a raw material for the production of cyclohexanol, cyclohexanone and caprolactam, which are, in turn, used for the production of polyamide fibres and threads. Cyclohexane is produced by catalytic hydrogenation of benzene, and the unreacted benzene has to be separated in order to isolate a pure product. Separation of small amounts of benzene from cyclohexane is an energy-demanding task complicated for implementation. ¹⁸⁵ Due to close boiling points (80.1 °C for benzene and 80.7 °C for cyclohexane), conventional distillation is inapplicable. Therefore, azeotropic and extractive distillation processes are used, ^{214 – 216} which decreases the environmental and economic efficiency of the processes. The methods of benzene separation from cyclohexane by pervaporation using polymer membranes appear relatively promising from the engineering standpoint. ^{217 – 220} In recent years, considerable attention of researchers has been devoted to the adsorption separation of benzene and cyclohexane using silica gels, ²²¹ mesoporous carbon ²²² and also organic cavitands. ^{223, 224}

The first studies dealing with inclusion compounds formed upon encapsulation of benzene into MOF pores appeared rather long ago. In 2007, Shimomura et al. ²²⁵ reported the results of an X-ray diffraction study of the compound {[Zn(μ₄-TCNQ-TCNQ)(bpy)] · 1.5 PhH}_n (TCNQ-TCNQ is 1,2-bis[4-(dicyanomethylene)cyclohexa-2,5-dien-1-yl]ethane-1,1,2,2-tetracarbonitrile), which showed that benzene molecules are strongly held by C–H...π interactions (Fig. 63).

Zoom In Zoom Out Reset image size

One of the first examples of comparative study of MOF adsorption of benzene and cyclohexane was reported by Tan et al. ²²⁶ in 2012. For the three-dimensional MOF {[Cu₂(obb)₂(bpy)_0.5(DMF)] · 2 DMF}_n , the uptake capacity for benzene adsorption from the gas phase was 20.6 mass % and that for cyclohexane was 14.9 mass %. As specific features of the process, the authors noted a wide hysteresis in the adsorption isotherms and incomplete desorption related to the lock effect of the framework pores towards the adsorbed molecules.

Basic amide groups and C–H...π and π – π interactions were considered ²²⁷ to be responsible for the increased benzene – cyclohexane selectivity of [Cu₂(L⁵)₄(NO₃)₄(H₂O)]_n MOF [L⁵ is di(4-pyridyl)-5-tert-butylisophthalamide). The sorbent can take up ∼2 benzene molecules per formula unit when adsorbed from the gas phase, with the selectivity over cyclohexane being 8.

Zhou et al. ²²⁸ made an attempt to separate benzene and cyclohexane by sorption under dynamic conditions with the hcp-UiO-66 MOF based on {Zr₁₂O₈(OH)₁₄} clusters and terephthalic acid. The sorbent more efficiently retains benzene with the selectivity coefficient of 2.39, which is sufficient for separation of cyclohexane from mixtures with relatively low contents of benzene.

Some MOFs exhibit inverse selectivity, that is, they better retain cyclohexane than benzene. For example, the [Zn(eim)₂] metal azolate framework, where eim is 2-ethylimidazolate anion, (MAF-6) has a high concentration of ethyl groups on the pore surface, which make it highly hydrophobic. As a result, during gas chromatographic separation, more lipophilic cyclohexane (lipophilicity: log P = 3.6) is held more strongly than benzene (log P = 2.2). ²²⁹

Logvinenko et al. ²³⁰ compared the thermal stability of benzene and cyclohexane inclusion compounds into {[Zn₂(bdc)₂(DABCO)] · 4PhH}_n and {[Zn₂(bdc)₂(DABCO)] · 3.5 C₆H₁₂}_n MOFs. The compound with cyclohexane was thermally more stable, despite the fact that the boiling point and the enthalpy of vaporization of cyclohexane are virtually equal to those of benzene. The higher kinetic stability of the cyclohexane inclusion compound was attributed to the entropy factor, related to the conformational flexibility of the cyclohexane molecule.

As in the case of xylene adsorption, the presence of coordinatively unsaturated sites in MOF often has a crucial effect on the selectivity of benzene sorption. Liu et al. ²³¹ investigated sorption of benzene by the [Mg(dobdc)]_n framework (Mg-MOF-74). This framework has channels with a hexagonal cross-section with an inscribed circle diameter of ∼15 Å; each magnesium atom coordinates five oxygen atoms in a square-pyramidal environment. Studying the adsorption isotherms of benzene from a vapour revealed a high uptake capacity (8.2 mmol g⁻¹) and adsorbed benzene density of 815 kg m⁻³, which is comparable with the density of liquid benzene (870 kg m⁻³) and attests to a liquid-like state of the adsorbed benzene molecules. When sorption was carried out from an equimolar mixture of benzene and cyclohexane vapour, only benzene was shown to be sorbed, and the diffusion coefficient of pure cyclohexane in the sorbent was 1 – 3 orders of magnitude higher than that for benzene. The molecular dynamics simulation of the intermolecular interactions of benzene and cyclohexane with the framework surface indicates that benzene molecules are mainly adsorbed near magnesium ions, with the average distance from the centre of mass of the molecule to the nearest magnesium ion being 3.78 Å. Meanwhile, the low-polarity cyclohexane molecules have a low affinity to coordinatively unsaturated magnesium ions, with the average distance between them being 4.65 Å. ²³¹

Benzene and cyclohexane sorption was studied for a series of frameworks with various coordinatively unsaturated sites [M(dobdc)]_n , where M = Mn, Ni, Mg, Cu, Zn, Co, Fe (the codes are M-MOF-74). The benzene adsorption capacity decreased in the series from manganese(II)- to iron(II)-containing framework; this series coincides with the order of decreasing ionization potentials of the metals. For the Mn-MOF-74 framework, the saturation capacity for benzene reached 9.38 mmol g⁻¹, while that for cyclohexane was only 0.25 mmol g⁻¹. The calculation of the benzene – cyclohexane selectivity using the ideal absorbed solution theory (IAST) for Mn-MOF-74 gives a very high value of about 10⁵. ²³²

Kondo et al. ²³³ reported the MOF [Cu(bpp)₂(BF₄)₂]_n [bpp is 1,3-di(4-pyridyl)propane], capable of solvent-induced reversible transformation from the close-packed non-porous 3D structure into a porous 1D structure. For sorption from the gas phase, the uptake capacity for benzene was 248 mg g⁻¹ (∼2 benzene molecules per formula unit), while the uptake of cyclohexane was negligibly low. It is noteworthy that the adsorbed benzene molecules are held very strongly by the MOF and are not released even when the relative pressure drops to 10⁻².

Modification of the inner surface of channels in the two-dimensional MOF {[CuI(L⁶)] · (DMF) · H₂O}_n [L⁶ is 4'-(anthacene-9-yl)-4,2':6',4''-terpyridine] by anthracene moieties provided high selectivity of benzene sorption from the gas phase with the benzene – cyclohexane selectivity of ∼42. ²³⁴

The polar groups on the pore surface of the [Zn(pbda)(dpni)]_n framework (pbda is 4,4'-[(2-tert-butyl)-1,4-phenylene)bis(oxy)]dibenzoic acid, dpni is N,N'-di(4-pyridyl)napthalene-1,4,5,8-tetracarboxydiimide) were indicated ²³⁵ to be responsible for the selectivity of benzene sorption (uptake capacity of ∼4.5 mmol g⁻¹) over cyclohexane sorption (the uptake is negligibly low).

With the goal of generating π-electron-deficient sites on the surface of MOF channels, Manna et al. ²³⁶ proposed a new ligand, 4-(4,6-diamino-1,3,5-triazin-2-yl)benzoic acid. The reaction of this ligand with copper nitrate gave a 2D MOF, the layers of which were combined into a three-dimensional structure with 6.71×7.08 Å channels through hydrogen bonding. Due to the π – π stacking with electron-deficient triazine rings, the π-electron density-rich benzene molecules were efficiently adsorbed by the obtained MOF with an uptake capacity of 1.5 mmol g⁻¹, whereas the uptake of cyclohexane was only 0.2 mmol g⁻¹. The IAST calculation of the sorption selectivity for an equimolar benzene – cyclohexane mixture resulted in the value of ∼200. ^{236, 237}

The role of π-acceptor 1,3,5-triazine moieties in increasing the benzene – cyclohexane adsorption selectivity via π – π interactions was also demonstrated in relation to the frameworks {(Et₄ N)₃[In₃(tatb)₄] · (DEF)₁₆ · (H₂O)₁₁}_n , where tatb is 4,4',4''-s-triazine-2,4,6-triylbenzoate, DEF is diethylformamide, ²³⁸ (code FJI-C1) and {[Cd(ataia)] · 4H₂O}_n , where ataia is 5-[(4,6-diamino-1,3,5-triazin-2-yl)amino]isophthalate. ²³⁹ When adsorption was carried out from saturated vapour at 298 K, the uptake capacity of FJI-C1 for benzene was about 13.4 mass %, which was 111 times higher than the capacity of the same MOF for cyclohexane. ²³⁸ For the cadmium-organic framework the selectivity coefficient was 6.5. ²³⁹

The ligand flexibility and the presence of Lewis π-acid sites on the pore surface of the three-dimensional MOF [Zn₂(L⁷)₂(DMF)₂]_n [L⁷ is 2,2'-(1,3,5,7-tetraoxo-5,7-dihydropyrrolo[3,4-f]isoindole-2,6(1H,3H)-diyl)dipropiolate] with one-dimensional channels provide selective adsorption of benzene, although the uptake capacity is moderate (one benzene molecule per formula unit). Cyclohexane is virtually not adsorbed by this framework. ²⁴⁰

The competing sorption of benzene and cyclohexane from liquid and gaseous equimolar mixtures was studied ²⁴¹ in relation to a series of isostructural MOFs described as [Zn₁₂(tdc)₆(gly)₆(DABCO)₃]_n (H₂tdc is thiophene-2,5-dicarboxylic acid, gly is alkane-1,2-diol dianion). These MOFs contain channels with a diameter from 1.4 to 4.8 Å, depending on the alkyl chain length of the diol. The alkyl chain length also affects the selectivity for benzene over cyclohexane; it is maximum for ethane-1,2-diol and amounts to 20 : 1 for sorption from the gas phase and 92 : 1 for sorption from the liquid phase. Elongation of the alkane-1,2-diol chain is accompanied by a decrease in the selectivity and, in the case of butane-1,2- and pentane-1,2-diols, the selectivity is reversed, i.e., cyclohexane is adsorbed predominantly. The MOF with the pentane-1,2-diolate anion had a cyclohexane selectivity over benzene equal to 5 for the gas phase and 2.5 for the liquid phase. ²⁴¹

The sorption selectivity can be controlled by changing the pore size in MOFs. Wang et al. ²⁴² investigated the adsorption of benzene and cyclohexane from their equimolar liquid mixture by a series of MOFs obtained by ion exchange: NH₄@ZnPzC, Me₃NH@ZnPzC and Et₃NH@ZnPzC (PzC is 1H-pyrazole-4-carboxylate). On going from the trimethylammonium cation to the triethylammonium cation, the adsorption selectivity for benzene, which has a smaller molecular size than cyclohexane, increased from 3.1 to 14, which is consistent with decreasing pore size of the framework. ²⁴² A similar trend was observed for a MOF based on copper(II) ions, CAT@[Cu₃-(μ₃-OH)(μ₃-PzC)₃]_n (where , Et₃N⁺, Li⁺): the selectivity coefficient increased from 5 to 12 on going from the ammonium cation to the lithium cation. ²⁴³

The correlation between the kinetic diameter of the benzene molecule (5.85 Å) and the pore size of MOF [Zn(ip)(bpa)]_n , where ip is isophthalate, bpa is 1,4-di(4-pyridyl)acetylene, (CID-23) endows the framework with an enhanced selectivity for benzene; the benzene – cyclohexane selectivity coefficient for adsorption from an equimolar gas-phase mixture is 25. ²⁴⁴

The possibility of template tuning of MOF for attaining high sorption properties was demonstrated. ²⁴⁵ It is known that the reaction of 4-(3,5-dimethyl-1H-pyrazol-4-yl)benzoic acid (H₂mpba) with zinc(II) ions may give diverse metal-organic frameworks. In the presence of benzene, the MOF {[Zn₂(Hmpba)₂(mpba)] · 2PhH}_n is formed with zigzag-like channels filled with benzene molecules. Guest exchange may allow the preparation of inclusion compounds with other guests, particularly cyclohexane. X-Ray diffraction data indicate that the framework undergoes a substantial restructuring as one guest is replaced by another one. The obtained MOF was used as the stationary phase for gas chromatographic separation of benzene and cyclohexane with a selectivity coefficient of 2.85. ²⁴⁵

The first example of using MOFs for separating benzene and cyclohexane (or other hydrocarbons) by liquid chromatography was reported by Nuzhdin et al., ²⁴⁶ who used a composite sorbent based on the framework HKUST-1 and aerogel SiO₂ as the stationary phase. In this sorbent, MOF ensures the selectivity towards unsaturated hydrocarbons and silica gel improves the hydrodynamic properties of the column.

Thus, the diversity of MOF building blocks allows for the fabrication of porous materials with high selectivity for the desired xylene isomer or benzene, while the high porosity of MOFs ensures the uptake capacity values that cannot be attained for traditional sorbents. The progress in the methods of MOF synthesis makes the frameworks more and more economically accessible, and some of them can even now compete with materials applied in the existing industrial processes of hydrocarbon separation.

The tritopic clathrochelate ligands 54, 55 based on arylboronic acids 56 – 58 and dioxime 59a shown in Scheme 4 ²⁵⁵ were used to prepare two larger-size type coordination cages with Pd...Pd distances of up to 4.2 nm in the presence of palladium complex 60.

Scheme 4 

Zoom In Zoom Out Reset image size

The effect of the size and geometry of clathrochelate ligands on the coordination-controlled association of coordination cages of various types was studied both theoretically and experimentally. ^{256, 257} It was found that, instead of the tetragonal barrel-like (A) cages, the reaction of mononuclear clathrochelate synthons with 2 equivalents of (dppp)M(OTf)₂ [61: M=Pd (a), Pt (b)] gives gyrobifastigium-like (B) type cages, which are shown in Scheme 5 ²⁵⁷ in relation to synthon 62a as an example. Meanwhile, analogous synthons 62b–e and complexes 61b – e were converted to both barrel-like cages and assemblies with molecular weight of >23 kDa and diameter of ∼4.5 nm. The geometry of compounds obtained in this way depends on the structure of complexes 61b – e. Indeed, when synthons 62d,e react with complexes 61d,e, the tetragonal barrel-like cages A are formed; synthons 62b,c,e and complex 61b react to give gyrobifastigium-like cages B; pentagonal barrel-like cages C are formed in the reaction of synthon 62c with complex 61d; and the reaction of synthon 62c with complex 61c affords assembly D (Scheme 6). ²⁵⁷ The binuclear clathrochelate ligands 62a,b react with complex 61b to give pentagonal barrels 63a,b (Scheme 7). ²⁵⁷

Scheme 5 

Zoom In Zoom Out Reset image size

Scheme 6 

Zoom In Zoom Out Reset image size

Scheme 7 

Zoom In Zoom Out Reset image size

Cecot et al. ²⁵⁸ showed that during the formation of coordination cages, lateral groups can direct the coordination-controlled self-assembly of clathrochelate ligands by promoting the formation of a structure that enables energetically favourable (i.e., associated with minor steric restrictions) packing of bulky groups. Clathrochelate ligand 62f, containing four pyridyl groups in the apical positions and three phenanthrene groups in the lateral chelating moieties, was reacted with complexes 61b,c to give coordination cages with a tetragonal prismatic geometry (Scheme 8). ²⁵⁸ In the case of clathrochelate ligand 61c with aliphatic n-butyl lateral groups, gyrobifastigium-like cages were obtained. ²⁵⁸

Scheme 8 

Zoom In Zoom Out Reset image size

Jansze et al. ²⁵⁹ reported the synthesis of bent ditopic clathrochelate ligands 64a,b and 65a,b with terminal 3-pyridyl groups. They were obtained by template condensation of a mixture of the corresponding aryldiboronic (66, 67)and 3-pyridylboronic (56) acids on an iron(II) ion as the template (Scheme 9). ²⁵⁹ These macrobicyclic synthons were used for the coordination-controlled self-assembly of cages in the presence of palladium complex 60. It was shown ²⁵⁷ that cages with nioxime (cyclohexane-1,2-dione dioxime) substituents [R–R=(CH₂)₄] in the lateral chelating moieties are more stable than their dimethylglyoximate analogues (R = Me).

Scheme 9 

Zoom In Zoom Out Reset image size

Binuclear zinc and cobalt(II) clathrochelates 68 – 71, containing cyano groups at the periphery of encapsulating ligands (Scheme 10), were prepared ²⁶⁰ by two pathways:

Scheme 10 

Zoom In Zoom Out Reset image size

• —  
  using the template polycondensation of CN-containing arylboronic acids (e.g., compound 72) and the appropriate phenoldioximate ligand synthons 73a,b on zinc or cobalt(II) ion as the template (for complexes 68, 69);
• —  
  post-synthetic cross-coupling of polybrominated zinc(II) clathrochelates 74a,b, obtained from bromides 75a,b, with 4-cyanophenylboronic acid (72).

The reactions of clathrochelate ligands 68 – 71, containing terminal cyano groups, with silver salts resulted in the formation ²⁶⁰ of heterometallic (Zn²⁺, Ag⁺ and Co²⁺, Ag⁺) coordination polymers.

Bila et al. ²⁶¹ reported the formation of microporous thermally and hydrolytically stable polymer networks with permanent porosity based on ligands 76a – 79a (Fig. 64). These ligands were obtained in two ways. The first method was based on the Suzuki – Miyaura covalent cross-coupling of pre-synthesized iron(II) dibrominated clathrochelates 80a,b with reactive halogen atoms of compound 75b and the corresponding diboronic acids 81, 82 (Scheme 11) and the other method comprises Sonogashira – Hagihara polycross-coupling of clathrochelate precursors 83a,b, containing terminal C≡C bonds in the apical substituents, with 1,3,5-tribromobenzene (Scheme 12). The clathrochelate precursor 80b, derived from enantiomerically pure α-dioximate ligand 59c, was used ²⁶¹ to prepare chiral polymeric products, one of which adsorbed mainly D- tryptophan in comparison with the L-epimer.

Zoom In Zoom Out Reset image size

Scheme 11 

Zoom In Zoom Out Reset image size

Scheme 12 

Zoom In Zoom Out Reset image size

Fe ^II ,Fe ^III -heterobinuclear clathrochelates were obtained by template condensation of 4-pyridylboronic acid (84)with the corresponding phenol dioxime ligand synthon 73 in the presence of iron(III) and/or iron(II) ions (Scheme 13). Their macrobicyclic ligands can encapsulate iron ions in two redox states to give either homobinuclear clathrochelates 85a,b with two encapsulated iron(II) cations or their heterobinuclear Fe ^II ,Fe ^III -analogues 86a,b, which easily undergo reversible chemical or electrochemical oxidation or reduction.

Scheme 13 

Zoom In Zoom Out Reset image size

Using a mixture of zinc or iron salts, Bila et al. ²⁶² prepared Zn ^II ,Fe ^III -heterobinuclear clathrochelate 87 (Scheme 14). Studying ion exchange reactions of encapsulated metal cations in these binuclear clathrochelates showed ²⁶² that Fe ^II ,Fe ^III -heterobinuclear cage complexes 86a,b are thermodynamically more stable than analogous binuclear complexes of zinc (88) and cobalt(II) (89) (Scheme 15).

Scheme 14 

Zoom In Zoom Out Reset image size

Scheme 15 

Zoom In Zoom Out Reset image size

Palladium-catalyzed polycross-coupling of Fe ^II ,Fe ^III -heterobinuclear clathrochelate 90 with the terminalbromine atoms in the encapsulating ligand and benzene-1,4-diboronic acid (81) afforded ²⁶² thermally and hydrolytically stable coordination polymer with permanent porosity (Scheme 16).

Scheme 16 

Zoom In Zoom Out Reset image size

The template condensation of the corresponding phenol dioxime ligand synthon with 3-(acrylamidophenyl)- or 3-(methacrylamidophenyl)boronic acids on cobalt or manganese(II) ion templates yields binuclear boron-containing macrobicyclic complexes 91a,b with reactive apicalsubstituents. ²⁶³ It was found that the type of supramolecular clathrochelate network in the crystals of the product depends on the nature of terminal groups at the crosslinking boron atoms and that transition from 2D to 3D supramolecular assemblies can be accomplished using appropriate guests (Fig. 65). ²⁶³

Zoom In Zoom Out Reset image size

Iron(II) clathrochelates 92a – d with terminal 4-pyridyl groups are converted ²⁶⁴ to paddle-wheeled metal-organic frameworks with N,N'-di(4-pyridyl)napthalene-tetracarboxydiimides via ligand exchange reactions.

The first hybrid bi- and trinuclear Fe ^II ,Zr ^IV - and Fe ^II ,Hf^IV-porphyrinatoclathrochelates with one (93a,b) or two apical meso-tetraphenylporphyrinato moieties (94a,b) wer obtained under mild conditions by transmetallation of mono- (95) or di-antimony-containing (96) macrobicyclic precursors on treatment with dichlorozirconium(IV)- or dichlorohafnium(IV)-meso-tetraphenylporphyrins 97a,b as Lewis acids (Scheme 17). ²⁶⁵ It was reported ²⁶⁶ that the reactive antimony-containing iron(II) clathrochelate 98 with a terminal vinylgroup shown in Scheme 18 also undergoes exchange reaction of the cross-linking group (transmetallation) under the action of the hafnium(IV) phthalocyanine complex (HfPc, 99a) as a stronger Lewis acid to give hybrid binuclear HfPc-cross-linked iron(II) phthalocyaninoclathrochelate 100.

Structures 91 

Zoom In Zoom Out Reset image size

Structures 92 

Zoom In Zoom Out Reset image size

Scheme 17 

Zoom In Zoom Out Reset image size

Scheme 18 

Zoom In Zoom Out Reset image size

Antimony-containing iron and nickel(II) tris(pyridineoximates) 101a,b with the reactive triethylantimony crosslinking group were prepared by template condensation of 2-acetylpyridineoxime and triethylantimony(V) dibromide using the corresponding metal ion as the template. ²⁶⁷ These labile complexes were readily transmetallated on treatment with zirconium and hafnium(IV) phthalocyanine complexes 99a,b as Lewis acids to give binuclear iron (102a,b) and nickel(II) (102c,d) tris(pyridineoximates) with a phthalocyanine cross-linking group (Scheme 19, pathway a). Subsequently, Dudkin et al. ²⁶⁸ proposed a more convenient and efficient one-step preparation method of hybrid complexes 102a – d in high yields, which consists in template condensation of 2-acetylpyridineoxime with phthalocyanines 99a,b in the presence of 3d metal salts (pathway b in Scheme 19).

Scheme 19 

Zoom In Zoom Out Reset image size

The aggregation of four hybrid binuclear iron and nickel(II) tris(pyridineoximates) shown in Scheme 19 with metal phthalocyanine cross-linking moieties 102a – d was studied by spectral methods both in solutions and in fluoro-plastic thin films. ²⁶⁹ It was shown that these complexes form H-type dimers in solutions or in poly(vinylidene fluoride) – hexafluoropropylene (PVDF-HFP) copolymer matrix, while the complexes immobilized on the surface of polymers exist mainly as single molecules.

Similar pseudo-macrobicyclic cobalt(III) tris(pyridineoximates) 102e,f with phthalocyanine cross-linking moieties were prepared ²⁷⁰ according to Scheme 19 (pathway b) by direct template reaction of 2-acetylpyridineoxime with zirconium and hafnium(IV) phthalocyanine complexes 99a,b and cobalt(II) salt. During this reaction, the central cobalt(II) ion was oxidized to Co³⁺. According to single crystal X-ray diffraction, ²⁷⁰ the CoN₆ polyhedron of the cross-linked binuclear ZrPc complex has a truncated trigonal pyramidal geometry. The encapsulated cobalt(III) ion occurs nearly at the centre of this polyhedron, which is typical of low-spin cobalt(III) complexes: this cation is only slightly shifted along the molecular Co...Zr pseudo-C₃ symmetry axis of the semiclathrochelate cation.

The monopropargylamine iron(II) cage complex 103, obtained in a high yield by nucleophilic substitution reaction of monochlorinated clathrochelate precursor 104 with propargylamine, was introduced in the copper-catalyzed click-reaction with o-carborane-12-methyl azide, resulting in the formation of hybrid iron(II) monocarborane clathrochelate 105 (Scheme 20; BH groups are located at the dodecahedron vertices). ²⁷¹

Scheme 20 

Zoom In Zoom Out Reset image size

Iron(II) n-butylborate hexachloroclathrochelate 106 showed ²⁷² high cytotoxicity in vitro against human promyelocytic leukemia (HL-60) cell line [the half-maximal inhibitory concentration (IC₅₀) was 6.5 μmol L⁻¹]. This result was correlated with the substantial (17-fold with respect to the control) increase in the intracellular oxidative stress. Relying on the reported results, ²⁷² this effect can be attributed to the alkylation (shown in Scheme 21) of glutathione (107) with highly reactive electrophilic clathrochelate 106. This led to inhibition of the antioxidant defense system in cells and to catalytic generation of reactive oxygen species by the products of alkylation of this macrobicyclic complex, e.g., compound 108.

Scheme 21 

Zoom In Zoom Out Reset image size

A number of iron and cobalt(II) mono- and bis-clathrochelates were tested in vitro in the transcription system of Taq DNA polymerase (DNAP), which uses polymerase chain reaction. ²⁷³ All of these complexes were found to be micromolar transcription inhibitors in this system, but only a few clathrochelates with terminal biorelevant groups showed IC₅₀ values in a low micromolar range (5 to 40 μmol L⁻¹). The highest in vitro inhibitory activity was found ²⁷³ for homodifunctionalized iron(II) dicarboxyl monoclathrochelates. Meanwhile, their analogues such as C–C conjugated bis(clathrochelates) and heterofunctionalized morpholine-containing cage complexes with one terminal carboxylgroup and iron(II) monoclathrochelates with two carboxyl-terminated alkyl sulfide groups and alkylamine ribbed substituents containing terminal pyridine or morpholine groups are much less active transcription inhibitors of the DNAP system.

Rhenium (III) semiclathrochelates 109 and 110, containing a biorelevant apical substituent, were synthesized from rhenium(III) solvate complex 111 (Scheme 22). ²⁷⁴ The template condensation of 111 with nioxime and 3-formylphenylboronic acid gives rise to rhenium(III) tris(nioximate) 109 with a terminal formyl group, which is subsequently subjected to H⁺-catalyzed condensation with isoniazid (112) to give rhenium(III) semiclathrochelate 113 with an apical functionalizing vector substituent. Rhenium(III) semiclathrochelate 110 with an apical hexadecyl group at the bridging boron atom was prepared ²⁷⁴ by template condensation of nioxime and hexadecylboronic acid, with rhenium(III) ion serving as the template. In addition, crosslinking of pre-synthesized non-macrocyclic rhenium(III) tris(nioximate) 114 on treatment with cholesteryl boronic acid 115 resulted in the synthesis of rhenium(III) semiclathrochelate 116 with a cholesterol apical moiety (see Scheme 22).

Scheme 22 

Zoom In Zoom Out Reset image size

Zelinskii et al. ²⁷⁵ described monoribbed-functionalized iron(II) cage complexes, the molecular design of which implied their efficient supramolecular binding to nucleic acids, and determined their biologically important ADMET characteristics. ^†

Heterodifunctionalized iron(II) clathrochelates 117 and 118, containing terminal biorelevant carboxyl groups and reactive C≡C bonds, were obtained as depicted in Scheme 23. ²⁷⁶ The two-step synthetic procedure included stepwise nucleophilic substitution of dichloro clathrochelate precursor 119 under the action of p-carboxyphenylthiolate anion (to give intermediate 120) and the subsequent treatment with 2-propargylamine. Clathrochelates 121 and 122, prepared by the click reaction between alkyne 123 and azides 124 and 125, were proposed as promising precursors of macrobicyclic iron(II) complexes. ²⁷⁶ These molecules contain an optically active (first of all, luminescent) reporter group or a radioactive label to visualize their cellular uptake and/or biodistribution in a specified biological system.

Scheme 23 

Zoom In Zoom Out Reset image size

Kovalska et al. ²⁷⁷ discovered the appearance of induced chirality of metal clathrochelates 126a–m and 127, caused by their supramolecular or covalent binding to both biological macromolecules and optically active organic molecules, chirality inducers. In particular, ribbed-functionalized iron(II) clathrochelates, when acting as guests, were found to suppress the CD response (CD is circular dichroism) upon the supramolecular binding to an appropriate protein macromolecule as the host. Metal tris(dioximate) clathrochelates have a Russian doll-like structure (shown in Fig. 66): owing to the distorted trigonal prismatic (TP) – trigonal antiprismatic (TAP) geometry, the MN₆ coordination polyhedra have no inversion centre and possess chirality, but due to equal probabilities of the formation of levo-(ʌ) and dextrorotatory (Δ) forms, they give no CD signal. The selective fixation of one of C₃ -distorted conformations produces a CD signal in the visible region, where intense charge transfer bands (CTBs), d(M)→π*(L), are also observed. The supramolecular interaction of the initially CD-inactive iron(II) clathrochelates containing terminal carboxyl groups with bovine serum albumin (BSA) induces ²⁷⁷ molecular asymmetry. Hence, the CD spectra of these clathrochelate-centred assemblies exhibit intense signals in the 350 – 600 nm range, whereas their methyl esters and amide derivatives remain virtually CD-inactive. It was shown that the supramolecular interaction of carboxyl-containing clathrochelates with BSA, or their covalent binding to (R)-(+)-1-phenylethylamine depend both on the solvent polarity and on the structural isomerism of the clathrochelate CD reporter.

Zoom In Zoom Out Reset image size

The induced circular dichroism (ICD) spectra of clathrochelate-centred isomeric macrobicyclic iron(II) complexes with six terminal carboxyl groups was used ²⁷⁸ to determine the structural and conformational changes of serum albumins. These clathrochelate CD reporters are able to discriminate between transport albumins with similar structure such as BSA and human serum albumin (HSA), the binding to which gives clathrochelate-centred structures with intense ICD spectra. Variation of the shape and intensity of the bands in these spectra was also used to detect ²⁷⁸ changes in the tertiary structure of albumin macromolecules. It was shown that the structural isomerism of CD-inducing hexcarboxyl clathrochelates significantly affects both the ICD response upon binding to the above-mentioned proteins and the thermodynamic parameters of the clathrochelate – albumin supramolecular assembly. These iron(II) clathrochelates were proposed ²⁷⁸ as promising three-dimensional molecular platforms for the design and synthesis of sensitive CD reporting molecules able to efficiently recognize specific elements on the surface of protein macromolecules.

Structures 126, 127 

Zoom In Zoom Out Reset image size

Iron(II) clathrochelates with two terminal carboxyl groups were tested ²⁷⁹ as potential CD reporters for studying the structures of various proteins. The authors obtained structural isomers of heterodifunctionalized iron(II) clathrochelate with two non-equivalent ribbed carboxyphenyl sulfide groups. They studied the interaction of these clathrochelates and their homodifunctionalized isomers with serum albumins, lysozyme, β-lactoglobulin (BLG), trypsin and insulin using quenching of intrinsic fluorescence of the protein and ICD spectra of clathrochelate-centred structures. High-intensity ICD signals were observed upon the formation of 1 : 1 supramolecular assemblies of these clathrochelates with albumins. In the presence of BLG, various structural isomers of dicarboxyl-containing iron(II) clathrochelate demonstrated ICD bands with opposite signs, which indicates that opposite (Λ or Δ) configurations are stabilized in these assemblies upon the supramolecular binding to the mentioned globular proteins.

Functionalization of terminal carboxyl groups in iron(II) carboxyaryl sulfide clathrochelate isomers 128a – c was used to prepare their macrobicyclic carboxymethyl (129a – c) and carboxyamide (130a – c, 131a – c) derivatives (Scheme 24). ²⁸⁰ Quenching of the intrinsic luminescence of the protein was used to investigate the reaction of compounds 129 and 130 with BSA. It was shown that this modification of the terminal groups of iron(II) clathrochelates reduces quenching of the intrinsic luminescence of this protein by an order of magnitude and also alters the type of influence of the clathrochelate isomerism on the quenching of the BSA intrinsic luminescence. The most pronounced luminescence quenching was observed for complexes 128 and, among them, by isomer 128c, which quenches the intrinsic luminescence of BSA 16 times more strongly than 128a,b. The complexes with ester (129) and amide (131) substituents affect to a much lesser extent the protein luminescence intensity. In the case of amide-containing clathrochelates 131, the most pronounced quenching is observed for the ortho-isomer 131c, which quenches the protein luminescence almost two times more efficiently than complexes 131a,b. ²⁸⁰

Scheme 24 

Zoom In Zoom Out Reset image size

Shykin et al. ^{281, 282} reported efficient photochemical splitting of water on exposure to visible light catalyzed by iron(IV) oxalyl hydrazonate clathrochelate 132. The possible catalytic cycle for this fast and efficient homogeneous photochemical decomposition of water to O₂ is depicted in Scheme 25 and includes the formation of Fe ^V -containing intermediate under catalytic conditions.

Scheme 25 

Zoom In Zoom Out Reset image size

The first 'classical' iron(I) dioximate 133 was prepared as shown in Scheme 26 ²⁸³ and was characterized by X-ray diffraction. The chemical reduction of the macrobicyclic precursor — iron(II) hexachloroclathrochelate 106 — with potassium graphite in the presence of 18-crown-6 ether (134) results in a clathrochelate with an encapsulated iron(I) ion. Complex 133 and its cobalt(I)-encapsulating analogues were highly chemically stable to strong H-acids such as trifluoroacetic acid.

Scheme 26 

Zoom In Zoom Out Reset image size

Ozvat et al. ²⁸⁴ studied the effect of encapsulation of the cobalt(II) ion by the dinitrosarcophagine ligand on the temperature dependence of the ⁵⁹ Co NMR chemical shifts; this chemical shift can be used to determine the local temperature (so-called NMR thermometry). It was shown that encapsulation of this low-spin ion (s = 1/2) increases its temperature sensitivity. The temperature dependence of the chemical shift (δ) in the ⁵⁹ Co NMR spectrum (Δδ/ΔT ratio) for the series of low-spin cobalt(II) complexes 135 – 139 increases with increasing degree of encapsulation of cobalt(II) ion: hexaammoniate 135 with monodentate ligands had a value of 1.44(2) ppm °C⁻¹, while the value for macrobicyclic dinitrosarcophagine complex 139 was 2.04(2) ppm °C⁻¹. ²⁸⁴

The temperature-dependent NMR spectroscopy can serve for highly accurate determination of the anisotropy of the magnetic susceptibility tensor (Δχ) for paramagnetic d-metal complexes. This approach was successfully implemented ²⁸⁵ to study cholesteryl-containing cobalt(II) pseudo-clathrochelates 140, 141 with a high anisotropy of the magnetic susceptibility tensor (Scheme 27). The conformational rigidity and large size of the apical cholesterol moiety, which contains a large number of magnetically nonequivalent nuclei of hydrogen and carbon, ensured a high convergence of the obtained experimental and theoretically calculated¹H and ¹³ C NMR chemical shifts. This allowed for highly accurate determination of Δχ and the zero-field splitting energy for cobalt(II) complexes of this type.

Structures 135 – 139 

Zoom In Zoom Out Reset image size

Scheme 27 

Zoom In Zoom Out Reset image size

Paramagnetic complexes 140 and 141 were also proposed ²⁸⁵ as promising single-molecule magnets and paramagnetic labels for determining the structure of biological macromolecules (in particular, cholesterol-binding proteins) and as model compounds for determining chirality using NMR spectroscopy.

2.5.1.1. Keplerates containing {(Mo)Mo₅} type pentagonal units

The reaction of keplerate 142 with a solution of FeCl₃ · 6H₂O at pH 3 is accompanied by replacement of the dimolybdenum {Mo₂O₄}²⁺ bridges by the {Fe(H₂O)}³⁺ groups, This gives rise to compound [Mo₇₂Fe₃₀O₂₅₂(AcO)₁₂{Mo₂O₇(H₂O)}₂{H₂Mo₂O₈(H₂O)}(H₂O)₉₁] · 150 H₂O (143) (the condensed formula is {Mo₇₂Fe₃₀}), which contains neutral discrete spherical molecules. The {Mo₂O₇(H₂O)} and {H₂Mo₂O₈(H₂O)} moieties (Fig. 68), which are coordinated in the keplerate cavity in the tetradentate fashion, are disordered. ²⁸⁹

Zoom In Zoom Out Reset image size

The Raman spectra for the solution and crystals of compound 143 are identical, which attests to the stability of the spherical shell upon dissolution. Magnetochemical measurements reveal weak antiferromagnetic exchange interactions between the high-spin Fe³⁺ ions. ²⁸⁹ Owing to the high spin ordering and high total magnetic moment, the molecule of keplerate 143 is one of the most highly symmetrical magnetic bodies. ²⁹⁰ In dilute aqueous solutions, the molecules of 143 are rapidly assembled into large spherical aggregates with a hydrodynamic radius of 34 nm, comprising approximately 600 {Mo₇₂Fe₃₀} spherical molecules (Fig. 69). The aggregate size can be controlled by varying the keplerate concentration. ^{291, 292} The influence of the solution pH on the assembly was studied. When pH < 2.9, keplerate 143 is neutral and exists in solution as single species, whereas in the pH range of 2.9 – 6.6, self-assembly takes place. ²⁹³ Complexes of nanosized aggregates with dimethyldioctadecylammonium (DODA) cation were prepared; this allowed the aggregates to be transferred into organic phase without degradation. Upon the coordination of DODA, the hydrodynamic radius of the spherical aggregates increases to 60 nm. ²⁹⁴

Zoom In Zoom Out Reset image size

A decrease in the pH is accompanied by the formation of the neutral keplerate [Mo₇₂Fe₃₀O₂₅₂(AcO)₁₀{Mo₂O₇(H₂O)}.{H₂Mo₂O₈(H₂O)}₃(H₂O)₉₁] · 140 H₂O (144), which has two acetate ligands less and the charge of which is counterbalanced by an additional {H₂Mo₂O₈(H₂O)}^2– oxomolybdate unit. When the crystals of keplerate 144 are dried in air, the solid-state condensation reaction takes place to give Fe–O–Fe bonds between the keplerates (Fig. 70). The compound thus formed has a network structure with the general formula {[H₄Mo₇₂Fe₃₀O₂₅₄(AcO)₁₀{Mo₂O₇(H₂O)}.{H₂Mo₂O₈(H₂O)}₃(H₂O)₈₇] · 80 H₂O}_n (145). ^{295, 296}

Zoom In Zoom Out Reset image size

The reaction of iron(II) sulfate with a sodium molybdate solution acidified with sulfuric acid gives rise to the partially reduced keplerate (146). ²⁹⁷ The cavity of anion 146, unlike that of keplerate 143, is occupied by sulfate ligands, while the 20 keplerate pores confined by the {Fe₃Mo₃O₆} rings are occupied by the K⁺ ions.

The reaction of a sodium molybdate solution acidified with a considerable amount of glacial acetic acid with iron(III) chloride gives the keplerate Na₄[{Mo₆O₁₉}⊂{Mo₇₂Fe₃₀O₂₅₂(AcO)₂₀(H₂O)₉₂}] · 120 H₂O (147). ²⁹⁸ Unlike keplerates 143 and 144, complex 147 contains, in the internal cavity, the hexamolybdate [Mo₆O₁₉]²–with the Lindqvist structure, which is not linked to the keplerate shell by covalent bonds. The addition of salts FeCl₂ · 4H₂O or FeCl₃ · 6H₂O to an aqueous solution containing [PMo₁₂O₄₀]^3– in air, some of Keggin anions are destroyed to give the {Mo₆O₂₁(H₂O)₆} building blocks, which react with iron ions. ^{299, 300} The remaining Keggin anions act as templates for the assembly, resulting in the O_Keggin...O_keplerate shortest distances of approximately 2.6 Å, which is typical of hydrogen bonding (Fig. 71).

Zoom In Zoom Out Reset image size

When FeCl₂ · 4H₂O is used, the reaction gives [H_x PMo₁₂O₄₀⊂Mo₇₂Fe ^III ₃₀(AcO)₁₅O₂₅₂(H₂O)₁₀₂] · 120 H₂O (148), which is a keplerate molecule with the encapsulated partially reduced [H_x PMo₁₂O₄₀]^3– anion. In the case of FeCl₃ · 6H₂O, fully oxidized keplerate [PMo₁₂O₄₀⊂Mo₇₂Fe^III ₃₀ (AcO)₁₅O₂₅₂(H₂O)₁₀₂] · 120 H₂O (149) is formed. Similar compounds, Na₆[SiMo₁₂O₄₀⊂Mo^VI ₆₈Mo^V ₄Fe^III ₃₀O₂₅₂(AcO)₁₆(H₂)₁₀₀] · 120 H₂O (150) ²⁹⁷ and (151), ³⁰¹ were obtained with the Keggin anions [SiMo₁₂O₄₀]^4– and [BW₁₂O₄₀]^5–.

When iron(II) reacts with a solution of sodium molybdate in the presence of acetic acid, the compound (152) is formed. ³⁰² The {Fe ^III (H₂O)₂} and {Mo ^V O(H₂O)} bridges in this product are randomly distributed over thirty positions.

The partial oxidation of keplerate 142 with air oxygen in a solution with pH 2.5 in the presence of NaCl gives dark blue crystals of [{Mo₆O₂₁(H₂O)₄(AcO)}₁₂{Mo ^V O(H₂O)}₃₀]·150 H₂O (153). Apart from the thirty Mo ^V metal centres, which constitute the {MoO(H₂O)}³⁺ bridges, the structure contains six more Mo ^V atoms, which are delocalized and fall within the {Mo₆O₂₁(H₂O)₆} pentagonal units. The presence of 36 Mo ^V centres was confirmed by the results of cerimetric titration. An interesting feature of the O=Mo ^V (H₂O) bridge is that the terminal oxygen atoms point inside the spherical anion. ³⁰³

The reaction of chromium(III) chloride with sodium molybdate in acidified aqueous solution with high acetic acid concentration gave the keplerate [{Na(H₂O)₁₂}⊂{Mo₇₂Cr^III ₃₀O₂₅₂(AcO)₁₉(H₂O)₉₄}] · 150 H₂O (154). ³⁰⁴ An interesting feature of this complex is the occurrence of the {Na(H₂O)₁₂}⁺ associate in the keplerate anion cavity, with the oxygen atoms of {Na(H₂O)₁₂}⁺ forming an icosahedron. The associate is stabilized owing to the hydrophobic environment formed by acetate ligands inside the shells. The behaviour of complex 154 in solution resembles that of the iron-containing keplerate 143: at low pH (<2.7) the compound exists as discrete species, while the self-assembly occurs in the pH range from 2.7 to 7.0. ³⁰⁵

The addition of vanadyl sulfate to an acidified solution of sodium molybdate in the presence of potassium cations results in the formation of the keplerate anion [K₁₀⊂{Mo₆O₂₁(SO₄)(H₂O)₃}₁₂{(V ^IV O)₃₀(H₂O)₂₀}]^26–({Mo₇₂V₃₀}) (155), which was isolated as the mixed salts K₁₄Na₈(VO)₂ · [155] · 150 H₂O (Ref. 306) and Na₈K₁₆(VO)(H₂O)₅ · [155] · 150 H₂O. ³⁰⁷ The assembly of keplerate anions 155 in solution was studied by Kistler et al. ³⁰⁵ Unlike neutral keplerates 143 and 154, the keplerate {Mo₇₂V₃₀} has a high negative charge and exists in aqueous solutions as discrete anions. The addition of acetone to a solution of anion 155 induces the formation of spherical aggregates, the size of which increases with increasing concentration of acetone. The reduction of aqueous solutions of sodium molybdate and vanadyl sulfate with sodium dithionite in the presence of potassium cations gave the product K₂Na₆[K₂₀⊂{Mo₆O₂₁(SO₄)(H₂O)₃}₁₂.{V ^IV O(H₂O)}₂₂{Mo ^V O(H₂O)}₈] · 140 H₂O (156). ³⁰⁸ Each {(Mo)Mo₅} pentagonal unit coordinates the sulfate anion in the tridentate mode, while twenty potassium cations plug the keplerate pores, being linked to six oxygen atoms.

Two mixed-metal keplerate complexes, K₂₃Na₄[{Mo^V/VI ₆O₂₁(H₂O)₃(SO₄)}₁₂(V^IVO(H₂O))₁₅(Mo^VO(H₂O))₈(Fe^III(H₂O)₂)₇] · 150H₂O (157) and K₂₁Na₂[{Mo^V/VI ₆O₂₁(H₂O)₃(SO₄)}₁₂(V^IVO(H₂O))₁₁(Mo^VO(H₂O))₈(Fe^III(H₂O)₂)₁₁] · 150 H₂O (158), were prepared by adding vanadyl and iron(II) sulfates to an acidified solution of sodium molybdate. ³⁰⁹ The [Fe ^III (H₂O)₂]³⁺, [V ^IV O(H₂O)]²⁺ and [Mo ^V O(H₂O)]³⁺ bridges present in this structure are statistically disordered over 30 positions.

It was shown that Fe³⁺ ions in keplerate 143 can be partly replaced by larger cations, such as Ce³⁺ and Pr³⁺. The addition of the corresponding lanthanide chlorides resulted in the synthesis of [{Mo₁₈O₆₆Ln₂(H₂O)_n }⊂{Mo₇₂Fe₂₄Ln₆O₂₅₂(H₂O)₁₀₅}] · 200 H₂O, where Ln = Ce (159) and Pr (160). ³¹⁰ The six lanthanide bridging ions are disordered over 12 positions and are slightly elevated above the keplerate surface. The structure of the moiety designated as {Mo₁₈O₆₆Ln₂(H₂O)_n } and encapsulated in the keplerate cavity could not be reliably determined because of its high disorder.

2.5.1.2. Keplerates containing {(W)W₅} type pentagonal units

The keplerate anion [(NH₄)₂₀⊂{W₆O₂₁(SO₄)}₁₂{Fe(H₂O)}₃₀.(SO₄)₁₃(H₂O)₃₄]^12– (161) is formed upon the reaction of sodium tungstate with iron(II) sulfate in the presence of ammonium chloride. ³¹¹ The pores of keplerate 161 confined by the {Fe₃W₃O₆} rings are occupied by twenty ammonium cations.

The reaction of a tungstate solution acidified with sulfuric acid with vanadyl sulfate in the presence of potassium cations affords the keplerate K₁₄(VO)₂[K₂₀⊂{W₆O₂₁(SO₄)}₁₂(VO)₃₀(SO₄)(H₂O)₆₃] · 150 H₂O (162). ³¹² Twelve out of the thirteen sulfate ligands located in the internal cavity are coordinated to twelve {(W)W₅} units in the tridentate fashion, while one more sulfate ligand links one V atom and two W atoms and is disordered over eight positions. Each of the 20 pores confined by the {V₃W₃O₆} rings accommodates a potassium cation (Fig. 72).

Zoom In Zoom Out Reset image size

One more method that can be used to prepare keplerate anions {W₇₂(VO)₃₀} comprises the reaction of sodium tungstate with ammonium vanadate in the presence of potassium sulfite and hydrazine sulfate. The vanadyl cations are formed in solution in situ and constitute the structure of keplerate 163, which was isolated as the complex K₂Na₁₄[Na₁₄K₆⊂{W₆O₂₁(H₂O)₃(SO₄)}₁₂{VO(H₂O)}₃₀]·90 H₂O. In this case, no additional disordered sulfate ligand is present in the cavity, and the pores are occupied by sodium and potassium cations. ³¹³

2.5.2.1. Keplerates containing {(Mo)Mo₅} type pentagonal units

As has already been noted, the first known keplerate (142) was obtained from a solution of ammonium molybdate in the presence of acetate ligands. ²⁸⁶ An important role in this reaction belongs to pH of the solution, because it is necessary to create conditions for the formation of the {Mo ^VI O₇} pentagonal bipyramidal groups, which form the basis for the pentagonal building blocks. Upon the addition of hydrazine sulfate, molybdenum centres are partly reduced and are converted to bridging groups, which are stabilized due to the high concentration of bidentate acetate ligands. The assembly of keplerate 142 was studied in a water – acetone solvent mixture. It was shown that the presence of counter ions is of crucial importance, because both the assembly and destruction of aggregates depend on the absolute magnitude of the charge of keplerate: neutral species and highly charged anions do not form aggregates. The aggregate size can be controlled by changing the component ratio in the water – acetone system. ³¹⁴ The reaction of an aqueous solution of keplerate 142 with a solution of DODA gives the corresponding salt (DODA)₄₀(NH₄)₂[(H₂O)₅₀⊂Mo₁₃₂O₃₇₂(AcO)₃₀(H₂O)₇₂] (164), which immediately migrates to the organic phase. ^{315, 316}

When a solution of keplerate 142 is kept in air in the presence of Na⁺ cation, the product (165) is formed. ³¹⁷ The basket-shaped structure of the anion can formally be derived from the structure of anion 142 by removing one group (Fig. 73). Compound 165 is substantially more readily soluble in water than complex 142, owing to the easy access of solvent molecules into the cavity.

Zoom In Zoom Out Reset image size

The keplerate anions [Mo₁₃₂O₃₇₂(HCO₂)₃₀(H₂O)₇₂]^42–(166), ³¹⁸ [Mo₁₃₂O₃₇₂(ClCH₂CO₂)₃₀(H₂O)₇₂]^42– (167) ³¹⁸ and [Mo₁₃₂O₃₇₂(EtCO₂)₃₀(H₂O)₇₂]^22– (168) ³¹⁹ were prepared by reaction of hepta-ammonium molybdate with formic, monochloroacetic and propionic acids in the presence of hydrazine sulfate. When butyric acid was used, the [Mo₁₃₂O₃₇₂(PrⁿCO₂)₂₄(H₂O)₈₄]^16– keplerate anion (169) was formed, with 24 butyrate ligands being coordinated in the cavity, which is caused by steric reasons. ³²⁰ In the inner cavity of anion 166, self-organization of water molecules takes place to form onion type hydrogen-bonded cluster. Six H₂O molecules form a nearly trigonal antiprism with a 3.5 Å distance from the centre of the onion. The second shell (with 6.2 – 6.9 Å size) contains approximately 35 closely packed H₂O molecules. The third shell (8.2 – 8.7 Å) consists of a smaller number of water molecules, which occupy positions between bidentate ligands in the pore region. This unusual arrangement is obviously possible only in the presence of relatively small formate ligands, which coordinate the groups.

A specific feature of anions 168 and 169 is that the associate with the O–N distance of 2.7 Å, disordered over nine possible positions, is incorporated in each of the twenty pores of the shell. These keplerates were shown to encapsulate molecules of hydrophobic alcohols into the cavity.

One more method for preparing new keplerate anions is replacement of labile acetate ligands in the cavity of keplerate 142. For example, the reaction with oxalic acid (H₂C₂O₄) yielded keplerate (NH₄)₄₂[{Mo₆O₂₁(H₂O)₆}₁₂·{Mo₂O₄(O₂CCO₂H)}₃₀] · {300 H₂O+2 H₂C₂O₄} (170), in which the acetate ligands have been replaced by oxalate ligands. ³²¹ When a solution of compound 170 is kept at room temperature, partial deprotonation and dissociation of oxalate ligands take place to give the [{Mo₆O₂₁·(H₂O)₆}₁₂{Mo₂O₄(O₂CCO₂H)}₈{Mo₂O₄(O₂CCO₂)}₁₆{Mo₂O₄.(H₂O)₂}₆]^52– keplerate anion (171). X-Ray diffraction data confirm the presence of 24 partially deprotonated oxalate ligands, which are statistically distributed over thirty positions, and six {Mo₂O₄}²⁺ linkers each containing two water molecules as ligands.

The introduction of a reactive azide group into the keplerate cavity was attained by replacing the acetate ligands in compound 142 in the presence of excess azidoacetic acid, which gave [{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄.(N₃CH₂CO₂)}₁₃{Mo₂O₄(H₂O)₂}₁₇]^25– (172). ³²² The reaction of complex 172 with excess propiolic acid at room temperature afforded a mixed-ligand keplerate with coordinated azidoacetate and propiolate groups over a period of several minutes. ³²² On standing of the solution, the Huisgen reaction product, keplerate [Mo₁₃₂O₃₇₂(triazole)₁₁(H₂O)₁₁₀]^23–(173), where triazole-H is 4-carboxy- or 5-carboxy-1,2,3-triazole-1-acetic acid, was isolated in a quantitative yield. The NMR spectrum of compound 173 exhibited no signals for the coordinated azidoacetic or propiolic acid, but showed signals for free 4-carboxy- and 5-carboxy-1,2,3-triazole-1-acetic acids in 4 : 1 ratio relative to the keplerate.

The possibility of replacing acetate groups in the keplerate anion with isobutyrate, pivalate and benzoate ligands was studied by¹H NMR spectroscopy. ³²³ For this purpose, keplerate 142 was several times treated with a 4 M solution of sodium acetate buffer; this gave compound 174, which was described as Na₃₄[{Mo₆O₂₁(H₂O)₆}₁₂{(Mo₂O₄)₃₀.(AcO)₂₂(H₂O)₁₆}] · 300 H₂O according to the data of elemental analysis. The data of¹H NMR spectroscopy also confirmed the presence of 22 acetate ligands per each keplerate. When aliquot portions of acids RCO₂D (R = Prⁱ, Bu^t, Ph) were added to a solution of compound 174 in D₂O, the substitution took place only in the case of isobutyric and pivalic acids, but not with benzoic acid. Whereas for R = Prⁱ, the equilbrium in the substitution reaction is attained as soon as within 17 min, in the case of R=Bu^t, this requires 45 days. This is attributable to the size of substituent at the carboxyl group, which is smaller than the keplerate pores in the former case. For larger Bu^t group, entering the cavity requires a minor change in the bond lengths and angles in the {Mo₉O₉} group and/or other reversible changes in the coordination environment of the Mo metal centres.

The acetate ligands can also be replaced by other oxygen-containing anion groups. The reaction of keplerate 142 with sodium hypophosphite at pH 2 resulted in the isolation of product (NH₄)₄₂[{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(H₂PO₂)}₃₀]·300 H₂O (175), in which the keplerate core was preserved without changes, except that all thirty acetate groups have been replaced by hypophosphite groups. ³¹⁷ The addition of keplerate 142 to a solution of a mixture of sodium dihdrogen phosphate and ammonium chloride at pH 5 afforded the anionic complex [{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(HPO₄)₃₀}]^72– (176). ³²⁴ It is of interest that in this case, the solution pH had to be increased (by adding, e.g., NH₄OH) rather than decreased, as in other cases. If this reaction is carried out at pH 2, partial decomposition of keplerate 142 takes place, giving rise to some free molybdate. The latter enters the cavity as {Mo ^VI O₃H} groups, each being bonded to three phosphate ligands surrounding the {Mo₉O₉} pores from the inside. In view of geometric considerations, up to eight molybdenum groups can be accommodated in the cavity; however, only five {MoO₃H} groups were actually found, which corresponds to the formation of [{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(HPO₄)₃₀}(MoO₃H)₅]^67– (177). ³²⁴

Refluxing of a solution of keplerate 142 in the presence of excess (NH₄)₂SO₄ at pH 1 is accompanied by replacement of the acetate ligands by thirty sulfate groups, thus giving keplerate [(H₂O)_81–n +(NH₄)_n ⊂{Mo₆O₂₁(H₂O)₆}₁₂·{Mo₂O₄(SO₄)}₃₀]^(72–n)– (178). ³²⁵ A similar reaction at pH 3 leads to partial replacement of acetate ligands to yield the anionic complex [{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₁₈.{Mo₂O₄(AcO)}₁₂]^60– (179). ³²⁵ The addition of CeCl₃ during the synthesis of anion 178 induces the formation of keplerate Ce₃(NH₄)₅₄[Ce₃⊂{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₃₀]·300 H₂O (180). ³²⁴ Three cerium cations reside in the cavity of this compound, each of them being bound to the coordinated sulfate ligand in the bidentate fashion. In the presence of excess urea and sodium chloride, [{(NH₂C(O)NH₃)₂₀+Na₁₃}⊂{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₃₀]^39– (181) is formed. ³²⁴ In this compound, 20 protonated urea cations can fit inside the {Mo₉O₉} rings, which confine the pores, and they are hydrogen-bonded to the oxygen atoms, while each of the 13 sodium cations inside the cavity is linked to six oxygen atoms, three of which belong to sulfate groups and the other three belong to the aqua ligands (Fig. 74).

Zoom In Zoom Out Reset image size

When keplerate 142 reacts with sodium sulfate at pH 2.5 and at room temperature, the acetate ligands are not completely replaced, but the process stops after the formation of the [Na₆+(H₂O)_n ⊂{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₂₄.{Mo₂O₄(AcO)}₆]^60– anion (182). ³²⁵ Meanwhile, if the reaction is carried out at pH 4.5 with long-term holding of the reaction mixture at 30 °C, the [{Mo₆O₂₁(H₂O)₆}₁₂.{Mo₂O₄(HSO₄)}₁₈{Mo₂O₄(AcO)}₁₂]^42– keplerate anion (183), containing hydrogen sulfate groups, is formed. ³¹⁷

The reaction of keplerate 142 with Cs₂SO₄ carried out with heating (80 °C) gives the product Cs₁₀(NH₄)_52–n ·[{Cs₁₀+(H₂O)_81–n +(NH₄)_n }⊂{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₃₀]·200 H₂O (184). ³²⁴ The caesium cations are coordinated to keplerate pores, but due to too small size, they are not located at the centre, but are slightly shifted towards the sides. The reaction of complex 178 with calcium chloride resulted in the synthesis of Ca₁₆{Ca₂₀(H₂O)₆₀}⊂{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₃₀] (185). ³²⁶ Twenty Ca²⁺ cations are arranged in the inner cavity of the keplerate, being bound to the pores from the inside, while the other are coordinated in the pores and on the outer surface of the shells.

When CO₂ is bubbled through a neutral solution of keplerate 142 in the presence of formamidinium chloride, the anion [(HC(NH₂)₂)₂₀⊂{Mo₆O₂₁(H₂O)₆}₁₂.{Mo₂O₄(CO₃)}₃₀]^52– (186) is formed. ³²⁷ In compound 186, all 20 pores are plugged with formamidinium cations, which are hydrogen-bonded to the oxygen atoms of the {Mo₉O₉} rings. The ammonium salt of the same anion — (NH₄)₇₂[{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(CO₃)}₃₀] · 260 H₂O (186') — was prepared in a similar way. At room temperature, salts 186 and 186' gradually decompose, which is associated with the release of coordinated CO₂; salt 186 decomposes much more slowly due to the stabilizing effect of closed pores.

It was demonstrated that the hypophosphite ligands in (NH₄)₄₂[{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(H₂PO₂)}₃₀] · 300 H₂O(175) can be partially substituted. ³¹⁷ On short-term heating (90 °C) of a solution of compound 175 with guanidinium sulfate, the guanidinium salt of the keplerate anion [{[(NH₂)₃C]}₂₀⊂{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₄(SO₄)}₁₀.{Mo₂O₄(H₂PO₂)}₂₀]^32– (187) is formed in a nearly 100% yield. ³²⁸ In this product, 20 guanidinium cations are coordinated to keplerate pores via hydrogen bonds (Fig. 75).

Zoom In Zoom Out Reset image size

Apart from the dimolybdenum bridging group {Mo₂O₂(μ-O)₂}²⁺, there are other known examples of binuclear linkers. A sulfide analogue — keplerate Na₃₈K_3.5(NMe₄)_3.5[{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₂S₂(AcO)}₃₀]·(AcO)₃ · 200 H₂O (188) ³²⁹ — was obtained by the reaction of ammonium molybdate with a source of the {Mo₂O₂(μ-S)₂}²⁺ cations (Fig. 76). ³³⁰ A similar reaction carried out in the presence of sulfate anions yielded the keplerate (NMe₂H₂)₂₇(NH₄)₁₅[{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₂S₂}₃₀.(SO₄)₁₅(H₂O)₃₀] · 90 H₂O (189), ³³¹ in which the sulfate ligands were disordered over thirty possible positions in the {Mo₂O₂S₂} groups. Recrystallization of compound 188 gave a structure with the [{Mo₆O₂₁(H₂O)₆}₁₂{Mo₂O₂S₂}₃₀(AcO)₁₅(H₂O)₃₀]^27– anion (190) characterized by a similar disorder of the coordinated acetate ligands. ³²⁹

Zoom In Zoom Out Reset image size

2.5.2.2. Keplerates containing {(W)W₅} type pentagonal units

The first keplerate with dimolybdenum bridges containing type pentagonal units was reported in 2009. ³³² It was synthesized by adding a slight excess of the binuclear complex to a solution of sodium tungstate acidified with acetic acid. Apart from thirty acetate ligands coordinated to the dimolybdenum bridging moiety, the structure contains six disordered acetate ligands, which are coordinated to pentagonal units and bind the central W atom to a peripheral one. Thus, the keplerate anion has the formula [{W₆O₂₁(H₂O)₅(AcO)_0.5}₁₂{Mo₂O₄(AcO)}₃₀]^48– (191). In solutions of compound 191 containing 30 – 75% acetone, spherical aggregates are formed, and their size increases with increasing acetone concentration. This result is consistent with the data that were obtained earlier for analogous molybdenum keplerate 142. ³¹⁴

When an acidified solution of keplerate 191 reacts with ammonium selenate, thirty coordinated acetate ligands are replaced by selenate ligands. ³³³ Six additional acetate ligands are also substituted, but by water molecules. On the basis of elemental analysis data, this keplerate was described by the formula (NH₄)₇₂[{W₆O₂₁(H₂O)₆}₁₂.{Mo₂O₄(SeO₄)}₃₀] · 150 H₂O· 6(NH₄)₂SeO₄ (192). Recrystallization of compound 192 in the presence of dimethylammonium cations affords the keplerate (NH₄)₂₀(NH₂Me₂)₃₂[{W₆O₂₁(H₂O)₆}₁₂{Mo₂O₄}₃₀(SeO₄)₂₀.(H₂O)₂₀] · 150 H₂O (193). Some of the selenate ligands are lost during recrystallization, with their sites being occupied by coordinated water molecules.

Using dimolybdenum monosulfide building block, [Mo₂O₂(μ-O)(μ-S)]²⁺ (Ref. 334), the keplerate [{W₆O₂₁(H₂O)₆}₁₂{Mo₂O₃S(AcO)}₃₀]^42– (194) was prepared. ³³⁵ When a solution of this anion was acidified with sulfuric acid to pH 2 in the presence of methylammonium chloride, the product Na₄(MeNH₃)₆₈[{W₆O₂₁(H₂O)₆}₁₂.{Mo₂O₃S(SO₄)}₃₀] · 230 H₂O (195), containing sulfate ligands, was obtained. Each of the thirty dimolybdenum moieties is disordered with 50% probability between the bridging oxide and sulfide ligands; this gives rise to four types of pores on the keplerate surface — {W₃Mo₆O₉}, {W₃Mo₆O₈S}, {W₃Mo₆O₇S₂} and {W₃Mo₆O₆S₃} — differing in both the size and reactivity (Fig. 77).

Zoom In Zoom Out Reset image size

The reaction of sodium tungstate with the {Mo₂O₂(μ-S)₂}²⁺ cation ³³⁰ gives rise to [{W₆O₂₁(H₂O)₆}₁₂{Mo₂O₂S₂(AcO)}₂₀{Mo₂O₂S₂(H₂O)₂}₁₀]^32– (196). ³³⁶ Upon the addition of CoCl₂ · 6H₂O to a solution of anion 196, the crystals of [{W₆O₂₁(H₂O)₆}₁₁{W₆O₂₁.(H₂O)(H₂Mo₂O₈(H₂O))}(Mo₂O₂S₂)₃₀(AcO)₁₀(H₂O)₄₀]^24– (197) are formed. The {H₂Mo₂O₈(H₂O)} inner group, which was detected previously in keplerate 143, ²⁸⁹ is coordinated to five oxygen atoms of the {(W)W₅} unit and is disordered over four pentagonal units of this type.

The diversity of keplerates is notable '...not only from an aesthetic point of view, but also because they show properties of interest for different areas of science', ³³² and this will be clear from the examples given below.

2.5.3.1. Conduction of chemical reactions within the shell

2.5.3.2. Catalytic reactions involving keplerates

2.5.3.3. Production of materials based on keplerates

2.5.3.4. Potential biomedical applications of keplerates

Actually, an attempt to find ways for real integration of analytical and supramolecular chemistry initiated the formation of a new branch of analytical chemistry. ³⁶⁸ Simultaneously, the fundamental questions related to supramolecular analytical chemistry were formulated as follows:

• —  
  which types of chemical reactions should be classified as supramolecular interactions and whether the general definition proposed by J.M.Lehn ¹ fully covers this subject;
• —  
  whether there is any threshold in energy or in the time of intermolecular interaction to distinguish between supramolecular and other types of interactions, for example, covalent exchange interactions;
• —  
  whether the supramolecular nature is inherent in spontaneous selective reactions of complex formation (Werner coordination) between metal ions and organic reagents, widely used by analytical chemists, and whether these reactions should be assigned to supramolecular chemistry;
• —  
  whether dynamic covalent interactions based on the formation and exchange of reversible covalent bonds can be assigned, together with the classical non-covalent supramolecular interactions, to supramolecular analytical chemistry; ³⁶⁹

The pioneering publication of Anslyn ³⁶⁸ presents a lot of examples of applying host – guest pH systems, some metal ions, inorganic and organic anions and various uncharged organic compounds and enantiomers as optical and electrochemical sensors. The use of arrays of non-selective sensors such as electronic nose and electronic tongue, in particular for the recognition of the odour and taste of food and other volatile products, was quite rightly attributed to supramolecular processes.

More appropriate definition of the term 'supramolecular analytical chemistry', further development of this concept and classification and analysis of recent achievement in this field were surveyed by Anslyn and co-workers. ³⁶⁹ The supramolecular analytical chemistry is defined as a field that explores the molecular recognition and self-assembly of chemical structures using dynamic interactions that create assemblies which result in signal modulations upon addition of analytes. ³⁶⁹ The authors proposed to subdivide sensory interactions into three types: sensing of single analytes, sensing based on competing interactions of substances and chirality sensing. For each of these types, several practical approaches to the design of supramolecular sensors are distinguished:

• —  
  direct detection based on the selective interaction of the host (receptor) and analyte (substrate) molecules;
• —  
  indirect detection associated with displacement of the indicator (probe) molecule by analyte molecule from the cavity of the host;
• —  
  competitive detection requiring an array of nonselective sensors functioning as an electron nose or electron tongue and subsequent use of chemometric methods for the processing of analytical signal with a complex geometric shape.

Spanish researchers ^{375 – 379} paid attention to another aspect of analytical application of host – guest systems and self-organized assemblies formed mainly by diphilic micelle-forming surfactant molecules, that is, their application in the separation and preconcentration and in determination methods coupled to them. The interaction of receptor molecules and polymolecular assemblies with an analyte is of supramolecular nature, i.e., it is based on self-assembly via non-covalent electrostatic, hydrophobic and donor – acceptor interactions and hydrogen bonding. Analysis of the literature shows that cyclodextrins and calixarenes, which are much more readily available and less expensive than spherands, cavitands or carcerands, are used most often in chromatographic analysis, capillary electrophoresis and extraction with receptor molecules acting as hosts.

The so-called micellar nanosized objects or systems are used even more widely as analytical reagents for separation, preconcentration and determination of compounds. These reagents include direct and reverse micelles and microemulsions, liposomes and vesicles formed by surfactant molecules and even Langmuir – Blodgett films. ^{380 – 383} The term 'micellar systems' covering all these objects was derived from the name of the simplest representatives of these self-organized systems, that is, surfactant micelles. ^{382, 383} Indeed, the presence of diphilic surfactant molecules as the major constituent of a supramolecular assembly, the way of their spontaneous formation and similarity of their properties are common features of all organized systems of this type. ^{382, 383} Generally, the media in which the major solvent contains receptor molecules or micellar systems are called organized media. ^{380 – 383} These media are usually transparent, optically isotropic solutions in which the major bulk of the solvent (aqueous or non-aqueous one) contains nanosized pseudophase systems able to solubilize (co-dissolve) substances insoluble in this solvent. ³⁸³ Thus, liquid organized media are homogeneous and are single-phase on a macroscale, but are nanoheterogeneous and two-phase on a nanoscale.

Extensive use of supramolecular organized media started back in the 1980s. ^{380, 381} In the 21st century, Spanish researchers ^{375 – 379, 384} proposed to call these systems supramolecular solvents (SUPRAS), because, from the standpoint of the authors, they replaced the classical aqueous, aqueous organic and non-aqueous media by biosimilar ones. According to these views, supramolecular analytical chemistry was redefined as 'an emergent science covering the study and application of supramolecular systems (host molecules and molecular aggregates) in analytical processes'. ^{375–379, 384}

It should be also borne in mind that the supramolecular chemistry is closely connected with nanochemistry and nanotechnologies, which are also based on bottom-up self-organization processes, and, hence, with nanoanalytical chemistry, which uses both nano-objects and self-assembly in various stages and in various fields of chemical analysis. ^{385 – 388}

The classical chemical analysis is not reduced to mere detection and quantitative determination of substances, but also includes the preceding sequence of interdependent processes (operations). Therefore, supramolecular analytical chemistry also should not be restricted to the recognition effect, i.e., qualitative analysis, use of host – guest systems and studying supramolecular assemblies as solvents of a new type ^{375 – 379} or nanoreactors. ³⁸³ It is necessary to consider all stages of analysis that involve supramolecular effects and that provide data for qualitative and quantitative determination of a sample component. These operations include sampling, sample conservation and preservation of the sample composition, sample preparation (including separation of components of a complex mixture and analyte preconcentration), conduction of the analytical reaction (process) under supramolecular conditions and calculation of the content of the analyte. In some cases, e.g., in strip tests and sensors, some of the indicated stages may be absent.

This part of the review is meant to briefly consider the role of supramolecular effects in particular stages of chemical analysis on the basis of both literature data and our own research. Examples of using supramolecular effects and systems for sample preparation, separation and preconcentration, as media for conducting chemical reactions or as nanoreactors are presented. In addition, we describe self-assembly phenomena that underlie the generation of the analytical signal (effect) and are implied in the procedures of quantitative determination in various methods of chemical analysis. We did not claim to cover all phenomena, systems, approaches and methods that can be attributed to supramolecular analytical chemistry, but we intended to outline the main trends, goals and achievements of this new, emerging field of analytical chemistry. The number of publications in which supramolecular effects and systems were utilized for the development of analytical procedures is several thousand, some of the original papers were integrated in reviews and monographs. ^{368, 369, 375 – 383, 389 – 398}

Analysis of the literature shows that two types of supramolecular objects (systems) are employed in the stages of chemical analysis (sample preparation, separation, preconcentration and determination). The first type are solutions of pre-organized receptor molecules (cyclodextrins, calixarenes, cavitands, cyclophanes, cyclopeptides, etc.), which act as hosts with respect to analytes. The second type includes solutions of self-organized micelle-like assemblies, which serve as new media for solubilization, assembly and bring together the components of analytical reactions, i.e., like receptor molecules, they actually function as nanoreactors. ³⁸³ Micellar systems can also be formed in non-aqueous solvents and on solid surfaces, being organized as hemimicelles and admicelles (adsorption bilayers).

Thus, supramolecular analytical chemistry, unlike nanoanalytics, is rather a new tool for conducting analytical reactions than a new branch of analytical chemistry. This tool is based on the use of receptor molecules and self-assembled molecular aggregates (micellar systems) as media (nanoreactors) for reaction components to approach each other and for changing local properties of the microenvironments of analytical systems, i.e., for creating conditions for the whole sequence of reactions to occur. It is noteworthy that both the analytical reactions and processes taking place in supramolecular media (cavities of receptor molecules) have supramolecular nature, because components interact according to the self-assembly principle and serve as sources of the analytical signal. In this connection, we will give a number of examples of supramolecular interactions and reactions of the supramolecular nature that have been widely used in analytical chemistry for approximately 40 years.

The simplest example of a supramolecular process used in chemical analysis is solubilization, i.e., an analyte migrates from the main bulk of the solvent into the receptor cavity or into the nanosized space of the micellar assembly. Solubilization results in increased solubility of poorly water soluble hydrophobic organic reagents and analytes and changes the polarity of their environment and their physicochemical properties even at the stage of sample preparation for the analysis and then throughout the whole analytical determination process. ^{380, 383, 387} Solubilization extends the range of compounds that can be determined in aqueous solutions by titrimetric, ^{380 – 382} photometric, ^{380 – 383, 389} luminescence, ^{389 – 391, 399} electrochemical, ³⁹² chromatographic ^{394, 395, 400 – 403} and electrophoretic ^{403 – 405} methods and by atomic absorption and atomic emission spectrometry. ⁴⁰⁶ The solubilization of participants of an analytical process in the nanopseudophase of an organized system (micelle, microemulsion, liposome, vesicle or cyclodextrin/calixarene cavity) considerably changes their hydrophobic properties, hydration, hardness and conformation of the molecule, which modifies physicochemical, spectroscopic, electrochemical, adsorption, protolytic, tautomeric, complexing, redox and other properties. ^{383, 385, 387, 388, 407, 408} This also changes the pattern of charge distribution in the molecule, the efficiency of intra- and intermolecular transfer of the electronic excitation energy and electron transfer, particle distribution among the phases, solubility and also the rate, route and equilibrium state of analytical reactions. As a result, the analytical signal intensity markedly increases, the spectroscopic, electrochemical and chromatographic parameters of the signal and conditions of electrophoretic separation are optimized and, as a consequence, the sensitivity and selectivity of analytical determination is enhanced.

A fundamental difference of microheterogeneous organized media from common homogeneous solutions (aqueous, aqueous organic and non-aqueous solutions) is that local effects play the crucial role in this case. ³⁸³ This means that changes in the properties of compounds solubilized in organized systems are caused by changes only in the state of their microenvironment, i.e., in their solvation shell, but not in the whole bulk of the solution. ⁴⁰⁹

It is possible to distinguish a number of consequences of solubilization and formation of ionic associates of organic reagents with surfactant counter-ions that are common to analytical systems: ^{381 – 383} , 389, 409, 410

• —  
  the ability to bring together and preconcetrate components of an analytical reaction in the nanopseudophase of an organized system, even if the components substantially differ in hydrophobicity;
• —  
  multicentre and multifunctional interactions (electrostatic, donor – acceptor, van der Waals and hydrophobic interactions and hydrogen bonding) of components or parts of the nanopseudophase with the solubilized substrate, with hydrophobic interactions being predominant;
• —  
  clear-cut oriented sorption and cavity effect, in which the nature and geometric matching between the host and the guest are the decisive factors for analyte binding to the substrate;
• —  
  significant microheterogeneity of the medium within the nanopseudophase along the direction from the interface with water (solvent) towards the centre, manifested as a sharp change in the dielectric constant, microviscosity, micropolarity, microacidity and other physicochemical properties of the medium.

An important advantage of micellar systems and receptor molecules is the possibility of controlling their composition and the properties of assemblies formed in solution (charge, hydrophobicity, micropolarity, etc.), and, hence, solubilization of reagents and analytes. This control is directly related to the supramolecular nature of assemblies and can be accomplished in several ways: ³⁸³

• (1)  
  targeted choice of the type of organized system: direct or reverse micelles, cyclodextrins, microemulsions [oil-in-water (o/w) and water-in-oil (w/o)];
• (2)  
  adjusting the nature of surfactants that form the organized systems, in particular varying the hydrophilic group (cation, anion, oxyethylene chain length) or hydrophobic moiety (aliphatic, aromatic, oxypropylene); varying the hydrophilic – lipophilic balance of surfactant molecules and analytes and the sizes of receptor cavity and the analyte;
• (3)  
  adjusting the properties of the organized nanosized assemblies, in particular:

The occurrence of solubilization not only considerably improves characteristics of the known methods for separation, preconcentration and determination of many inorganic and organic compounds, but has also resulted in the development of some new methods. The most well known among them is micellar extraction (called also 'cloud point extraction' or 'supramolecular solvents'), proposed in 1977 by Ishii and Watanabe. ⁴¹¹ It is successfully combined with determination of compounds, by, for example, gas ⁴¹² and liquid ^{394, 413} chromatography and capillary electrophoresis ^{406, 413} and with photometric, ^{380, 381, 406, 414} luminescence, ^{393, 415} atomic absorption and atomic emission ⁴⁰⁶ determination methods. There are two dozens of reviews dealing with the micellar extraction alone in sample preparation, spectrometric determination of metal ions and inorganic anions, organic compounds and biomolecules.

Other examples of the effective use of supramolecular effects include micellar liquid chromatography (MLC) ⁴¹⁶ and micellar thin layer chromatography, ⁴¹⁷ which were proposed by Armstrong and co-workers. Micellar TLC has been widely used in Russia and in Ukraine. ^{400, 418 – 422} One of the numerous reviews on MLC including analysis of both author's own works and studies of foreign colleagues ⁴⁰¹ was published in 1999. The most widespread among these methods are micellar electrokinetic chromatography (MEKC), developed by Terabe et al., ⁴²³ and its microemulsion version (MEEKC), proposed by Watarai. ⁴²⁴ The use of micelles and microemulsions provided an unusually high separation efficiency for both ionic and neutral compounds, because the number of theoretical plates reached in some cases 3 – 5 million. ⁴²⁵ The results of application of MEKC, MEEKC and capillary electrophoresis with cyclodextrin mobile phases are comprehensively integrated in a monograph. ⁴²⁵ The use of micelles and cyclodextrins allowed simultaneous separation of hydrophilic and hydrophobic, charged and neutral, optically active and inactive molecules, eliminated the necessity of regeneration of chromatographic column and made it possible to use electrochemical detector with gradient elution and to perform direct injection of biological fluids as micellar solutions into the column.

Organized media have found use almost in all types of luminescence spectrometry employed in chemical analysis: fast, photoinduced, sensitized fluorescence, room temperature phosphorescence (RTP) and sensitized RTP (s-RTP), immunofluorescence and chemiluminescence; this is reflected in a number of books and fundamental reviews. ^{380, 381, 426, 427, 382, 383, 389 – 391, 393, 399, 415} Spectral changes and analytical effects were found to be more significant in luminescence than in photometry. In some cases, 100- and 1000-fold increase in the analytical signal intensity was attained, and the limit of detection decreased to a nano-, pico- or femtogramme level. ^{390, 391, 426, 427} New results were obtained in phosphorimetric analysis, because it became possible to observe an analytical signal in micelles, microemulsions and cyclodextrins at room temperature rather than at liquid nitrogen temperature (77 K). ^{390, 391, 426} The main conditions for observing phosphorescence at room temperature are the presence of a heavy atom and removal of oxygen from the solution. ^{428 – 430} The sensitized RTP based on the triplet – triplet (T–T) transfer of the electron excitation energy proposed in our study appears most interesting. This procedure substantially increased the selectivity of the phosphorimetric determination of polycyclic aromatic hydrocarbons; this was the first implementation of the idea of using analytical organic reagents in the triplet state. ^{431 – 433}

One more type of luminescence based on the supramolecular transfer of electron excitation energy that is affected by organized media is the sensitized fluorescence of lanthanides. In this case, upon light absorption, a chromophoric chelating ligand passes from a singlet to a triplet excited state via intersystem crossing, and the energy transfer from the triplet level of the ligand to the lowerenergy resonance level of the lanthanide ion is accompanied by characteristic sensitized fluorescence. This conversion of the energy absorbed by one or several ligands to lanthanide ion emission as a result of intramolecular transfer of the excitation energy was called 'antenna effect'. ^{390, 399, 427, 434, 435} The chelate formation brings about one more effect, namely, displacement of lanthanide-coordinated water molecules by organic ligands, which results in quenching of the metal excitation energy by these water molecules. Therefore, mixed-ligand complexes are often used instead of complexes with one ligand. This allows displacement of more lanthanide-coordinated water molecules. The use of micellar media is favourable for the removal of coordinated water, increases the stability constant of chelates, changes the rates of photophysical processes and increases the efficiency of energy transfer and the sensitized fluorescence intensity. This is associated with the change in the microenvironment of the solubilized chelate and brings together the reaction components in a nanosized space. ^{436 – 438} A combination of mixed-ligand complex formation and introduction of a second lanthanide ion into the micellar system (sensitized co-fluorescence effect) proved to be especially efficient, resulting in an almost 1000-fold increase in the fluorescence intensity. ⁴³⁹ An unusually high 3800-fold enhancement of the sensitized fluorescence signal was found for a system containing warfarin, an anticoagulant used in agriculture as a rodenticide and pesticide. ⁴⁴⁰ It is noteworthy that supramolecular nature is inherent in not only micellar media employed for increasing the luminescence intensity, but also in the processes of energy transfer from the excitation energy donor to an emitting acceptor in a micelle. ³⁹⁰

Chemical sensors make use of dynamic interaction of a sensing element with an analyte in solution or in the gas phase. Their major benefit is the possibility of obtaining direct information (detection and recognition) about the presence and amount of one (pH of the solution, contents of glucose, oxygen, biogenic amine, etc.) or several components using an array of electrochemical or optical sensors. The reversibility of interaction between the receptor layer and the analyte is the main factor of supramolecular nature. The number of publications on chemical sensors is a few tens of thousands; some examples can be found in reviews. ^{368, 369, 441 – 443} The sensors that appeared in the last 10 – 15 years comprise, as a rule, various nanoparticles, which participate in supramolecular reactions. ^{369, 444, 445} Supramolecular nature is characteristic of numerous immune and enzymatic reactions meant for determining biologically active compounds. ⁴⁴⁶

Thus, it can be stated that analytical chemistry would further extend the range of supramolecular systems and interactions that would be utilized to develop methods of analysis of various specimens, especially new materials and biological fluids.

3.1.1.1. Synthesis and properties of amphiphilic sulfonate calix[n]arenes

The first calixarene 202 with surface-active properties was synthesized by Shinkai et al. ⁴⁵⁸ in 1984 by alkylation of sulfonate calix[6]arene with n-hexyl bromide in the presence of NaOH (Scheme 28). In 1986, a series of amphiphilic sulfonate calix[6]arenes was complemented with hexamethyl- and hexadodecyl-substituted macrocycles 203, 204, ⁴⁴⁹ and two years later, n-butyl-substituted calix[4(6,8)]arenes 205 – 207 were prepared (see Scheme 28). ^{459, 460}

Scheme 28 

Zoom In Zoom Out Reset image size

Several novel derivatives of calix[n]arene 208 – 215 bearing propyl sulfonate groups at the lower rim and various alkyl groups in the para-position of the aromatic ring were synthesized by Shinkai et al. ⁴⁶⁰ in 1989. Calix[n]arenes unsubstituted at the lower rim reacted with propane sultone in the presence of sodium hydride (Scheme 29).

Scheme 29 

Zoom In Zoom Out Reset image size

The pioneering works of Shinkai and co-workers concerning sulfonate calixarenes were continued by other researchers. Thus, Reinaud and co-workers ⁴⁶¹ proposed an original method to use precursors 216, 217 in the synthesis of sulfonate calixarene 218 (Scheme 30) via the regioselective ipso-chlorosulfonylation followed by alkaline hydrolysis. Imidazole moieties deactivated the electrophilic attack, due to their ability to be protonated; as a result, the reaction proceeded selectively on the anisole moieties of the macrocycle.

Scheme 30 

Zoom In Zoom Out Reset image size

A study of the aggregation of macrocycle 218 in water showed that despite the absence of the typical amphiphilic architecture, it can form polydisperse multilamellar vesicles of the pH-dependent size (450 nm at pH 8.5 and 50 nm at pH 6.5). ⁴⁶² The addition of Ag ^I ions was found to affect the aggregation process, resulting in the formation of compact micelles of 2.5 nm diameter; this was related to a change in the macrocycle packing parameter as a result of complex-induced contraction of the macrocycle lower rim (Fig. 80).

Zoom In Zoom Out Reset image size

Modification of calix[n]arenes with propane sultone was proposed by Puntoriero and co-workers, ⁴⁶³ who succeeded in reacting p-tert-butylcalix[5]arene (219) with 1,4-butane sultone in the presence of sodium hydride to obtain water-soluble sulfonate 220 (Scheme 31). Using the diffusion-ordered NMR studies and fluorescence spectroscopy, it was demonstrated that macrocycle 220 can form a supramolecular amphiphilic structure with n-dodecylammonium chloride, due to encapsulation of the latter into the macrocyclic cavity, and also to form micelles of 2.6 nm diameter (Fig. 81).

Scheme 31 

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Liu and co-workers ⁴⁶⁴ carried out chlorosulfonylation of calixarene 221 with chlorosulfonic acid followed by the treatment with NaOH in anhydrous ethanol and thus obtained the tetrasodium salt of p-sulfonatocalix[4]arene tetraheptyl ether 222 in a good yield (Scheme 32).

Scheme 32 

Zoom In Zoom Out Reset image size

Macrocycle 222 can co-assembly with high amounts (46%) of antipsychotic drug Chlorpromazine to afford joint nanoaggregates in aqueous solutions (Fig. 82). Moreover, due to electrostatic interactions, the resulting aggregates successively react with trimethylated chitosan, which can actively transfer nanoparticles across the blood – brain barrier via absorption-mediated transcytosis.

Zoom In Zoom Out Reset image size

The same research team ⁴⁶⁵ prepared functional liposomes from tetra-O-butyl-substituted sulfonylated calix[4]arene 205 and a zwitterionic phospholipid, dipalmitoyl phosphatidylcholine (DPPC), for living cell imaging and targeted drug delivery (Fig. 83). Due to electrostatic interactions, the liposome was functionalized with biotinylated pyridinium salt as a targeting agent and a fluorescein-containing pyridinium salt as a fluorescent probe. The liposomes thus obtained were internalized into the MCF-7 cancer cells through the receptor-mediated endocytosis, thus demonstrating great prospects for their therapeutic and diagnostic applications.

Zoom In Zoom Out Reset image size

Stoikov and co-workers ⁴⁶⁶ successfully implemented Shinkai's approach to modification of calix[n]arenes with propane sultone to functionalize p-tert-butylcalix[4]arene (223). Propanesulfonate 224 (Scheme 33) proved to be able to form colloidally stable fractal hybrid nanodendrites of 95 nm size with silver ions. Later, the same research team ⁴⁶⁷ carried out the reaction of 1,3-propane sultone or 1,4-butane sultone with calix[4]arene 225 bearing tertiary amino groups to produce sulfobetaines 226a,b (see Scheme 33) in high yields. These macrocycles can form stable submicron-sized aggregates with silver ions.

Scheme 33 

Zoom In Zoom Out Reset image size