Alexander K. Shveyd

Room-temperature ferroelectricity in supramolecular networks of charge-transfer complexes

Materials exhibiting a spontaneous electrical polarization that can be switched easily between antiparallel orientations are of potential value for sensors, photonics and energy-efficient memories. In this context, organic ferroelectrics are of particular interest because they promise to be lightweight, inexpensive and easily processed into devices. A recently identified family of organic ferroelectric structures is based on intermolecular charge transfer, where donor and acceptor molecules co-crystallize in an alternating fashion known as a mixed stack: in the crystalline lattice, a collective transfer of electrons from donor to acceptor molecules results in the formation of dipoles that can be realigned by an external field as molecules switch partners in the mixed stack. Although mixed stacks have been investigated extensively, only three systems are known to show ferroelectric switching, all below 71 kelvin. Here we describe supramolecular charge-transfer networks that undergo ferroelectric polarization switching with a ferroelectric Curie temperature above room temperature. These polar and switchable systems utilize a structural synergy between a hydrogen-bonded network and charge-transfer complexation of donor and acceptor molecules in a mixed stack. This supramolecular motif could help guide the development of other functional organic systems that can switch polarization under the influence of electric fields at ambient temperatures.

A catenated strut in a catenated metal-organic framework

Mechanically interlocked molecules (MIMs), in the form of donor–acceptor [2]catenane-containing struts of exceptional length, have been incorporated into a three-dimensional catenated metal–organic framework (MOF) at precise locations and with uniform relative orientations. Catenation is expressed simultaneously within the struts and the framework.

Enabling tetracationic cyclophane production by trading templates

The time taken to produce the ubiquitous and promiscuous receptor, which has become known as the little “blue box”, has been halved as a result of using a pH-responsive derivative of 1,5-diaminonaphthalene to displace the template employed during its synthesis. The fact that this surrogate trades places within the cavity of the “blue box” is the key to the time-saving for the simple reason that it leaves the lock, so-to-speak, as soon as it becomes protonated by a strong acid. During the subsequent full characterisation of the “blue box”, it was discovered that the tetracationic cyclophane exists at least in two different polymorphs in the solid state.

Lock-arm supramolecular ordering: A molecular construction set for cocrystallizing organic charge transfer complexes

Organic charge transfer cocrystals are inexpensive, modular, and solution-processable materials that are able, in some instances, to exhibit properties such as optical nonlinearity, (semi)conductivity, ferroelectricity, and magnetism. Although the properties of these cocrystals have been investigated for decades, the principal challenge that researchers face currently is to devise an efficient approach which allows for the growth of high-quality crystalline materials, in anticipation of a host of different technological applications. The research reported here introduces an innovative design, termed LASO-lock-arm supramolecular ordering-in the form of a modular approach for the development of responsive organic cocrystals. The strategy relies on the use of aromatic electronic donor and acceptor building blocks, carrying complementary rigid and flexible arms, capable of forming hydrogen bonds to amplify the cocrystallization processes. The cooperativity of charge transfer and hydrogen-bonding interactions between the building blocks leads to binary cocrystals that have alternating donors and acceptors extending in one and two dimensions sustained by an intricate network of hydrogen bonds. A variety of air-stable, mechanically robust, centimeter-long, organic charge transfer cocrystals have been grown by liquid-liquid diffusion under ambient conditions inside 72 h. These cocrystals are of considerable interest because of their remarkable size and stability and the promise they hold when it comes to fabricating the next generation of innovative electronic and photonic devices.

A neutral redox-switchable [2]rotaxane

A limited range of redox-active, rotaxane-based, molecular switches exist, despite numerous potential applications for them as components of nanoscale devices. We have designed and synthesised a neutral, redox-active [2]rotaxane, which incorporates an electron-deficient pyromellitic diimide (PmI)-containing ring encircling two electron-rich recognition sites in the form of dioxynaphthalene (DNP) and tetrathiafulvalene (TTF) units positioned along the rod section of its dumbbell component. Molecular modeling using MacroModel guided the design of the mechanically interlocked molecular switch. The binding affinities in CH(2)Cl(2) at 298 K between the free ring and two electron-rich guests--one (K(a) = 5.8 × 10(2) M(-1)) containing a DNP unit and the other (K(a) = 6.3 × 10(3) M(-1)) containing a TTF unit--are strong: the one order of magnitude difference in their affinities favouring the TTF unit suggested to us the feasibility of integrating these three building blocks into a bistable [2]rotaxane switch. The [2]rotaxane was obtained in 34% yield by relying on neutral donor-acceptor templation and a double copper-catalysed azide-alkyne cycloaddition (CuAAC). Cyclic voltammetry (CV) and spectroelectrochemistry (SEC) were employed to stimulate and observe switching by this neutral bistable rotaxane in solution at 298 K, while (1)H NMR spectroscopy was enlisted to investigate switching upon chemical oxidation. The neutral [2]rotaxane is a chemically robust and functional switch with potential for applications in device settings.

Tayi et al . reply

Organic ferroelectric materials operating at room temperature are in demand in the emerging field of lightweight, flexible and environmentally friendly electronics. Tayi et al.1 reported roomtemperature ferroelectricity in organic mixed-stack charge-transfer crystals, produced using a supramolecular design concept—the lock-arm supra molecular ordering (LASO)—that synergistically combines intermolecular charge transfer and hydrogen bonds2,3. Here we present an independent experimental investigation that found no evidence for ferroelectricity in one of the LASO compounds described in ref. 1 and, together with theoretical calculations, demonstrates that a possible ferroelectric behaviour is not of electronic origin as proposed in ref. 1. We therefore question the reproducibility of the roomtemperature ferroelectricity claimed for LASO systems1. There is a Reply to this Comment by Tayi, S. A. et al. Nature 547, http://dx.doi. org/10.1038/nature22802 (2017). Room-temperature ferroelectricity has been claimed for three charge-transfer crystals formed by the same electron acceptor (A) and three different donors (D) using LASO1. The ferroelectric behaviour was ascribed to a sizeable charge ρ transferred from D to A molecules arranged in non-centrosymmetric structures characterized by polar dimers [...(D+ρA−ρ) (D+ρA−ρ)...], as observed in other mixed-stack charge-transfer crystals, albeit at cryogenic temperatures4,5. Tayi et al.1 supported their finding with a combination of structural, vibrational and dielectric measurements that we replicate here for their system 1·2 (see Fig. 1a). We note that the X-ray diffraction data of the three LASO crystals have been recently re-examined by the same authors, assigning all of them to centro-symmetric space groups3, which is incompatible with ferroelectricity. We synthesized compounds 1 and 2 and grew single crystals of compound 1·2 in different solvent mixtures under strictly anhydrous conditions. Our independent X-ray structural determination confirms the growth of the same polymorph reported in ref. 1. We also find, as in ref. 3, that a better structural refinement is obtained for the non-polar space group P1 than for the P1 group initially proposed in ref. 1. Density functional theory (DFT) calculations consistently indicate that the non-polar phases are the most stable for the three LASO compounds. Vibrational spectroscopy provides information on both electronic and structural properties of mixed-stack charge-transfer crystals6. Totally symmetric molecular vibrations offer an unambiguous probe of the dimerization of the lattice: these Raman-active modes show up in infrared spectra polarized along the stack axis only in dimerized phases, where they modulate the asymmetric flow of electronic charge, acquiring a much larger intensity than that of other vibrations. Being sensitive to local static or dynamic disorder, vibrational spectroscopy offers information complementary to X-ray diffraction data, which probes only long-range order. Our absorption infrared and Raman spectra in Fig. 1b do not present coincident peaks, firmly excluding stack dimerization, whether local or nonlocal, static or dynamic. The frequency coincidences in infrared and Raman spectra reported by Tayi et al.1 are probably accidental, owing to the presence of many bands in their limited-quality infrared spectra. We assume their infrared spectra were obtained by applying the Kubelka–Munk transformation to the singlepoint reflectance spectra of single crystals, although this transformation is appropriate only for the diffuse reflectance spectra of powders7. Infrared spectroscopy also allows us to estimate the ionicity ρ, through the frequency shift of properly chosen ‘charge sensitive’ vibrational modes, as is well established for the carbonyl (C= O) stretching of chloranil complexes8,9. Tayi et al.1 attributed sizeable charge transfer to the three LASO compounds (ρ = 0.67 for 1·2, ρ = 0.89 for 1·3 and ρ = 0.43 for 1·4), based on the frequencies of the carbonyl stretching of 1, and using a calibration procedure employing tetracyanoquinodimethane complexes that is not clear to us. The inconsistencies of this approach have been pointed out in ref. 10. Here we remark that, according to chemical intuition, experimental literature9, and DFT calculations8, the frequency of carbonyl modes is expected to decrease strongly upon negatively charging 1, and not to increase slightly, as proposed in ref. 1. Without any need of calculation or calibration, Fig. 1c compares the infrared spectra of 1 and 2 with those of the 1·2 crystal, polarized perpendicularly to the stack axis, where the C= O stretching bands appear. The spectrum of 1·2 is very similar to the superposition of the spectra of its components in the whole spectral range and specifically in the C= O stretching region, as is also confirmed by Raman spectra (see Extended Data Fig. 1). This proves that the 1·2 complex is essentially neutral (ρ ≈ 0). A similar conclusion can be drawn for the LASO compounds 1·3 and 1·4, based on the marginal shifts of the carbonyl stretching1. Hubbard model calculations10, and our DFT results (see Extended Data Table 1) further support the conclusion, pointing to largely neutral states for all three systems.