The tetrapyridyl ligand bbpya (bbpya=N,N-bis(2,2′-bipyrid-6-yl)amine) and its mononuclear coordination compound [Fe(bbpya)(NCS)₂] (1) were prepared. According to magnetic susceptibility, differential scanning calorimetry fitted to Sorai’s domain model, and powder X-ray diffraction measurements, 1 is low-spin at room temperature, and it exhibits spin crossover (SCO) at an exceptionally high transition temperature of T_1/2=418 K. Although the SCO of compound 1 spans a temperature range of more than 150 K, it is characterized by a wide (21 K) and dissymmetric hysteresis cycle, which suggests cooperativity. The crystal structure of the LS phase of compound 1 shows strong NH⋅⋅⋅S intermolecular H-bonding interactions that explain, at least in part, the cooperative SCO behavior observed for complex 1. DFT and CASPT2 calculations under vacuum demonstrate that the bbpya ligand generates a stronger ligand field around the iron(II) core than its analogue bapbpy (N,N′-di(pyrid-2-yl)-2,2′-bipyridine-6,6′-diamine); this stabilizes the LS state and destabilizes the HS state in 1 compared with [Fe(bapbpy)(NCS)₂] (2). Periodic DFT calculations suggest that crystal-packing effects are significant for compound 2, in which they destabilize the HS state by about 1500 cm⁻¹. The much lower transition temperature found for the SCO of 2 compared to 1 appears to be due to the combined effects of the different ligand field strengths and crystal packing.

The crystal structure and in vitro cytotoxicity of the amphiphilic ruthenium complex [3](PF₆)₂ are reported. Complex [3](PF₆)₂ contains a Ru−S bond that is stable in the dark in cell-growing medium, but is photosensitive. Upon blue-light irradiation, complex [3](PF₆)₂ releases the cholesterol–thioether ligand 2 and an aqua ruthenium complex [1](PF₆)₂. Although ligand 2 and complex [1](PF₆)₂ are by themselves not cytotoxic, complex [3](PF₆)₂ was unexpectedly found to be as cytotoxic as cisplatin in the dark, that is, with micromolar effective concentrations (EC₅₀), against six human cancer cell lines (A375, A431, A549, MCF-7, MDA-MB-231, and U87MG). Blue-light irradiation (λ=450 nm, 6.3 J cm⁻²) had little influence on the cytotoxicity of [3](PF₆)₂ after 6 h of incubation time, but it increased the cytotoxicity of the complex by a factor 2 after longer (24 h) incubation. Exploring the unexpected biological activity of [3](PF₆)₂ in the dark elucidated an as-yet unknown bifaceted mode of action that depended on concentration, and thus, on the aggregation state of the compound. At low concentration, it acts as a monomer, inserts into the membrane, and can deliver [1]²⁺ inside the cell upon blue-light activation. At higher concentrations (>3–5 μm), complex [3](PF₆)₂ forms supramolecular aggregates that induce non-apoptotic cell death by permeabilizing cell membranes and extracting lipids and membrane proteins.

Liposomes capable of generating photons of blue light in situ by triplet–triplet annihilation upconversion of either green or red light, were prepared. The red-to-blue upconverting liposomes were capable of triggering the photodissociation of ruthenium polypyridyl complexes from PEGylated liposomes using a clinical grade photodynamic therapy laser source (630 nm).

Liposomes capable of generating photons of blue light in situ by triplet–triplet annihilation upconversion of either green or red light, were prepared. The red-to-blue upconverting liposomes were capable of triggering the photodissociation of ruthenium polypyridyl complexes from PEGylated liposomes using a clinical grade photodynamic therapy laser source (630 nm).

The interaction between the ruthenium polypyridyl complex [Ru(terpy)(dcbpy)(H₂O)]²⁺ (terpy=2,2′;6′,2“-terpyridine, dcbpy=6,6′-dichloro-2,2′-bipyridine) and phospholipid membranes containing either thioether ligands or cholesterol were investigated using UV–visible spectroscopy, Langmuir–Blodgett monolayer surface pressure measurements, and isothermal titration calorimety (ITC). When embedded in a membrane, the thioether ligand coordinated to the dicationic metal complex only when the phospholipids of the membrane were negatively charged, that is, in the presence of attractive electrostatic interaction. In such a case coordination is much faster than in homogeneous conditions. A two-step model for the coordination of the metal complex to the membrane-embedded sulfur ligand is proposed, in which adsorption of the complex to the negative surface of the monolayers or bilayers occurs within minutes, whereas formation of the coordination bond between the surface-bound metal complex and ligand takes hours. Finally, adsorption of the aqua complex to the membrane is driven by entropy. It does not involve insertion of the metal complex into the hydrophobic lipid layer, but rather simple electrostatic adsorption at the water–bilayer interface.

Photocatalytic systems often suffer from poor quantum yields due to fast charge recombination: The energy-wasting annihilation of the photochemically created charge-separated state. In this report, we show that the efficiency of photoinduced electron transfer from a sacrificial electron donor to positively charged methyl viologen, or to negatively charged 5,5′-dithiobis(2-nitrobenzoate), increases dramatically upon addition of charged phospholipid vesicles if the charge of the lipids is of the same sign as that of the electron acceptor. Centrifugation and UV/Vis titration experiments showed that the charged photosensitizers adsorb at the liposome surface, that is, where the photocatalytic reaction takes place. The increased photoelectron transfer efficiency in the presence of charged liposomes has been ascribed to preferential electrostatic interactions between the photosensitizer and the membrane, which prevents the formation of photosensitizer–electron-acceptor complexes that are inactive towards photoreduction. Furthermore, it is shown that the addition of liposomes results in a decrease in photoproduct inhibition, which is caused by repulsion of the reduced electron acceptor by the photocatalytic site. Thus, liposomes can be used as a support to perform efficient photocatalysis; the charged photoproducts are pushed away from the liposomes and represent “soluble electrons” that can be physically separated from the place where they were generated.

The new ruthenium complex [Ru(terpy)(dcbpy)(Hmte)](PF₆)₂ ([2](PF₆)₂; dcbpy=6,6′-dichloro-2,2′-bipyridine, terpy=2,2′;6′,2“-terpyridine, Hmte=2-(methylthio)ethanol) was synthesized. In the crystal structure, this complex is highly distorted, revealing steric congestion between dcbpy and Hmte. In water, [2]²⁺ forms spontaneously by reacting Hmte and the aqua complex [Ru(terpy)(dcbpy)(OH₂)]²⁺ ([1]²⁺), with a second-order rate constant of 0.025 s⁻¹ M⁻¹ at 25 °C. In the dark, the RuS bond of [2]²⁺ is thermally unstable and partially hydrolyzes; in fact, [1]²⁺ and [2]²⁺ are in an equilibrium characterized by an equilibrium constant K of 151 M⁻¹. When exposed to visible light, the RuS bond is selectively broken to release [1]²⁺, that is, the equilibrium is shifted by visible-light irradiation. The light-induced equilibrium shifts were repeated four times without major signs of degradation; the RuS coordination bond in [2]²⁺ can be described as a robust, light-sensitive, supramolecular bond in water. To demonstrate the potential of this system in supramolecular chemistry, a new thioether–cholesterol conjugate (4), which inserts into lipid bilayers through its cholesterol moiety and coordinates to ruthenium through its sulfur atom, was synthesized. Thioether-functionalized, anionic, dimyristoylphosphatidylglycerol (DMPG), lipid vesicles, to which aqua complex [1]²⁺ efficiently coordinates, were prepared. Upon exposure of the Ru-decorated vesicles to visible light, the RuS bond is selectively broken, thus releasing [1]²⁺ that stays at the water-bilayer interface. When the light is switched off, the metal complex spontaneously coordinates back to the membrane-embedded thioether ligands without a need to heat the system. This process was repeated four times at 35 °C, thus achieving light-triggered hopping of the metal complex at the water-bilayer interface.

In this study, we show that 1) different isomers of the same mononuclear iron(II) complex give materials with different spin-crossover (hereafter SCO) properties, and 2) minor modifications of the bapbpy (bapbpy=N6,N6′-di(pyridin-2-yl)-2,2′-bipyridine-6,6′-diamine) ligand allows SCO to be obtained near room temperature. We also provide a qualitative model to understand the link between the structure of bapbpy-based ligands and the SCO properties of their iron(II) compounds. Thus, seven new trans-[Fe{R₂(bapbpy)}(NCS)₂] compounds were prepared, in which the R₂bapbpy ligand bears picoline (9–12), quin-2-oline (13), isoquin-3-oline (14), or isoquin-1-oline (15) substituents. From this series, three compounds (12, 14, and 15) have SCO properties, one of which (15) occurs at 288 K. The crystal structures of compounds 11, 12, and 15 show that the intermolecular interactions in these materials are similar to those found in the parent compound [Fe(bapbpy)(NCS)₂] (1), in which each iron complex interacts with its neighbors through weak NH⋅⋅⋅S hydrogen bonding and π–π stacking. For compounds 12 and 15, hindering groups located near the NH bridges weaken the NS intermolecular interactions, which is correlated to non-cooperative SCO. For compound 14, the substitution is further away from the NH bridges, and the SCO remains cooperative as in 1 with a hysteresis cycle. Optical microscopy photographs show the strikingly different spatio-temporal evolution of the phase transition in the noncooperative SCO compound 12 relative to that found in 1. Heat-capacity measurements were made for compounds 1, 12, 14, and 15 and fitted to the Sorai domain model. The number n of like-spin SCO centers per interacting domain, which is related to the cooperativity of the spin transition, was found high for compounds 1 and 14 and low for compounds 12 and 15. Finally, we found that although both pairs of compounds 11/12 and 14/15 are pairs of isomers their SCO properties are surprisingly different.