Rare-earth based tetrapyrrolic sandwiches: chemistry, materials and applications

Abstract

The unique properties of natural tetrapyrrolic compounds have inspired the rapid growth of research interest in the design and synthesis of artificial porphyrinoids and their metal complexes as a basis of modern functional materials. A special role in the design of such materials is played by sandwich complexes formed by tetrapyrrolic macrocycles with rare earth elements, especially lanthanides. The development of synthetic approaches to the functionalization of tetrapyrrolic compounds and their rare earth complexes has facilitated the intensive development of new applications over the last decade. As a way of expanding the functionalities of rare earth complexes, sophisticated examples have been obtained, including mixed-ligand complexes, π-extended analogues, covalently linked and fused sandwiches, complexes with less-common tetrapyrrols, sandwiches with non-tetrapyrrolic macrocycles and even complexes containing up to six stacked ligands. This review intends to offer a general overview of the preparation of such sophisticated REE tetrapyrrolic sandwiches over the last decade as well as emphasizes the current challenges and perspectives of their application in areas such as single-molecule magnetism (SMM), organic field-effect transistors (OFET), conductive materials and nonlinear optics (NLO).

---

Alexander G. Martynov

Alexander G. Martynov graduated from the Higher Chemical College of the Russian Academy of Sciences in 2006. He received his PhD and Dr Hab. degrees in 2008 and 2020. In 2022 he was awarded the Moscow government prize for young scientists. In 2022 he became a Professor of the Russian Academy of Sciences. Currently he is a leading researcher at the Frumkin Institute of Physical Chemistry and Electrochemistry RAS. His research interests are in the areas of spectroscopy of tetrapyrrolic macrocycles and tetrapyrrol-based materials with tuneable properties including REE sandwich complexes for magnetic applications, agents for photodynamic therapy and carbene transfer catalysts.

Yoji Horii

Yoji Horii was born in Gumma, Japan in 1989. He received his Bachelor's (2012), Master's (2014) and PhD (2017) degrees at Tohoku University under the supervision of Professor Masahiro Yamashita. He worked on magnetism and electronic states of phthalocyaninato-lanthanide single-molecule magnets. He moved to Osaka University as the postdoctoral fellow in April 2017. In 2019, he moved to Tohoku University as a postdoctoral fellow. Since April 2020, he has been working as an assistant professor at Nara Women's University. His current research interests include molecular magnetism, two dimensional nanomaterials and soft materials.

Keiichi Katoh

Keiichi Katoh is currently an associate professor in the Department of Chemistry of Josai University where he is leading research activities in inorganic chemistry and magnetism. He received his PhD in 2003 from the Graduate University for Advanced Studies (Institute for Molecular Science). After being a post-doctoral fellow at Hokkaido University (Research Institute for Electronic Science) until 2004, he worked as a post-doctoral fellow at Kyoto University (Institute for Chemical Research) until 2008, and he joined the Prof. Yamashita group as a post-doctoral fellow at Tohoku University in 2008. From 2009 to 2020, he was appointed to an assistant professor in the Department of Chemistry of Tohoku University. He works on the development of solid state spin systems based on molecular magnetic materials.

Yongzhong Bian

Yongzhong Bian was born in Shanxi, China. He received his BSc (1998) and PhD (2005) (with Prof. Jiangzhuang Jiang) from Shandong University. He carried out a JSPS postdoctoral fellowship (2005–2007) at the University of Tokyo with Prof. Takuzo Aida. He became an associate professor at Shandong University in 2007, and moved to the University of Science and Technology Beijing as a full professor in 2009. His current research centers on porphyrin chemistry, encompasses applications in Fluorescent chemosensing and bioimaging, photodynamic therapy, photochemical catalysis and molecular recognition.

Jianzhuang Jiang

Jianzhuang Jiang was born in Heilongjiang, China. He received his BSc (1985), MSc (1988), and PhD (1993) (with Tsinglien Chang) from the Peking University. During his doctoral study (1990–1992), he obtained a Fellowship from the Ministry of Culture, Science, and Sport of Japan and carried out his PhD work at the Osaka University under the guidance of Kenichi Machida and Ginya Adachi. He became a Postdoctoral Fellow at Peking University with Tsinglien Chang (1993–1994), a Visiting Scholar at The Chinese University of Hong Kong with Dennis K. P. Ng and Thomas C. W. Mak (1995–1996), and a Postdoctoral Fellow at the Queensland University of Technology with Dennis P. Arnold (1998–2000). He joined Shandong University and the University of Science and Technology Beijing in 1996 and 2008, respectively. He has been working on porphyrin- and phthalocyanine-based materials for more than thirty years to develop new single-molecule magnets, organic field-effect transistors, optical materials, energy materials, and porous materials.

Masahiro Yamashita

Masahiro Yamashita received his BSc degree in 1977, MSc in 1979, and DSc in 1982 from Kyushu University. After his graduation, he joined the Institute for Molecular Science (IMS). In 1983, he was appointed as an Assistant Professor at Kyushu University. In 1988, he was appointed as an Associate Professor at Nagoya University, and he was promoted to a full Professor at the same university in 1998. He was a full Professor at Tokyo Metropolitan University from 2000 to 2004. He moved to Tohoku University as a full professor in 2005. He has been honored with the Inoue Scientific Award (2002), the Chemical Society of Japan Award for Creative Work (2005), the Award of Japan Society of Coordination Chemistry (2014), Mukai Award (2019), and the Chemical Society of Japan Award (2020). He is now an Associate Member of the Science Council of Japan. He is a Fellow of the Royal Society of Chemistry (FRSC).

Yulia G. Gorbunova

Yulia G. Gorbunova is a Full member of Russian Academy of Sciences (2022) and main researcher at the Russian Academy of Sciences (Kurnakov Institute of General and Inorganic Chemistry RAS (IGIC RAS) and Frumkin Institute of Physical Chemistry and Electrochemistry RAS). She graduated in Chemistry from the Lomonosov Moscow State University and received her PhD (1995) and habilitation (2006) from IGIC RAS. Her current research interests are related to the chemistry of tetrapyrrolic compounds as the basis of functional materials, including molecular switchers, nonlinear optics, electrochromic materials, SMM, sensors and hybrid organo-inorganic nanomaterials. She was a recipient of the Award of European Academy of Science for young researcher (2001), the Russian Government prize in the field of science and techniques (2002) and the Chugaev Award in coordination chemistry (2010). In 2016 Prof. Gorbunova was awarded by order “Chevalier dans l’Ordre national des Palmes Academiques” by the Government of France.

---

1. Introduction

Just as it is difficult to overestimate the role of natural porphyrins in the functioning of living beings, it is also difficult to disagree with the fact that synthetic porphyrinoids have greatly contributed to the development of modern materials science. In particular, this has been made possible by a variety of unique photophysical, electrochemical, catalytic, magnetic, and other properties provided by both the tetrapyrroles themselves and the metal ions placed in their cavities.

Among various coordination compounds formed by tetrapyrrole macrocycles, the complexes with rare earth elements (REE) attract special attention.^1–6 Due to large ionic radii and high coordination numbers, these elements form sandwich double- and triple-decker complexes with porphyrins (Por), phthalocyanines (Pc) and related macrocycles (Fig. 1), which, with some exceptions, are not observed for the other metals of the periodic table. Tightly coupled π-systems of stacked aromatic macrocycles provide these sandwiches with unique optical and electrochemical properties, and the presence of 4f-electron shells in paramagnetic lanthanide (Ln) metal centres expands the area of their application to the actively developing and promising fields of molecular magnetism and spintronics.^7–11

Fig. 1 Double- and triple-decker sandwich complexes formed by REE ions with tetrapyrrolic macrocycles, X = C-Ar, C-Alk, and N.

Interestingly, the somewhat detective chemistry of sandwich complexes with tetrapyrrolic ligands started almost simultaneously with rise of phthalocyanines, although the first element to form double-decker complexes – tin(IV) – did not belong to rare earth elements. The synthesis of its bisphthalocyaninate, Sn(Pc)₂, was reported in 1936 by Linstead et al.¹² The authors used the reaction of (Pc)SnCl₂ and Na₂(Pc), and this raise-by-one-story approach much later became the main way to various heteroleptic and mixed ligand sandwiches.

Much later, in 1965 Kirin and Moskalev et al. reported seminal work on the template reaction between REE acetates and melted phthalonitrile.¹³ The authors isolated surprisingly stable green and blue complexes, where metal ions were sandwiched between two Pc ligands – M(Pc)₂; however, at that time the details of their electronic structure and the nature of the intriguing redox behaviour remained unclear. In particular, thin films formed by bisphthalocyaninates were able to reversibly change colour when a potential was applied or when these films were treated with volatile oxidants and reducing agents. The electrochromic behaviour of bisphthalocyaninates was explained in 1979 by Corker et al.¹⁴ who used UV-Vis and EPR spectra to show that M(Pc)₂ undergoes ligand-centred redox-triggered interconversion between three forms of trivalent metal complexes – anionic [M³⁺(Pc²⁻)₂]⁻, neutral [(Pc⁻˙)M³⁺(Pc²⁻)]⁰ and cationic [M³⁺(Pc⁻)₂]⁺ forms. Due to the pronounced difference in spectral signatures these three forms possessed different colours, and for many years they were known as blue, green and red forms, respectively. Apart from changes in UV-Vis spectra, switching between redox states caused modulation of NIR absorbance, and the nature of NIR bands of neutral green complexes, has been discussed in the works of Simon et al.,¹⁵ Aroca et al.,¹⁶ Tomilova and Lukyanets et al.¹⁷

Importantly, under harsh conditions of template condensation, apart from double-decker complexes, REE ions can also form blue triple-decker sandwiches M₂(Pc)₃,^18,19 where Pc ligands are isoelectronic with those in blue forms of bisphthalocyaninates. Because of the similarity of their optical properties and lack of available mass-spectral characterization these complexes were sometimes mistaken for each other. These contradictions have been eliminated by single-crystal X-ray assignment performed in 1980 for neodymium bisphthalocyaninate, Nd(Pc)₂ by Kasuga et al.,²⁰ and in 1999 for dilutetium tris-(tetra-15-crown-5-phthalocyaninate), Lu₂[(15C5)₄Pc]₃ by Troyanov et al.²¹

The survey for milder conditions where double- and triple-deckers could be selectively obtained, afforded several convenient protocols, which are commonly used in synthetic practice nowadays. These protocols include template condensation of phthalonitriles and REE salts in high boiling alcohols in the presence of DBU as a phthalonitrile activator as suggested by De Cian et al.,²² and also direct reactions of pre-synthesized phthalocyanine ligands with metal salts as suggested by Takahashi et al.^23,24 While the former method yields double-decker sandwiches, the latter method furnishes both double- and triple-decker complexes, and later studies showed that the selectivity of complex formation can be governed either by the presence or absence of DBU, respectively.^25,26

In the case of phthalocyanines both template and direct methods are equally applicable for the synthesis of REE sandwiches, however, in the case of porphyrins only direct method can be used to synthesize double- and triple-decker complexes. First example of homoleptic REE bisporphyrinates was reported in 1977 by Buchler et al. with the example of quadrivalent cerium complex Ce(TTP)₂ formed upon the reaction of Ce(acac)₃ with tetra-p-tolylporphyrin, H₂(TTP) in refluxing 1,2,4-trichlorobenzene (TClB).²⁷ The authors noticed a remarkable analogy between the structure and the redox-properties of this complex and the “special pair” formed by chlorophyl molecules in the photosynthetic systems of green plants.²⁸ Later more double- and triple-decker porphyrinates were isolated and characterized,^29–31 and it was found that porphyrin ligands tended to form sandwiches only with early and middle lanthanides, while Pc ligands could form sandwiches with all REE ions, including the relatively small Sc³⁺ ions.^32,33

Finally, REE complexes provided the marriage of porphyrins and phthalocyanines in mixed ligand sandwiches, where both types of ligands are combined and strongly coupled due to intramolecular orbital interactions providing them with peculiar optical properties of panchromatic absorbers. Synthesis and crystallographic characterization of the first example of such complexes were reported in 1986 by Moussavi et al. with the example of mixed ligand triple-decker complexes with tetra-p-anisylporphyrin (Pc)M(Pc)M(TAnP), where M = Nd, Eu, and Gd.³⁴

The rapid development of the chemistry of sandwich REE complexes was driven both by advances in spectroscopy and physicochemical characterization of these challenging objects and breakthrough discoveries which affected many areas of modern chemistry. Among them we should inevitably mention the first organic semiconductor – lutetium bisphthalocyaninate, Lu(Pc)₂, described for the first time in 1987 by Turek et al.,³⁵ and especially the first example of a 4f single molecule magnet – terbium(III) bisphthalocyaninate, Tb(Pc)₂, reported in 2003 by Ishikawa et al.³⁶

Because of this, the largely fundamental interest in these compounds was gradually replaced by a focus on applied research where archetypal double- and triple-decker complexes M(Pc)₂ and M₂(Pc)₃ coexist with more sophisticated compounds, including mixed-ligand complexes, π-extended analogues, covalently linked and fused sandwiches, complexes with less-common tetrapyrrols, sandwiches with non-tetrapyrrolic macrocycles and even complexes containing up to six stacked ligands obtained by coordination or supramolecular bonding. Moreover, these complexes have been recognized as molecular switches whose properties can be altered by the reversible self-assembly processes, protonation, redox reactions, etc., opening the way to the creation of “smart” materials.

These sophisticated complexes began to attract the most attention over the last decade; thus, the aim of the present review is mainly to summarize the advances in the chemistry of REE tetrapyrrolic sandwiches over this period to supplement the previously published comprehensive reviews^1–6 with new data. Another goal of this review is to emphasize the current challenges and perspectives of sandwich complex application in various fields such as single-molecule magnetism (SMM), organic field-effect transistors (OFET) and conductive materials, and nonlinear optics (NLO).

2. Advances in the synthesis of functionalized multiple-decker sandwiches

2.1 New homo- and heteroleptic double- and triple-decker complexes bearing symmetrically substituted tetrapyrolic ligands

From the synthetic viewpoint, sandwich complexes with symmetrically substituted tetrapyrrolic ligands are the most available objects, particularly due to facile synthesis and separation procedures. The following section will include various classes of homo- and heteroleptic sandwiches with symmetrical ligands. In the latter case, ligands belong to the same classes of tetrapyrrolic macrocycles.

2.1.1 Aryloxy- and aryl-substituted sandwich phthalo- and naphthalocyaninates. Over the last decade among other sandwich complexes with symmetrical ligands special attention was paid to aryloxy-substituted phthalocyaninates, particularly due to high solubility, which contributes to the processing of thin-film materials and devices. Aryloxy-groups reveal electron-withdrawing behaviour,³⁷ making their reduction easier and their oxidation more difficult. Finally, these substituents participate in the molecular packing of complexes governing conductivity and other useful properties of materials based on them.

For example, Kong et al. used template condensation of 4,5-di(α-naphthoxy)phthalonitrile with Eu(acac)₃ and DBU in refluxing n-PentOH to make a double-decker complex Eu[(NaphO)₈Pc]₂ (Scheme 1) which was used as a component of a flexible ambipolar OFET,³⁸ and it also revealed sensing properties towards NO₂ and NH₃ (cf. Section 4).³⁹ To synthesize a triple-decker complex Eu₂[(NaphO)₈Pc]₃, raise-by-one-story approach was applied, namely octa-naphthyloxy-phthalocyanine was condensed with the double-decker complex (Scheme 1).⁴⁰ The resulting trisphthalocyaninate could form self-assembled microstructures whose morphology and n-/p-mobilities could be controlled by solvent–vapour annealing procedure (cf. Section 4.2). Similar approaches were used to synthesize various examples of other aryloxy-substituted sandwiches, including heteroleptic complexes, which were designed as high-performance air-stable ambipolar organic thin film transistors,^41–44 optical limiters^45–48 and single molecule magnets.⁴⁹

Scheme 1 Synthesis of α-naphthyloxy-substituted Eu(III) bis- and trisphthalocyaninates.^38,40

Phenyl-substituted phthalo- and naphthalocyanines were used by Dubinina et al. for the synthesis of REE complexes.^50,51 The selectivity of complex formation upon direct metalation of these ligands depended on conditions affording either monophthalo- and naphthalocyaninates or sandwich complexes (Scheme 2). While double-decker complexes were obtained for both types of macrocycles, triple-decker complexes could be obtained by only using phthalocyanine ligands.

Scheme 2 Synthesis of lanthanide complexes with phenyl-substituted phthalo- and naphthalocyanines.^50,51

Single-crystal XRD studies of Lu(Ph₈Pc)₂ revealed intramolecular edge-to-face π–π stacking interactions between peripheral phenyl groups, reinforcing the overall intramolecular interactions between Pc ligands (Fig. 2).⁵⁰ In turn, it resulted in a strong hypsochromic shift of NIR bands in UV-Vis-NIR spectra of bisphthalocyaninates – for example the intervalence band shifted from 1575 nm in the spectrum of Tb(tBu₄Pc)₂ to 1433 nm in the spectrum of Tb(Ph₈Pc)₂. In contrast, benzo-annelation going from phthalo- to naphthalocyaninates caused bathochromic shifts of intervalence bands by ca. 200 nm.

Fig. 2 Top view of the X-ray structure of Lu(Ph₈Pc)₂ (CCDC PIXGUA). Dark blue lines show edge-to-face π-stacking interactions.⁵⁰

Measurements of conductivity of thin drop-casting films formed by phenyl-substituted terbium(III) phthalocyaninates showed that in the series of mono-, bis- and trisphthalocyaninates the complex Tb(Ph₈Pc)₂ had the largest conductivity of 5 × 10⁻¹⁰ S cm⁻¹, which is however much smaller than the values obtained for unsubstituted double-deckers.⁵¹ Optical microscopy data showed that such a low conductivity value was partially caused by film inhomogeneity. AFM studies of films, formed by triple-decker complex Tb₂(Ph₈Pc)₃ showed that it forms ring-shaped nanostructures – nano-coffee rings, which appear due to an increase in the surface tension of the solutions of this complex in comparison with mono- and bisphthalocyaninates. This feature can be used for nano-patterning of various substrates.

2.1.2 Crown-substituted sandwich phthalocyaninates. Crown-phthalocyanines still attract continuing interest, although they are known since late 1980s,^52,53 and lutetium(III) bis(tetra-15-crown-5)phthalocyaninate, Lu[(15C5)₄Pc], was actively studied in 1990s as a potential component of ionoelectronic materials.^54,55 In subsequent years, homoleptic crown-substituted double-deckers were synthesized by Gorbunova et al. for almost all rare earth elements including the smallest Sc³⁺ ions.^56,57 The availability of this series allowed the smooth changes of major physical-chemical properties following f-contraction to be revealed,⁵⁸ although discrepancy was observed in the case of cerium complexes which can exist in tri- and quadrivalent states of metal centres.⁵⁹

Attention to crown-substituted sandwiches was attracted mainly in the context of fabrication of electronic devices due to their ability to form thin films with well-defined morphology.^58,60 While investigating the formation of ultrathin Langmuir films by crown-substituted bisphthalocyaninates, Selektor et al. discovered the phenomenon of orientation-induced redox isomerism in planar supramolecular systems with Ce[(15C5)₄Pc]₂ as the example.⁶¹ Depending on the surface pressure at the water/air interface the complex underwent valence tautomeric interconversion: Ce⁴⁺[L²⁻]₂ ↔ [L²⁻]Ce³⁺[L˙⁻]; moreover, films with different redox states of metal centres could be transferred from water to solid surfaces to characterize the valence state by various spectroscopic techniques. The phenomenon was also found to be general for most of bisphthalocyaninates, which are formed by lanthanides exhibiting di- and trivalent states – Sm, Eu, Er, Tm, and Yb.^62,63

Homonuclear trisphthalocyaninates of strongly paramagnetic lanthanides – Ln₂[(15C5)₄Pc]₃, Ln = Tb, Dy, Ho, and Tm were studied by variable-temperature NMR showing the possibility of their application as thermosensors in the physiological temperature range (303–323 K)^64,65 due to the sensitivity of lanthanide-induced shifts to temperature. The Tb(III) complex was found to have the highest sensitivity of this shift among the reported paramagnetic complexes of 4f and d elements with polydentate ligands in organic media.

Several types of crown-substituted complexes were synthesized to study single-molecule magnetism,^66,67 which is particularly interesting for this type of REE sandwiches since it can be tuned by the interaction of crown-ether macrocycles with alkali metal cations. Such tuning was demonstrated for the first time by Horii et al.⁶⁷ with heteroleptic terbium(III) complex [(15C5)₄Pc]Tb(Pc) synthesized from (Pc)Tb(acac), 4′,5′-dicyanobenzo-15-crown-5 and DBU in hexanol as the example (cf. Section 3.3.1 for more details on SMM studies).

Alternative one-pot strategy towards this complex was proposed by Martynov et al.⁶⁸ It was based on the generation of (Pc)Tb(acac) by transmetalation of a lanthanum(III) complex La(Pc)₂ with Tb(acac)₃·nH₂O and the reaction of this intermediate with tetra-15-crown-5-phthalocyanine, H₂[(15C5)₄Pc], in the presence of DBU in refluxing 1-chloronaphthalene (ClN) and OctOH mixture (Scheme 3a). The yield of the complex was 24%, and the presence of DBU was shown to be crucial for the formation of the double-decker complex – in the absence of DBU the mixtures of homonuclear triple-decker complexes (Pc)Tb[(15C5)₄Pc]Tb(Pc) and [(15C5)₄Pc]Tb[(15C5)₄Pc]Tb(Pc) were obtained.^69,70

Scheme 3 Transmetalation of La(Pc)₂ in the synthesis of heteroleptic crown-substituted bisphthalocyaninates (a) and heteronuclear trisphthalocyaninates (b).^66,68,71,72

Original variant of the raise-by-one-story approach was proposed to synthesize heteronuclear triple-decker complexes, [(15C5)₄Pc]M*[(15C5)₄Pc]M(Pc), using La(Pc)₂ as a donor of unsubstituted phthalocyaninate-dianion (Scheme 3b).^71,72 Thus, its reaction with crown-substituted bisphthalocyaninates, M*[(15C5)₄Pc]₂ and M(acac)₃·nH₂O, in refluxing ClN resulted in remarkably rapid – 10–15 min – formation of the target heteronuclear complexes with a wide range of pairs of dia- and paramagnetic metal centers (M ≠ M* = Nd, Sm, Eu, Gd–Yb, Y) in high yields. SMM properties of homonuclear di-terbium(III) complex and heteronuclear terbium(III)/yttrium(III) complexes were studied and the effects of intramolecular ferromagnetic interactions together with the impact of coordinating surrounding were revealed.⁶⁶

A heteroleptic octopus-type Tb(III) bisphthalocyaninate bearing a 15-crown-substituted ligand and a ligand with eight thioacetate-terminated flexible tentacles was synthesized by Shokurov et al. via the raise-by-one-story approach starting from tetrahydropyranyl-protected phthalocyanine and half-sandwich Tb(III) crown-phthalocyaninate followed by the cleavage of protective groups, the introduction of iodine atoms and their nucleophilic substitution with thioacetate residues (Scheme 4).⁷³ The resulting octopus formed a self-assembled monolayer (SAM) on gold with a face-to-face orientation of the molecules, serving as an anchor for further surface modification, namely the binding of additional crown phthalocyanine molecules via potassium ion bridges. The enhanced redox-switching ability of the bilayer thus formed was demonstrated (cf. Section 4.3).

Scheme 4 Synthesis of heteroleptic octopus-type Tb(III) bisphthalocyaninate.⁷³

2.1.3 Chiral sandwich phthlocyaninates. An interesting example of aryloxy-substituted sandwiches was synthesized by Zhou et al. via template condensation of phthalonitrile, annelated with enantiomerically-pure R- or S-binaphthyloxy (BNPO) groups (Scheme 5a).⁷⁴ Studies of the resulting complexes Y[(R-/S-BNPO)₄Pc]₂ using CD-spectroscopy revealed efficient transfer of chiral information from periphery to phthalocyanine chromophores, and the efficiency of this transfer was much more pronounced in comparison with another example of chiral bisphthalocyaninates, M[(MenthO)₄Pc]₂, bearing D- or L-menthyl groups (Scheme 5b).⁷⁵ The latter complexes were prepared by Lv et al. via direct metalation of the corresponding methyl-substituted phthalocyanines with M(acac)₃·nH₂O (M = Eu, Lu, Y) in n-PentOH. The latter example also emphasized the role of intramolecular π–π interaction since the monomeric chiral ligand H₂[(D-/L-MenthO)₄Pc] was CD-inactive.

Scheme 5 Synthesis of chiral bisphthalocyaninates on the examples of Y[(BNPO)₄Pc]₂ (a, only S-enantiomer is shown),⁷⁴ and menthyl-substituted bisphthalocyaninates M[(MenthO)₄Pc]₂ (b).⁷⁵

Another naturally occurring chiral terpene, (−)-myrtenal, was used by Zheng et al. to synthesize the dicyanobenzo-annelated pinene derivative which underwent template condensation to furnish double-decker complexes M[(pinene)₄Pc]₂, M = Eu, Er, and Lu (Scheme 6).⁷⁶ Introduction of bulky conformationally rigid hydrocarbon substituents was made with the aim to decrease both inter- and intramolecular interactions between Pc ligands. Such tuning of intermolecular interactions decreased the aggregation of complexes and enhanced their film-forming ability. Decrease of intramolecular interactions resulted in a significant red shift of the intervalence band in the NIR region in comparison with other double-decker complexes, improving the electrochromic properties of compounds thus modified. Finally, chirality of complexes was manifested in their CD-spectra.

Scheme 6 Key steps leading to pinene-annelated phthalonitrile and bisphthalocyaninates (DMAD – dimethyl acetylene dicarboxylate). Only one isomer of double-decker complexes is shown.⁷⁶

Terbium(III) bisphthalocyaninate substituted with chiral (S)-2-(dodecyloxy)propoxy groups was synthesized by Gonidec et al. as a liquid crystalline material with tuneable magnetic properties.⁷⁷ This double-decker complex was synthesized by the direct metalation of the corresponding phthalocyanine (Scheme 7) with lithium bis(trimethylsilyl)amide as a basic reagent promoting complex formation. Similar to mentyl-substitued complexes, the metal-free ligand was CD-silent, while the double-decker complex showed a pronounced CD signal in the Q-band region. Also, the complex showed mesogenic behaviour and both crystalline and disordered states of the same specimen could be kinetically trapped upon measurements of SMM behaviour at low temperatures. Both structural states showed markedly different magnetization dynamics revealing important influence of the structural environment of a single-molecule magnet taking advantage of its phase behaviour.

Scheme 7 Synthesis of mesogenic terbium(III) bisphthalocyaninate, bearing (S)-2-(dodecyloxy)propoxy-substituents.⁷⁷

2.1.4 Sandwich phthalocyaninates bearing strong electron donors. High sensitivity of phthalocyanines to electronic effects of peripheral substituents allows the tuning of the properties of their complexes, thus over the last decade extreme samples have been reported, bearing particularly strong electron donors or acceptors.⁷⁸ Among the former, the introduction of di-n-butylamino-groups had a profound effect on the electronic structure of phthalocyanines – for example, octa-substituted metal-free ligand H₂[(Bu₂N)₈Pc] revealed a significant decrease by 0.26–0.35 V for the HOMO–LUMO gap in comparison with H₂(Pc) due to the effective p–π conjugation between the peripheral nitrogen atoms and the Pc chromophore.⁷⁹ This phthalocyanine was also used to synthesize various double-decker complexes (Scheme 8). Its direct metalation with M(acac)₃·nH₂O, in the TClB/OctOH (5 : 1 v/v) mixture yielded double-decker complexes in mediocre yields.⁸⁰

Scheme 8 Synthesis of bisphthalocyaninates based on octa-(di-n-butylamino)phthalocyanine.^81,82

The complex Tb[(Bu₂N)₈Pc]₂ revealed excellent SMM behaviour which could be further improved by the synthesis of a heteroleptic complex [(Bu₂N)₈Pc]Tb(Pc) (cf. Section 3.2.2).⁸¹ This complex was synthesized by the raise-by-one-story approach involving the metalation of H₂[(Bu₂N)₈Pc] with Tb(acac)·nH₂O in refluxing OctOH and in situ reaction of the generated Tb(III) monophthalocyaninate with Li₂(Pc). In the same manner, erbium(III) complex [(Bu₂N)₈Pc]Er(Pc) was also synthesized and its anionic form was a first sandwich-type phthalocyanine-based erbium(III) compound with field-induced slow relaxation.⁸²

Chiral heteroleptic complex was synthesized in a similar manner by the reaction of [(Bu₂N)₈Pc]Tb(acac) with H₂[(BNPO)₄Pc],⁸³ and it also revealed excellent SMM characteristics. Such superior SMM behaviour of amino-substituted complexes was explained using DFT calculations in terms of the electrostatic effects of electron-rich ligands.

2.1.5 Sandwich phthalocyaninates bearing strong electron acceptors. Apart from spectacular red–green–blue colour change upon oxidation and reduction of sandwich Pcs rendering them as components of electrochromic materials, switching between redox states of sandwiches can be used to tune their magnetic properties.^84,85 Thus, introduction of strong electron acceptors into double-decker complexes was used to stabilize their reduced forms and study their SMM performance. Starting from a series of dicyanophthalimides Gonidec et al. synthesized electron-deficient double-decker complexes stabilized in their anionic forms Tb[(R-imide)₄Pc]₂⁻TBA⁺ (Scheme 9).⁸⁶ Their SMM properties were virtually identical to those of the parent Tb(Pc)₂⁻TBA⁺ complex. However, while the anionic forms of unsubstituted double-decker can be readily oxidized, often not on purpose, the imide-substituted compounds were stable in the anionic forms in a wide range of potentials, rendering them as components of hybrid magnetic materials without undesired charge transfer interactions between the substrates and SMMs.

Scheme 9 Synthesis of the imide substituted double-decker complexes.⁸⁶

Another extremely electron-deficient Tb(III) bis(octa-isopropyl)phthalocyaninate where all hydrogen atoms were replaced with fluorine atoms was studied by Gonidec et al.⁸⁷ It could be electrochemically reduced to di- and trianionic forms, and magnetic characteristics of these forms were studied by low-temperature MCD (cf. Section 3.4).

2.1.6 Halogenated sandwich phthalocyaninates. Several examples of halogenated sandwich complexes were synthesized by template condensation of the corresponding phthalonitriles in the presence of metal salts. For example, the short-run reaction of 4-iodophthalonitrile with lanthanide acetates in melt at 380 °C for 5–6 min yielded a series of iodinated complexes M(I₄Pc)₂, M = Sm-Dy in 13–25% yields; the ability of complexes to absorb gaseous NH₃ was studied by Şenocak et al.⁸⁸ The yield of Tb(I₄Pc)₂ could be improved to 42% when 4-iodophthalonitrile was condensed in the presence of Tb(acac)₃·nH₂O and DBU in refluxing PentOH. Under the same conditions, condensation of this phthalonitrile with terbium(III) monophthalocyaninate [(tBuPhO)₈Pc]Tb(acac) yielded the heteroleptic complex (I₄Pc)Tb[(tBuPhP)₈Pc]. Functionalization of iodinated sandwich using the Sonogashira reaction with 4-pentyn-1-ol was also studied.⁸⁹

Neodymium(III) bis(octachlorophthalocyaninate), Nd(Cl₈Pc)₂, was synthesized by Kuzmina et al. via two alternative strategies. Template condensation of 4,5-dichlorophthalonitrile with Nd(OAc)₃·H₂O and DBU in refluxing PentOH yielded only 10% of the target complex; however, its yield could be improved to 87% using the raise-by-one-story reaction between H₂(Cl₈Pc) and (Cl₈Pc)Nd(OAc) and MeOLi in a refluxing mixture of trichlorobenzene and cetyl alcohol. This sandwich complex revealed improved optical limiting behaviour in comparison with nonhalogenated derivatives due to the heavy atom effect, although halogenated monophthalocyaninates turned out to be better optical limiters in comparison with sandwich complexes.⁹⁰ Interestingly, 4,5-dibromophthalonitrile did not yield any traces of double-decker complexes under the conditions of template synthesis.⁹¹

2.1.7 Sandwich phthalocyaninates with substituents in nonperipheral positions. All the aforementioned sandwiches contained substituents in the peripheral positions of the Pc rings, while nonperipherally-substituted complexes still remain relatively unexplored, although sterical hindrance caused by proximate nonperipheral substituents can have a profound influence on the structure and properties of such complexes, including particularly strong basicity of meso-nitrogen atoms and near-IR absorbance.⁹²

Sandwiches based on tetra-α-substituted Pc ligands were reported for the first time by Wang et al. in 2005 with tetra-[α-(3-pentyloxy)]phthalocyaninates, M[(α-Et₂CHO)₄Pc]₂, M = Eu, Lu, and Y as examples (Fig. 3a).⁹³ These complexes were synthesized in a straightforward manner by the reaction of metal-free phthalocyanine with the corresponding M(acac)₃·nH₂O in refluxing OctOH, and the major C_4h isomer of europium(III) complex was characterized using XRD, which showed its elegant pinwheel-like structure. SMM behaviour of terbium(III) complex was also studied, showing the clear dependence of the magnetic properties of the complex on its redox state (cf. Section 3.4).⁹⁴

Fig. 3 Homoleptic nonperipherally-substituted bisphthalocyaninates: (a) C_4h isomer of M[(α-C₅H₁₁O)₄Pc]₂ and X-ray structure of the Y(III) complex (CCDC NAZTEO),⁹³ and (b) meso-protonated complex HM[(α-BuO)₈Pc]₂ with the X-ray structure of the Eu(III) complex (CCDC XUXCOJ01).⁹⁵ Substituents in the X-ray structures are truncated to oxygen atoms; hydrogen atoms are omitted.

The synthesis of complexes with octa-α-substituted ligands (Fig. 3b) turned out to be more challenging. In 2010 Gao et al. showed that the reaction of H₂[(α-BuO)₈Pc] with M(acac)₃·nH₂O and DBU in OctOH did not yield any traces of sandwich complexes; however, it was found that crown-ethers catalyse this reaction, and the best yield of the complex HM[(α-BuO)₈Pc]₂ – 36.5% – was obtained in the presence of dibenzo-18-crown-6.⁹⁵ Crystallographic characterization of this complex revealed severe distortion of Pc ligands because of the sterical hindrance, caused by the eight proximate butoxy-groups as well as the presence of a proton at one of meso-nitrogen atoms. Localization of this proton in solution was shown by Damjanović et al. using the ¹H-NMR study of the paramagnetic terbium(III) complex HTb[(α-BuO)₈Pc]₂,⁹⁶ and the impact of protonation on the SMM behaviour of this complex was also demonstrated by Horii et al. (cf. Section 3.5).⁹⁷

2.2 Sandwich complexes with low-symmetry phthalocyanine ligands

Lowering the symmetry of phthalocyanines is a powerful tool to tune their physical-chemical properties,⁹⁸ which is particularly important in the field of nonlinear optics as well as in the fabrication of hybrid materials, where it allows the introduction of anchor groups which afford immobilization of complexes on various supports.

2.2.1 Rational template synthesis of A₇B-type bisphthalocyaninates. While low-symmetry phthalocyanines are typically synthesized by template cross-condensation of two different phthalonitriles, this approach has only limited applicability in the case of sandwich complexes since such condensation leads to enormous number of various products; however, it still can be applied when only the A₇B scaffold is required. This was showed for the first time with A₇B-type complex [(15C5)Pc]Lu(Pc) as the example, which was synthesized in a small yield of 3% by cross-condensation of unsubstituted and 15-crown-5-substituted phthlonitriles.⁹⁹

This approach was optimized by Alpugan et al. with the example of europium(III)-templated cross-condensation of 4,5-di(n-hexylthio)phthalonitrile as component A with phthalonitriles, bearing one or two 5-hydroxypentylthio-groups as component B (Scheme 10).¹⁰⁰ Using the stoichiometric ratio of precursors yielded a mixture of six products with laborious separation issues. However, due to similar reactivities of both components, the distribution of products is expected to be close to statistical, thus increasing the ratio of A : B the yield of products, containing more than one B unit can be decreased, and the optimal ratio of 10 : 1 was selected to obtain two main products – complexes Eu(A₄)₂ and (A₄)Eu(A₃B), whose separation was facilitated by increased polarity of the asymmetric compound. Reproducible 10–12% yields were obtained for A₇B-type complexes based on hydroxylated phthalonitrile. Further functionalization of complexes was achieved by their mesylation and substitution of mesyl groups with azide as a possible substrate for click-chemistry.

Scheme 10 Selective synthesis of a functionalized A₇B-type europium(III) bisphthalocyaninate.¹⁰⁰

2.2.2 Direct synthesis of homoleptic complexes with low-symmetry tetrapyrrolic ligands. Application of template reactions for the preparative synthesis of sandwich phthalocyaninates with more than one low-symmetry ligand is prohibitive, thus their synthesis is performed by the direct metalation of low-symmetry ligands.

Direct metalation of A₃B-type ligand was attempted by Oluwole et al. for the synthesis of homoleptic europium(III) bis- and trisphthalocyaninates bearing diethyleneglycol chains with terminal OH groups (Scheme 11).¹⁰¹ Unexpectedly, direct metalation of this ligand with Eu(OAc)₃ in refluxing OctOH neither in the absence nor in the presence of DBU failed to produce target sandwiches, likely due to side processes associated with the interaction of diethyleneglycol groups with Eu³⁺ ions. However, protection of OH-groups with dihydropyran in starting phthalocyanine prior to complex formation allowed smooth conversion of the protected derivative into double-decker and triple-decker complexes, and the selectivity was controlled by the addition of DBU. Deprotection of THPO-groups yielded OH-substituted sandwiches, which were immobilized on the surface of quantum dots to give the first examples of hybrid optical limiters based on sandwich REE complexes.

Scheme 11 Synthesis of sandwich europium(III) complexes bearing diethyleneglycol substituents with terminal OH-groups.¹⁰¹

A similar necessity to introduce protective groups was observed by Pushkarev et al. with the example of metalation of the phthalocyanine ligand bearing one phenolic OH group conjugated with the Pc system (Scheme 12a).¹⁰² Degradation of the unprotected ligand was observed upon metalation with europium and lutetium acetylacetonates in high boiling solvents; however, the benzylated derivative smoothly produced double-decker complexes where benzyl protections were removed by treatment with sulphuric acid.

Scheme 12 Synthesis of sandwich complexes starting from hydroxy-substituted phthalocyanine ligands.^102,103

This approach allowed Pedrini et al. to synthesize homoleptic terbium(III) bisphthalocyaninate with A₃B-type ligands bearing undec-10-en-1-yloxy groups, which were used to obtain covalent 1D architectures using diene metathesis (Scheme 12b).¹⁰³ This reaction turned out to be concentration-dependent – the product of intramolecular cyclisation of two ω-alkenyl chains was obtained when metathesis was performed in dichloromethane solution at 2.4 mM concentration. Increasing the concentration by an order of magnitude afforded the mixture of oligomers containing species from dimer to pentamer according to MALDI TOF mass-spectrometry. Importantly, the magnetic properties of this mixture were only slightly modified after the formation of the bond between molecules, highlighting the potential of this reaction scheme to form multinuclear magnetic assemblies.

Studies of metalation of covalently bonded dimeric phthalocyanines in the synthesis of sandwich complexes revealed that this reaction can proceed either in the intra- or intermolecular mode leading to mononuclear complexes or oligomeric assemblies (Scheme 13).

Scheme 13 Examples of sandwich complexes formed with dimeric tetrapyrrols: (a) intra-¹⁰⁴vs. intermolecular^105,106 binding of lanthanides with clamshell-phthalocyanine H₄[c-Pc₂]; (b) 24-crown-8-interlocked bisphthalocyaninates M[(24C8)Pc₂];¹⁰⁷ and (c) binaphthol-strapped chiral bis(porphyrinato)cerium(IV) complexes.¹⁰⁸

For example, clamshell-type diphthalocyanine H₄[c-Pc₂] with an o-xylyl bridge studied by Pushkarev et al. underwent mainly intracavity metalation upon interaction with lanthanide acetylacetonates (Scheme 13a), although some oligomers were detected by MALDI TOF mass-spectrometry.¹⁰⁴ To form well-defined dimeric bisphthalocyaninates the starting clamshell ligand can be treated with lanthanide monophthalocyaninates.^105,106 With a diterbium complex (tBu₄Pc)Tb[c-Pc₂]Tb(tBu₄Pc) as the example ferromagnetic intermetallic interactions were observed leading to elongation of magnetic relaxation times in comparison with the monomeric terbium complex Tb(tBu₄Pc)₂ (cf. Section 3.3.1).¹⁰⁵

The balance of intra- vs. intermolecular complexation was studied by Birin et al. for butoxy-substituted diphthalocyanine with 24-crown-8 bridge H₄[(24C8)Pc₂].¹⁰⁷ Its interaction with M(acac)₃, M = La, Ce, and Eu, and DBU in ClN yielded mixtures of mononuclear complexes and olygomers, and the yields of monomeric intercavity complexes (Scheme 13b) increased with the decrease of the metal size, gaining 40% for the Ce(IV) complex. The latter complex revealed unusual solvatochromic behaviour when going from chloroform to toluene, which was not observed in the case of homoleptic symmetrical Ce[(15C5)₄Pc]₂. Such behaviour was also studied by NMR spectroscopy which revealed solvent-induced changing of sandwich conformation.

Intracavity complexes of cerium(IV) with diporphyrins with chiral BINOL bridges Ce[(BINOL-n)Por₂], n = 1 or 2 (Scheme 13c), were synthesized by Lu et al. using metalation of the corresponding ligands with cerium(III) acetylacetonate in trichlorobenzene in the presence of metal lithium, which was added to activate the porphyrin.¹⁰⁸ CD measurements evidenced that (R)- and (S)-BINOL-strapped complexes formed perfect mirror spectra with negative and positive signs in the Soret band regions respectively. This observation suggests that the C₂-chirality of a BINOL strap is transcribed to the double-decker core as a defined chiral twist in the inter-porphyrin arrangement. The intensity of signals also depended on the length of the aliphatic bridge – (CH₂CH₂)_n between BINOL and CePor₂ units, indicating the possibility to tune the efficiency of the intramolecular chirality transfer.

Homoleptic complexes with low-symmetric ABAB-type Pc ligands have been designed as NLO materials, bearing in mind the concept of octupolar molecules, which can be approximated by a cube with alternating donor and acceptor groups in the corners (Scheme 14). With this aim, double-deckers were synthesized by direct metalation of the corresponding Pcs bearing either four hexylthio-groups – H₂[(SHex)₄Pc],^109,110 or alternating pairs of α-PentO and β-Cl substituents H₂[(α-PentO)₄Cl₄Pc].^111,112

Scheme 14 Synthesis of octupolar bisphthalocyaninates based on ABAB-type ligands.^109–112

Homoleptic terbium(III) bis-porphyrinate HTb(DADAPor)₂ was synthesized by the reaction of metal-free tetra-meso-substituted porphyrin bearing alternating donor- and acceptor p-diphenylamino-phenyl and p-carboxymethylphenyl-ethynyl groups (Scheme 15).¹¹³ Although X-ray structure of this complex was not obtained, DFT calculations suggested that intramolecular interactions between donor and acceptor fragments should stabilize distorted cuboid conformation of this sandwich with a twist angle near 17°. This structural feature provided the studied complex with the largest off-resonant first hyperpolarizability as evaluated by harmonic light scattering measurements. Details on NLO studies of synthesized octupolar sandwiches are given in Section 5.1.

Scheme 15 Synthesis of octupolar bisporphyrinate with ABAB-type ligands.¹¹³

2.2.3 Heteroleptic complexes with low-symmetry phthalocyanine ligands. Heteroleptic complexes bearing only one low-symmetry ligand are convenient components of covalent or supramolecular assemblies with nanomaterials. Their synthesis is typically accomplished via the raise-by-one-story approach, where monophthalocyaninate with one ligand reacts either with another ligand or with phthalocyanine precursors.

For example, bisphthalocyaninates bearing one pyrene-containing A₃B ligand were synthesized by Klyatskaya et al. starting from Li₂(Pc) which was treated with M(acac)₃ in ClN generating (Pc)M(acac) species, M = Tb, Dy, and Ho.^114,115 Their further interaction with dilithium complex of pyrenyl-substituted phthalocyanine Li₂[(Pyrene)Pc] yielded 20–24% of the target heteroleptic complexes [(Pyrene)Pc]M(Pc), which were separated from homoleptic complexes by column chromatography (Scheme 16). Isolation of the heteroleptic triple-decker terbium(III) complex with one terminal unsubstituted ligand and two adjacent pyrenyl-substituted ligands was also reported; the order of ligands was determined by collision induced dissociation using Fourier transform ion cyclotron resonance and an Orbitrap mass spectrometer.¹¹⁶

Due to the presence of pyrenyl group in the double-decker terbium(III) complex, it could be anchored to the surface of carbon nanotubes via π–π interactions. The anisotropy energy barrier and the magnetic relaxation time of the resulting conjugate are both increased compared with those of pure crystalline [(Pyrene)Pc]Tb(Pc).¹¹⁴ Also this conjugate behaved as an original spin-valve device, where localized magnetic moments lead to a magnetic field dependence of the electronic transport through single-walled carbon nanotubes, resulting in magnetoresistance ratios up to 300% at temperatures lower than 1 K (cf. Section 3.8).¹¹⁷

Scheme 16 Synthesis of pyrenyl-substituted bisphthalocyaninates [(Pyrene)Pc]M(Pc).^114,115

Interestingly, the ability of the pyrenyl group to participate in stacking interactions governed the crystal packing of the [(Pyrene)Pc]Tb(Pc) complex – its molecules formed head-to-tail dimers with the shortest intermolecular distances of 3.345–3.369 Å which is in good agreement with the general values observed for π–π stacking (Fig. 4).¹¹⁵

Fig. 4 Structure of a head-to-tail supramolecular dimer, formed in crystal packing by molecules of [(Pyrene)Pc]Tb(Pc) (CCDC LOGBEQ).¹¹⁵ Hexyl substituents, hydrogen atoms and solvate molecules are omitted for clarity. The shortest intermolecular distances are shown with blue dotted lines.

Heteroleptic bisphthalocyaninates, bearing one fullerene-appended ligand, were synthesized by Ballesteros et al. starting from a A₃B-type ligand with one benzylic alcohol group H₂[(BnOH)Pc] whose reaction with acetylacetonates in o-dichlorobenzene yielded the corresponding monophthalocyaninates [(BnOH)Pc]M(acac), M = Sm, Eu and Lu (Scheme 17).¹¹⁸ These complexes were used as templates in reactions with phthalonitrile to construct unsubstituted Pc ligands yielding heteroleptic sandwiches which were esterified with fulleropyrrolidine carboxylic acid C₆₀-COOH. Photophysical characterization of the resulting complexes [(BnOCO-C₆₀)Pc]M(Pc) showed that their components totally lack electronic interactions in the ground state; however, photoexcitation of the C₆₀ fragment at 387 nm lead to a fast and long-lived charge transfer from M(Pc)₂ to C₆₀ π-systems. In contrast, the short-lived nature of the bisphthalocyaninate excited states fails to trigger analogous charge-transfer reactions upon photoexcitation of this moiety.

Scheme 17 Synthesis of fullerene-substituted bisphthalocyaninates [(BnOCO-C₆₀)Pc]M(Pc).¹¹⁸

Monophthalocyaninates can act as templates in cross-condensation of pairs of different phthalonitriles A and B paving the way to heteroleptic complexes where one of the ligands is represented by various combinations A_nB_4−n, n = 0–4. This one-pot approach was used by Pan et al. to synthesize a series of tetrathiafulvalene(TTF)-fused heteroleptic europium(III) double-deckers [(TTF)_nPc]Eu(Pc).¹¹⁹ With this aim (Pc)Eu(acac) reacted with the mixture of unsubstituted and TTF-fused phthalonitriles (Scheme 18) and the mixture of the resulting complexes could be successfully separated by size-exclusion chromatography together with chromatography on silica to separate the cis- and trans-isomers of [(TTF)₂Pc]Eu(Pc). Introduction of the TTF units onto double-decker complexes had strong effects on their UV-Vis spectra and electrochemical properties, revealing the significant contribution of these units to the electronic structure of double-deckers.

Scheme 18 One-pot synthesis of a series of tetrathiafulvalene (TTF)-fused bisphthalocyaninates [(TTF)_nPc]Eu(Pc), n = 0–4.¹¹⁸

Unprecedent heteroleptic complexes with open-chain triisoindole-1-one ligands (R-TIO)Lu(Pc), R = Me or iPr, were unexpectedly isolated by Wang et al. from the conventional raise-by-one-story synthesis of heteroleptic bisphthalocyaninates via template condensation of bulky diaryloxy-phthalonitriles with lutetium(III) monophthalocyaninate and DBU in refluxing pentanol (Scheme 19).¹²⁰ It was found that the target [(R₂PhO)₈Pc]Lu(Pc) complexes were sole products of template condensation only when it was performed under strictly anaerobic conditions. However, the presence of even 5 ppm of oxygen in inert gas media terminated the reaction at the stage of uncyclized isoindole oligomeric derivatives rather than the phthalocyanine chromophores. The yields of TIO-containing sandwiches could be increased up to 25.8% when the reactions were performed in a pure oxygen atmosphere. Experiments with labeled ¹⁸O₂ showed that the carbonyl group in the TIO ligand contains ¹⁶O atom, suggesting that it comes from the solvent. TIO ligands were stabilized both by bulky aryloxy-groups as well as by half-sandwich part of the molecule, and demetalation of these complexes resulted in complete degradation of triisoindole-1-ones.

Scheme 19 Synthesis of heteroleptic double-deckers with triisoindole-1-one ligands.¹²⁰

2.3 Mixed ligand sandwich complexes

In hierarchy of sandwich complexes mixed ligand double- and triple-deckers combining tetrapyrrolic macrocycles from different families (Por, Pc, Nc, etc.) attract special attention due to wider possibilities of tuning of both structural and functional characteristics in comparison with homoligand complexes. Among the most important features in the context of the present review we should mention specific acid–base equilibria in double-decker Por-Pc complexes (cf. Section 3.5). Panchromatic absorbance appearing from orbital interactions between stacked ligands is particularly important for the elaboration of broadband light harvesting materials.¹²¹ Combination of ligands with different sterical demands in sandwich complexes also affords precise control of twist angles between these ligands which in turn rules the shape of the coordination polyhedron of the metal centre bound to these ligands and affects its SMM behaviour (cf. Section 3.2.1).

2.3.1 Mixed complexes containing porphyrin and phthalocyanine ligands. Mixed ligand sandwiches containing both porphyrin and phthalocyanine macrocycles have been known for several decades starting from the seminal work of Moussavi et al. in 1989.³⁴ A new round of interest in these complexes began in 2012 when Sakaue et al. reported the synthesis, structural characterization and magnetic relaxation studies of homo- and heteronuclear triple-deckers (TAnP)M*(Pc)M(Pc), M* = M = Tb and M* ≠ M = Tb/Y.¹²² These complexes were prepared in modest yields of 2.3–4.2% by generation of REE monoporphyrinates from tetra-p-anisyl-porphyrin H₂(TAnP) and M*(acac)₃ with subsequent interaction with M(Pc)₂ in refluxing trichlorobenzene (Scheme 20). XRD characterization of complexes revealed different symmetries of coordination surrounding of metal centres M and M*, which has a profound effect on the dynamics of magnetization relaxation (cf. Section 3.2.1).

Scheme 20 Synthesis of homo- and heteronuclear mixed ligand complexes (TAnP)M*(Pc)M(Pc), M* = M = Tb or M* ≠ M = Tb/Y.¹²²

Interest in studies of mixed ligand sandwiches requires efficient synthetic pathways with improved yields, which can be realized either by optimization of experimental protocols, or by application of auxiliary functionalities in macrocycles, governing the regioselectivity of complex formation.

With the aim to synthesize early lanthanide complexes based on tetra-meso-aryl-porphyrins and crown-phthalocyanine Birin et al. proposed one-pot procedures implying reactions between corresponding porphyrins, M(acac)₃ and 15-crown-5-phthalonitrile, the latter undergoes template condensation with simultaneous formation of sandwich complexes (Scheme 21).^123–126 The balance of concurrent stages was studied¹²⁷ and the optimal conditions were found leading either to mixtures of double- and triple-decker complexes or selectively to triple-decker complexes with a Por–Pc–Por ligand arrangement. The crucial role of electron-donating crown-ether groups in the synthesis of such triple-deckers was demonstrated, when unsubstituted phthalonitrile was used instead of crown-substituted precursor; the synthesis terminated at the step of mixed-ligand double-decker complex.¹²⁸ Going from early to late lanthanides drastically decreased the yield of triple-deckers, thus, to synthesize a terbium(III) complex it was necessary to use the raise-by-one procedure with the addition of terbium(III) porphyrinates to pre-synthesized double-decker complexes.¹²⁹

Scheme 21 Synthesis of crown-substituted mixed ligand sandwich complexes.^123–129

Classical conditions of the raise-by-one-story approach to Por–Pc–Pc mixed ligand complexes via the interaction of monoporphyrinates with bisphthalocyaninates in high-boiling solvents turned out to be inapplicable in the synthesis of complexes with phthalocyanines bearing bulky substituents. To circumvent this drawback Jin et al. used microwave activation, the thus optimized conditions allowed the synthesis of series of complexes in high yields (Scheme 22).¹³⁰ Importantly, MW activation allowed the speeding-up of the synthetic procedures, thereby minimizing possible thermolysis of labile compounds. This benefit allowed the synthesis of complex with free meso-positions in porphyrin ligands, while under classical solvothermal synthesis of similar complexes starting from 5,15-diphenylporphyrin its degradation was observed under prolonged refluxing of reaction mixtures.¹²⁸

Scheme 22 Microwave-mediated synthesis of lanthanide porphyrin/phthalocyanine triple-deckers bearing bulky substituents.¹³⁰

Straightforward synthesis of mixed ligand triple-deckers leading to the regioselective Por–Pc–Por arrangement of ligands was achieved by González-Lucas et al.¹³¹ who applied a diporphyrin with a flexible decamethylene spacer (Scheme 23). This pincer ligand sequentially captures lanthanide ions and phthalocyanines to efficiently form desired triple-decker complexes in high yield when it reacts with stoichiometric amounts of La(acac)₃ and H₂(Pc). In a similar manner, complexes with Pr(III), Nd(III), Sm(III) and Eu(III) were synthesized; however, attempts to synthesize complexes with a smaller metal centre – Dy(III) resulted in the formation of a double-decker complex linked with a free-base porphyrin residue. Reaction of this complex with La(acac)₃ resulted in the formation of an intermediate heteronuclear La(III)/Dy(III) complex which however underwent complete transmetalation with the formation of a homonuclear lanthanum(III) complex.

Scheme 23 Synthesis of sandwich mixed ligand complexes with spacered diporphyrin and X-ray structure of dilanthanum(III) complex (CCDC LUVWUX). Only one of the symmetry-imposed disorder components is shown. Hydrogen atoms and solvent molecules have been omitted for clarity.¹³¹

To broaden the range of useful properties of mixed ligand complexes, they can be functionalized with redox-active groups or hydrogen-bond donors which can form supramolecular aggregates of various architectures via self-assembling processes.

A series of ferrocene-decorated mixed ligand sandwiches were synthesized by Zhu et al. using the raise-by-one-story approach starting from 5,15-diferrocenylporphyrin in 20–5% yields for double-decker complexes and 70% for a triple-decker sandwich (Fig. 5).¹³² Electrochemical studies of these complexes revealed two consecutive Fc-based one-electron oxidation waves, suggesting the effective electronic coupling between these units which could be finely tuned by structural modifications of the sandwich framework. Thus, the separation between these waves increased with the decrease of the REE ionic radius, but it decreased when going from a double- to triple-decker complex.

Fig. 5 Ferrocene-decorated double- and triple-decker mixed ligand complexes.¹³²

Mixed ligand europium(III) double-deckers bearing from one to four n-octylamino-groups (Fig. 6) were synthesized and used as tectons for self-assembled nanostructures, formed upon injection of concentrated solutions of complexes in toluene into excess of methanol.^133,134 The morphology of the resulting aggregates was studied by various methods, and it was demonstrated that it governed their functional characteristics. For example, in the case of (A₁-TPP)Eu(Pc) and (A₄-TPP)Eu(Pc) this procedure resulted in the fabrication of uniform well-defined nano-rods and nano-sheets with conductivities of 3.85 × 10⁻⁵ and 2.48 × 10⁻⁵ S m⁻¹, respectively. This difference was attributed mainly to the more effective intermolecular π-electron delocalization in nano-sheets which was concluded from the analysis of X-ray diffraction patterns of the nanostructures.¹³⁴

Fig. 6 Amino-substituted mixed ligand complexes (A₁-TPP)Eu(Pc): R₁ = NHOct, R₂ = R₃ = H; (A₂-TPP)Eu(Pc): R₁ = R₂ = NHOct, R₃ = H; (A₄-TPP)Eu(Pc): R₁ = R₂ = R₃ = NHOct.^133,134

2.3.2 Mixed ligand complexes with naphthalocyanines. Introduction of naphthalocyanines into mixed ligand complexes affords strong dissymmetrisation of electronic density due to anisotropic expansion of the π-system, which can be used for grafting of molecules in a well-defined fashion on various supports. For example, Komeda et al. absorbed (Nc)Tb(Pc) on the Au(111) surface in segregated Pc- or Nc-up modes to study Kondo resonance originating from interactions between conduction electrons and a localized spin (cf. Section 3.7).¹³⁵ Katoh et al. showed that the cationic complex (Nc)Tb(Pc)⁺PF₆⁻ forms tetragonal crystals with perfect collinear arrangement of metal centres with efficient intermolecular ferromagnetic coupling (cf. Section 3.3.3).¹³⁶

Mixed-ligand Pc–Nc complexes can be synthesized using the raise-by-one-story approach. Thus, neutral complex (Nc)Tb(Pc) was synthesized in a mediocre yield by template condensation of 2,3-naphthalonitrile in the coordination sphere of terbium(III) monophthalocyaninate in the presence of DBU and ammonium molybdate (Scheme 24i).¹³⁶ High quality crystals of its cationic complex were obtained by the electrochemical oxidation of neutral (Nc)Tb(Pc)⁰ in dichloromethane in the presence of NBu₄PF₆ electrolyte.

Scheme 24 Template (i) vs. direct (ii) synthesis of complexes containing phthalo- and naphthalocyanine ligands.^43,136,137

Scheme 25 Synthesis of double-decker complexes with octaethylporphyrazine and the X-ray structure of the complex Tb(OEPz)₂ (CCDC QOQPIX).¹⁴⁰ Hydrogen atoms and solvate molecules are not shown for clarity.

The interaction of naphthalocyanine ligands with monophthalocyaninates using lithium methoxide in the mixture of TClB and cetyl alcohol allowed Dubinina et al. to optimize the synthesis of mixed ligand phthalo-/naphthalocyaninates (Scheme 24ii).^43,137 It afforded a series of complexes containing various substituents in both Pc and Nc rings, and the yields of sandwiches were high even for the late lanthanides. AFM studies of thin films formed by lutetium(III) Pc–Nc complexes evidenced that their tendency to aggregate depended on the substitution patterns, gaining maximum for (Ph₈Nc)Lu(Ph₈Pc), and in general the series of complex behaved as semiconductors with small activation energy.⁴³

A mixed ligand europium(III) complex with decylthio-substituted naphthalocyanine and push-pull porphyrin (Fig. 7a) was synthesized by Stefak et al., who used the template condensation of 6,7-di(n-decylthio)-2,3-naphthalonitrile in the presence of the corresponding monoporphyrinate.¹³⁸ This complex was suggested as a rotor which can be absorbed on the Cu(111) surface in a porphyrin-up fashion, forming a network where molecules undergo simultaneous and coordinated rotational switching under the influence of an STM tip-induced electric field by applying biases above 1 V at 80 K (Fig. 7b).¹³⁹

Fig. 7 (a) Donor–acceptor porphyrinate rotor mounted on a naphthalocyaninato-europium(III) complex.¹³⁸ (b) Behaviour of a supramolecular network of a donor–acceptor double-decker complex in monolayer, formed by complex on the Cu(111) surface.¹³⁹

2.4 Sandwiches with less-common tetrapyrrolic ligands

Advances in the synthesis of novel tetrapyrrolic ligands were often followed by the introduction of these ligands into the chemistry of sandwich REE complexes. The following section will include examples of double- and triple-deckers with porphyrazines including hetero-annulated derivatives, tetrabenzoporphyrins, and core-modified macrocycles – tetrabensotriazaporphyrins (TBTAP), corroles and N-confused porphyrins. Most of these complexes are yet to be comprehensively explored in terms of applicability, thus this section can give a summary inspiring their further study.

2.4.1 Sandwich complexes with porphyrazine ligands. In contrast to porphyrins and especially phthalocyanines attention to porphyrazines in context of sandwich REE complexes still remains very modest.

A series of neutral double-decker complexes with an octaethylporpyrazine ligand (OEPz) with Tb(III), Dy(III), Gd(III) and Y(III) metal centres was reported in 2014 by Giménez-Agulló et al. Sandwiches were synthesized starting from H₂(OEPz) using microwave activation.¹⁴⁰ Interestingly, the synthetic protocol included the necessity to add tetrabutylammonium benzoate as an auxiliary reagent – in its absence double-decker complexes could not be obtained. The authors claimed that this salt was needed to stabilize the intermediate – anionic form [M³⁺(OEPz²⁻)₂]⁻TBA⁺, which underwent oxidation to the neutral radical form during reaction workup.

X-ray structure of the complex showed that the molecule adapts staggered conformation with SAP surrounding of the metal centre which is a prerequisite of SMM behaviour of Tb(III) and Dy(III) complexes. Indeed, AC magnetometry showed that Tb(OEPz)₂ is characterized by high value of magnetization relaxation barrier comparable with the value of Tb(Pc)₂.

Sandwich REE complexes with heterocycle-annelated macrocycles were reported for the first time by Tarakanova et al. in 2014¹⁴¹ with complexes with tetradiazepinoporphyrazine (DZPz) ligands as examples. High CH-acidity of the peripheral diazepine rings made them efficient hydrogen bond donors thus the macrocycles themselves were found to form highly stable dimers in solutions.¹⁴² Staggered conformations of these dimers reminded the arrangement of tetrapyrrols in REE sandwiches; therefore, additional stabilization of a double-decker scaffold could be expected.¹⁴³ And indeed, it was shown that direct metalation of H₂[(tBuPh)₈DZPz] with M(acac)₃ and DBU (Scheme 26a) smoothly yielded double-decker complexes at the temperature of refluxing o-dichlorobenzene (180 °C), where classical phthalocyanines selectively form only monophthalocyaninates.^{141,143–145} Importantly, diazepine rings were susceptible to hydrolysis in alkaline media, which made template synthesis of REE complexes with DZPz ligands less efficient than direct metalation.¹⁴¹ DFT modelling of the resulting double-deckers suggested the presence of intramolecular hydrogen bonds between endocyclic CH-groups of diazepine rings and meso-nitrogen atoms of opposite ligands (Scheme 26b).

Scheme 26 (a) Synthesis of double-decker complexes with tetradiazepinoporphyrazine ligands. (b) Optimized structure of Lu(Ph₈DZPz)₂⁻ in top and side views, showing hydrogen bonds between CH₂ groups of diazepine rings and meso-N atoms. Other hydrogen atoms are omitted for clarity. Model was built using Cartesian coordinates, provided in the ESI in ref. 143.

Among other differences in the reactivity of DZPz and Pc ligands the stabilization of anionic forms of double-decker complexes should be emphasized. In contrast to M(Pc)₂⁻ which readily undergoes aerobic oxidation to neutral forms, the synthesized complexes M[(tBuPh)₈DZPz]₂⁻ remained in this form during chromatographic purification, and the addition of strong oxidants like bromine was needed to generate neutral forms which revealed characteristic intervalence bands in the NIR region.¹⁴³ The cerium complex fell out of this trend and its double-decker complex was stabilized in the neutral form with a quadrivalent metal centre.¹⁴⁵ Its electrochemical reduction to trivalent state was characterized by slow electron transfer kinetics which was explained by strong structural rearrangement of the sandwich molecule following the electron transfer step as well as shielding of the metal center by the bulky macrocycles.

Interestingly, solutions of anionic double-deckers with DZPz ligands were stable for several months showing no evidence of aggregation or decomposition; however, in the case of neutral forms evolution of their UV-Vis spectra was observed, depending on the nature of the metal centre. First of all, upon storage of solutions of M[(tBuPh)₈DZPz]₂⁰˙ in CH₂Cl₂ the hypsochromic shift and broadening of Q-bands were observed, and DLS studies of aged solutions revealed the formation of aggregates with a hydrodynamic radius of 20 nm. The tendency of neutral forms to aggregate was explained by the possibility of inversion of diazepine moieties leading to the reorientation of the intramolecular hydrogen bonds in the double-decker followed by the formation of intermolecular hydrogen bonds. In turn, it weakened the stabilizing effect of intramolecular H-bonding, and it resulted in lowering of stability in early lanthanide complexes – slow demetalation of La[(tBuPh)₈DZPz]₂⁰˙ was observed even after 1 day of solution storage, while lutetium(III) complex was stable towards demetalation ever after five days.¹⁴³

In the case of mixed ligand complexes [(tBuPh)₈DZPz]M(Pc), M = La, Eu and Lu, spontaneous aggregation due to inversion of diazepine ring was observed even for anionic forms.¹⁴⁶ In contrast to homoleptic complexes these double-deckers could form only dimeric species, which were observed by NMR spectroscopy. Tendency to aggregation followed the growth of the ionic radius of lanthanide ion due to an increase of interligand distance, weakening of intramolecular CH…N interactions and facilitation of diazepine ring inversion. The same trend was observed in the case of neutral forms generated during spectroelectrochemical studies. Thus, the intermolecular self-assembly of sandwich complexes containing DZPz decks can be controlled by changing the nature of both peripheral substituents and REE ions.

Soluble tetrathieno[2,3-b]porphyrazine annelated with cyclohexane rings was used by Dubinina et al. to synthesize lutetium(III) complexes expecting that the presence of sulphur-containing heterocycles can provide the resulting complexes with improved NLO properties.¹⁴⁷ Starting metal-free ligand was synthesized from tiophenedicarbonitrile by Zn-templated condensation and demetalation. Reactivity of the resulting ligand was quite unusual – upon its reaction with Lu(OAc)₃·4H₂O and MeOLi in refluxing mixture of TClB and cetyl alcohol only mono- and triple-decker complexes were obtained, which were separated by preparative TLC (Scheme 27). Although the formation of double-decker was detected by mass-spectrometry, it could not be isolated because of its low stability associated with the presence of π-radical.

Scheme 27 Synthesis of lutetium(III) complexes with substituted tetrathieno[2,3-b] porphyrazine.¹⁴⁷

2.4.2 Sandwich complexes with tetrabenzoporphyrin and tetrabenzocorrin ligands. Tetrabenzoporphyrins (TBP) were known since 1938;¹⁴⁸ however, their complexes with REE were described for the first time only in 2012. The studies were inspired by interest in the properties of porphyrin complexes with an expanded aromatic system. These and more recent results are summarised in Scheme 28.

Scheme 28 Synthetic pathways to sandwich complexes with benzoporphyrin and corrin ligands.

Thus, interaction of tetra(bicyclo[2.2.2])octadienoporphyrin with Gd(OAc)₃ in refluxing TClB was studied by Xu and Liu et al.¹⁴⁹ Under these conditions metalation of tetrapyrrolic macrocycles was accompanied by retro Diels-Alder aromatization producing TBP ligands. Heating of the reaction mixture for 15–17 h resulted in the formation of a double-decker complex Gd(TBP)₂, and increase of the reaction time to 45–48 h resulted in the predominant formation of a triple-decker complex Gd₂(TBP)₃.^150,151 Magnetic properties of complexes were studied revealing antiferromagnetic interactions between π- and f-electrons in Gd(TBP)₂ and between two Gd³⁺ ions in the case of Gd₂(TBP)₃.

Interaction of tetrabenzoporphyrin H₂(TBP) itself with M(acac)₃, M = La, Nd, Gd, Dy, and Lu was used by Galanin et al. to generate half-sandwich complexes (TBP)M(acac) which reacted with Li₂(Pc) to form mixed ligand double-deckers (TBP)M(Pc) in 48–74% yields.¹⁵² A similar strategy was applied to synthesize complexes starting from tert-butyl-substituted benzoporphyrin H₂[(tBu)₄TBP].¹⁵³

Original strategy leading to homoleptic complexes M(TBP)₂, M = Gd, Er, and Lu, was proposed using a dipyrrolic precursor (PhIm)₂ prepared from readily available phthalimide PhIm.¹⁵⁴ This procedure compares favourably with the time-consuming protocol using tetra(bicyclo[2.2.2])octadienoporphyrin due to the availability of starting compounds.

Application of acyclic olygopyrrolic precursors affords more options in the synthesis of various porphyrinoid derivatives. Condensation of 2-methylquinoline with a tripyrrolic compound (PhIm)₃ prepared from phthalimide afforded meso-(quinolin-5-yl)tetrabenzoporphyrin H₂(QnlTBP) which was also used to synthesize mixed ligand complexes.¹⁵⁵ The aforementioned dipyrrolic compound (PhIm)₂ can be dimerized in the presence of lanthanide acetates and hydrazine with the formation of complexes with the 1-tethyltetrabenzo-octadehydrocorrin ligand.¹⁵⁶ Depending on the ratio of reagents, either half-sandwiches or homoleptic double-decker complexes could be obtained. The reaction of the former complex with phthalonitrile can be used to synthesize phthalocyanine-corrin mixed ligand double- and triple-decker complexes.

2.4.3 Sandwich complexes with tetrabenzotriazaporphyrin ligands. The tetrabenzotriazporphyrin (TBTAP) ligands are intermediate macrocycles between classical porphyrins and phthalocyanines, possessing a permanent dipole moment, which may be responsible for the emergence of new properties. REE complexes with these ligands were reported by Pushkarev et al. on the examples of homo- and heteroleptic lutetium(III)^157,158 and europium(III) double-decker complexes which were synthesized by direct metalation of meso-phenyl-TBTAP macrocycle using lithium methoxide as a basic reagent (Scheme 29a). Apart from a homoleptic double-decker complex, small amount of triple-decker sandwich was isolated in the course of synthesis of Eu(Ph-TBTAP)₂, and moreover, prolonged heating of reaction mixtures resulted in partial dephenylation of TBTAP ligands (Scheme 29b), which was never observed previously in the chemistry of meso-aryl-substituted tetrapyrroles.

Scheme 29 (a) Synthesis of homo- and heteroleptic complexes with meso-phenyl-tetrabenzotriazaporphyrin, H₂(Ph-TBTAP).^157,158 (b) Side products of dephenylation. (c) X-ray structure of the complex Eu(Ph-TBTAP)₂ (CCDC ZUNJID).¹⁵⁸ Hydrogen atoms are not shown for clarity, meso-phenyl groups are highlighted with blue colour. (d) Anti- vs. syn-conformers of M(Ph-TBTAP)₂ complexes.

Comparison of UV-Vis spectra of M(Pc)₂, (Ph-TBTAP)M(Pc) and M(Ph-TBTAP)₂ revealed that replacement of one or two meso-nitrogen atoms with C-Ph groups resulted in gradual perturbation of electronic systems, leading to complicated spectral patterns with split Q- and intervalence bands, which is not typical for symmetrical complexes. Electrochemical studies of new complexes evidenced that the introduction of two bulky phenyl rings has a profound effect on frontier orbital levels due to weakening of the intramolecular interaction between macrocyclic ligands.

Single-crystal X-ray characterization of Eu(Ph-TBTAP)₂ showed that the molecule adapt syn-conformation (Scheme 29c and d), although conformational analysis performed by DFT calculations suggested that anti-form should be more stable and this form is observed as a sole state in solution according to NMR studies. Thus, a less stable conformer is formed in the solid state due to the advantage in energy of the molecular packing.

2.4.4 Sandwich complexes with corrole ligands. Corroles as ring-contracted porphyrinoids have been known since 1964; however, their fascinating properties were revealed mainly in the 1990s when straightforward synthetic procedures towards this class of macrocycles were proposed by the groups of Gross and Paolesse.^159,160

First examples of REE complexes with corroles were synthesized only in 2013 by Buckley et al. either by metalation of tris-meso-aryl-corrole with M[N(SiMe₃)₂]₃, M = La and Tb or by the reaction of lithium corrolate with GdCl₃.¹⁶¹ In both the cases half-sandwiches were isolated, where the coordination sphere of lanthanide ions was saturated with neutral ligands – dimethoxyethane or trimethyltriazanonane. A similar procedure was used to synthesize cerium complexes with corrole ligands and axial ligand-dependent self-assembling of the resulting half-sandwiches was studied.¹⁶²

Mixed ligand complexes containing corroles and phthalocyanine ligands were synthesized by the reaction of pre-synthesized monophthalocyaninates with corroles in the presence of DBU in octanol (Scheme 30).^163–168 Several substituted macrocycles were introduced into this reaction, but invariantly on substituents and REE ions in all cases mixed ligand triple-deckers were isolated with inner Cor and terminal Pc ligands. Moreover, reaction of half-sandwich corrole complex with metal-free phthalocyanine and one-pot reaction of corrole, metal-free Pc and metal salts both yielded the identical triple-deckers and their yields were systematically larger than that when the first method was used.¹⁶⁶

Scheme 30 (a) Synthesis of mixed ligand complexes with phthalocyanine and corrole macrocycles. (b) X-ray structure of the complex Eu₂[(OctO)₈Pc]₂[(p-ClPh)₃Cor] (CCDC HOKQOP).¹⁶³ Hydrogen atoms and substituents in top views are not shown for clarity.

Some of the synthesized complexes were characterized by XRD revealing that in contrast to other symmetrical triple-deckers two outer phthalocyanine ligands were not staggered, and coordination polyhedra of two metal centres were slightly different, although in all cases they were close to distorted square prisms.^163,164,167 This distortion of coordination surrounding manifested in fast relaxation of magnetization because of QTM (the quantum tunnelling of the magnetisation; cf. Section 3.1), as evidenced from AC magnetic susceptibility measurements performed for dysprosium(III) complexes Dy₂[(PentO)₈Pc]₂[(p-XPh)₃Cor], X = F or Cl.¹⁶⁴

The presence of three protons in corroles suggests that the closed-shell form of triple-decker complexes should be anionic in nature, which makes them different from trisphthalocyaninates, where the charge of two trivalent REE ions is balanced by the charge of three dianionic ligands. Thus, to stabilize the negative charge of the (M³⁺)₂(Cor³⁻)(Pc²⁻)₂ species a counterion is needed. NMR studies of the europium(III) complex suggested that the charge is balanced by a proton, which is likely to be delocalized over the isoindoline nitrogen atoms of one of the phthalocyanine ligands.¹⁶³ Air oxidation of the protonated forms with the formation of paramagnetic complexes with delocalized π-electron was observed, which was followed by disappearance of NMR spectra and appearance of intervalence bands in UV-Vis-NIR spectra at ca. 2000 nm.¹⁶⁷

Rich electrochemistry of studied complexes was demonstrated, and up to five oxidized states and three reduced states were observed rendering them as potential molecular materials for information storage devices.^{163,166–168} The potentials of redox transitions could be finely tuned both by variation of metal centres and substituents in the para-positions of meso-phenyl groups, due to linear correlations between the sums of Hammett constants σ of these substituents and electrochemical characteristics of complexes.¹⁶⁸

Comparative studies of self-assembling of two europium(III) complexes Eu₂(R₈Pc)₂[(p-ClPh)₃Cor], where R = H or octyl revealed that substituents in Pc ligands affected the morphology of aggregates formed by the phase transfer method upon mixing of solutions of complexes in chloroform with excess of methanol.¹⁶⁵ TEM and SEM studies of the resulting aggregates showed that the complex with unsubstituted Pc ligands forms nanobelts of 1.5–2.0 μm length and 100 nm width, while the octyloxy-substituted complex forms large scale sheet-like nanostructures with an average length of ca. 500 nm and a width of 300 nm. NLO studies of these complexes showed significant nonlinear reverse saturation absorption and self-defocusing behaviour.

Altogether, these results suggest that these corrole-containing mixed-ligand complexes can be considered as promising candidates for future nano-electronic and optical device applications.

2.4.5 Sandwich complexes with N-confused porphyrin ligands and heteroanalogues of porphyrins. Among core-modified derivatives of porphyrins special attention is paid to N-confused (or inverted, iPor) analogues with endocyclic CH and exocyclic NH groups. They have been known since the seminal work of Furuta et al. reported in 1994,¹⁶⁹ and these macrocycles found application in the construction of multinuclear assemblies due to their versatile binding modes.¹⁷⁰

First REE sandwiches with N-confused porphyrins were synthesized in 2012 by Cao et al. on the examples of mixed ligand double-deckers of iPor-Pc type using the typical raise-by-one-story approach with monophthalocyaninate and metal-free N-confused porphyrin (Scheme 31a).¹⁷¹ For the comparison of spectral properties, N-methylated iPor complex was also synthesized.

Scheme 31 (a and b) Synthesis of heteroligand double-decker complexes with N-confused ligands. (c) X-ray structures of [(tBuPh)₄iPorH]Lu(Pc) (CCDC YEFZIU), [(tBuPh)₄iPorH]Eu(Pc)·MeOH (CCDC YEFZAM)¹⁷¹ and [(p-ClPh)₄iPorH]Pr[(tBuPh)₄Por] (CCDC BODFIL).¹⁷² Solvate molecules and most of hydrogen atoms are not shown for clarity.

In 2014 cross-condensation of common and N-confused porphyrins with La(III) and Pr(III) acetylacetonates was reported, leading to heteroleptic double-decker complexes of Por-iPor type (Scheme 31b).¹⁷² Homoleptic bisporphyrinates were isolated as side products, while formation of homoleptic complexes with N-confused ligands was not observed. Interestingly, heteroleptic complexes exhibited NIR band at ca. 850 nm, which was absent in homoleptic bisporphyrinates.

Another type of porphyrin core modification, namely replacement of one of pyrrolic nitrogen atoms with oxygen or sulphur was used for spectacular tuning of magnetic properties of mixed ligand double-deckers. This was demonstrated on the examples of dysprosium(III) complexes with phthalocyanine and oxa- or thiaporphyrin ligands [(tBuPh)₄XPor]Dy(Pc), X = O or S (Scheme 32).¹⁷³ It was demonstrated that although Dy–X bonds are longer than Dy–N bonds, the symmetries of coordination polyhedra did not change upon going from classical porphyrin to its oxa- and thia-analogues. However, the replacement of one of heteroatoms significantly enhanced the anisotropy of the Dy(III) ion providing the complex with one of the largest barriers ever reported for dysprosium(III) complexes with tetrapyrrolic ligands (cf. Section 3.2.2).

Scheme 32 Synthesis of mixed ligand double-decker complexes with oxa- and thiaporphyrin ligands and X-ray structure of [(tBuPh)₄SPor]Dy(Pc) (CCDC AHOPUK).¹⁷³ Hydrogen atoms and solvent molecules are not shown for clarity.

2.5 Complexes based on half-sandwiches with axially coordinated non-tetrapyrrolic ligands

Due to large coordination numbers of REE ions, their half-sandwich complexes tend to have labile coordination spheres. This phenomenon is at the heart of the raise-by-one-story approach leading to the aforementioned heteroleptic and mixed ligand sandwiches, but it is possible to combine ligands of completely different natures in a molecule as part of this approach.

Thus, the present section will include sandwich REE complexes, where porphyrin or phthalocyanine ligands coexist with ligands of non-tetrapyrrolic nature. Although they only resemble sandwich structures of archetypal double- and triple-decker complexes, we included them in this section due to the increasing interest in such type of objects providing pathways to novel fluorescent materials, optical limiters, molecular magnets, etc. Finally, some of the structures are curious by themselves, broadening the rich coordination chemistry of both REE and tetrapyrrolic macrocycles.

2.5.1 Sandwich complexes with Schiff bases axially coordinated to REE monophthalocyaninates. Starting from 2013 the scope of sandwich-type tetrapyrrole-based rare-earth molecular materials was extended by monophthalocyaninates with axially-coordinated Schiff base (SB) ligands. The summary of the reaction studied on the example of the ligand, based on o-phenylenediamine and methoxy-substituted salicylic aldehyde, H₂(SB) is given in Scheme 33.

Scheme 33 Synthesis of mixed-ligand complexes based on half-sandwich phthalocyaninates and Schiff base H₂(SB) with X-ray structures of the corresponding complexes – from the left to the right: (Pc)₂Dy₂(SB)₂Ca(MeOH)₂ (CCDC FEVDOB),¹⁷⁴(Pc)Dy(SB)K(MeOH) (CCDC FEVDUH),¹⁷⁴ (Pc)Tb(SB)Zn(Py)₂NO₃ (CCDC YIMMUE)¹⁷⁵ and (Pc)₂Dy₂(CB)(H₂O) (CCDC XILTOD).¹⁷⁶

First examples of such complexes were prepared by Wang et al. from (Pc)M(acac), M = Dy and Y and H₂(SB).¹⁷⁴ Depending on the base used for deprotonation of H₂(SB) complexes of different nuclearities were obtained. Performing this reaction with Ca(OH)₂ afforded a quadruple-decker complex where the [Ca(MeOH)₂]²⁺ group bridged two (Pc)M(SB)⁻ units (Scheme 33a), while dinuclear double-deckers (Pc)M[(SB)K(MeOH)] was isolated in the presence of KOH (Scheme 33b). The latter complex dissociated with the formation of double-decker species upon treatment with K₂CO₃ (Scheme 33c). Fast QTM relaxation was observed in zero-field AC magnetic susceptibility measurements for dysprosium(III) complexes of both types, although long-distance magnetic interactions between two paramagnetic centres in quadruple-decker complex was detected (cf. Section 3.2.2).

Later, Ma et al. synthesized terbium(III) complexes with all alkali metal cations [(SB)M(MeOH)]Tb(Pc), M = Li⁺ to Cs⁺ and the possibility to exchange these cations with the Zn²⁺ ion was demonstrated (Scheme 33d), providing an interesting example of polyfunctional complex with tuneable properties.¹⁷⁵

Slow diffusion of the acetonitrile solution of H₂(SB) into the solution of (Pc)M(acac), M = Gd and Dy, in chloroform afforded another type of triple-decker complex with bridging (SB)²⁻ unit and chemically non-equivalent REE ions – M₂(Pc)₂(SB)(H₂O) (Scheme 33e).¹⁷⁶ One of metal centres was in octahedral surrounding and another one was hepta-coordinated by four isoindole nitrogen atoms of (Pc)²⁻, two oxygen atoms of a Schiff-base ligand, and one oxygen atom of water molecule. Homoleptic binuclear dysprosium(III) complex with the H₂(SB) ligand had the same binding mode; however, incorporation of the phthalocyanine ligand had a profound effect on SMM properties of the mixed-ligand sandwich in comparison with Dy₂(SB)₃(H₂O) in terms of ferromagnetic coupling between paramagnetic ions and energy barrier, U_eff.

Analogous binuclear complexes were synthesized by Li et al. based on a family of chiral ligand prepared by the condensation of 1,2-diphenylethylenediamine and 5-chlorosalicylic aldehyde. Photophysical characterization of such complexes showed their potential in future nonlinear optical applications such as wavelength conversion and optical switching.^177,178

By varying the ratio of half-sandwich phthalocyaninate and Schiff base, complexes with different stoichiometries of ligands can be obtained. This was shown by Gao et al. on the example of the reaction between one equivalent of dysprosium(III) phthalocyaninate and two equivalents of tetrathiafulvalene-fused Schiff base H₂(SB-TTF) yielding a binuclear complex Dy₂(Pc)(SB-TTF)₂(MeOH) (Scheme 34a).¹⁷⁹ Due to the intrinsic dipole moment, this complex could be absorbed on the surface of HOPG either in Pc-up or S-TTF-up fashion depending on the applied potential, which was demonstrated by scanning tunneling microscopy at the single molecule level, rendering this compound as a component of functional materials fabricated by self-assembly.

Scheme 34 (a) Synthesis of a mixed-ligand complex with tetrathiafulvalene-fused Schiff base with the X-ray structure of complex Dy₂(Pc)(SB-TTF)₂(MeOH) (CCDC WONKIV).¹⁷⁹ (b) Synthesis of sandwiches based on closed-macrocyclic Schiff base and REE monophthalocyaninates with the X-ray structure of complex Dy₃(Pc)₂(mSB)(OMe)₂ (CCDC IKAQAO, only one of two crystallographically independent molecules is shown).¹⁸⁰ Hydrogen atoms are omitted for clarity.

Gao et al. used dysprosium(III) or erbium(III) monophthalocyaninates generated in situ as templates for condensation of 2,6-diformyl-4-methylphenol and 1,3-diaminopropane to form unusual trinuclear triple-decker sandwiches with a macrocyclic Schiff ligand and bridging acetate and methoxide counterions M₃(Pc)₂(mS)(OAc)(OMe)₂ (Scheme 34b).¹⁸⁰ Interestingly, in the crystal lattice these complexes existed as pairs of symmetrically unrelated molecules with slightly different symmetries of coordination polyhedra. This feature resulted in multiple magnetic relaxation modes found by AC magnetic susceptibility measurements of the dysprosium(III) complex. Pronounced intramolecular interactions between proximal metal centres was observed – ferromagnetic in the case of dysprosium and antiferromagnetic in the case of erbium (cf. Section 3.2.2).

2.5.2 Monoporphyrinato- and monophthalocyaninato-complexes with axially coordinated cyclic ligands. Sandwiches formed by heavy lanthanide monoporphyrinates¹⁸¹ and monophthalocyaninates¹⁸² with cyclen ligands coordinated to metal centres were reported by Santria and Kizaki et al. These complexes with well-defined coordination spheres of metal ions were synthesized by simple reactions of (TPP)MCl or (Pc)MCl, M = Tb–Yb and Y with excess of cyclen in methanol or THF, respectively (Scheme 35), and single crystals of all complexes were characterized by XRD to show that outer-sphere chloride anions do not bind to metal ions, thus, they adapt a tetragonal SAP symmetry of coordination polyhedra formed solely by pairs of N₄ ligands. The smooth changes of structural parameters of both families of complexes depending on the ionic radii of M³⁺ were demonstrated.

Scheme 35 Synthesis of REE porphyrinates and phthalocyaninates with axially-coordinated cyclen ligands with X-ray structures of [(TPP)Tb(Cyclen)]⁺Cl⁻ (CCDC LEBMUD) and [(Pc)Tb(Cyclen)]⁺Cl⁻ (CCDC WENBID).^181,182 Hydrogen atoms are omitted for clarity.

Ground multiplet states of a series of porphyrin complexes [(TPP)M(Cyclen)]⁺Cl⁻ were determined by NMR and magnetic susceptibility data together with CASSCF calculations, giving insight into their electronic structures.¹⁸³ Intramolecular magnetic interactions between localized 4f-electronic systems and photoexcited Por or Pc π-systems in terbium(III) Por/cyclen and Pc/cyclen complexes was studied by variable-temperature variable-field magnetic circular dichroism spectroscopy providing experimental data to validate the theoretical models describing such interactions.^184,185

Interesting sandwich complexes with η⁵-Cp and η⁵-Me₅Cp ligands sitting-atop scandium(III) phthalocyaninate were synthesized from (Pc)ScCl by Platel et al. (Scheme 36).¹⁸⁶ As a first step of this work, template condensation of phthalonitrile in the presence of ScCl₃ in ClN was performed, yielding inseparable mixture of target (Pc)ScCl with ca. 15% of unreacted ScCl₃. Treatment of this mixture with LiCH(TMS)₂ in THF afforded the formation of a soluble binuclear complex (Pc)ScCl₂Li(THF)₂ which could be crystallized as an individual compound. Its reaction with NaCp or LiCp* in THF provided (Pc)Sc(Cp) and (Pc)Sc(Cp*), respectively, as first representatives of metallophthalocyanine-ocenes.

Scheme 36 Synthesis of scandium phthalocyaninates – (Pc)ScCl₂Li(THF)₂ (CCDC DUDTED) and (Pc)Sc(Cp) (CCDC DUDTIH).¹⁸⁶ Hydrogen atoms are omitted for clarity.

Scheme 37 Synthesis of complexes with hemiporphyrazine ligand^188,189 and X-ray structures of (HHp)Eu(Pc) (CCDC GAWCAL) and (HHp)Eu(Hp) (CCDC GAVXUZ).¹⁸⁸ Hydrogen atoms are omitted for clarity.

Another class of macrocyclic ligands which was recently introduced into sandwich REE chemistry is hemiporphyrazin, Hp. Synthesis of this 20π-electron antiaromatic relative of azaporphyrins was reported for the first time in 1952 by Elvidge and Linstead.¹⁸⁷ However, REE complexes with Hp were reported for the first time only in 2017.¹⁸⁸ The complexes (HHp)M(Pc), M = Eu, Lu, were obtained by Liu et al. using reactions of the corresponding (Pc)M(acac) with H₂(Hp) and DBU in refluxing pentanol. Metalation of hemiporpyrazine with M(acac)₃·nH₂O under the same conditions yielded double-decker complex (HHp)M(Hp). ESR and NMR studies of the resulting sandwiches evidenced that they were diamagnetic in contrast to most bisphthalocyaninates existing as free radicals. The proton needed for such state was localized by NMR spectroscopy, XRD and quantum-chemical calculations – all methods evidenced that the proton is bound to the pyridine nitrogen atom. Quantum-chemical calculations of shielding surfaces of both types of complexes revealed that in homoleptic complexes hemiporphyrazine ligands retained their antiaromatic nature; however, combination of antiaromatic Hp and aromatic Pc species provided molecules with overall aromatic behaviour.

XRD characterization of both types of Hp complexes revealed nearly ideal SAP surrounding of metal centres, although somewhat distorted because of the protonation of one of the pyridine groups. Nevertheless, measurements of magnetic properties of (HHp)Dy(Pc) and (HHp)Dy(Hp) revealed excellent SMM properties (cf. Section 3.2.2).

2.5.3 Heteronuclear complexes with metalloligands. Half-sandwich porphyrinates and phthalocyaninates were used to form heteronuclear complexes with metalloligands. For example, in 2013 Gao et al. reported solvothermal one-pot reactions of metal-free H₂(TPP) or H₂(Pc) with Gd(III), Tb(III), Dy(III) and Ho(III) acetylacetonates, and sodium salt of Kläui's, Na[(η⁵-Cp)Co{P(=O)(OEt)₂}₃].¹⁹⁰ These reactions yielded seven-coordinate complexes with axial tripodal ligands (Scheme 38a and b). Tb(III) and Dy(III) complexes showed field-induced slow relaxation of magnetization; these complexes were convenient models demonstrating the correlation between the magnetic relaxation properties and subtle distortions of the coordination geometry of the paramagnetic lanthanide ions (cf. Section 3.2.2).

Scheme 38 Synthesis of heteronuclear complexes based on REE half-sandwiches. (a and b) Synthesis of Dy(III) porphyrinate (CCDC LIMDOC) and phthalocyaninate (CCDC LIMCUH) with a Kläui ligand. (c) Synthesis of Tb(III) phthalocyaninate bearing a polyoxotungstate ligand (Pc)M(PW₁₁O₃₉)⁶⁻(NBu₄⁺)₆ (CCDC GUZFOZ), and the X-ray structure of anionic part of the complex (Pc)Tb[PW₁₁O₃₉]⁶⁻. Solvate molecules, hydrogen atoms and counterions are omitted for clarity.

Later, half-sandwich Yb(III) with porphyrin derivatives and axial Kläui's ligands attracted great interest as highly near-IR emissive complexes with unprecedented quantum yields up to 63%,¹⁹¹ which contributes to the design of molecular probes for near-IR imaging.¹⁹²

In 2020 Sarwar et al. reported the synthesis of Tb(III), Dy(III) and Y(III) phthalocyaninates with a monolacunary α-Keggin polyoxotungstate ligand [P^VW^VI₁₁O₃₉]⁷⁻.¹⁹³ With this aim, the corresponding monophthalocyaninates (Pc)M(OAc) reacted with (NBu₄)₄H₃[PW₁₁O₃₉] in the presence of triethylamine and NBu₄Br under mild conditions, yielding anionic complexes with tetrabutylammonium counterions (Pc)M(PW₁₁O₃₉)⁶⁻(NBu₄⁺)₆ (Scheme 38c). Their characterisation included high resolution mass spectrometry, synchrotron-based single-crystal X-ray diffraction and magnetic studies revealing slow relaxation of the magnetisation for the Dy(III) derivate (cf. Section 3.2.2).

2.5.4 Inverted sandwich complexes. Final class of complexes in the following section – inverted sandwiches do not refer to sandwich complexes in the common sense, since they contain only one phthalocyanine ligand as a bridge between two M(dpm)₂⁺ groups, where dpm⁻ – dipivaloylmethanate anion. The term “inverted” refers to the fact that one macrocycle is sandwiched between two REE ions, as opposed to the conventional sandwiching of metal ions between two macrocycles.

Thus, these complexes were reported for the first time in 1983 in a short communication by Sugimoto et al., describing the reaction between Li₂(Pc) and M(dpm)₃ in refluxing THF under anaerobic conditions.¹⁹⁴ This reaction afforded a series of (μ-Pc)[M(dpm)₂]₂, where M = Sm to Yb and Y. The X-ray structure of the samarium(III) complex with benzene solvate molecules was determined (Fig. 8), showing the extremely rare representative of binuclear complexes formed by a single tetrapyrrolic macrocycle. A bit later magnetic properties of these unique complexes were studied by Maeda et al., showing the first evidence of intramolecular spin–spin exchange interaction between trivalent REE ions.^195,196 However, for many years that followed, these complexes attracted no attention.

Fig. 8 X-ray structure of (μ-Pc)[Sm(dpm)₂]₂ (CCDC CALWIV).¹⁹⁴ Solvate molecules and hydrogen atoms are not shown.

Only in 2014 detailed characterization of magnetic properties of (μ-Pc)[M(dpm)₂]₂, M = Tb and Dy, was performed by Tsuchiya et al., showing their striking difference from common binuclear triple-deckers.¹⁹⁷ Unexpectedly, antiferromagnetic interaction between metal centres was found in contrast to ferromagnetic coupling in the corresponding trisphthalocyaninates. NMR studies showed that aromatic proton resonances were shifted down-field – again, it was in sharp contrast with M₂(Pc)₃ with large up-field lanthanide-induced shifts. These intriguing results were explained assuming the ligand field potential of the inverted-sandwich complexes gave strong easy-plain-type magnetic anisotropy, where the J_z value of the lowest sublevel has a small value, while trisphthalocyaninates have a strong easy-axis-type magnetic anisotropy. The benzene solvates (μ-Pc)[M(dpm)₂]₂·2C₆H₆ showed interesting behaviour with respect to desolvation and exchange of solvation molecules to dichloromethane, affording the pathway for fine tuning of magnetic properties (cf. Section 3.6).^198–200

Interesting properties of inverted sandwiches could encourage synthetic studies to broaden this class of REE complexes; however, to date, only dpm ligands could stabilize their unusual architecture. To the best of our knowledge, the sole reported attempt to exchange these anions was made by Ge et al. when the complex (μ-Pc)[Dy(dpm)₂]₂ was treated with 3-methoxy- or 3-ethoxysalicyllic aldehydes H(LR), R = OMe, OEt.²⁰¹ However, it resulted in the deterioration of the inverted sandwich scaffold with the formation of a new type of triple-deckers (μ-LR)₂Dy₂(Pc)₂(H₂O) where two (DyPc)⁺ units were bridged by a pair of coplanar deprotonated LR⁻ ligands (Scheme 39a). Although at the first glance both complexes had similar structures invariant to substituents in bridging ligands, their magnetic properties were markedly different. Magnetic studies indicated that the MeO-substituted complex behaved as a zero-field single-molecule magnet with a higher energy barrier, while EtO-substituted complex exhibited a fast tunneling relaxation process (cf. Section 3.2.2). Detailed analysis of the XRD data revealed that one of MeO-groups coordinated to dysprosium(III) ion, while oxygen atoms of both ethoxy-groups remained in the non-coordination mode (Scheme 39b). This difference in the substitution of ligands in the middle layer leads to perturbation of the main magnetic axes, consequently changing the coupling interaction and modulating the magnetic behaviour for this Dy2 system.

Scheme 39 (a) Transformation of inverted sandwich (μ-Pc)[Dy(dpm)₂]₂ into a triple-decker mixed ligand complex (μ-LR)₂Dy₂(Pc)₂(H₂O) upon interaction with 3-alkoxysalicylic aldehydes, H(LR), R = OMe, OEr. (b) Structure of the middle layers formed by deprotonated ligands LR⁻, coordinated to dysprosium ions (CCDC MESMIJ, R = Me and MESMOP, R = Et).²⁰¹

2.6 Sandwich complexes that go beyond triple-deckers

Simple electrostatic considerations suggest that triple-deckers formed by doubly-deprotonated porphyrins and phthalocyanines with a pair of trivalent REE cations are the largest possible electroneutral species. However, there is interest in larger sandwiches, which can fill in the gap between discrete double-/triple-decker complexes and infinite lattices formed by these complexes in solid functional materials. The following section will describe some approaches which were developed over the last decade to obtain discrete multideckers, avoiding electrostatic limitations.

2.6.1 Discrete sandwiches containing from four to six decks bridged through REE(III) and Cd(II) metal centres. From the electrostatic viewpoint electroneutral trisphthalocyaninates can be considered as complexes formed by anionic double-deckers [M(Pc)₂]⁻ and cationic half-sandwiches. Thus, binding of two [M(Pc)₂]⁻ species with an appropriate divalent cation could be used as a pathway to a quadruple-decker complex. In 2010 this idea was realized by Fukuda et al. on the example of the reaction between [Lu(Pc)₂]⁻(NBu)₄ and cadmium acetate (Scheme 40a).²⁰² Heating of solid reagents at 400 °C followed by chromatographic separation afforded 3.4% of a quadruple-decker complex (Pc)Lu(Pc)Cd(Pc)Lu(Pc), which was characterized by elemental analysis, mass-spectrometry and NMR spectroscopy. The complex underwent eight reversible one-electron redox-processes – four oxidation reactions and four reduction reactions, while the parent lutetium bisphthalocyaninate showed only two oxidation reactions and three reduction reactions in the same potential range; moreover, in this case the gap between first reduction and oxidation potentials was wider than in the case of a quadruple-decker complex, revealing the stacking effect which stabilized the redox forms of the extended complex. Spectroelectrochemical studies also revealed exceptionally low-lying π–π* excited states of the oxidized form of a quadruple-decker complex – the corresponding absorbance bands appeared at ca. 1000 cm⁻¹, where normally only finger-print vibration bands are observed in IR spectra.²⁰³ These experimental observations were interpreted using molecular orbital calculations.²⁰⁴

Scheme 40 (a–c) Synthesis of discrete quadruple-decker complexes Lu₂Cd(Pc)₄,²⁰²Dy₂Cd[(BuO)₈Pc]₄,²⁰⁵ and Eu₂Cd[(HexS)₈Pc]₄²⁰⁶via connection of two double-decker molecules with Cd(II) ions. (c) Structure of complex Dy₂Cd[(BuO)₈Pc]₄ (CCDC IYOCIJ),²⁰⁵ the structure of only one of two crystallographically independent molecules is shown, hydrogen atoms and solvent molecules are omitted.

Both stability and yields of quadriphthalocyaninates could be improved when electron-rich BuO-substituted ligands were used instead of unsubstituted phthalocyaninates. Thus, Wang et al. showed that the reaction of neutral Dy[(BuO)₈Pc]₂ with Cd(OAc)₂·2H₂O in refluxing TClB afforded target substituted quadruple-decker complex Dy₂Cd[(BuO)₈Pc]₄ in yield as high as 43% (Scheme 40b).²⁰⁵ In a similar manner, Eu(III) complex with n-hexylthio-substituents Eu₂Cd[(HexS)₈Pc]₄ was synthesized starting from the corresponding double-decker in 84% yield (Scheme 40c).²⁰⁶ Both the OFET and gas sensing properties were studied on the self-assembled film of the latter complex (cf. Section 4).

Single crystals of the Dy(III) complex were obtained by Wang et al., and XRD provided a first ultimate proof of a quadruple-decker structure (Scheme 40d).²⁰⁵ Separation between Dy(III) ions – ca. 6.65 Å was expectedly much larger than in triple-decker complexes; however, it was still enough to detect notable intramolecular interaction between paramagnetic metal centres, which was responsible for the fast quantum tunnelling of magnetization in a zero DC field (cf. Section 3.2.4). Interestingly, introduction of alkoxy and alkylthio groups provided complexes with different conformations. In the case of the BuO-substituted complex both Dy(III) and Cd(II) metal centres were in distorted prismatic surrounding with a twist angle of ca. 20°,²⁰⁵ while in HexS-substituted complex Eu(III) and Cd(II) had lightly deviated SAP coordination geometries with twist angles of ca. 43 and 39°.²⁰⁶

The capability of cadmium bridge to connect pairs of different double-deckers was tested by Wang et al. on the example of homoleptic bisphthalocyaninates M(Pc)₂ and mixed-ligand complexes (TClPP)M(Pc), M = Eu and Y.²⁰⁷ In principle, this reaction could lead to isomeric quadruple-decker complexes with different sequences of tetrapyrrols – Pc–Pc–Pc–Por or Pc–Pc–Por–Pc, but both NMR and X-ray crystallography showed that only one type of complex with terminal porphyrin ligand was formed. The complexes were isolated in 9–14% yields. Similarly to homoleptic quadriphthalocyaninates, the synthesized mixed-ligand complex had rich electrochemistry associated with four ligand-centred oxidation and three reduction processes, whose potentials showed dependence on the metal centre.

No surprise, extended sandwiches became attractive objects to study long-distance interactions between paramagnetic metal ions in the context of elaboration of SMM materials and understand mechanisms or magnetization relaxation. Thus, the approach to quadriple-deckers was extended to even higher sandwiches. Wang et al. showed that reactions of M[(BuO)₈Pc]₂ with one or two equivalents of H₂[(BuO)₈Pc] and Cd(OAc)₂·2H₂O allowed the introduction of one or two additional phthalocyanine layers between REE-containing moieties, affording therefore quintuple- and sextuple-decker sandwiches (Scheme 41).^208,209 Moreover, this strategy allowed the synthesis of heteroleptic quintuple-decker sandwiches with precise sequences of unsubstituted and octa-pentyloxy-substituted ligands – either Pc–Pc–Pc*–Pc–Pc or Pc–Pc*–Pc*–Pc*–Pc.²¹⁰ Single-crystal XRD studies of complexes evidenced of smooth elongation of distances between REE ions going from triple- to sextuple-deckers, providing more objects for studies of long-range magnetic interactions (cf. Section 3.2.4).^211–214

Scheme 41 Synthesis of discrete quintuple-decker complexes M₂Cd₂(Pc)₅²⁰⁸ and sextuple-decker complexes M₂Cd₃[(BuO)₈Pc]₆.²⁰⁹ Butoxy-groups in resulting multi-deckers are not shown.

In principle, this approach can be used to obtain multi-decker complexes with any number of tetrapyrrolic ligands, since further sequential addition of “CdPc” fragments does not violate the electroneutrality requirement in the resulting extended sandwiches. However, the problems with isolation of required multi-decker from the mixture of complexes are highly likely to arise.

2.6.2 Sandwich complexes based on fused dimeric tetrapyrrols. Synthesis of binuclear fused phthalocyanines [f-Pc₂] sharing common benzene rings was reported for the first time by Kobayashi et al. in 1994 via the cross-condensation of diiminoisoindolines derived from phthalonitriles and pyromellitic tetranitriles.²¹⁵ Almost 20 years later Wang et al. recognized f-Pc ligand as a convenient scaffold to form biradical sandwich complexes which can be used to clarify the role of radicals in 2p–4f integrated systems and study the magnetic properties of polynuclear systems (cf. Section 3.3.2).²¹⁶

Thus, in 2013 Wang et al. showed that Y(III)- or Dy(III)-templated cross-condensation of bis(diiminoisoindoline) with 4,5-dibutoxyphthalonitrile in refluxing pentanol in the presence of DBU furnished target fused bisphthalocyaninates (Pc)M[f-Pc₂]M(Pc)^OBu in 2.0–4.3% yields (Scheme 42a).²¹⁶ In the same manner Tb(III) complex was synthesized by Morita et al. in 2.5% yield using hexanol as a solvent for template condensation.²¹⁷ Using 4,5-di(ethylthio)phthalonitrile in this procedure allowed Wei et al. to increase the yield of the fused bisphthalocyaninate (Pc)M[f-Pc₂]M(Pc)^SEt to 24.8% (Scheme 42b).²¹⁸

Scheme 42 Synthesis of fused sandwich phthalocyaninates. (a and b) Binuclear complexes (Pc)M[f-Pc₂]M(Pc)^XR, XR = OBu or SEt. (c) Tetranuclear complexes (Pc)₂M₂[f-Pc₂]M₂(Pc)₂^OBu. (d) Tri- and tetranuclear complexes (Pc)₂M₂[f-Pc₂]M(Pc)^SEt and (Pc)₂M₂[f-Pc₂]M₂(Pc)₂^SEt. Superscripts ^OBu and ^SEt indicate that all accessible peripheral positions of the benzene rings contain the corresponding substituents.

Fused complexes with higher nuclearity could be synthesized using variants of the raise-by-one-story approach. Thus, Motita and Katoh et al. treated metal-free BuO-substituted fused phthalocyanine H₄[f-Pc₂]^OBu with double-deckers and Tb(III) or Dy(III) acetylacetonates in TClB to synthesize tetranuclear Tb(III)²¹⁹ and Dy(III)²²⁰ complexes (Pc)₂M₂[f-Pc₂]M₂(Pc)₂^OBu (Scheme 42c). Alternatively, Wei et al. treated SEt-substituted fused Eu(III) bisphthalocyaninate with H₂[(SEt)₈Pc] and Eu(acac)₃ to synthesize tri- and tetranuclear complexes (Pc)M[f-Pc₂]M₂(Pc)₂^SEt and (Pc)₂M₂[f-Pc₂]M₂(Pc)₂^SEt which could be separated by chromatography (Scheme 42d).²¹⁸ In all cases the yields of isolated complexes were relatively small, yet, it was enough to characterize their properties and demonstrate their applicability.

Fused bi-, tri- and tetranuclear substituted complexes were used as redox-active components in sensors to NO₂ and NH₃ revealing excellent sensitivity to these gasses.^218,221 The film formed by the tetranuclear SEt-substituted complex was one of the best gas sensors obtained so far for all solution-processed organic semiconductor films in dry air at room temperature.²¹⁸ Binuclear SEt-substituted complex was also used as a component of an electrochemical sensor for simultaneous detection of acetaminophen and ascorbic acid.²²² SHex-substituted tetranuclear complex was found to form well-defined single crystal microsheets upon self-assembling in solution, which were used to fabricate OFET with high hole and electron mobilities of 18 and 0.3 cm² V⁻¹ s⁻¹, respectively, representing the best result for phthalocyanine-based OFETs at that time (cf. Section 4).²²³

A unique example of fused bisporphyrinate synthesized from β-β/meso–meso/β-β triply-linked diporphyrin H₄[f-Por₂] was reported by Lee et al. in 2019.²²⁴ Starting ligand was treated with LiHMDS to generate a tetralithium intermediate which reacted with half-sandwiches (TPP)M(acac) to furnish fused sandwiches (TPP)M[f-Por₂]M(TPP), M = Tb and Y in small yields (Scheme 43). Complexes were isolated in neutral states exhibiting an open shell biradical character, as determined from magnetic susceptibility measurements and theoretical calculations. The Tb(III) complex exhibited rich electrochemistry, having more than eight clear peaks in the voltammetry curves; moreover, it has a very small HOMO–LUMO energy gap as evidenced from optical (3700 nm, 0.33 eV) and electrochemical measurements (3400 nm, 0.36 eV). These features together with high stability render these biradicals as promising candidates for use in molecular electronics, optics, and spintronics.

Scheme 43 Synthesis of fused sandwich bisporphyrinates (TPP)M[f-Por₂]M(TPP).²²⁴

3. Development of the functionalities of sandwich tetrapyrrole rare-earth single-molecule magnets

In this section, we will provide a comprehensive survey of several types of sandwich tetrapyrrole rare-earth (RE) single-molecule magnets (SMMs) that use a supramolecular approach, combined with a discussion of their functionalities. Importantly, the properties of these molecules are changed significantly by the formation of associations among such molecules. The chemical and physical properties of molecular assemblies composed of a finite number of molecules are also affected by various intermolecular interactions. Currently, researchers are also conducting studies on applying self-assembly strategies that actively apply coordination bonds to molecular magnetic materials. In 2002, Wernsdorfer et al. reported that the introduction of hydrogen bonds and van der Waals forces into a tetranuclear Mn complex SMM (abbreviated as [Mn₄O₃Cl₄(O₂CEt)₃(py)₃] = Mn₄) resulted in dimers of the type Mn₄⋯Mn₄, which exhibit superexchange interactions and suppress the quantum tunnelling of magnetisation (QTM; cf. Section 3.1).²²⁵ This phenomenon was achieved via the superexchange interaction between the SMMs, which act as an exchange bias (H_bias ≈ ±M₂J/gμ_B) on the QTM that occurs at a zero magnetic field, resulting in a spin-sublevel energy mismatch. Thus, the magnetic interaction acts as a perturbation between the SMMs that affects the QTM that occurs at a near-zero magnetic field. In other words, it is also possible to control the properties of SMMs using intra/intermolecular magnetic interactions between SMMs. Moreover, to control the magnetic interactions between SMMs, it is possible to use not only hydrogen bonds but also π–π interactions, CH–π interactions, and inclusion (host–guest) interactions associated with ion and molecular recognition. Controlling the spin relaxation time (τ) by utilizing the interactions between the infinite number of intramolecular vibrational modes of SMMs and spin excitation is essential for spintronics. SMMs (molecular crystal) have weak intermolecular interaction and low phonon energy in the acoustic mode. Therefore, the spin-acoustic mode interaction is restricted. On the other hand, by introducing a strong intermolecular interaction into the SMMs, the phonon energy in the acoustic mode can be increased, and interaction with the spin can be expected. In other words, from the viewpoint of controlling magnetic relaxation, the approach of introducing a molecular assembly structure into SMMs is one method (vide infra).

Furthermore, ligand modifications, protonation–deprotonation reactions, and redox reactions also cause dramatic changes in SMM characteristics by inducing substantial changes in their coordination environments. Sandwich tetrapyrrole rare-earth-type SMMs have attracted attention for potential use in the fields of spintronic devices and/or quantum technology (QT). However, the performance of these devices depends on the nature of the ground-state electronic structures of the SMM units. By using supramolecular approaches to prepare sandwich tetrapyrrole rare-earth-type SMMs, their ground-state electronic structures and functionality can potentially be controlled (Fig. 9).

Fig. 9 Brief overview of rare-earth sandwich tetrapyrrole SMMs. Double-decker complexes with various kinds of porphyrinoids act as SMMs. The magnetic properties of the double-decker SMMs can be reversibly tuned by redox reactions and protonation–deprotonation reactions. Various approaches for connecting/arranging SMMs have been used to introduce Ln–Ln interactions and tune the coordination environment.

3.1 Brief outline of single-molecule magnets

In 1993, Sessoli, Gatteschi, and co-workers reported that the cluster [Mn₁₂O₁₂(AcO)₁₆(H₂O)₄]·2AcOH with S = 10 (henceforth abbreviated as Mn₁₂; Fig. 10a), which contains twelve Mn ions (eight Mn³⁺ with S = 2 and four Mn⁴⁺ with S = 3/2), exhibited superparamagnetic properties; they referred to Mn₁₂ as an SMM because it exhibited both single molecule properties and ferromagnetic properties (Fig. 10c).^226,227 However, REE-type SMMs are fundamentally different from transition-metal (TM) SMMs, because their SMM behaviour originates from the f electrons (Fig. 10b). Since the orbital angular momentum L does not disappear for f electrons, ground-state multiplets exhibit a total angular momentum J due to spin–orbit coupling (SOC). When the ground-state multiplet is placed in a ligand field (LF) environment, the degeneracy splits into the 2J + 1 sublevels (Fig. 10d). As a result, double-well potentials form, and SMM behaviour emerges (vide infra).^36,226 In general, the sublevel splitting in REE-type SMMs (10²–10³ cm⁻¹)⁵ is much larger than that of TM-type SMMs (ca. 70 cm⁻¹),²²⁸ indicating that REE-type SMMs are more useful in terms of their SMM characteristics, such as large magnetic anisotropy and U_eff (Fig. 10c and d).^229,230

Fig. 10 Molecular structures of (a) Mn₁₂ and (b) Tb(Pc)₂ based on the cif files from ref. 231 and 232. Schematic energy diagrams of (c) Mn₁₂ and (d) Tb(Pc)₂. In the case of Mn₁₂, eight Mn³⁺ and four Mn⁴⁺ ions with d-orbital splitting (LF splitting) couple with each other (exchange interactions) to form an S = 10 state. The L of Mn³⁺ and Mn⁴⁺ is quenched due to the large LF splitting. The zero-field splitting (D = –0.46 cm⁻¹) of S = 10, which mainly originates from second-order SOC, results in a U_eff of ca. 40 cm⁻¹ for spin reversal between m_S = +10 and m_S = −10. In the case of Tb(Pc)₂, S = 3 and L = 3 couple with each other (first-order SOC) to form ⁷F₆. LF splitting of the Tb³⁺ centre results in a U_eff of ca. 400 cm⁻¹ for spin reversal between m_J = +6 and m_J = −6. m_Jvs. energy plots were constructed from ref. 233.

In the last 20 years, SMMs with high operating temperature have been synthesized by controlling the LF and/or crystal field (CF) around the metal ion, even in single-ion magnets (SIMs).²³⁴ Since the Mn₁₂ cluster complex has been shown to exhibit the SMM phenomena, 3d to 5f metal complexes have been investigated as SMMs (henceforth, the term SMM will be used to refer to SIMs as well).^{228,230,235–244} SMMs have a fixed number of metal ions and an ordered magnetic ion structure, and exhibit uniaxial magnetic anisotropy due to the introduction of an LF and SOC.^{226,227,229,245} A brief description of magnetic anisotropy is given here. When a magnetic material is magnetised in various directions, there are directions in which it is easy to magnetise (easy axis of magnetisation) and difficult to magnetise (hard axis of magnetisation). When the direction is in one direction (uniaxial), it is called “uniaxial magnetic anisotropy”. This magnetic anisotropy is caused by the orientation of magnetic moments of the atoms which have the CF and SOC that make up the magnetic material. That is, the stable direction of orbital angular momentum in the crystal affects the direction of spin, which is the source of magnetism. As described later, the 4f orbital shape of rare earth metal ions has a very strong directivity, so it is possible to control the magnetic anisotropy by controlling the appropriate CF (cf. Section 3.2). A characteristic of Mn₁₂ is that the Jahn-Teller axis of Mn³⁺ ions with magnetic anisotropy is almost parallel to the S₄ axis, showing large uniaxial magnetic anisotropy as a single molecule.²²⁶ It is possible to control the magnetic properties of metal complexes, such as their electronic state, LF, and magnetic anisotropy, by changing the metal ions and coordinating ligands. In particular, it has been shown that the LF directly affects the magnetic properties. Classical magnets exhibit large magnetisation via three-dimensional interactions among the electron spins of the metal ions, but SMMs involve the uniaxial magnetic anisotropy of one molecule. It is thought that SMMs are also “single-domain magnets”, as intermolecular magnetic interactions are ignored.

Using the effective spin Hamiltonian, we discuss the uniaxial magnetic anisotropy and U_eff in SMMs. The concept of effective spin Hamiltonian was introduced by M. H. L. Pryce which has only the spin operator being considered. When incorporating the CF and SO into the effective spin Hamiltonian, it can be expressed by D (uniaxial crystal field constant) and E (biaxial crystal field constant). Here, the spin Hamiltonian is given by with the D. In the absence of an external DC magnetic field (H = 0; H and H_dc are used interchangeably), the Zeeman term goes to zero, so Dm_S² (m_S is the z-component of S) is the splitting energy at zero magnetic field (zero-field splitting, ZFS). For the reasons mentioned above, the splitting width of the ground and highest energy levels is defined as U_eff = |D|S² for TM-type SMMs, where the D value is the “zero-field splitting constant” and the S is the spin quantum number of the ground state, which mainly originates from second-order SOC (the U_eff of REE-type SMMs is explained in the next section). For SMMs, the degeneracy of the total spin state S_total is split into 2S + 1 by LF and SOC, a quantized double-well potential is formed by the uniaxial magnetic anisotropy, and the spin inversion between spin-up and spin-downstate caused by the spin sublevel shows Arrhenius-like relaxation (τ = τ₀ exp(U_eff/T), Fig. 10c). Here, τ represents the magnetic relaxation time, τ₀ is the frequency factor, U_eff is the energy barrier associated with spin inversion, and T is the temperature. At temperatures higher than U_eff associated with spin inversion, the SMM behaves as a paramagnetic material, but when the temperature decreases and the U_eff cannot be overcome, the magnetic moment (spin) is frozen. The temperature at which the τ reaches 100 s is defined as the blocking temperature (T_B). SMMs behave like a magnet which shows magnetic hysteresis with coercivity (H_C) below T_B due to long τ values. In addition, the field cooled (FC) and zero-field cooled (ZFC) measurements of the SMM diverge at T_B. The magnitude of magnetic anisotropy is possible to estimate from the M versus HT⁻¹ plot at different T values. If SMMs have large uniaxial magnetic anisotropies and/or the population of low-lying excited states, M versus HT⁻¹ shows non-superposed magnetisation curves.

The temperature change in the spin system reaches thermal equilibrium through heat exchange between the spin and the lattice, and finally moves to the heat bath. Therefore, the information for slow magnetic relaxation can be obtained from alternating current (AC) magnetic measurements. The measurement data are analyzed by using a generalized Debye model containing the dispersion coefficient α. It is possible to estimate values of τ. The magnetic relaxation undergoes a single relaxation process at α = 0, a Cole–Cole plot (χ′′ versus χ′ plot) is semicircular.²⁴⁶ In addition, magnetic relaxation is a phenomenon associated with spin–phonon coupling, in which the lattice vibration of the molecular crystal is mainly dominated by intramolecular vibration. SMMs undergo three types of magnetic relaxation processes due to spin–lattice interactions, i.e., (1) direct, (2) Raman, and (3) Orbach processes.²⁴⁷ The direct process is caused by the absorption or emission of a phonon (ħω), while the Raman process is due to phonon scattering between the ground state and the virtual transition to higher energy levels.^248–250 Numerical simulation of the T dependence of τ is carry out using a combination of the magnetic relaxation processes.^247,251,252

Furthermore, at extremely low temperatures (e.g. 40 mK), spin inversion via QTM occurs due to the quantum tunneling effect when the energies of the spin sublevels are equal, and the steps in the magnetic hysteresis curves (M versus H) are observed using a micro-SQUID (Fig. 12c).^227,245,253 The origin of QTM in SMMs can be understood via an energy level diagram for a system with a tunnelling gap (ΔE_tunnel) by the hyperfine (hf) interaction between the electronic spin and the nuclear spin I. QTM is complicated due to the interactions of the ground state with the I, which cause energy level crossing (Fig. 12b).^253–255 It is possible to explain QTM using the Landau–Zener–Stückelberg (LZS) model to the ground state ±m_S levels.²⁴⁵ Quantum phenomena, such as the QTM observed in SMMs, have been studied in relation to the field of quantum logic (e.g., quantum spin coherence,²⁵⁶ quantum phase transition,²⁵⁷ Berry phase interference,²⁵⁸ and quantum computers²⁵⁹). Interest in the use of molecules as spin qubits is growing rapidly, and several perspective articles and reviews focusing on such molecules have been published recently.^260–263 Using SMMs in information storage devices would allow memory capacity to be expanded quite dramatically. Thus, the development of multi-functional SMMs can be expected to continue for several years to come.²⁶⁴ In addition, single-crystal X-ray diffraction analysis and/or theoretical calculations (DFT and ab initio; e.g. CASSCF/RASSI-SO/SINGLE_ANISO calculations)²⁶⁵ for the structure optimisation and the ground state are the most effective methods for investigating the coordination environment of Ln³⁺ ions in SMMs. Additionally, electron spin resonance (ESR), nuclear magnetic resonance (NMR), inelastic neutron scattering, and muon spin relaxation (μSR) are useful in the investigation of the SMM phenomena.²⁴⁵ It is possible to evaluate the SMM characteristics by using a variety of measurements together with magnetic measurements. Thus, the field of molecular magnetism, which includes SMMs, has been extensively studied recently in physics and nanotechnology contexts, particularly in the areas of molecular spintronics and quantum technology (QT) (Fig. 11).²⁶⁶

Fig. 11 Quantum technology actively using SMMs. Rare-earth sandwich tetrapyrrole SMMs (Tb(Pc)₂) have excellent chemical and physical stability, which make it possible to use them in quantum devices. Reprinted with permission from (a) ref. 267. Copyright 2010, American Chemical Society. Reprinted from (b) ref. 268. Copyright 2016, Royal Society of Chemistry. Reprinted from (c) ref. 269. Copyright 2011, MDPI. Reprinted with permission from (d) ref. 270. Copyright 2014, American Association for the Advancement of Science.

For more information, several books and reviews suggested by the authors are included in the reference. In particular, D. Gatteschi, R. Sessoli and J. Villain have published a groundbreaking book on SMMs in 2006, which covers everything from basic principles of physical phenomena to measurement techniques.²⁴⁵

3.2 Brief overview of rare-earth sandwich tetrapyrrole single-molecule magnets

Phthalocyanines feature prominently in diverse research fields, including molecular conductors, solar cells, liquid crystals, pharmaceuticals, and biological chemistry,^271,272 as they are highly soluble in several organic solvents, chemically stable and robust, as well as commercially available.^273–275 In addition, in recent years, tetrapyrrole has attracted attention as a ligand for SMMs.^9,36

In order to identify an underlying cause for SMM properties, we review the coordination environment and SMM properties of Ln(Pc)₂ with a D_4d symmetry. The LF potential around a Tb³⁺ ion (⁷F₆) with a total angular momentum (J) of 6 splits the ground-state multiplet so that the lowest sublevel has the largest z-component of J (m_J = ±6) and a large energy gap (ca. 400 cm⁻¹) to the next-closest sublevel (m_J = ±5), which is correlated with U_eff (Fig. 10d).³⁶ In 2003, Ishikawa et al. first reported the SMM properties of (NBu₄⁺)[Tb(Pc)₂]⁻, which contains a single rare-earth 4f ion (U_eff = 230 cm⁻¹; τ₀ = 6.25 × 10⁻⁸ s; H_C ≈ 0.01 T@0.04 K).^36,276 For the neutral (one-electron-oxidized) species with an unpaired π-radical (S = 1/2) on the Pc ligands, the U_eff is 410 cm⁻¹.²⁷⁷ In comparison to (NBu₄⁺)[Tb(Pc)₂]⁻ and , the cationic species Tb[(EtO)₈Pc]₂⁺ has a much larger U_eff (550 cm⁻¹).⁸⁴ The cationic species has a shorter interligand distance than the neutral and anionic ones because the electrons are removed from antibonding molecular orbitals.⁸⁴ In addition, structural changes in Tb[(BuO)₈Pc]₂⁺ are induced by oxidation from its neutral/anionic species, whereby the interligand distance decreases upon oxidation (cf. Section 3.4).²⁷⁸ Thus, LF contractions cause strong LF splitting, which directly affects the U_eff value. This result quantified LF contractions in the cationic species. Later, the group of complexes that exhibits SMM characteristics with a mononuclear d/f metal ion also became known as single-ion magnets (SIMs). In particular, Ln(Pc)₂-based SMMs are attractive complexes that can be developed using various approaches, because it is possible to control the distance between the Ln³⁺ ions, the coordination environment, and the electronic state by chemically modifying the Pc ligands.^3,4,279 It should be mentioned here that the relationship between the magnitude of the LF splittings and the U_eff is not straightforward. In case of Tb(Pc)₂, the U_eff coincides with the energy difference between ground m_J = ±6 and 1st excited m_J = ±5 states (Fig. 10d). The reason why the magnetic relaxations via 2nd or higher LF level as seen in Mn₁₂ do not occur in the series of Tb(Pc)₂ remains unclear to date. However, recent research studies on the SMMs indicate the importance of the vibrations of atoms and molecules in the crystalline lattice.^248,280,281 For instance, SMMs with a high U_eff (1541 cm⁻¹) and record T_B value (80 K) were reported by Guo et al. in 2018, which was achieved by modifying the vibrational modes upon ligand modification of the [Dy(Cp)₂]⁺ (Cp = cyclopentadienyl) complex.^234,282,283 The high U_eff of this complex originates from the Orbach process via 4th excited LF levels, meaning that those via lower excited states are effectively suppressed. The control of the coupling between spin and vibrations is therefore extremely important for achieving high-performance SMMs.

There have been many attempts, both experimental and theoretical, to elucidate the relationship between coordination geometry and LF parameters, which have a great influence on the ground-state multiplet structure and magnetic anisotropy (cf. Sections 3.2.1 and 3.2.2).²²⁹ Contractions of square antiprismatic (SAP) coordination environments affect the LF of Ln³⁺ ions, thus bringing about highly significant changes in the SMM properties.^{36,84,233,242,252,276,284–287} In general, the LF (or CF) parameter in the Hamiltonian is defined as follows:

where B^q_k is the LF parameter and Ô^q_k is the Stevens operator.^288–291 There are noticeable differences in SMM properties due to the presence or absence of particular off-diagonal LF terms (B^q_k, k = 2, 4, 6; q ≠ 0), which are the non-axial LF parameters. As a specific example, it has been reported that the LF parameters of the square prismatic (SP) coordination structure with D_4h symmetry and those of the ideal SAP coordination structure with D_4d symmetry are different. In the SAP geometry, only the axial anisotropy LF parameters occur (B^q_k, k = 2, 4, 6; q = 0).^{252,276,284,288,291–293} On the other hand, the SP geometry contributes not only to the axial LF parameters, but also to the transverse anisotropy LF parameters B⁴₄ and B⁴₆. In the case of [Tb(Pc)₂]⁻, the LF parameters (B^q_k/cm⁻¹) are as follows: real geometry (B⁰₂ = 414, B⁰₄ = −228, B⁰₆ = 33, B⁴₄ = 10)²⁵² and ideal geometry (B⁰₂ = 574, B⁰₄ = −116, B⁰₆ = 20).²⁹² In the case of [Dy(Pc)₂]⁻, the B^q_k are as follows: real geometry (B⁰₂ = 395, B⁰₄ = −220, B⁰₆ = 40, B⁴₄ = 10).^284,287 The fourth-range extradiagonal parameters (B⁴₄ and/or B⁴₆) exert a greater influence on both the hyperfine structure of the ground state and the ground-state multiplet structure.^{122,276,294–296} Moreover, as the structure deviates from D_4d symmetry, the twist angle (θ) and the zenithal angles (φ) between the C₄ axis and the direction of the Ln³⁺–N_iso coordination bond exert great influence on the ground state and the first excited state of the octacoordination geometry of the Ln(Pc)₂ complex as determined using the theoretical approach.^286,287 Several groups have reported the m_J ground states for different φ and Ln³⁺–N_iso distances. It is clear that the Tb³⁺ ground-state doublet will be |m_J| = 6 at φ < 55°. However, taking a Ln³⁺–N_iso distance of 2.3 Å as an example, the |m_J| = 15/2 ground state of Dy³⁺ will no longer be the lowest at φ > 49°.²⁸⁶ Previous studies have reported that the lowest |m_J| ground state of the Tb(Pc)₂ system is the |m_J| = 6 state, and that of the Dy(Pc)₂ system is the |m_J| = 13/2 state.²³³ These high-|m_J| ground states have high magnetic anisotropy, which is advantageous for SMM design. Consequently, the distortion of the coordination environment is reflected in the SMM characteristics of rare-earth sandwich tetrapyrrole SMMs. The protonation–deprotonation reactions of the double-decker SMMs cause significant changes in the coordination environment around the Ln³⁺ ions, enabling reversible switching of the SMM properties (cf. Section 3.5).²⁹⁷ The SMM properties of Ln³⁺ sandwich complexes are also sensitive to the nuclear spins of the Ln³⁺ ions (cf. Section 3.2.3)²⁷⁶ and Ln–Ln magnetic interactions (f–f interactions; cf. Sections 3.2.4 and 3.3).²⁹⁸ Various approaches are available to induce f–f interactions to improve SMM properties (cf. Sections 3.3.1–3.3.4).

Ln(Pc)₂ exhibits excellent thermal stability and is attracting attention as an SMM to which the vacuum deposition method can be applied; they can be deposited on various substrates in an ultra-high vacuum (UHV) environment. Ln(Pc)₂-based SMMs have excellent physical and chemical stability, and many groups, including ours, are investigating these SMMs as molecular-memory and spintronic device materials (cf. Section 3.8).^{117,135,232,267,299–306} The unpaired π-radical (S = 1/2) on the Pc ligands in Tb(Pc)₂ interacts with the conduction electrons, and Kondo resonance has been observed using scanning tunnelling spectroscopy (STS).³⁰² Furthermore, the Kondo resonance largely depends on the twist angle (θ) of the tetrapyrrole ligand and the electronic structure due to the delocalized π spin state, demonstrating the manipulation of a single spin with a conduction electron.^135,302,307 In addition, Ln-Pc-based SMMs are attractive materials for quantum mechanical devices that can be developed based on the separation between the ground-state doublet m_J and the excited states, as well as the hyperfine (hf) interaction between the electronic spin and the nuclear spin I, because it is possible to control the hyperfine QTM (hf-QTM).^{269,270,303,308–315} The important point here is that each spin (S, J, I) in Tb(Pc)₂ shows a cascade phenomenon through spin interactions. Here, Tb(Pc)₂ incorporated in a transistor structure is in contact with the source, drain, and gate electrode. The various spins (S, J, I) interact with the injected conduction electrons via the delocalized π electrons, and the conduction electrons recognize changes in the hf-QTM state during magnetic field sweeps.²⁷⁰ This represents a readout mechanism for the nuclear spin state, in which the transport properties of Tb(Pc)₂ are controlled by the conduction electrons. Thus, Tb(Pc)₂ SMMs have all the properties necessary to serve as a functional component of quantum-mechanical devices (Fig. 12).

Fig. 12 Molecular qubit based on a Tb(Pc)₂ SMM. (a) Schematic diagram of the qubit detection. Tb(Pc)₂ is placed between Au electrodes. (b) In the Zeeman diagram of Tb(Pc)₂, the strong hyperfine coupling between the m_J = ±6 and the nuclear spin (I = 3/2) forms four avoided level crossings with a tunnel gap (ΔE ≈ 1 μK) and four different qubit states (|–3/2〉, |–1/2〉, |+1/2〉, and |+3/2〉). (c) Quantum tunnelling phenomenon due to hyperfine coupling in which the nuclear spin is observed in the magnetisation curve of Tb(Pc)₂. Reprinted from ref. 308 with permission from the Royal Society of Chemistry.

3.2.1 Effect of the coordination environment around the Ln³⁺ ions on the SMM properties. In this section, we summarize the theoretical and experimental research that has elucidated the relationship between the coordination geometries around the Ln³⁺ ions and the properties of the resulting SMMs. Ishikawa et al. reported that the wavefunctions of the LF sublevels, which are strongly correlated with the SMM properties, are sensitive to the coordination geometry around the Ln³⁺ ions.²³³ In double-decker complexes, a Ln³⁺ ion is coordinated by the eight pyrrole N atoms of the porphyrinoids, and the coordination geometries are evaluated via the twist angle θ between the upper and lower ligands. In this case, there are two extreme coordination geometries, i.e., the square-prism (SP) geometry (θ = 0°) and square antiprism (SAP) geometry (θ = 45°). In the case of the SP geometry, the LF Hamiltonian Ĥ_LF contains non-zero B⁴₄Ô⁴₄ and B⁴₆Ô⁴₆ terms:

Ĥ_LF = B⁰₂Ô⁰₂ + B⁰₄Ô⁰₄ + B⁴₄Ô⁴₄ + B⁰₆Ô⁰₆ + B⁴₆Ô⁴₆

where the B^q_k terms represent the LF parameters and the Ô^q_k terms represent the Stevens operators. Stevens operators with non-zero q values act to mix |m_J〉 and |m_J ± q〉. Therefore, the B⁴₄Ô⁴₄ and B⁴₆Ô⁴₆ terms mix the |m_J〉 and |m_J ± 4〉; this mixing enhances the QTM. In contrast, the Ĥ_LF of the ideal SAP geometry does not contain an operator that gives off-diagonal terms between |m_J〉 and :

Ĥ_LF = B⁰₂Ô⁰₂ + B⁰₄Ô⁰₄ + B⁰₆Ô⁰₆.

Therefore, there is no mixing between |m_J〉 and in the ideal SAP geometry, suppressing QTM. It is not easy to construct the SP geometry in double-decker complexes without replacing the Ln³⁺ ions.³¹⁶ Instead, SP and SAP geometries around Ln³⁺ ions have been realized by using porphyrinato–phthalocyaninato triple-decker complexes (Fig. 13).

Fig. 13 X-ray structures of mixed ligand complexes, showing coordination surrounding symmetries for different metal centres: (a) (Pc)Tb(Pc)Tb(TAnP) (CCDC SAZLIQ);¹²² (b) (TPP)Tb(Pc)Tb(Pc) (CCDC IJOSUX)³¹⁷ and (c) (TClPP)Dy[(PhO)₈Pc]Dy[(PhO)₈Pc] (CCDC MIKSEG).³¹⁸ Hydrogen atoms and solvate molecules are omitted for clarity.

In 2012, Sakaue et al. reported the crystal structure and SMM properties of the heteroligand triple-decker complex (TAnP)Ln(Pc)Ln*(Pc), Ln* = Ln = Tb, or Ln* ≠ Ln = Tb, Y (Scheme 20 and Fig. 13a).¹²² The metal center sandwiched by two Pc ligands (Ln*) shows SAP geometry, whereas that sandwiched by Pc and TAnP, namely, Ln, shows SP geometry. It is possible to synthesize heterometal complexes (TAnP)Y(Pc)Tb(Pc) and (TAnP)Tb(Pc)Y(Pc), in which one of the metal ions is replaced with a diamagnetic and isostructural Y³⁺ ion. The Tb³⁺ ion of (TAnP)Y(Pc)Tb(Pc) adopts SAP geometry, whereas that of (TAnP)Tb(Pc)Y(Pc) adopts SP geometry. The maxima of the imaginary part of the AC magnetic susceptibility for (TAnP)Y(Pc)Tb(Pc) occur at much lower frequencies than those of (TAnP)Tb(Pc)Y(Pc) (Fig. 14b). In other words, (TAnP)Y(Pc)Tb(Pc) shows much longer τ values than (TAnP)Tb(Pc)Y(Pc). These differences have been ascribed to the differences in the coordination geometry around the Tb³⁺ ions. As the Tb³⁺ ion of (TAnP)Y(Pc)Tb(Pc) adopts SAP geometry, mixing between LF sublevels and the accompanying QTM is suppressed, which results in long τ values.

Fig. 14 (a) vs. frequency plots for (TAnP)Y(Pc)Tb(Pc) (upper) and (TAnP)Tb(Pc)Y(Pc) (lower). (b) Coordination geometry and θ vs. U_eff plots for the series of Tb³⁺ triple-decker complexes. Reproduced from (a) ref. 122 and (b) ref. 317 with permission from the Royal Society of Chemistry.

In 2016, Katoh et al. reported a detailed study on the crystal structures and the SMM properties of triple-decker complexes composed of two Tb³⁺ ions, TPP ligands, and Pc.³¹⁷ In case of the triple-decker complex (TPP)Tb(Pc)Tb(TPP), the coordination geometry of the two Tb³⁺ ions is close to the SP geometry (θ = 4°), and the U_eff is ca. 50 cm⁻¹. Another triple-decker complex, (Pc)Tb(Pc)Tb(TPP), showed two different types of coordination geometry around the Tb³⁺ ions (Fig. 14b), and thus two peaks in the vs. T plots. The coordination geometry around the Tb³⁺ (Fig. 13b) sandwiched by two Pc ligands is similar to the SAP geometry (θ = 38°), whereas that of the Tb³⁺ sandwiched by a TPP and a Pc ligand is similar to the SP geometry (θ = 14°). The θ value of the phthalocyaninato triple-decker complex Tb₂[(BuO)₈Pc]₃ is 32°.³¹⁹ By comparing the magnetic properties of this series of triple-decker complexes, Katoh et al. demonstrated the existence of a linear relationship between the U_eff and θ values (Fig. 14b). This result is consistent with the degree of mixing of the off-diagonal LF terms (B⁴₄Ô⁴₄ and B⁴₆Ô⁴₆) due to the deviation from the SAP geometry.

The trend for Dy³⁺ complexes is similar to that for Tb³⁺ complexes. In 2013, Kan et al. reported a series of Dy³⁺ triple-decker complexes composed of phthalocyaninato and porphyrinato ligands (Fig. 13c).³¹⁸ The θ value of the Ln³⁺ between a phthalocyaninato and a porphyrinato ligand is narrow (10°), whereas the θ value of Ln³⁺ surrounded by two phthalocyaninato ligands is wide (25°). The mono-Dy³⁺ complex with a wide θ value, (TClPP)Y[(PhO)₈Pc]Dy[(PhO)₈Pc], shows slow magnetic relaxation without a bias DC magnetic field. In contrast, (TClPP)Dy[(PhO)₈Pc]Y[(PhO)₈Pc] does not show slow magnetic relaxation even in a DC field. The homonuclear Dy³⁺ complex (TClPP)Dy[(PhO)₈Pc]Dy[(PhO)₈Pc] shows field-induced SMM properties. These results demonstrate that wider θ values improve the τ values.

In 2012, Wang et al. reported the structural and magnetic properties of the Dy³⁺ double-decker complexes (TClPP)Dy(Pc) and (TClPP)Dy[(α-Et₂CHO)₄Pc] (Fig. 15).³²⁰ The θ value of (TClPP)Dy(Pc) is 44°, whereas that of (TClPP)Dy[(α-Et₂CHO)₄Pc] is 38°. (TClPP)Dy(Pc) shows clear peaks in the vs. T plots in comparison with (TClPP)Dy[(α-Et₂CHO)₄Pc], indicating that QTM is suppressed in (TClPP)Dy(Pc) due to the SAP coordination geometry.

Fig. 15 X-ray structures of mixed ligand complexes (TClPP)Dy(Pc) (CCDC MAVQIL)³¹⁷ and (TClPP)Dy[(α-Et2CHO)₄Pc] (CCDC MAVQOR)³²¹ together with the symmetries of coordination surrounding lanthanide ions. Hydrogen atoms and solvate molecules are omitted for clarity.

3.2.2 Effect of peripheral and coordinating atom substitution on the SMM properties. Atom substitution at the peripheral positions of the porphyrinoid ligands affects the SMM properties through changing the coordination geometry and electrostatic potential of the coordinating N atoms.

In 2013, Ganivet et al. reported a series of Tb³⁺ double-decker complexes composed of octa(p-tert-butyl phenoxy)phthalocyanine H₂[(tBuPhO)₈Pc], tetra(tert-butyl)phthalocyanine H₂(tBu₄Pc) and unsubstituted phthalocyanine H₂(Pc).⁴⁹ The heteroleptic ligand complexes in which one of the Pc ligands is unsubstituted show higher U_eff values (642–652 cm⁻¹) than the homoleptic ones (394–504 cm⁻¹). Ganivet et al. proposed that electron-donating substituents increase the Tb–N distances, thus pushing the Tb³⁺ toward the unsubstituted phthalocyanine and hence increasing the U_eff. This concept was subsequently extended by Chen et al. in 2017. These authors reported a high U_eff value (652 cm⁻¹) and a magnetic hysteresis of up to 30 K at a sweep rate of 200 Oe s⁻¹ for the double-decker complex [(Bu₂N)₈Pc]Tb(Pc), in which one of the Pc ligands is substituted with alkylamino groups (Scheme 8).⁸¹ DFT calculations of the optimized structures for [(Bu₂N)₈Pc]Tb(Pc) and the corresponding homoligand complex Tb[(Bu₂N)₈Pc]₂ revealed that the electrostatic potential of (Bu₂N)₈Pc is greater in the heteroligand Pc/(Bu₂N)₈Pc complex than in the homoligand (Bu₂N)₈Pc complex (Fig. 16a). This result supports the enhancement of U_eff in heteroligand complexes.

Fig. 16 (a) Electrostatic potential around Tb³⁺ in the Tb[(Bu₂N)₈Pc]₂ and [(Bu₂N)₈Pc]Tb(Pc) double-decker complexes. (b) Magnetic hysteresis curves for atom-replaced double-decker complexes. [(tBuPh)₄SPor]Dy(Pc) show larger magnetic hysteresis (right) than [(tBuPh)₄OPor]Dy(Pc) (left) at a sweep rate of 200 Oe s⁻¹. Adapted with permission from (a) ref. 81. Copyright 2017, American Chemical Society. Reproduced from (b) ref. 173 with permission from the Royal Society of Chemistry.

The SMM properties of the porphyrazine (Pz)-based double-decker complexes Tb(OEPz)₂ and Dy(OEPz)₂ were reported by Gimnez et al. (Scheme 25).¹⁴⁰ The U_eff values of Tb(OEPz)₂ (266 cm⁻¹) and Dy(OEPz)₂ (22 cm⁻¹) are similar to those of their Pc analogues. However, Pz complex benefits from its Pc counterpart due to easier processing. They have a higher solubility of up to 50 mM in CH₂Cl₂ and they can be readily sublimed under relatively mild conditions, that is, 360 °C and 10 mbar. Tb(OEPz)₂ and Dy(OEPz)₂ were evaporated using organic molecular beam epitaxy at 550 K and absorbed intact on the surfaces of Au(111), Ag(111) and Cu(111) in a flat-on fashion, forming hexagonal close-packed layers.³²² Systematic STM studies with multi- and monolayer XPS and complementary DFT calculations were performed showing the preservation of the electronic properties of both ligands and lanthanide centres upon adsorption. Persistence of slow dynamics in Tb(OEPz)₂ single molecule magnets embedded in conducting polymers was confirmed by muon spin relaxation on the example of thin films of PEDOT:PSS mixed with the complex, suggesting that the resulting hybrid material can be interesting for organic spintronics.³²³

Replacing the coordinating N atoms of the porphyrinoid with another element modulates the LF splitting of the Ln³⁺ ions. In 2015, Cao et al. reported the magnetic properties of [(tBuPh)₄OPor]Dy(Pc) and [(tBuPh)₄SPor]Dy(Pc), in which one of the coordinating N atoms of the porphyrin unit is replaced with an O or an S atom, respectively (Scheme 32 and Fig. 16b).¹⁷³ Importantly, the U_eff values of [(tBuPh)₄OPor]Dy(Pc) (95 cm⁻¹) and [(tBuPh)₄SPor]Dy(Pc) (135 cm⁻¹) are larger than those of the non-replaced analogue [(tBuPh)₄Por]Dy(Pc) (U_eff = 28 cm⁻¹). In addition, the U_eff of [(tBuPh)₄SPor]Dy(Pc) is the largest among the Dy³⁺-porphyrinoid sandwich complexes. The difference in the U_eff values between [(tBuPh)₄OPor]Dy(Pc) and [(tBuPh)₄SPor]Dy(Pc) are also evidenced by the magnetic hysteresis curve, in which the latter complex exhibits larger magnetic hysteresis than former complex does (Fig. 16c). Ab initio calculations revealed that the ground-state doublet of [(tBuPh)₄Por]Dy(Pc) has the second largest m_J value of ±13/2, whereas those of [(tBuPh)₄OPor]Dy(Pc) and [(tBuPh)₄SPor]Dy(Pc) have an m_J value of ±15/2. The coordination of the O and S atoms to Dy³⁺ is weaker than that of the N atoms, as evident from the experimental bond lengths and calculated charge distributions. The decrease in the LF along the Dy³⁺–X (X = O, S) direction stabilizes the m_J = ±15/2 states and thus enhances the SMM properties.

In 2018, the SMM properties of hemiporphyrazinato double-decker complexes (Scheme 37) were reported by Liu et al.¹⁸⁹ The homoleptic bis(hemiporphyrazinato) complex (HHp)Dy(Hp) and heteroleptic complex (HHp)Dy(Pc) show U_eff values of 56 cm⁻¹ and 40 cm⁻¹, respectively. These U_eff values are larger than those of Pc-based Dy³⁺ double-decker complexes. The lone electron pair on N in the pyridine in Hp is more localized in comparison with that of the isoindole N atom of Pc. As a result of the stronger LF of Hp compared to that of Pc, the SMM properties of the Hp-based complexes are superior to those of the Pc-based complexes.

SMMs that combine porphyrinoid and non-porphyrinoid ligands are rare. In 2013, Wang et al. reported triple-decker SMMs composed of Pc and the Schiff-base ligand SB. Reaction of the half-sandwich complex [(Pc)Dy(acac)] with H₂(SB) affords the triple-decker complex (Pc)₂Dy₂(SB)(H₂O) (Scheme 33), which shows a U_eff value of 19.8 cm⁻¹.¹⁷⁶ The two Dy³⁺ ions of (Pc)₂Dy₂(SB)(H₂O) are crystallographically independent, whereby one adopts a SAP geometry that comprises four N atoms of Pc and two N and two O atoms of SB, while the other adopts a seven-coordinated geometry constructed by four N atoms of Pc, two O atoms of SB, and one oxygen atom of a water molecule. In 2014, Gao et al. reported a similar triple-decker SMM, Dy₂(Pc)(SB-TTF)₂(MeOH) (Scheme 34a), with a U_eff value of 20 cm⁻¹, which forms assembled structures on highly oriented pyrolytic graphite (HOPG).¹⁷⁹ The same author reported the SMM properties of the trinuclear Dy³⁺ complex Dy₃(Pc)₂(mS)(OAc)(OMe)₂ (Scheme 34b) composed of two Pc ligands and macrocyclic Schiff base liagnd(mS), which exhibits zero-field SMM behavior with a U_eff value of 42 cm⁻¹ and ferromagnetic interactions. In contrast, its Er³⁺ analogue Er₃(Pc)₂(mS)(OAc)(OMe)₂ exhibts neither SMM properties nor ferromgnetic interactions.¹⁸⁰ A similar analogue of the dinuclear Dy³⁺ complexes (μ-LR)₂Dy₂(Pc)₂(H₂O) composed of salicyclic aldehydes H(LR) (R = OMe and OEt) and Pc reported by Ge et al. (Scheme 39) exhibits zero-field SMM behavior with the U_eff values ranging from 27 (R = OEt in the presence of a DC magnetic field) to 47 cm⁻¹ (R = OMe).²⁰¹ The MeO-substituted complex exhibits better SMM properties than the EtO-subsituted one does because coordination geometry of the latter one deviates from the SAP geometry, which is caused by a long bond length between the ethocy O atom and Dy³⁺ ion. As the result of difference in coordination geometry and LF, methoxy-substituted complex exhibits intramolecular ferromagnetic interactions between two Dy³⁺ ions. Reaction of the half-sandwich complex [Dy(Pc)(OAc)] with the polyoxotungstate [PW₁₁O₃₉]⁶⁻ affords (Pc)Dy[PW₁₁O₃₉]⁶⁻(NBu₄⁺)₆, which shows a U_eff value of 34 cm⁻¹ at a bias DC field of 500 Oe (Scheme 38c).¹⁹³ SMMs composed of Kläui's tripodal ligand and a porphyrinoid, (L^Co)Ln(Pc) and (L^Co)Ln(TPP) (Ln = Tb and Dy) (Scheme 38a and b), were reported by Gao et al.,¹⁹⁰ and their U_eff values ranged from 6 to 17 cm⁻¹.

3.2.3 Quantum tunnelling of the magnetisation driven by nuclear spin. Quantum tunnelling of magnetisation (QTM) is an emblematic phenomenon originating from the quantum mechanical properties of SMMs. QTM was observed for the first time in the Mn₁₂ cluster SMM.²⁵⁵ This cluster has a total spin quantum number S = 10, and its degree of degeneracy is 21, i.e., 21 of the magnetic states mix with each other at a certain magnetic field. The occurrence of QTM was observed as steps in the magnetic hysteresis curve of the Mn₁₂ cluster. The energy separation of 21 of the magnetic states of the Mn₁₂ cluster is only a few wavenumbers (Fig. 10c). Therefore, the application of a relatively weak magnetic field can induce the mixing of magnetic states, making it possible to observe the steps in the hysteresis curve of the Mn₁₂ cluster. In contrast, the energy separation of the LF sub-levels of [Tb(Pc)₂]⁻ (and [Dy(Pc)₂]⁻) is too great (several hundred wavenumbers) for mixing to be induced between the 13 (16 for [Dy(Pc)₂]⁻) m_J states (Fig. 10d). In contrast, Ishikawa et al. reported that doping excess amounts of the tetrabutylammonium (Bu₄N⁺) salts of [Tb(Pc)₂]⁻, [Dy(Pc)₂]⁻, and [Ho(Pc)₂]⁻ into a diamagnetic matrix ([Y(Pc)₂]⁻) led to the observation of steps in the magnetic hysteresis at dozens of millikelvin (Fig. 12c).^276,324 These steps originate from the QTM among the ground-state LF sublevels coupled with the nuclear spins of Ln³⁺ (hyperfine coupling). In the case of [Tb(Pc)₂]⁻, the lowest energy doublet |m_J〉 = |±6〉 couple with the nuclear spin I = 3/2 of ¹⁵⁹Tb (natural abundance of 100%) to form the |m_J,m_I〉 (m_I = ±3/2, ±1/2) states (Fig. 12b). The steps in the magnetic hysteresis of [Tb(Pc)₂]⁻ originate from the QTM among those eight hyperfine states. The steps in the magnetic hysteresis curves of [DyPc₂]⁻ are complicated and unclear compared to those of [TbPc₂]⁻ due to the seven isotopes of Dy, namely, ¹⁵⁶Dy, ¹⁵⁸Dy, ¹⁶⁰Dy, ¹⁶¹Dy, ¹⁶²Dy, ¹⁶³Dy, and ¹⁶⁴Dy. Moreno-Pineda et al. reported the magnetic hysteresis of (Et₄N⁺)[¹⁶³Dy_0.05Y_0.95(Pc)₂]⁻ (I = 5/2) and (Et₄N⁺)[¹⁶⁴Dy_0.05Y_0.95(Pc)₂]⁻ (I = 0).³¹¹¹⁶³Dy has a nuclear spin, whereas ¹⁶⁴Dy does not. The magnetic hysteresis curves of (Et₄N⁺)[¹⁶³Dy_0.05Y_0.95(Pc)₂]⁻ show clearer steps compared to those of the mixture of nuclear isomers (Fig. 17a). Moreover, the multiple steps in the hysteresis curves disappear in (Et₄N⁺)[¹⁶⁴Dy_0.05Y_0.95(Pc)₂]⁻ (Fig. 17b). This is direct evidence that nuclear spins correlate with the QTMs.

Fig. 17 Field dependence of the magnetisation at 0.03 K with the field applied parallel to the easy axis of magnetisation for (a) (Et₄N⁺)[¹⁶³Dy_0.05Y_0.95(Pc)₂]⁻ (I = 5/2) and (b) (Et₄N⁺)[¹⁶⁴Dy_0.05Y_0.95(Pc)₂]⁻ (I = 0). Insets show magnifications in the ±30 mT region. (c) Crystal structure of triple-decker complex (Pc)Tb(Pc)Tb[(Hex)₈Pc]. (d) Hysteresis loop of (Pc)Tb(Pc)Tb[(Hex)₈Pc] recorded with a field sweep of 0.140 T s⁻¹. (e) First field derivative for a field sweep from −1 to +1 T of the data in (d), and (f) Zeeman diagram with the field parallel to the magnetic easy axes. Adapted with permission from (a and b) ref. 311. Copyright 2017, Wiley-VCH. Adapted with permission from (c–f) ref. 314. Copyright 2018, American Chemical Society.

Nuclear-spin-driven QTM has also been observed in the Tb³⁺ triple-decker complex (Pc)Tb(Pc)Tb[(Hex)₈Pc], which contains two Tb³⁺ ions (Tb(1) and Tb(2)) (Fig. 17c–f).³¹⁴ As will be discussed later, magnetic interactions exist between the Tb³⁺ ions in the triple-decker complex (cf. Section 3.2.4). Therefore, QTM occurs among |m¹_J, m¹_I, m²_J, m²_I〉. The number of possible quantum states d for the triple-decker complex (d = (2I + 1)² = 16) is larger than that for [Tb(Pc)₂]⁻ (d = (2I + 1)¹ = 4). The strategy of using Ln³⁺–Ln³⁺ interactions is promising for the construction of multiple qubits, which is required for quantum computers.

3.2.4 f–f interactions elucidated using multiple-decker complexes. In addition to double-decker complexes, multiple-decker complexes that contain two lanthanoid ions Ln³⁺ also show SMM properties. The main differences between the double-decker and multiple-decker complexes are the existence of intramolecular Ln³⁺–Ln³⁺ interactions in the latter case. Magnetic interactions between Ln³⁺ ions in binuclear metal complexes were reported by Ishikawa et al.²⁹⁸ before the first report of rare-earth-based SMMs in 2003.³⁶ The authors synthesized a series of phthalocyaninato triple-decker complexes (Pc)Ln(Pc)Ln′[(BuO)₈Pc] (Ln = Ln′ = Tb, Dy, Ho, Er, Tm, Yb), denoted as [Ln,Ln′] (Fig. 18). A comparison of the DC magnetic properties of [Ln,Ln′], [Y,Ln′], and [Ln,Y] (in the latter two, one Ln³⁺ is replaced with the diamagnetic and isostructural Y³⁺ ion), enables the observation of Ln³⁺–Ln³⁺ interactions with the increase/decrease in Δχ_MT values:
where , and represent the χ_M values for [Ln,Ln′], [Y,Ln′], and [Ln,Y], respectively. Ishikawa et al. revealed that the behaviour of the Δχ_MT values can be explained solely by the magnetic dipolar term, indicating that the magnetic f–f interactions originate from magnetic dipole–dipole interactions. Increases in the Δχ_MT in Tb³⁺, Dy³⁺, and Ho³⁺ complexes with decreasing temperature indicate the existence of ferromagnetic interactions. On the other hand, the Er³⁺ and Tm³⁺ complexes show antiferromagnetic interactions, which were observed as decreases in Δχ_MT. The Δχ_MT values of the Yb³⁺ complex showed only negligible temperature dependence. This complicated behaviour can be interpreted in terms of dipole–dipole interactions. As discussed in the section on double-decker complexes (cf. Section 3.2), Tb³⁺, Dy³⁺, and Ho³⁺ ions show easy-axis magnetic anisotropy when sandwiched by Pc ligands. In other words, the magnetic dipoles of these ions in the [Ln,Ln′] complexes tend to be aligned along the C₄ axis (perpendicular to the Pc plane). Therefore, head-to-tail ferromagnetic arrangement of the magnetic dipoles is favored in Tb³⁺, Dy³⁺, and Ho³⁺ complexes. In contrast, Er³⁺ and Tm³⁺ sandwiched by Pc ligands show easy-plane magnetic anisotropy. As a result, parallel and antiferromagnetic arrangement of the magnetic dipoles is favored in Er³⁺ and Tm³⁺ complexes. The Δχ_MT of the Yb³⁺ complex is less temperature dependent than those of other [Ln,Ln′] complexes because of the weak magnetic anisotropy of Yb³⁺. The use of diamagnetic and isostructural Y³⁺ ions demonstrated by Ishikawa et al. is a smart method to accurately determine the effect of magnetic Ln³⁺–Ln³⁺ interactions that has been used for a wide variety of Ln³⁺-based SMMs.

Fig. 18 Molecular structures and Δχ_MT values of (Pc)Ln(Pc)Ln′[(BuO)₈Pc] denoted as [Ln,Ln′]. The red curves show theoretical data. Adapted with permission from ref. 298. Copyright 2003, American Chemical Society.

The next question is how long the Ln³⁺–Ln³⁺ distance must be in order to induce f–f interactions. The Ln³⁺–Ln³⁺ distance of the Pc triple-decker complex estimated using DFT calculations is 3.57 Å (this value is consistent with the crystal structure analyses for triple-decker analogues).²⁹⁸ Increasing the Ln³⁺–Ln³⁺ distance should affect the magnitude of the f–f interactions.

In 2010, Fukuda et al. reported the first example of a discrete quadruple-decker complex composed of four Pc ligands.²⁰² The reaction of the anionic Lu³⁺ double-decker complex [Ln(Pc)₂]⁻ with Cd²⁺ acetate afforded (Pc)Lu(Pc)Cd(Pc)Lu(Pc), which is referred to as [LnCdLn] (Fig. 19a). As the two Ln³⁺ ions are separated by a diamagnetic Cd²⁺ ion, the Ln³⁺–Ln³⁺ distance is longer than those in triple-decker complexes. Fukuda et al. investigated the Ln³⁺–Ln³⁺ interactions in [LnCdLn] (Ln = Tb, Dy, Er)³²⁵ using the same method as for the series of [Ln,Ln′] complexes²⁹⁸ and revealed the existence of f–f interactions, although their magnitude was low compared to those observed in [Ln,Ln′]. In the case of [TbCdTb] and [DyCdDy], an increase in Δχ_MT is observed, indicating intramolecular ferromagnetic interactions between the Ln³⁺ ions (Fig. 19b). As discussed in the section on [Ln,Ln′], the head-to-tail arrangement of the magnetic dipoles is stabilized when axially anisotropic Ln³⁺ ions (Tb³⁺ and Dy³⁺) are involved. On the other hand, [ErCdEr] shows a decrease in Δχ_MT due to the easy-plane magnetic anisotropy of the Er³⁺ ions.

Fig. 19 (a) Synthesis and crystal structures of a series of phthalocyaninato multiple-decker complexes; representative molecular crystal structures of the Dy³⁺ complexes Dy₂Cd[(BuO)₈Pc]₄, Dy₂Cd₂[(BuO)₈Pc]₅ and Dy₂Cd₃[(BuO)₈Pc]₆; n-butoxy groups in the crystal structures are omitted for clarity. Δχ_MT vs. T plots for the Dy³⁺ multiple-decker complexes were constructed using data from ref. 211, 213 and 326. (b) DC magnetic susceptibilities of quadruple-decker complexes that contain Tb³⁺, Dy³⁺ and Er³⁺ ions. Reproduced from (b) ref. 325 with permission from the Royal Society of Chemistry.

Wang et al. and Katoh et al. have reported the crystal structures of Dy³⁺ and Tb³⁺ quadruple-decker complexes composed of (BuO)₈Pc ligands, Tb₂Cd[(BuO)₈Pc]₄ and Dy₂Cd[(BuO)₈Pc]₄.^205,211 On account of the low solubility of the quadruple-decker complexes composed of unsubstituted Pc ligands, peripheral n-butoxy substitution is mandatory to obtain single crystals suitable for X-ray diffraction analysis. The strategy of using Cd²⁺ ions to elongate the stacked π-system was further extended by Wang et al. Their group succeeded in obtaining a series of stacked π-systems: the quintuple-decker complexes Ln₂Cd₂[(BuO)₈Pc]₅²⁰⁸ and sextuple-decker complexes Ln₂Cd₃[(BuO)₈Pc]₆.²⁰⁹ Horii et al. reported the crystal structures of the Dy³⁺ multiple-decker complexes composed of [(BuO)₈Pc] ligands (Fig. 19a) and confirmed that the Dy³⁺–Dy³⁺ distances are 9.8 Å for the quintuple-decker complex Dy₂Cd₂[(BuO)₈Pc]₅ and 13.0 Å for the sextuple-decker complex Dy₂Cd₃[(BuO)₈Pc]₆.^213,326 Very weak f–f interactions exist in both complexes, as demonstrated by the increase in their Δχ_MT values (Fig. 19a right). These f–f interactions can be used to alter their SMM properties, as discussed in the following section.

3.3 Using f–f interactions to improve SMM properties

In 2002, Wernsdorfer et al. reported the magnetic properties of a SMM dimer composed of the TM-based cluster Mn₄.²²⁵ The step at zero DC magnetic field in the magnetic hysteresis curve of the monomeric SMM Mn₄ disappears in the dimerized SMMs Mn₄⋯Mn₄ due to the suppression of QTM. The mechanism of the suppression of QTM can be explained as follows. If the probability of QTM between the up (|↑〉) and down spin (|↓〉) states of one SMM is P_QTM, the QTM probability between the ferromagnetically coupled states (|↑,↑〉 and |↓,↓〉) in the SMM dimer is expressed as P_QTM², because QTM between these states requires simultaneous spin reversal. Since P_QTM² is much smaller than P_QTM, QTM between the coupled spin states is a forbidden transition. Therefore, the coupling of two SMMs suppresses the QTM at zero bias DC magnetic field. This strategy (exchange bias effect) is important when considering SMMs as a possible candidate for information storage, because it can improve data stabilisation at zero DC field. The f–f interactions in multiple-decker complexes can also be used as a source of exchange bias. The effect of f–f interactions on the dynamic magnetic properties of rare-earth SMMs was reported by Ishikawa et al.³²⁷ They prepared the homodinuclear Tb³⁺ triple-decker complex (Pc)Tb(Pc)Tb[(BuO)₈Pc] (denoted as [Tb,Tb]) and the two heterodinuclear complexes (Pc)Y(Pc)Tb[(BuO)₈Pc] (denoted as [Y,Tb]) and (Pc)Tb(Pc)Y[(BuO)₈Pc] (denoted as [Tb,Y]), as summarized in Fig. 18. The mono-Tb³⁺ complexes [Y,Tb] and [Tb,Y] show clear DC magnetic field (H_dc) dependences in their vs. T plots; the peak maxima shift to a higher temperature with increasing H_dc magnitude (Fig. 20a). In contrast, the vs. T plots for [Tb,Tb] are H_dc-independent in the field range of 0 to 3000 Oe (the right side of Fig. 20a). The two peaks in the plots of [Tb,Tb] are due to the two inequivalent Tb³⁺ ions. The increase in the temperatures of the maxima of [Y,Tb] and [Tb,Y] corresponds to the suppression of QTM by applying an H_dc. The QTM occurs among the energetically degenerate states. However, applying an H_dc resolves the degeneracy of the ground-state doublet |±6〉 due to Zeeman splitting. Therefore, the QTM of [Y,Tb] and [Tb,Y] is suppressed with increasing H_dc, causing the remarkable H_dc dependences of . In contrast, the Tb³⁺ ions of [Tb,Tb] are affected by the magnetic dipolar field from the other Tb³⁺ ion. Therefore, the Tb³⁺ ions of [Tb,Tb] act as if a DC magnetic field is applied even in the absence of an external H_dc. Therefore, the QTM in [Tb,Tb] is suppressed without an H_dc, resulting in weak H_dc dependence in its vs. T plots.

Fig. 20 (a) DC magnetic field (H_dc) dependences of the vs. T plots at an AC frequency of 997 Hz for the triple-decker complexes [Y,Tb], [Tb,Y] and [Tb,Tb]. (b) Schematic illustration of the Zeeman diagrams of Tb³⁺ multiple-decker complexes. (c) H_dc dependences of the vs. ν plots for the series of butoxy-substituted Tb³⁺ multiple-decker complexes at 4 K based on the data from ref. 319, 328 and 211. Adapted with permission from (a) ref. 327. Copyright 2005, Wiley-VCH.

The Zeeman diagrams of the triple-decker complexes reported by Sakaue et al. aid in the understanding of the relaxation mechanism of [Tb,Tb].¹²²Fig. 20b is a schematic illustration of the Zeeman diagrams of various multiple-decker Tb³⁺ complexes.²¹² Due to the ferromagnetic dipole–dipole interactions between the Tb³⁺ ions in the triple-decker complex [Tb,Tb], the lowest energy states are |m_J, m_J〉 = |±6, ±6〉 at zero H_dc. The antiferromagnetically coupled doublet |±6, ∓6〉 is located at ca. 3 cm⁻¹. Due to the hyperfine coupling between J and I, the Zeeman diagram of |±6, ±6〉 and |±6, ∓6〉 shows multiple lines, which are depicted as bands in Fig. 20b. At zero DC magnetic field, |+6, +6〉 and |−6, −6〉 are degenerate. However, QTM between |+6, +6〉 and |−6, −6〉 is prohibited because this transition requires simultaneous spin reversal. The Zeeman diagram also predicts energy crossing between |±6, ∓6〉 and |+6, +6〉 at H_dc = ±3000 Oe, indicating that the QTM between |±6, ∓6〉 and |+6, +6〉 is enhanced at 3000 Oe.

In the case of the quadruple-decker complexes, the magnitude of the f–f interactions decreases and the H_dc dependences of τ are changed. Fukuda et al. reported the H_dc dependences of the τ values for homonuclear ([TbCdTb]) and heteronuclear ([TbCdY]) quadruple-decker complexes.³²⁹[TbCdTb] has f–f interactions, whereas [TbCdY] does not. At zero H_dc, [TbCdTb] shows longer τ values than [TbCdY]. However, at H_dc = 1000 Oe, the τ values of [TbCdTb] become faster than those of [TbCdY]. Further increasing the H_dc causes the τ of both complexes to become similar. This complicated modulation of τ can again be explained using a Zeeman diagram. Since the magnetic easy axis of the Tb³⁺ ion in the multiple-decker complex is parallel to the C₄ axis, the ferromagnetic head-to-tail arrangement of the magnetic dipoles is stabilized in [TbCdTb]. Therefore, the ground and first excited states of [TbCdTb] are |±6, ±6〉 and |±6, ∓6〉, respectively. Compared to the triple-decker complex, the energy difference between |±6, ±6〉 and |±6, ∓6〉 is smaller in [TbCdTb] due to the weaker f–f interactions. The ground-state doublet |±6, ±6〉 shows Zeeman splitting, whereas |±6, ∓6〉 does not. Due to the hyperfine interactions, |±6, ±6〉 and |±6, ∓6〉 are shown as crowded lines. At zero H_dc, the QTM in [TbCdTb] is suppressed due to the exchange bias effect, which prolongs the τ values. However, |+6, +6〉 and |±6, ∓6〉 cross at around 1000 Oe (middle plot in Fig. 20b). Therefore, the QTM between these is enhanced, decreasing the τ values of [TbCdTb]. This is the reason for the magnetic field dependences of the τ values in [TbCdTb] and [TbCdY]. A similar trend has also been observed in the Dy³⁺ analogues. Katoh et al. reported that the homonuclear Tb³⁺ triple-decker complex Tb₂[(BuO)₈Pc]₃ shows two peaks in the imaginary part of the AC magnetic susceptibility in the presence of a non-zero H_dc (3000 Oe) as summarized in the left side of Fig. 20c.³¹⁹ In contrast, quadruple-decker Tb₂Cd[(BuO)₈Pc]₄³²⁸ and quintuple-decker Tb₂Cd₂[(BuO)₈Pc]₅²¹¹ complexes that contain two Tb³⁺ ions show dual magnetic relaxations at zero H_dc. Importantly, the dual magnetic relaxations become a single one in the presence of a non-zero H_dc (2000 Oe). The complicated magnetic behaviour in these complexes originates from the differences in the magnitude of the f–f interactions. The dual magnetic relaxations of Tb₂[(BuO)₈Pc]₃ have been observed at 3000 Oe, where level-crossing between |+6, +6〉 and |±6, ∓6〉 occurs. Therefore, one of the dual magnetic relaxations in Tb₂[(BuO)₈Pc]₃ is attributed to the QTM and spin-lattice relaxations among |+6, +6〉 and |±6, ∓6〉. In the case of Tb₂Cd[(BuO)₈Pc]₄ and Tb₂Cd₂[(BuO)₈Pc]₅, |±6, ±6〉 and |±6, ∓6〉 show crossing at zero H_dc because of the weaker f–f interactions (small energy difference between |±6, ±6〉 and |±6, ∓6〉), as shown in Fig. 20b. Therefore, magnetic relaxations among these states occur and are observed as the dual magnetic relaxations at zero H_dc. Increasing H_dc resolves the degeneracy of |±6, ±6〉 and |±6, ∓6〉, resulting in the observation of single magnetic relaxation.

The effect of the f–f interactions on the dynamic magnetic properties has been reported in Dy³⁺ complexes. Katoh et al. reported the SMM properties of the triple-decker complexes Dy₂[(BuO)₈Pc]₃ and DyY[(BuO)₈Pc]₃.³³⁰ Homonuclear Dy₂[(BuO)₈Pc]₃ has intramolecular f–f interactions, whereas DyY[(BuO)₈Pc]₃ does not. The magnetic relaxation time τ of Dy₂[(BuO)₈Pc]₃ is 1.30 × 10⁻⁴ s at 1.82 K. This value is one order of magnitude greater than that for DyY[(BuO)₈Pc]₃ (3.20 × 10⁻⁵ s at 1.9 K), indicating that f–f interactions act as an exchange bias to suppress the QTM at zero H_dc.

Horii et al. reported the exchange bias effect in Dy³⁺ quadruple-decker (Dy₂Cd[(BuO)₈Pc]₄), quintuple-decker (Dy₂Cd₂[(BuO)₈Pc]₅), and sextuple-decker (Dy₂Cd₃[(BuO)₈Pc]₆) complexes.^213,326 Similar to the triple-decker complexes composed of two Dy³⁺ ions, the series of multiple-decker complexes shows longer τ values than the mono-Dy³⁺ complex due to the f–f interactions in the former complex. The exchange bias effect is observed in Dy₂Cd₃[(BuO)₈Pc]₆, in which the intramolecular Dy³⁺–Dy³⁺ distance is ca. 13 Å.

3.3.1 Connecting double-decker SMMs without using Cd²⁺ ions. The exchange bias effects of the multiple-decker complexes prolong the τ values. However, double-decker complexes show longer τ values than multiple-decker complexes. The short τ values in the latter case are associated with the changes in the coordination geometry around Ln³⁺ that are induced by the coordination of the metal ions. In the case of the triple-decker complex, the inner Pc ligand is coordinated by two Ln³⁺ centres. Therefore, the coordination geometry around Ln³⁺ becomes unsymmetrical, enhancing the QTM (P_QTM becomes large). Thus, it is desirable to induce the exchange bias effect while keeping the SAP geometry of the double-decker complex (i.e., keeping P_QTM small).

The use of the clamshell-type phthalocyanine³³¹ reported by Tolbin et al. enables the construction of a discrete stacked π-system without using metal coordination.¹⁰⁴ Horii et al. reported a clamshell-type quadruple-decker complex (Pc)Tb[c-Pc₂]Tb(Pc) in which the two Tb³⁺ double-decker units (tBu₄Pc)Tb(Pc) are connected by a benzyloxy linker (Scheme 13 and Fig. 21a).¹⁰⁵ Ferromagnetic f–f interactions are present between the two double-decker units, as evidenced by the increase in the χ_MT value at low temperatures because the head-to-tail arrangements of the magnetic dipoles are stabilized. The magnetic hysteresis of (Pc)Tb[c-Pc₂]Tb(Pc) at a zero DC magnetic field is wider than that of the monomer double-decker complex (tBu₄Pc)Tb(Pc), indicating that the exchange bias is effective in prolonging the τ values. The reaction of (Pc)Tb[c-Pc₂]Tb(Pc) with Cd²⁺ acetate affords the Cd²⁺-containing complex (Pc)Tb[c-Pc₂Cd]Tb(Pc). Although ferromagnetic interactions between the Tb³⁺ ions exist in (Pc)Tb[c-Pc₂Cd]Tb(Pc), it shows a butterfly-shaped magnetic hysteresis curve in which the curve is closed at zero H_dc. These results indicate that the QTM is enhanced (increased P_QTM value) by the coordination of a Cd²⁺ ion.

Fig. 21 (a) Molecular structures and magnetic hysteresis curves of the clamshell-type quadruple-decker complexes. (b) Supramolecular dimerisation of the crown-ether-substituted SMM [(15C5)₄Pc]Tb(Pc). The maxima in the vs. ν plots shift toward the low ν region upon dimerisation. (c) Dimerisation of the double-decker SMM H[(α-BuO)₈Pc]Tb(Pc) (CCDC FIJBEI) by Na⁺ ions yielding dimer Na₂{[(α-BuO)₈Pc]Tb(Pc)}₂ (CCDC FIJBIM). Reproduced from (a) ref. 105 with permission from the Royal Society of Chemistry. Adapted with permission from (b) ref. 67. Copyright 2018, Wiley-VCH.

Horii et al. used supramolecular assembly reported by Ishikawa et al.³³² to introduce f–f interactions.⁶⁷ The crown-ether-substituted double-decker complex [(15C5)₄Pc]Tb(Pc) forms a supramolecular quadruple-decker complex [K₄{[(15C5)₄Pc]Tb(Pc)}₂]⁴⁺ in the presence of excess potassium acetate (Fig. 21b). Quantitative conversion from [(15C5)₄Pc]Tb(Pc) to [K₄{[(15C5)₄Pc]Tb(Pc)}₂]⁴⁺ was confirmed by paramagnetic ¹H NMR analysis and absorption spectra. In addition, ferromagnetic Tb³⁺–Tb³⁺ interactions exist in [K₄{[(15C5)₄Pc]Tb(Pc)}₂]⁴⁺. At 1.8 K, the τ value of [K₄{[(15C5)₄Pc]Tb(Pc)}₂]⁴⁺ is 1000 times longer than that of [(15C5)₄Pc]Tb(Pc), indicating that the exchange bias suppresses QTM.

Recently, the crystal structures of the triple-decker complexes dimerized via analogous supramolecular associations using crown ethers and K⁺ ions were elucidated by Martynov et al.³³³ The triple-decker complex [(15C5)₄Pc]Tb(Pc)Tb(Pc) forms a supramolecular dimer in the presence of K⁺ ions (Fig. 22a). Such dimerization did not alter the symmetry of coordination surrounding of metal centres, and the Ln⋯Ln distances ranging from 3.5 to 13.2 Å renders these complexes as convenient models to study long-range interactions between paramagnetic metal centres aligned in a collinear fashion. In the case of [(15C5)₄Pc]Tb[(15C5)₄Pc]Tb(Pc), the reaction with four K⁺ ions does not induce dimerisation but instead changes the coordination geometry of Tb* from SAP to SP which was demonstrated by crystallographic characterization of isostructural Y³⁺ complex and its supramolecular inclusion compound (Fig. 22b).³³⁴ Therefore, the approach of using (15C5)₄Pc is promising not only for introducing f–f interactions, but also for directly changing the LF. Although the SMM properties of [(15C5)₄Pc]Tb(Pc)Tb(Pc) and [(15C5)₄Pc]Tb[(15C5)₄Pc]Tb(Pc) have not yet been reported, modulation of the magnetic anisotropy, which may be correlated with the LF, was demonstrated via NMR analyses.

Fig. 22 X-ray structure of the supramolecular assemblies formed by heteroleptic Y³⁺ crown-phthalocyaninates. (a) Dimer formed by [(15C5)₄Pc]Y(Pc)Y(Pc) (CCDC SULZIL) in the presence of KBPh₄. (b) Intercalation of potassium cations between crown-substituted ligands of [(15C5)₄Pc]Y[(15C5)₄Pc]Y(Pc) (CCDC ITUJEP and ITUGUC)^333,334 without the formation of supramolecular dimers. Tetraphenylborate ions, hydrogen atoms and solvent molecules are not shown for clarity.

The connection of the double-decker SMM H[(α-BuO)₈Pc]Tb(Pc) by Na⁺ was reported by Chen et al. (Fig. 21c).^335,336 Compared to the monomeric complex H[(α-BuO)₈Pc]Tb(Pc) (U_eff = 125 cm⁻¹), the corresponding dimer complex Na₂{[(α-BuO)₈Pc]Tb(Pc)}₂ shows a larger U_eff value (367 cm⁻¹). The DC magnetic measurements imply that the Tb³⁺–Tb³⁺ magnetic interactions in Na₂{[(α-BuO)₈Pc]Tb(Pc)}₂ are very weak. Therefore, the change in the stacking angle from θ = 28° to θ = 34° upon Na⁺ coordination is the main reason for the increase in U_eff.

3.3.2 Rare-earth-fused phthalocyaninato sandwich complexes. The distance between spins and their placement are important factors that influence SMM characteristics. QTM can be effectively suppressed by arranging the Tb(Pc)₂ units on the magnetic anisotropy axis (C₄ rotation axis) at the distance at which the magnetic dipole interactions are maximized (closest Tb⋯Tb distance: ca. 0.62 Å).^67,105 To put it simply, the vertical stacking of Tb(Pc)₂ improves SMM characteristics and suppresses the QTM. This raises the naive question, “How does the geometry of the Ln(Pc)₂/Ln₂(Pc)₃ units affect the SMM characteristics?” Klyatskaya et al. reported that effective through-space π–π interactions between pyrene-substituted unsymmetrical complexes [(Pyrene)Pc]Ln(Pc), Ln = Tb, Dy, and Ho lead to the formation of head-to-tail π dimers with closest Tb⋯Tb distance: ca. 13.08 Å (Fig. 4).¹¹⁵ We now summarize the research results related to the effects of through-bond interactions in Ln³⁺–Pc sandwich complexes on their SMM properties.

In 2013, Wang et al. prepared the first biradical dinuclear Dy³⁺ triple-decker SMM with a fused phthalocyanine ligand (Pc)Dy(f-Pc₂)Dy(Pc)^OBu (Scheme 42 and Fig. 23a).³²⁸ The distance between the Dy³⁺ ions coordinated to the fused ring Pc⁴⁻ is 10 Å or more and π–f interactions are observed. The π-bridge delocalized biradical antiferromagnetic interactions effectively suppress the QTM of the Dy³⁺-ion-based SMMs, and at temperatures of up to 5 K, butterfly-type magnetic hysteresis with H_C was observed.

Fig. 23 Crystal structure of (a) (Pc)Ln(f-Pc₂)Ln(Pc)^OBu (Ln³⁺ = Dy and Tb) and (d) (Pc)₂Ln₂(f-Pc₂)Ln₂(Pc)₂^OBu (Ln³⁺ = Tb and Dy) and the micro-SQUID-measurement magnetization decay curves (b) and hysteresis loops for oriented crystalline samples of (c) (Pc)Tb(f-Pc₂)Tb(Pc)^OBu and (e) (Pc)₂Dy₂(f-Pc₂)Dy₂(Pc)₂^OBu. Molecular structures of (a) and (d) based on the cif files from ref. 217 and 220. Protons and n-butoxy groups in (Pc)Ln(f-Pc₂)Ln(Pc)^OBu and (Pc)₂Ln₂(f-Pc₂)Ln₂(Pc)₂^OBu are omitted for clarity. Adapted with permission from (a and b) ref. 217. Copyright 2018, American Chemical Society. Adapted with permission from (d and e) ref. 220. Copyright 2018, Wiley-VCH.

In 2018, Morita et al. prepared the first dinuclear Tb³⁺ fused-Pc⁴⁻ triple-decker SMM (abbreviated as (Pc)Tb(f-Pc₂)Tb(Pc)^OBu).²¹⁷ The complex shows an open hysteresis loop with H_C = 180 Oe at 1.8 K (these values were obtained from the magnetic data for a magnetically diluted sample). From the crystal structure, the twist angle (θ) between the Pc ligands was determined to be 45°, which indicates an SAP coordination environment. The strict SAP coordination caused the absence of the off-diagonal LF terms (B^q_k; q ≠ 0), which eliminates the tunneling gap (ΔE) via level-crossings in the ground state, meaning that QTM processes are suppressed. In addition, solution ¹H NMR measurements of (Pc)Tb(f-Pc₂)Tb(Pc)^OBu show that the coordination environment of the Tb³⁺ ions affects the anisotropic magnetic susceptibility (χ_ax). Incorporating radicals improves the magnetic properties, as radical-bridged metal complexes exhibit slow magnetic relaxation dynamics due to suppressed QTM, exhibiting high T_B values due to strong exchange coupling J_ex in comparison with non-radical systems.^{239,337–339} However, it is not necessary to incorporate strong magnetic interactions between the Ln³⁺ ions and radical ligands, as shown by (Pc)Tb(f-Pc₂)Tb(Pc)^OBu. Although the induced intramolecular magnetic interactions using the molecular design of (Pc)Tb(f-Pc₂)Tb(Pc)^OBu are weak, these interactions play a role in suppressing QTM processes. This prediction was supported by DFT calculations, which were used to analyse the exchange couplings in the complex. The obtained results indicated the presence of magnetic interactions among the four spin centres, including indirect Tb³⁺–Tb³⁺ interactions through π-radicals. On the other hand, hyperfine couplings (0.015 cm⁻¹) were observed for an oriented crystalline sample in micro-SQUID measurements (Fig. 23b and c), indicating that the inter- and intramolecular magnetic interactions were very weak, with smaller couplings than that for Tb(Pc)₂⁻ (0.017 cm⁻¹).²⁷⁶ Magnetic hysteresis loops with H_c = 860 Oe at 1.8 K were observed for a crystalline sample of (Pc)Tb(f-Pc₂)Tb(Pc)^OBu (Fig. 23). Considering the relationship between the molecular arrangement and H_c, the weak intermolecular magnetic interactions are more effective in suppressing QTM. These results indicate that the collinearity of magnetic dipole–dipole interactions strongly affects the suppression of QTM.

In 2013, Morita et al. prepared the first tetranuclear Tb³⁺ fused-phthalocyaninato quintuple-decker SMM (Fig. 23c) by using a fused phthalocyaninato ligand to control the spatial arrangement of the Tb³⁺ ions in the SMM (Pc)₂Tb₂(f-Pc₂)Tb₂(Pc)₂^OBu.²¹⁹ In DC magnetic susceptibility measurements, ferromagnetic interactions among the four Tb³⁺ ions in (Pc)₂Tb₂(f-Pc₂)Tb₂(Pc)₂^OBu were observed, with two kinds of magnetic dipole–dipole interactions. (Pc)₂Tb₂(f-Pc₂)Tb₂(Pc)₂^OBu can be described as a weakly ferromagnetically coupled dimer of the SMM Tb₂[(BuO)₈Pc]₃³¹⁹ with strong dipole–dipole interactions in the triple-decker moieties and weak dipole–dipole interactions through the fused-Pc⁴⁻ linking the two Tb₂[(BuO)₈Pc]₃ units. Dual magnetic relaxation processes were observed in (Pc)₂Tb₂(f-Pc₂)Tb₂(Pc)₂^OBu, similar to those in other dinuclear Tb³⁺–Pc complexes.^211,317,340 This provides clear evidence that the magnetic relaxation phenomena depend heavily on the magnetic dipole interactions between the Tb³⁺ ions in (Pc)₂Tb₂(f-Pc₂)Tb₂(Pc)₂^OBu.

In 2018, Katoh et al. reported the magnetic properties and spin-relaxation processes of the tetranuclear Dy³⁺ fused-phthalocyaninato quintuple-decker SMM (Pc)₂Dy₂(f-Pc₂)Dy₂(Pc)₂^OBu and its crystal structure (Fig. 23d and e).²²⁰ The structure of (Pc)₂Dy₂(f-Pc₂)Dy₂(Pc)₂^OBu can be regarded as a dimer of Dy₂[(BuO)₈Pc]₃ SMMs³⁴¹ with different magnetic relaxation behaviour corresponding to the octacoordination-geometry sites Dy1 with a C₄ symmetry (θ₁ = 23°) and Dy2 with a D_4d symmetry (θ₂ = 45°). The lowest Kramers doublet (KD) of (Pc)₂Dy₂(f-Pc₂)Dy₂(Pc)₂^OBu is thought to be a mixed state of m_J = ±13/2 and ±11/2. At an H_dc of 1750 Oe and T range of 1.8–3.75 K, the QTM is suppressed, and the direct process is enhanced; these depend on the different coordination environments. Consequently, the coordination geometry plays a more significant role in the spin-relaxation phenomena of (Pc)₂Dy₂(f-Pc₂)Dy₂(Pc)₂^OBu than the dipolar bias between Dy³⁺ ions.^341,342 The different conclusions for the Tb³⁺ and Dy³⁺ systems are attributable to the ground m_J state for their coordination geometries.^{36,84,233,242,252,276,284–287}

These studies show that the SMM properties can be fine-tuned by introducing magnetic dipole–dipole interactions in a controlled SMM spatial arrangement. In addition, we can expect to obtain nuclear spin molecular qubits using the spin cascade |S = 1/2〉‖J = 6〉‖I = 3/2〉 of the Tb(Pc)₂ system by taking advantage of the more complex intramolecular magnetic interactions in (Pc)Tb(f-Pc₂)Tb(Pc)^OBu.^270,303,308

3.3.3 Crystal engineering of rare-earth-phthalocyaninato sandwich complexes for one-dimensional spin arrangement. Self-assembly via coordination bonds and supramolecular approaches are effective methods of controlling the geometric spin arrangement. It is possible to rationally design molecule-based magnetic materials with zero-dimensional (0D), one-dimensional (1D), two-dimensional (2D), and three-dimensional (3D) structures with a variety of properties, such as electronic and ionic conduction, gas adsorption, and stimulus-induced switching characteristics.³⁴³ This concept has been a huge success in improving the functionalities of Ln(Pc)₂-based SMMs.

can be used in molecular spintronics due to its SMM characteristics and its physical and chemical stability.³⁴⁴ Investigating the influence of the direction of the magnetic dipole interactions acting among the SMMs in the bulk solid on the magnetic relaxation behaviour is important for controlling its magnetic properties, which depend on the molecular arrangement.^345,346 In 2017, Yamabayashi et al. showed that the arrangement of the SMMs in crystals affects the ground state, and that QTM is suppressed at low temperatures.³⁴⁷ As a result, it has been suggested that QTM depends on the orientation angle (σ) of the uniaxial magnetic anisotropy with respect to the vector connecting the Tb³⁺ ions. In addition, QTM is suppressed in the Tb(Pc)₂ dimer when the Tb(Pc)₂ units are connected with a spacer acting as a H_biasvia through-space magnetic dipole interactions (dipolar bias).¹⁰⁵ Moreover, it is possible to eliminate the influence of the transverse field by arranging the Tb³⁺ ions in the direction of the C₄ axis.⁹⁷ The probability of QTM occurring in one Tb(Pc)₂ unit is P_QTM, and the probability of QTM occurring simultaneously in a Tb(Pc)₂ dimer is P_QTM², indicating that QTM is effectively suppressed.

Magnetic dipole interactions are dominant in quasi-1D molecular magnetic materials, in which (Nc)Tb(Pc) SMM units adopt a structure similar to that of Tb(Pc)₂ SMMs (Fig. 24b). In 2018, Katoh et al. reported that the magnetic properties of (Nc)Tb(Pc)^0/+ (neutral (Nc)Tb(Pc)⁰: U_eff = 652 cm⁻¹ with τ₀ = 2.5 × 10⁻¹² s; cationic (Nc)Tb(Pc)⁺: U_eff = 584 cm⁻¹ with τ₀ = 5.5 × 10⁻¹² s) with 1D structures are significantly different from those of a magnetically diluted sample ((Nc)Tb_0.1Y_0.9(Pc)⁰: U_eff = 342 cm⁻¹ with τ₀ = 1.1 × 10⁻⁸ s).¹³⁶ In particular, the τ of (Nc)Tb(Pc)⁺ (22 s at 1.8 K) in the low-temperature region is five orders of magnitude slower than that of (Nc)Tb_0.1Y_0.9(Pc)⁰ (1.2 ms at 1.8 K). Furthermore, the H_C of (Nc)Tb(Pc)⁺ was retained at temperatures of up to ca. 20 K, which is the T_B. The single-ion anisotropy of the Tb³⁺ ions in a 1D structure and the magnetic dipole interactions acting among the molecules determine the direction of the magnetic properties. These results show that the spin dynamics can be improved by manipulating the arrangement of the SMMs in the solid state. There are clear advantages to using quasi-1D SMM systems with magnetic dipole interactions such as (Nc)Tb(Pc)^0/+. This strategy could lead to dramatic changes in the spin-blocking phenomena, because τ is slower at a zero magnetic field. In other words, since the probability of QTM occurring in (Nc)Tb(Pc)⁺ is P_QTMⁿ (n > 2), the QTM is effectively suppressed in a 1D system. Therefore, this relatively simple molecular-design method can be used to construct a variety of Ln(Pc)₂-SMM-based functional molecular magnetic materials, and new functionalities, such as electronic and ionic conduction, gas adsorption, and various switching characteristics, can be incorporated by controlling the pore size.

Fig. 24 (a) Crystal structure of [(Nc)Tb(Pc)]⁺PF₆⁻ (CCDC YICQIN). Hydrogen atoms and the PF₆⁻ anion are omitted for clarity. (b) 1D structure of [(Nc)Tb(Pc)]⁺PF₆⁻ along the c-axis. (c) MH measurement hysteresis loop with a coercivity (H_C) of 150 mT (1500 Oe) for a single crystal of [(Nc)Tb(Pc)]⁺PF₆⁻ at 1.8 K (H//c-axis). The data were normalized to the value in a field of 7 T. (d) 1D structure of [Dy(Pc)₂]I_x along the c-axis. Hydrogen atoms are omitted for clarity. (e) MH curve of [Dy(Pc)₂]I_x for a polycrystalline sample at 2.2 K (top). Magnetoresistance (MR) measurement of [Dy(Pc)₂]I_x for a single crystal at 2.2 K. Adapted with permission from (a–c) ref. 136 and (d and e) ref. 348. Copyright 2018 and Copyright 2021, Wiley-VHC.

These Ln(Pc)₂-SMM-based materials are expected to exhibit galvanomagnetic effects due to the correlation between the localized f electrons and conduction electrons. There have been several reports of partially oxidized 1D [Ln(Pc)₂]I_x complexes (Ln = Ce, Pr, Sm, Ho, and Yb).^349–353 In 2021, a 1D arrangement of a partially oxidized Dy³⁺ double-decker SMM, [Dy(Pc)₂]I_x (1.95 < x < 2.26), was reported by Sato et al. (Fig. 24d).³⁴⁸ Comparison of the Arrhenius plots with those of a magnetically diluted sample [Dy_0.02Y_0.98(Pc)₂]I_x revealed that the τ for [Dy(Pc)₂]I_x is four times slower at 1.8 K ([Dy(Pc)₂]I_x: 0.0050 s; [Dy_0.02Y_0.98(Pc)₂]I_x: 0.0013 s), and that its QTM at low temperatures is suppressed by the dipolar bias in the 1D array. In [Dy(Pc)₂]I_x, the Pc ligands were partially oxidized, and the I₃⁻ ion was present as a counter ion. Based on the polarised reflection spectra of [Dy(Pc)₂]I_x, the π-radicals were delocalized in a 1D column and behaved as a conductor even at low temperatures. Given that conductivity was observed at temperatures lower than T_B, the negative magnetoresistance (MR) effects of ca. 1% corresponding to the magnetic hysteresis of the SMM were observed when the galvanomagnetic characteristics were investigated at 2.2 K. This indicated that the galvanomagnetic properties between the SMMs and the electrical conduction characteristics due to the π–f interactions could be observed. Our groups are continuing to investigate the galvanomagnetism due to π–f interactions in partially oxidized [Ln(Pc)₂]I_x 1D complexes. This strategy may pave the way for the construction of electro-conductive SMMs as molecular spintronic devices via a spin-cascade system.

3.3.4 Cocrystals of rare-earth phthalocyaninato sandwich complexes and fullerenes. In this section, we review the effects of supramolecular assembly with fullerenes on the SMM properties. Supramolecular complexes between porphyrin (Por²⁻) moieties and (C60-Ih)[5,6]fullerene (C₆₀), and/or fullerene derivatives, have been investigated in detail with respect to host–guest chemistry and electronic structures.^354–359 In fullerene–porphyrin systems, a close contact between the molecules and selective supramolecular interactions is generated. C₆₀ regioselectively binds at the centre of the porphyrin complex via a molecular recognition process supported by the non-planar deformation of the porphyrin ligand, which matches the convex shape of C₆₀. Since the LF around the Ln³⁺ ions affects their charge density, researchers have focused on ways to control the magnetic properties of Ln³⁺ SMMs via supramolecular assembly with a fullerene. The coordination geometry significantly affects the spin-relaxation phenomena of Ln³⁺–Pc/Por multiple-decker complexes.^{105,122,317,341} In particular, Ln₂(Pc)₃ complexes have an important advantage over Ln(Pc)₂ complexes as the magnetic relaxation processes correspond to the Zeeman splitting when two 4f spin systems are present.^{67,211,219,220,319,326,330,340–342}

In 2014, Wang et al. reported cocrystals of C₆₀ with the SMM (Pc)Dy(TCIPP), along with the relationship among the coordination environment, intermolecular interactions, and the SMM behaviour.³⁶⁰ Focusing on the twist angle θ between the ligands, it has been reported that the initial angle (43.60° for (Pc)Dy(TCIPP)) does not change dramatically after co-crystallisation (41.22°/44.45° for (Pc)Dy(TCIPP)·0.5C₆₀, 44.41° for (Pc)Dy(TCIPP)·C₆₀, and 44.35° for (Pc)Dy(TCIPP)·2C₆₀). However, the frequency (ν) dependences of the alternating current (AC) measurements changed due to the different molecular arrangements, depending on the number of incorporated C₆₀ molecules.

Combining a triple-decker SMM with C₆₀ can be expected to reduce the overall symmetry and thus reveal the effects of the LF term. In 2020, Katoh et al. reported the dinuclear Tb³⁺ triple-decker complex (TPP)Tb(Pc)Tb(TPP), which contains porphyrin units on the outside to bind with C₆₀ (henceforth abbreviated as [(TPP)Tb(Pc)Tb(TPP)]·C₆₀).³²¹ The supramolecular complex [(TPP)Tb(Pc)Tb(TPP)]·C₆₀ (Fig. 25a and b) was prepared by assembling C₆₀ with the dinuclear Tb³⁺ triple-decker complex (TPP)Tb(Pc)Tb(TPP) with quasi-D_4h symmetry to investigate the relationship between the coordination symmetry and SMM properties (Fig. 25e and f). The two Tb³⁺ sites of (TPP)Tb(Pc)Tb(TPP) are equivalent, and θ was determined to be 3.62° (Fig. 25d). On the other hand, the two Tb³⁺ sites of [(TPP)Tb(Pc)Tb(TPP)]·C₆₀ are not equivalent. The θ values for sites Tb1 and Tb2 were determined to be 3.67° and 33.8°, respectively, due to a change in the coordination symmetry of (TPP)Tb(Pc)Tb(TPP) upon association with C₆₀ (Fig. 25c). At 1.8 K, (TPP)Tb(Pc)Tb(TPP) and [(TPP)Tb(Pc)Tb(TPP)]·C₆₀ undergo different magnetic relaxations, and the changes in the ground state affect the spin dynamics. (TPP)Tb(Pc)Tb(TPP) and [(TPP)Tb(Pc)Tb(TPP)]·C₆₀ relax via QTM at zero H_dc, and the H dependences of their τ values are similar at H > 1500 Oe. On the other hand, at H < 1500 Oe, their τ values show different behaviour, since the off-diagonal terms (B^q_k; q ≠ 0) affect the magnetic relaxation mechanism. Based on the temperature and H dependences of τ, the spin dynamics can be explained by spin–phonon interactions together with the direct and Raman mechanisms. Thus, a supramolecular approach can be used to control the magnetic anisotropy along the C₄ rotation axis and the spin-dynamic properties in Ln³⁺–Pc multiple-decker complexes.

Fig. 25 Cocrystals of rare-earth phthalocyaninato sandwich complexes and fullerene derivatives. (a–c) Crystal structure of [(TPP)Tb(Pc)Tb(TPP)]·C₆₀. The inclusion of C₆₀ changes the coordination environment of the Tb³⁺ ion and affects the SMM characteristics. (d) Coordination environment of (TPP)Tb(Pc)Tb(TPP) before the subsumption of C₆₀. The micro-SQUID-measurement hysteresis loops (e) and alternating current (AC) magnetic susceptibility data (f) of [(TPP)Tb(Pc)Tb(TPP)]·C₆₀. (g) The cocrystal [(Pc)Tb(OEP)]₃(Li⁺@C₆₀·PF₆⁻)₂ forms a quasi-kagome lattice due to the two kinds of isosceles triangles (A and B). Protons and crystal solvents of [(TPP)Tb(Pc)Tb(TPP)]·C₆₀ and [(Pc)Tb(OEP)]₃(Li⁺@C₆₀·PF₆⁻)₂ are omitted for clarity. Adapted with permission from (a–f) ref. 321. Copyright 2020, Wiley-VHC. Reproduced from (g) ref. 361 with permission from the Royal Society of Chemistry.

In 2020, Iwami et al. reported the cocrystallisation of a lithium-ion-encapsulating fullerene (Li⁺@C₆₀), C₆₀ and C₇₀ with a (Pc)Tb(OEP) SMM (Fig. 25g).³⁶¹ The OEP unit interacted with the fullerene derivatives to form five different cocrystals: [(Pc)Tb(OEP)]₃(Li⁺@C₆₀·PF₆⁻)₂, [(Pc)Tb(OEP)](Li⁺@C₆₀·BF₄⁻), [(Pc)Tb(OEP)]C₇₀·C₆H₁₄, [(Pc)Tb(OEP)]C₆₀·C₆H₁₄, and [(Pc)Tb(OEP)]C₆₀·Et₂O. For the cocrystals containing Li⁺@C₆₀ ([(Pc)Tb(OEP)]₃(Li⁺@C₆₀·PF₆⁻)₂ and [(Pc)Tb(OEP)](Li⁺@C₆₀·BF₄⁻)), the use of different counter anions resulted in different crystal packings. In particular, [(Pc)Tb(OEP)]₃(Li⁺@C₆₀·PF₆⁻)₂ was found to have a unique quasi-kagome lattice packing that induces intermolecular ferromagnetic interactions. Depending on the crystallisation conditions, different packings, i.e., 1D column packings for [(Pc)Tb(OEP)](Li⁺@C₆₀·BF₄⁻) and [(Pc)Tb(OEP)]C₇₀·C₆H₁₄ and AABB-type arrays of (Pc)Tb(OEP) and C₆₀ dimers for [(Pc)Tb(OEP)]C₆₀·C₆H₁₄ and [(Pc)Tb(OEP)]C₆₀·Et₂O, were formed. In addition, the magnetic properties of the (Pc)Tb(OEP) units were modulated via cocrystallisation.

The relationship between the changes in the coordination geometry of the same molecule and the SMM properties induced via a supramolecular approach was also investigated. The corresponding results showed that the magnetic anisotropy along the C₄ rotation axis of the metal sites in the dinuclear Ln³⁺–Pc multiple-decker complexes could be controlled by forming a supramolecular complex, which could lead to switchable SMMs.

3.4 Correlation between redox states and SMM properties

Porphyrinoids are redox-active ligands due to their extended π-systems and high redox stability. Therefore, sandwich complexes composed of porphyrinoids are also redox active. In this section, we will explain the relationship between redox states and SMM properties. The Ln³⁺ ions (except for Ce³⁺) are in essence redox inert. Therefore, the redox reactions of the sandwich complexes presented here occur on the porphyrinoid ligands.

Ishikawa and Takamatsu et al. reported that the U_eff values of the anionic forms of the double-decker complexes (NBu₄⁺){Tb[(EtO)₈Pc]₂}⁻ and (NBu₄⁺){Dy[(EtO)₈Pc]₂}⁻ are enhanced in their corresponding cationic forms, {Tb[(EtO)₈Pc]₂}⁺(SbCl₆⁻)⁸⁴ and {Dy[(EtO)₈Pc]₂}⁺(SbCl₆⁻).³⁶² The mechanism of the structural changes due to the redox reactions can be explained using molecular orbital theory³⁶³ (Fig. 26a and b). As was briefly introduced in Section 3.2, the oxidized species of Tb³⁺–Pc and Dy³⁺–Pc double-decker complexes tend to show improved SMM properties compared to the reduced ones due to the short ligand-to-metal distance in the oxidized form, which result in greater LF splitting and U_eff values. The molecular orbitals of the sandwich complexes are constructed from the combination of the Pc ligands. The stacked dimerisation shown in the double-decker complexes constructs the antibonding HOMO orbital and the bonding HOMO−1 orbital (Fig. 26a). The removal of electrons (oxidation of the complex) increases the bond order between the ligands, causing axial compression and a decrease in the interligand distance. Although this mechanism is widely accepted, recent ab initio calculations based on the CASSCF/RASSI-SO method have indicated that the magnitude of the LF splitting does not strongly depend on the structural changes induced by the redox reactions, and that the calculated energy levels of the first excited doublets (|m_J〉 = |±5〉) for Tb³⁺ double-decker complexes are ca. 300 cm⁻¹.^293,364,365 Therefore, the 8% increase in the U_eff of (NBu₄⁺){Tb[(EtO)₈Pc]₂}⁻ due to oxidation may arise from thermally assisted QTM via higher excited LF sublevels.

Fig. 26 (a) Molecular orbitals of [Ln(Pc)₂]⁻. Removal of two electrons from the antibonding HOMO affords the cationic form [Ln(Pc)₂]⁺ with a shorter Ln³⁺–Pc distance. (b) vs. T plots for anionic ((NBu₄⁺){Tb[(EtO)₈Pc]₂}⁻) and cationic ({Tb[(EtO)₈Pc]₂}⁺(SbCl₆⁻)) Tb³⁺ double-decker complexes. (c) Magnetic hysteresis behaviour of a Tb³⁺ double-decker complex in a series of redox states ([Tb(iprPc)₂]^−/0/+) as captured using MCD spectroscopy. (d) Structure of the perfluoro double-decker complexes [Tb(F64Pc)₂]^{−/2−/3−}. Adapted with permission from (a) ref. 363. Copyright 2006, Elsevier Ltd. Adapted with permission from (b) ref. 84 (c) ref. 366 and (d) ref. 87. Copyright 2007, Copyright 2010 and Copyright 2013, American Chemical Society.

In 2010, Gonidec et al. reported the magnetic properties of the various redox states of the double-decker complex [Tb(iprPc)₂]^−/0/+ (iprPc = 2,3,9,10,16,17,23,24-octa(isopropylidenedioxy)phthalocyaninato) as detected using magnetic circular dichroism (MCD) spectroscopy (Fig. 26c).³⁶⁶ The advantage of MCD is the measurement of the magnetic properties of the specimen in frozen solution states. Therefore, the effects of the intermolecular magnetic interactions and crystal packing, which could potentially affect the magnetic properties, can be minimized. The MCD intensity vs. magnetic field plots at 1.5 K show magnetic hysteresis for all redox states due to magnetisation blocking. [Tb(iprPc)₂]⁻ and [Tb(iprPc)₂]⁺ show butterfly-shaped hysteresis curves, and the opening of the curve is larger in [Tb(iprPc)₂]⁺. Importantly, [Tb(iprPc)₂]⁰ (a ligand π-radical complex) shows clear opening of the hysteresis curve at zero DC magnetic field. These results clearly indicate that the ligand π-radical suppresses QTM.²⁷⁷

In 2010, the same group reported the magnetic properties (MCD spectra) of the highly reduced double-decker complex [Tb(F₆₄Pc)₂]^{−/2−/3−} (F₆₄Pc = 1,4,8,11,15,18,22,25-octakis-fluoro-2,3,9,10,16,17,23,24-octakis-perfluoro(isopropyl)phthalocyaninato) stabilized by electron-withdrawing fluoro groups (Fig. 26d).⁸⁷ Magnetic hysteresis behaviour was observed in all the reduced states. Among these, [Tb(F₆₄Pc)₂]²⁻ showed the largest magnetic hysteresis. The non-π-radical complexes ([Tb(F₆₄Pc)₂]⁻ and [Tb(F₆₄Pc)₂]³⁻) showed narrowing of the hysteresis curve at zero DC magnetic field, whereas the π-radical species [Tb(F₆₄Pc)₂]²⁻ did not show such behaviour. Accordingly, the results of this study corroborate suppression of the QTM at the zero DC field due to a ligand π radical.

The first crystal structure characterisation of the dianionic double-decker complex [Cryptand(Na⁺)]₂[Tb(Pc)₂]²⁻ was reported by Konarev et al. in 2019.⁸⁵ The existence of the ligand π-radical was confirmed based on near-IR absorption and EPR analyses. The interligand distance in [Cryptand(Na⁺)]₂[Tb(Pc)₂]²⁻ (3.258 Å) is shorter than that (3.294 Å) in the monoanionic form (PPN⁺)[Tb(Pc)₂]⁻ (PPN⁺ = bis(triphenylphosphoranylidene)ammonium). Interestingly, [Cryptand(Na⁺)]₂[Tb(Pc)₂]²⁻ does not show slow magnetic relaxation or magnetic hysteresis at 1.9 K, which may be related to the relatively strong intermolecular antiferromagnetic interactions confirmed by the DC magnetic susceptibility measurements.

In 2020, Horii et al. reported a comprehensive study on the structural and magnetic properties of multiple-decker Tb³⁺-(BuO)₈Pc complexes (triple- (Tb₂[(BuO)₈Pc]₃), quadruple- (Tb₂Cd[(BuO)₈Pc]₄), and quintuple-decker (Tb₂Cd₂[(BuO)₈Pc]₅) complexes) in highly oxidized states.²¹⁴ Crystal structure analyses and DFT calculations revealed that the multiple-decker complexes show decreasing interligand distances with successive oxidations. More importantly, paramagnetic NMR analyses in the solution state revealed that the axial component of the magnetic susceptibility (χ_ax) also decreases with successive oxidation (Fig. 27a). This behaviour stands in clear contrast to that of the double-decker complexes, in which χ_ax increases with oxidation.²⁷⁸ In other words, the Tb³⁺ ions in the multiple-decker complexes show decreased magnetic anisotropy upon oxidation. A similar trend was confirmed in the AC magnetic measurements of the oxidized species characterized by X-ray analyses (Fig. 27b). In the case of the triple-decker complex Tb₂[(BuO)₈Pc]₃, the oxidation from the neutral to the two-electron-oxidized state {Tb₂[(BuO)₈Pc]₃}²⁺(SbCl₆⁻)₂ shifts the maxima of the vs. AC frequency (ν) plots to higher ν. In addition, the U_eff decreases from 233 cm⁻¹ (Tb₂[(BuO)₈Pc]₃) to 167 cm⁻¹ ({Tb₂[(BuO)₈Pc]₃}²⁺(SbCl₆⁻)₂) upon oxidation. In the case of the multiple-decker complexes, oxidation causes a decrease in the interligand distances as well as a decrease in the intermetal distances (Fig. 27c). Due to the approach of the positively charged metal ions along the axial direction, the axial LF is decreased. This is the reason for the decrease in U_eff upon oxidation of the multiple-decker complexes.

Fig. 27 (a) χ_ax values for the oxidation states of a (BuO)₈Pc multiple-decker complex. (b) vs. frequency (ν) plots for the redox series of double and triple-decker complexes (Tb₂[(BuO)₈Pc]₃ and {Tb₂[(BuO)₈Pc]₃}²⁺(SbCl₆⁻)₂). (c) Crystal structures of Tb₂[(BuO)₈Pc]₃ and {Tb₂[(BuO)₈Pc]₃}²⁺(SbCl₆⁻)₂ from the cifs in ref. 319 and 214. n-Butoxy groups are omitted for clarity. The triple-decker complex shows a decrease in axial magnetic anisotropy upon oxidation due to the decrease in the LF with the approach of a positively charged Ln³⁺ ion along the axial direction. Adapted with permission from (a) ref. 214. Copyright 2020, Wiley-VHC.

3.5 Proton-induced switching of the SMM properties

Switching of the SMM properties via external stimuli has great importance for the use of SMMs as information-storage devices due to the requirement to eventually erase stored information.

Tanaka and Inose et al.^297,367 reported the switching of SMM properties via the protonation/deprotonation of the Tb³⁺-porphyrinato double-decker complex HTb(TPP)₂. In the protonated form of HTb(TPP)₂, H⁺ coordinates to one pyrrole N atom (Fig. 28a). The distance between Tb³⁺ and the protonated N atom is 2.840 Å; this value is significantly longer than the bond lengths between Tb³⁺ and the non-protonated N atoms (2.430–2.627 Å). In other words, the Tb³⁺ ion in HTb(TPP)₂ adopts a heptacoordinated structure. The significant deviation from the SAP geometry in HTb(TPP)₂ results in the absence of SMM behaviour. When the coordinated proton is removed by DBU (1,8-diazabicyclo[5.4.0]undec-7-ene), the deprotonated form (HDBU⁺)[Tb(TPP)₂]⁻ is obtained. (HDBU⁺)[Tb(TPP)₂]⁻ crystallizes in the space group I₄/mmm, and the eight N atoms coordinating to the Tb³⁺ ion are crystallographically equivalent. Therefore, the Tb³⁺ of (HDBU⁺)[Tb(TPP)₂]⁻ adopts an octacoordinated SAP geometry, which is one of the most ideal symmetries for SMMs. As a result, (HDBU⁺)[Tb(TPP)₂]⁻ shows slow magnetic relaxations and butterfly-shaped hysteresis, which are characteristics of SMMs. The conversion from (HDBU⁺)[Tb(TPP)₂]⁻ to HTb(TPP)₂ is achieved by the addition of acetic acid. Therefore, the protonation/deprotonation process is reversible, enabling the reversible switching of the SMM properties by a single proton.

Fig. 28 (a) Proton-induced on/off switching of SMM properties (HTb(TPP)₂ and (HDBU⁺)[Tb(TPP)₂]⁻). (b) Protonation/deprotonation of the porphyrinato–phthalocyaninato Tb³⁺ double-decker complex H(Pc)Tb(TPP) and (c) phthalocyaninato–phthalocyaninato double-decker complex HTb[(α-BuO)₈Pc]₂. (d) Magnetic hysteresis curves of the protonated (HTb[(α-BuO)₈Pc]₂) and deprotonated form ((HDBU⁺){Tb[(α-BuO)₈Pc]₂}⁻) of the double-decker complex. Reproduced from (a) ref. 297 with permission from the Royal Society of Chemistry. Adapted with permission from (d) ref. 97. Copyright 2018, American Chemical Society.

The protonation/deprotonation of the phthalocyaninato-porphyrinato heteroligand double-decker complex H(Pc)Tb(TPP) was reported by Tanaka et al. (Fig. 28b).³⁶⁸ This complex is obtained as the protonated form H(Pc)Tb(TPP), in which the proton coordinates to the meso-N of the Pc ligand. Since the protonation/deprotonation occurs on the non-coordinating N atom, the changes in the SMM properties induced by deprotonation are not significant compared to those of the homo-porphyrinato complexes. However, the protonated form shows additional magnetic relaxations at low temperatures.

The protonated forms of homo-phthalocyaninato Ln³⁺ double-decker complexes (Fig. 3) were obtained using the [(α-BuO)₈Pc] ligand,⁹⁵ which has eight n-butoxy groups at the nonperipheral positions of Pc, causing a saddle-shaped distortion of the π-structure (Fig. 3). This distortion causes the meso-N atoms of [(α-BuO)₈Pc] to be more basic than those in [(β-BuO)₈Pc]. In 2018, Horii et al. reported the modulation of the SMM properties of HTb[(α-BuO)₈Pc]₂via a protonation/deprotonation reaction (Fig. 28c).⁹⁷ The protonation of the meso-N atoms in HTb[(α-BuO)₈Pc]₂ causes asymmetric distortion as well as a low-symmetry structure, whereas deprotonation using DBU affords (HDBU⁺){Tb[(α-BuO)₈Pc]₂}⁻, which is distorted in a highly symmetric windmill-like fashion. The deprotonated form shows larger magnetic hysteresis than the protonated one does. Similar trends were observed in magnetically diluted samples, indicating that the differences between the magnetic properties of the protonated and deprotonated forms originate from the molecular structures rather than the intermolecular magnetic interactions.

The following two examples also show changes in their SMM properties as a result of deprotonation. However, oxidation from the anionic (and protonated) form to the neutral (and deprotonated) form is the driving force for the deprotonation. In 2019, Chen et al. reported the magnetic properties of HTb[(α-C₅H₁₁O)₄Pc]₂ with 3-pentyloxy-substituents (Fig. 3a).⁹⁴ The U_eff value for HTb[(α-C₅H₁₁O)₄Pc]₂ is 300 cm⁻¹. On the other hand, the U_eff value for the neutral and deprotonated form Tb[(α-C₅H₁₁O)₄Pc]₂ obtained using 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) is 324 cm⁻¹, indicating that the SMM properties are enhanced by deprotonation. In contrast to the previous examples, the presence of the ligand π-radical in Tb[(α-C₅H₁₁O)₄Pc]₂ was confirmed by absorption spectroscopy. Therefore, Chen et al. concluded that the enhancement of the SMM properties stems from the radical–f interactions and enhanced LF due to the shorter ligand-to-metal distances in the neutral form.

In 2019, Sun et al. reported the negative effect of deprotonation on SMM properties by comparison of the chiral (phthalocyaninato)(porphyrinato) Dy³⁺ double-decker complexes (R)/(S)-H[(BNPO)₂Pc]Dy(TClPP) and the corresponding neutral unprotonated species (R)/(S)-[(BNPO)₂Pc]Dy(TClPP) (Fig. 29).³⁶⁹ The AC magnetic susceptibilities of the protonated form (R)/(S)-[(BNPO)₂Pc]Dy(TClPP) show clear frequency dependences, whereas (R)/(S)-[(BNPO)₂Pc]Dy(TClPP) shows no frequency dependence even under a bias DC field (H_dc = 2000 Oe). DFT calculations of (R)/(S)-H[(BNPO)₂Pc]Dy(TClPP) and (R)/(S)-[(BNPO)₂Pc]Dy(TClPP) revealed that the ionic part of the N–Dy bond decreases due to oxidation (deprotonation). This decrease in the electrostatic interactions between the ligand and Dy³⁺ decreases the uniaxial magnetic anisotropy of (R)/(S)-[(BNPO)₂Pc]Dy(TClPP).

Fig. 29 Molecular structures of Dy[(BNPO)₂Pc](TClPP).

3.6 Influence of solvation effects in the crystal lattice on SMM properties

The benzene solvates (μ-Pc)[Ln(dpm)₂]₂·2C₆H₆ showed interesting behaviour with respect to desolvation and exchange of solvation molecules. Thus, heating of solvates with Ln = Sm, Tb, Dy¹⁹⁸ and Er¹⁹⁹ at 100 °C was shown to cause desolvation without disintegration of crystals. This procedure had dramatic influences on the structural and magnetic properties of complexes which were demonstrated on a dysprosium analogue. The solvated complex contains one crystallographically independent metal center, which exhibits a single magnetic relaxation process with an energy barrier (U_eff) of 39 cm⁻¹. When desolvated, the complex had two different types of metal centres with slightly different surrounding, and it exhibited field-induced SMM behaviour with two thermally activated magnetic relaxation processes with U_eff which reach 44 cm⁻¹ and 76 cm⁻¹, respectively. Soaking of desolvated crystals in benzene recovered (μ-Pc)[Ln(dpm)₂]₂·2C₆H₆ which was confirmed by X-ray powder diffractions and elemental analyses.

This work confirmed that the solvated molecules can finely tune the magnetic relaxation mechanisms. Further development of this approach by Ge et al. showed that benzene can be replaced with dichloromethane in the course of “crystal to crystal” transformation.²⁰⁰ Again, the change of solvation state had influence on the magnetic properties, which was demonstrated on the example of (μ-Pc)[Er(dpm)₂]₂·2C₆H₆ undergoing transformation to (μ-Pc)[Er(dpm)₂]₂·2CH₂Cl₂.²⁰⁰ Magnetic susceptibility measurements showed that the two solvates exhibited dramatically different relaxation behaviour (Fig. 30). While benzene solvate exhibited fast relaxation and an estimated energy barrier of 2.6 cm⁻¹ under 600 Oe DC field, the dichloromethane solvate behaved as a field-induced SMM with a higher energy barrier (U_eff) of 34.3 cm⁻¹. The source of these magnetic differences was tentatively assigned to subtle structural differences caused by the change of the crystal lattice, as confirmed by ab initio calculations.

Fig. 30 Interconversion between solvates (μ-Pc)[Er(dpm)₂]₂·2C₆H₆ (CCDC 1494900) and (μ-Pc)[Er(dpm)₂]₂·2CH₂Cl₂ (CCDC 1494901) causing drastic change of magnetic relaxation as seen from the corresponding out-of-phase (χ′′) AC susceptibility plots under 600 Oe DC field. Adapted with permission from ref. 200. Copyright 2017, American Chemical Society.

Ruan et al. showed that crystallization of Tb(Pc)₂ from CHCl₃/MeOH or CH₂Cl₂/MeOH mixtures afforded orthorhombic crystals Tb(Pc)₂ (space group P2₁2₁2₁) and Tb(Pc)₂·CH₂Cl₂ (Pnma) solvates respectively.³⁷⁰ Although the structural differences in magnetically active sandwich molecules were only subtle, absorption of CH₂Cl₂ in the crystal lattice had a profound effect on magnetic relaxation, resulting in a butterfly-shaped hysteresis loop in the case of solvate up to 10 K, while Tb(Pc)₂ showed almost no opening even at 2 K. These results are consistent with the conclusions of Yamabayashi et al., shown in Section 3.3.3.³⁴⁷

In the context of this result, an important study by Dailey and Besson on the selective crystallization of four possible polymorphs of M(Pc)₂ complexes should be mentioned.³⁷¹ The authors proposed reproducible and selective crystallization techniques affording pure α-, γ- and δ-phases together with the CH₂Cl₂ solvate of complexes depending on the solvent, applied for crystallization. Systematic crystallographic and UV-Vis characterization of complexes allowed some of the confusion over chemical composition in previously published literature associated with interconversion between neutral and ionic forms of bisphthalocyaninates to be explained.

3.7 Magnetic studies on surfaces

The properties of molecules significantly change not only through aggregation but also through the interface of organic–inorganic hybrid materials. One of the goals of the research involving sandwich-type tetrapyrrole rare-earth metal–inorganic hybrid materials is their application in spintronic devices.³⁴⁴ In the case of , the magnetic species are composed of the f electron of the central Tb³⁺ ion and the π radical on the Pc ligands. The Pc ligands are almost flat, and the direction of the magnetic moment (axis of easy magnetisation) of the Tb³⁺ ion is perpendicular to the Pc ligand with uniaxially anisotropy, indicating that they are single-molecule magnets.³⁶ Furthermore, since they are thermally stable, they can be deposited on various substrates in an ultra-high vacuum (UHV). Several groups have reported on the structures and physical properties of sandwich-type tetrapyrrole rare-earth complexes on various substrates. In this section, we give an overview of the morphologies and magnetic properties of double-decker and triple-decker complexes on different substrates. Since 2008, many examples of Ln(Pc)₂ and/or Ln₂(Pc)₃ complexes on various substrates, including non-magnetic metals (Cu(100),²⁶⁷ Cu(111),^372,373 Cu(001),³⁷⁴ Au(111),^232,302 Ag(111),³⁷⁵ Ir(111),³⁷⁶ MgO/Ag(100),³⁷⁷ CuO/Cu(110)³⁷⁸), magnetic metals (Co/Ir(111),³⁷⁹ Co/Cu(100),³⁸⁰ Co/Au(111),²⁹⁹ Ni/Cu(100),³⁸¹ Ni/Ag(100),³⁸¹ Fe/Cu(100),³⁸² FeMn/Cu(100),³⁸² CoO/Ag(100),³⁸³ Mn/Ag(100),³⁸³ Ni(111)³⁸⁴), carbon-based substrates (HOPG,^385–387 graphite(0001),³⁸⁸ graphene/SiC(0001),³⁸⁹ graphene/Ni(111)³⁹⁰), silicon-based substrates (SiO₂(100)³⁹¹), perovskite (LaSrMnO₃/SrTiO₃,³⁸⁰ SrVO₃(001)³⁹²), and composite substrates,^393,394 have been reported. Moreover, materials of sandwich-type tetrapyrrole rare-earth metal complexes combined with carbon nanotubes (CNT)/fullerene are shown in Chapter 3.8. In general, eight lobes for both Ln(Pc)₂ and/or Ln₂(Pc)₃ are observed in topographic STM images,³⁰¹ which reflects the electronic state of the Pc ligand on the vacuum side. On the other hand, for Ln(Pc)^x/Au(111), four lobes are observed in the topographic image.²³² In addition, the STM height profiles for Ln(Pc)^x, Ln(Pc)₂, and Ln₂(Pc)₃ on substrates reflect the molecular structure (crystal structure).³⁰¹

In recent years, physical properties of surfaces have been actively studied to advance the research on applications, and together with the progress of nanotechnology, the academic and social importance have increased. Precise control is possible at the atomic level, and organic–inorganic hybrid compounds can be prepared with the surfaces acting as reaction fields. Atoms and molecules adsorbed on surfaces can arrange into various dimensional structures (zero dimensional (0D), one dimensional (1D), two dimensional (2D), and three dimensional (3D)) which exhibit different properties. For example, 2D substances have multifunctional properties. Due to the electronic state of graphene, spin-polarised electron bands form for non-magnetic materials. In addition, graphene is used for molecular spintronics because it utilizes spin–orbit interactions which increases when the π–d orbitals mix at the interface with magnetic metal electrodes.³⁹⁵ At the interface, charge transfer bring about the alteration of the electronic structure, thereby affecting or even quenching the magnetic moment of the molecules. Recently, such spin current control at the interface at the molecular level has become known as spinterface.³⁹⁶ In principle, it is possible to use the spin for quantum information processing (QIP), and spin detection and evaluation are important for realising a quantum computer (QC) and/or spintronic devices.²⁵⁸ Electron spin resonance (ESR), reported by E. Zaboisky in 1945, is one such detection method.

At the same time, scanning tunnelling microscopy (STM) and scanning tunnelling spectroscopy (STS) can be used to identify one magnetic atom adsorbed on a substrate surface and directly observe the density of states (DOS). Thus, they have been attracting attention as a method for single spin detection. STM and STS play a major role in directly investigating the relationship between morphology and electronic structure. For example, for CoPc adsorbed on an Au surface, a symmetric Kondo resonance is observed because only the direct tunnel process from the probe to the substrate is suppressed by the Pc ligand.³⁹⁷ It is possible to study the interactions between magnetic atoms and non-magnetic/magnetic metals or those between magnetic atoms on an atomic scale.³⁹⁸ The Kondo effect, which is the principle of Kondo resonance, was proposed by Jun Kondo in 1964. As the temperature of a metal decreases, its electrical resistance decreases. However, the electrical resistance increases when the temperature is lowered below a certain temperature due to trace amounts of magnetic impurities contained in the metal. This is a phenomenon in which the spins of metallic conduction electrons and the spins of magnetic impurities finally form a Kondo–Yoshida singlet, which has a higher resistance than the metallic state and the electrical resistance remains constant. In this process, as a result of the spin exchange process between the conduction electrons and the magnetic impurities, a resonance state is observed as a sharp peak at the Fermi level (0 V). The temperature at which the electrical conductivity becomes constant at low temperatures is defined as the Kondo temperature (T_K), and in the case of STS, it is proportional to the peak width at half maximum of the Kondo resonance obtained from tunnelling current spectroscopy. By applying the Kondo effect to the conduction electrons, it is possible that the conductance changes, and this phenomenon has been attracting attention again recently because of its high utility in spintronics. As a result, the Kondo resonance should be observed even for M(Pc)₂, which has a spin source.

A Kondo resonance for a delocalized π radical on the ligand due to the electronic state of Tb(Pc)₂ adsorbed on a substrate has been reported, and it is clear that the Kondo resonance observed on Au(111) depends on the molecular conformation (Fig. 31a–c and 32a, b).^302,399 In other words, the Kondo resonance can be switched off by changing θ by <8°. In addition, when Cs is deposited on Tb(Pc)₂/Au(111), the electronic structure of the molecule changes.⁴⁰⁰ This is probably because charge transfer occurs when Cs is added, switching on the Kondo resonance. On the other hand, for Tb(Pc)₂/Ag(111), Kondo resonance is silent in the first layer of Tb(Pc)₂ but active in that of the second layer.³⁷⁵ The π radical recovers by decoupling at the interface. Since the Kondo resonance is silent even for Tb(Pc)₂/Cu(111), the interactions between the substrate and are important for understanding the Kondo resonance.³⁷² In other words, it is possible to control the spin state of via the interface. In addition, since Kondo resonances depend on the electronic state of ligands, the T_K values of delocalized π radicals of Nc²⁻ (naphthalocyaninato) and Pc²⁻ ligands of (Nc)Tb(Pc)/Au(111) differ (Fig. 31d–h).¹³⁵ Furthermore, for Tb(OEP)₂ (OEP²⁻ = 2,3,9,10,16,17,23,24-octaethylporphyrinato)/Au(111), the electronic structure changes significantly due to the addition and desorption of proton at the isoindole nitrogen (N_iso) position, meaning that the Kondo resonance could be switched on and off (Fig. 31i–n).³⁰⁷ Moreover, Kondo resonances have been detected for Tb(Pc)^x/Au(111) and Tb₂(Pc)₃/Ag(111) (Fig. 32d).^232,401 However, a Kondo resonance for Y₂(Pc)₃/Au(111) has not been observed under the same conditions as those used for Y(Pc)₂/Au(111) (Fig. 32c).³⁰² It has been shown that interactions with a Cu(111) substrate quench the unpaired π electron of , explaining the lack of a Kondo resonance. In addition, it has been shown that the unpaired π electron expected for an isolated molecule of is quenched upon adsorption onto Cu(001). In the case of Nd(Pc)₂/Cu(100), quenching occurs due to charge rearrangement at the interface.⁴⁰² On the other hand, a 4f metal centre of Dy(Pc)₂/Cu(001) contributes to the Kondo resonance.³⁷⁴ From density functional theory (DFT) calculations, the Kondo resonance occurs due to the coupling between the Dy³⁺ ion and the Cu electron bath which is mediated by strong hybridization between Dy d and f orbitals.

Fig. 31 STM image of the sandwich-type tetrapyrrole rare-earth metal complexes on the Au(111) surface. (a) STM image of an Isolated Tb(Pc)₂ molecule. (b) Molecular structure of Tb(Pc)₂ which is shown in the upper (lower) Pc ligand is coloured in blue (silver). (c) A simulated STM image of Tb(Pc)₂ obtained by using DFT calculations. (d) Molecular structure of (Nc)Tb(Pc). (e–h) STM and DFT simulated STM images of isolated Pc-up and Nc-up ligands of (Nc)Tb(Pc). White bars in (a), (e), and (g) are 1 nm. (i and j) A schematic model of the operation caused by the injection of tunnelling electrons into a (HOEP)Tb(OEP). (i) Proton (H) atom hopping and (j) H atom desorption. (k–n) STM and DFT simulated STM images of isolated (k and m) Tb(OEP)₂ and (l and n) (HOEP)Tb(OEP). Topographic images, which is similar to those in (k) and (l), observed by the operations in (i) and (j). Reproduced from (a–c) ref. 302. Copyright 2011, Nature Research. Adapted with permission from (d–h) ref. 135. Copyright 2013, American Chemical Society. Reproduced from (i–n) ref. 307. Copyright 2018, Royal Society of Chemistry.

Fig. 32 STS measurement of the sandwich-type tetrapyrrole rare-earth metal complexes on a metal substrate near the Fermi level. (a) Change in the Kondo peak height of Tb(Pc)₂/Au(111) when the tip position is moved along the blue line (bottom) in the STM topography panel. The different colours correspond to the dI/dV intensity, described above the figure. (b) dI/dV spectra measured at the upper Pc ligand of Tb(Pc)₂/Au(111) at positions I (centre) and II (ligand) of the molecule. (c) dI/dV spectra and STM topography of Y(Pc)₂/Au(111) (I) and Y₂(Pc)₃/Au(111) (II). (d) STM topography (i) and topography with staggered three Pc ligands (iii) of Tb₂(Pc)₃/Ag(111). (ii) Map of the fitted Kondo temperature (T_K) on a single Tb₂(Pc)₃/Ag(111). (iv) Averaged point spectra showing the Kondo peak (red x's on ii) on a single Tb₂(Pc)₃/Ag(111). Reproduced from (a–c) ref. 302. Copyright 2011, Nature Research. Reproduced from (d) ref. 401. Copyright 2018, Royal Society of Chemistry.

Research on the Kondo resonance induced at the interface between a light rare-earth tetrapyrrole sandwich compounds has been reported. For (Pc)Ce(TPP)⁰ and , the formal charge of Ce⁴⁺ resulting from the donation of four negative charges to both Pc²⁻ and TPP²⁻ anions causes a 4f⁰ electronic configuration.^403,404 In this case, no Kondo resonance is expected. However, (Pc)Ce(TPP)/Ag(111) exhibits a Kondo resonance but (Pc)Ce(TPP)/Cu(111) does not.⁴⁰³ Thus, a formal 3 +/4+ oxidation state would result in one electron in the 4f orbital and one unpaired π electron in a radical state (˙−)/2− in the ligand for a valence fluctuating system, such as , in which the Ce ion has a valence charge of about 3.57 estimated from Ce L_III-edge X-ray absorption near edge structure spectrum (XANES).⁴⁰³ These systems with fluctuating valences could lead to a ligand-spin induced Kondo resonance similar to those observed for Cu(Pc)/Ag(100) and Tb(Pc)^x/Au(111).⁴⁰⁵ Intermediate valence states make it difficult to explain the presence of a Kondo resonance. Strong hybridization of the molecular orbitals and substrate could result in the quenching of the spin at the interface, eliminating the Kondo resonance. In addition, ligand–ligand interactions in Pc based molecules have been shown to reduce molecule–substrate coupling via the modification of the molecule/substrate distance or charge transfer and therefore to reduce or even quench T_K.^135,405 For example, the competition between the Kondo effect and Ruderman–Kittel–Kasuya–Yosida (RKKY) coupling in 1D and 2D lattices has been reported.^135,406 Thus, theoretical and experimental results are needed to fully understand the spin states at the interface involved in the Kondo process.

The magnetic anisotropy and coercivity can be investigated using soft X-ray magnetic circular dichroism (XMCD) measurements on Ln(Pc)₂ (Ln³⁺ = Tb and Dy) on a substrate (Fig. 33).²⁶⁷ Moreover, the magnetic properties of rare-earth metal ions three-dimensionally surrounded by organic ligands in an element-selective manner can be analysed, and the origin of the magnetic anisotropy of SMMs can be determined from the viewpoint of their electronic states. Soft XMCD can be used on a single atomic layer when synchrotron radiation is used. The uniaxial magnetic anisotropy of the Tb³⁺ ion of Tb(Pc)₂ can be determined by measuring the change in the XMCD spectrum when the magnetic field is applied in different directions (vertical and in-plane directions to the substrate). By performing sum-rule analysis on the obtained spectrum, magnetic hysteresis can be obtained from the quantitative evaluation of the Tb 4f magnetic moment and the magnetisation curve for each element. In addition, by applying the sum rule to the XMCD (and/or XAS) spectra, the orbital magnetic moment L_z and spin magnetic moment 2S_z of the Tb³⁺ ions can be calculated separately. Studies on the magnetic properties of Ln(Pc)₂ (Ln³⁺ = Tb and Dy) on various substrates using XMCD have been reported.^{345,380,387,391} The ferromagnetic Ni layer forms out-of-plane (OP)/Cu(111) and in-plane (IP)/Ag(100), depending on the type of substrate (Fig. 33d–g).³⁸¹ When Tb(Pc)₂ is adsorbed on these Ni layers, the elemental decomposition hysteresis loop changes according to the magnetisation direction of the ferromagnetic Ni layer due to the magnetic interaction between Tb(Pc)₂ and Ni layer. Tb(Pc)₂/Fe/Cu(111) has the same antiferromagnetic relationship as the Ni system.³⁸² When Li is deposited, the ferromagnetic exchange interactions change probably because charge transfer occurs when Li is added. Furthermore, it has been reported that the SMM characteristics of Tb(Pc)₂ and Dy(Pc)₂ adsorbed on the MgO substrate are better (Fig. 33c).^377,407 The improvement in the SMM properties is believed to be due to the effects of spin-phonon coupling on the magnetic stabilities of surface-adsorbed SMMs. This suggests that the use of substrates with low phonon density of states is beneficial for the construction of molecular-based spintronics devices and for further increases in the blocking temperature (T_B) of surface-adsorbed Ln(Pc)₂ (Ln³⁺ = Tb and Dy).

Fig. 33 XMCD hysteresis curves for Tb(Pc)₂ on a metal substrate. (a) Schematic illustration of a Tb(Pc)₂ on an ultrathin MgO film on Ag(100). (b) STM image of a 2D film of Tb(Pc)₂ on a double-layer film of MgO on Ag(100). (c) Hysteresis loop obtained with XMCD at 3 K and the best fit calculations of magnetisation dynamics for Tb(Pc)₂/MgO/Ag(100) and Tb(Pc)₂/Ag(100). (d) Schematic illustration of the magnetic anisotropy axis of Tb(Pc)₂/Ni(OP)/Cu(100) and (f) elemental Ni and Tb resolved hysteresis loops from XMCD at 8 K. (e) Schematic illustration of the magnetic anisotropy axis of Tb(Pc)₂/Ni(IP)/Ag(100) and (g) elemental Ni and Tb resolved hysteresis loops from XMCD at 8 K. Adapted with permission from (a and b) ref. 377. Copyright 2016, Wiley-VCH. Reproduced from (c) ref. 407. Copyright 2019, Wiley-VCH. Reproduced from (d–g) ref. 305. Copyright 2016, Royal Society of Chemistry.

Spintronics of SMMs, like Tb(Pc)₂ on a substrate, has been prepared using the spin state control. With respect to Tb(Pc)₂ adsorbed on a ferromagnetic Co layer grown on an Ir(111), the spin-polarised tunnel current changes according to the application direction (parallel/antiparallel) of the external magnetic field (B_ext) (Fig. 34a).³⁷⁹ It depends on the magnetisation direction of the chip and the sample, and the pattern of spin polarisation of the molecular orbital is observed (Fig. 34b–d). It has been shown that the spin of Tb(Pc)₂ is antiferromagnetically coupled to the magnetisation of Co-islands grown on Au(111) regardless of the adsorption site of the Tb(Pc)₂.²⁹⁹ Both edge and flat-lying molecules exhibit a stable spin direction opposite to the magnetisation direction of the Co-islands to which they are attached (Fig. 34e). More importantly, the spin state of Tb(Pc)₂ can be switched independently of the Co island by applying B_ext (Fig. 34f). The magnetic asymmetries of edge and flat molecules are reversed due to the difference in the DOS due to the Tb(Pc)₂-Co-island interactions between the two different adsorption configurations. Stable spin polarisation is inferred from the substantial hysteresis observed in the magnetisation curve of the Tb(Pc)₂ SMM even with the frozen magnetization of the Co island (Fig. 34g). The observation of significant openings of the hysteresis in the magnetisation curves suggests suppression of the quantum tunnelling of the magnetisation (QTM) process.

Fig. 34 Spin state control of Tb(Pc)₂ on a ferromagnetic Co layer. (a) Schematic illustration of the SP-STM experimental setting of a Tb(Pc)₂/Co/Ir(111) by applying an external magnetic field (B_ext) which is parallel (right) and antiparallel to the magnetisation direction (left). (b) Schematic illustration of the spin-split on the LUMO of Tb(Pc)₂ on the ferromagnetic Co layer. (c) Spin-resolved point-mode spectroscopy. The parallel (blue line) and the antiparallel magnetisation directions (red line) (above). Difference in the spectroscopic curves (below). (d) The spin polarisation of the molecular orbitals which have opposite spin polarisations: red (U = +1.0 V) and blue (U = +1.3 V). (e) STM topographic image of Tb(Pc)₂/Co/Au(111), which is grouped in islands Co 1 and Co 2. (f) Conductance mapping for Tb(Pc)₂/Co/Au(111) obtained with V_s = −600 mV and B_ext = +1 T (i) and −1 T (ii). (iii) Difference mapping ΔG (+1 T, −1 T). The scale of dI/dV (top) and the ΔG (bottom) are shown in the colour bar scales. (g) dI/dV signal measured as a function of external magnetic field (B_ext) for Tb(Pc)₂/Co/Au(111), which are shown in (f). The hysteresis loops of dI/dV shown in both edge and flat-lying molecular positions. Adapted with permission from (a–d) ref. 379. Copyright 2012, Nature Research. Adapted with permission from (e–g) ref. 299. Copyright 2019, American Institute of Physics.

SMMs and/or molecular nano magnets (MNM) have potential for technological applications in magnetic storage and for implementing qubits in the quantum computing (QC).²⁶⁶ The application in magnetic storage requires high anisotropic energy barriers for spin reversal (U_eff) and QTM suppression, whereas QC requires a QTM process, such as a spin cascade |S = 1/2〉‖J = 6〉‖I = 3/2〉 in the system, for qubit reading and manipulation (the spin cascade mechanism for qubit is described in Section 3.2).²⁷⁰ Therefore, both applications can be achieved using organic–inorganic hybrid materials by developing viable means for coordinating the QTM process in a controlled manner. An approach that uses the hybrid materials to control and enhance the properties of sandwich-type tetrapyrrole rare-earth metal SMMs will accelerate further research towards the realisation of spintronic devices made of functional sandwich-type tetrapyrrole rare-earth metal complexes.

3.8 Rare-earth phthalocyaninato sandwich complex/carbon nanotube hybrid materials

The internal nanospace of carbon nanotubes (CNTs) can be regarded as a reaction vessel that is 1/1000 smaller than a microreactor.⁴⁰⁸ Therefore, these chemical reaction fields, which have a diameter range of 1–3 nm (100 nm to 0.1 mm in length), are expected to have a great influence on reaction chemistry and/or nanoelectronics research in the 21st century.^409–412 It has been shown that atoms and molecules can be contained in the nanospace of CNTs.^413–418 Focusing on the use of the stable one-dimensional nanospace of a CNT as a host for host–guest structures, various SMM–CNT hybrid materials (i.e., with single-walled CNTs (SWCNTs): internal diameter of ca. 3 nm and multi-walled CNTs (MWCNTs): internal diameters of (5–50 nm)) have been reported for nanoelectronics and spintronics research.^{344,419–421}

In 2009, Kyatskaya et al. first reported a [(Pyrene)Pc]Tb(Pc)–SWCNT hybrid material, in which a [(Pyrene)Pc]Tb(Pc) (one Pc ligand is substituted with three hexyl groups and one 4-(4-pyren-1-ylbutoxy) group) was attached to the outside of a SWCNT, along with its magnetic properties.¹¹⁴[(Pyrene)Pc]Tb(Pc)–SWCNT shows a Coulomb blockade in which a Tb(Pc)₂ derivative and a SWCNT act as a quantum dot in a field-effect-transistor (FET) device. In the spin valves of [(Pyrene)Pc]Tb(Pc)–SWCNT, the maximum magnetic resistance ratio between the parallel state and the antiparallel state is approximately 300% at submillikelvin temperatures.^117,312 In 2011, in a related study, Urdampilleta et al. reported hybrid devices that consist of chemical vapor deposition (CVD)-grown CNT transistors decorated with Tb(Pc)₂ SMMs.^269,309 The devices were achieved by tailoring the supramolecular π–π interactions between the Tb(Pc)₂ SMMs and CNTs (Tb(Pc)₂–CNT) (Fig. 35a). Magnetoresistance measurements revealed steep steps, which were related to the magnetisation reversal of the individual Tb(Pc)₂ SMMs. The electron-transport properties of these devices are strongly dependent on the relative magnetisation orientation of the grafted SMMs. In this material, the Tb(Pc)₂ SMMs act as a local spin polarizer and analyser on the CNT electron-conduction channel. As a result, reluctance ratios of up to several hundred percent were observed at submillikelvin temperatures. These results suggest that a spin valve with a larger magnetoresistance ratio can be prepared by using SMMs as the spin source instead of ferromagnets. These systems were demonstrated to detect differential conductance (dI/dV) by taking advantage of the switching of the QTM steps by the corresponding magnetic field (μ₀H_z < ±50 mT). In other words, readout of the nuclear spin state was realized using Tb(Pc)₂ SMMs.

Fig. 35 Supramolecular spin valves based on the Tb(Pc)₂–SWCNT hybrid device. (a) Schematic illustration of the device. (b) Conductance hysteresis loops of Tb(Pc)₂–SWCNT. dI/dV measured at 40 mK with an in-plane magnetic field applied along the easy axial direction (0°). Bias and temperature dependences of the conductance hysteresis loops of Tb(Pc)₂–SWCNT. (c) Scale map of the dI/dV hysteresis as a function of the in-plane magnetic field and source drain voltage (V_ds). (d) Temperature dependence of the conductance hysteresis loops. (e) Supramolecular structure of the Tb(Pc)₂@SWCNT hybrid material. TEM image and schematic illustration of Tb(Pc)₂@SWCNT. (f) STM image showing empty (red dot) and filled (black dot) regions. Comparative STS spectra measured at the red and black dotted points in STM image. The STS of Tb(Pc)₂ adsorbed on an Au(111) surface is illustrated as plot III (blue) in (f) as a reference. Reprinted from (a–d) ref. 269. Copyright 2011, MDPI. Reproduced from (e and f) ref. 422. Copyright 2021, Royal Society of Chemistry.

In 2019, Abd El-Mageed et al. reported a supramolecular structure formed by a Tb(BDP)₂–SWCNT hybrid material, in which Tb³⁺-5,15-bisdodecylporphyrin (BDP) double-decker SMMs were attached to the outside of an SWCNT.⁴²³ Scanning tunneling microscopy (STM) confirmed an ordered, self-organized spiral array on the SWCNT surface stabilized by non-sharing (π-stacking) interactions. In addition, atomic force microscopy (AFM) images exhibited a distinctive debundling effect due to the adsorption of the SMMs on the SWCNT surface with some helical array structures. The SMM properties of Tb(BDP)₂–SWCNT were investigated, and the hybrid materials interestingly showed larger butterfly-shaped hysteresis loops than the SMMs alone, indicating a slow τ for Tb(BDP)₂–SWCNT. So far, electron-transport properties have not yet been reported for Tb(BDP)₂–SWCNT.

In 2021, Katoh et al. first reported the encapsulation of Tb(Pc)₂ SMMs in the internal nanospace of SWCNTs (Tb(Pc)₂SMM@SWCNT).⁴²² The magnetic and electronic properties of the Tb(Pc)₂SMM@SWCNT hybrids (Fig. 35e) were investigated in detail using DC and AC magnetic susceptibility measurements, transition electron microscopy (TEM), scanning electron microscopy (SEM), STM, and STS. By arranging the Tb(Pc)₂ SMMs in the 1D internal nanospaces of the SWCNTs, it is possible to investigate the essential SMM characteristics of Tb(Pc)₂ without considering LF effects. Moreover, it appears that the electron correlation between Tb(Pc)₂ and the SWCNT can affect the electron transport and/or electromagnetic properties. Since bandgap modulation was observed due to the encapsulation of Tb(Pc)₂ in SWCNTs, these materials should be useful in FETs. Furthermore, as the stable internal nanospace of SWCNTs is used, it is thought that the density of the SMMs in the SMM–SWCNT hybrid material can be controlled, and the hybrids should be usable as spin valves. This strategy may pave the way for the construction of SMM–SWCNT hybrid materials for operable molecular spintronic devices using the quantum effects of individual SMMs.

4. Semiconducting materials based on REE sandwich complexes

Discovery of semiconducting properties of phthalocyanines reported by the group of J. Simon in 1987 on the examples of radical complexes LiPc and Lu(Pc)₂ is definitely one of important milestones not only for tetrapyrrolic compounds, but for all organic electronics in general.³⁵ This discovery nearly coincided with the first report on organic field effect transistors (OFETs) made in 1986 by Tsumura et al. with polytiophene conducting layer,⁴²⁴ thus no surprise that the same year Lu(Pc)₂ was introduced into OFET technology.

OFETs attract increasing research interest since they were first reported due to their potential for application in flexible displays, integrated circuits and low-cost electronic devices. A typical OFET device consists of an organic semiconducting layer, a dielectric layer, and three electrodes (source, drain, and gate electrodes). Three key parameters define the performance of an OFET device, namely charge carrier mobility (μ), current on/off ratio, and threshold voltage (V_T). The organic semiconducting layer is the most important part to determining the OFET performance, though other factors, such as the device configuration, the interfaces, and the materials for electrodes and insulators, also can affect the performance.^425,426

π-Conjugated tetrapyrrole derivatives, including porphyrins (Pors) and phthalocyanines (Pcs), are important organic semiconducting materials for OFETs because of their unique molecular and electronic structures.^427,428 For rare-earth based tetrapyrrolic sandwiches, the molecular energy levels are highly disturbed by the intramolecular π–π interactions between tetrapyrrolic rings, typically possessing destabilized HOMOs and stabilized LUMOs, thus having decreased molecular energy gaps in comparison to the constituent monomeric tetrapyrroles. This is desirable for improving the semiconducting properties of tetrapyrrolic materials, as demonstrated by Guillaud et al. for Lu(Pc)₂ and Tm(Pc)₂ based OFETs.⁴²⁹ The molecular energy levels of tetrapyrrolic sandwiches can be further tuned by choosing different ligands and rare-earth metal ions, adjusting the number of deckers, and incorporating various substitutes on tetrapyrrolic ligands.^1,7 The intermolecular interactions and solubility of sandwiches are also adjustable by such chemical modifications, which can provide further opportunity for the fabrication of solution-processable and high quality films for OFET devices.

Moreover, specific substitution of sandwich complexes may allow for the formation of supramolecular structures that provide intrinsic dielectric insulation at the molecular level. For example, ultrathin solid films of structurally similar octa-butoxy and tetra-crown-ether substituted Eu(III) bisphthalocyaninates, Eu[(BuO)₈Pc]₂ and Eu[(15C5)₄Pc]₂, were formed and their transversal conductivity and, consequently, electrochemical activity were studied in an aqueous electrolyte.⁴³⁰ Excellent transfer of charge through monomolecular ultrathin films of crown-substituted complex was observed enabling its redox-multistability, which can be utilized in molecular logic and information storage devices. However, transversal charge transfer was inhibited in the films formed by octa-butoxy-substituted complex, which might be promising for OFET applications.

Depending on the correlation between the molecular energy levels of the semiconducting layer and the work function of source–drain electrodes, holes or electrons can be preferentially injected from the source electrode to the semiconducting layer, and transported to the drain electrode, leading to p- or n-type transporting modes, respectively. If both types of charge carriers can be injected and transported, an ambipolar OFET device is achieved, which, however, is of great importance in complementary metal oxide semiconductor (CMOS) logic circuits.⁴³¹ Most tetrapyrrolic sandwiches can exhibit p-type OFET performance in air with usual source–drain electrodes (such as Au, with a work function of 5.1 eV), because of the suitable HOMO energy levels of sandwiches for hole injection (Table 1). However, unipolar n-type operations are often detected under an inert (or reducing) gas atmosphere, because of the relatively high LUMO energy levels of sandwiches, which cannot resist the air-derived electron traps. On the other hand, due to the versatility of tuning the HOMO and LUMO energy levels of tetrapyrrolic sandwiches by molecular design,⁴³² single-component air-stable ambipolar OFET devices based on tetrapyrrolic sandwiches have become the research focus in recent years (Table 2).

Table 1 Performance of sandwich complexes based p-type OFET devices

Compounds/deposition method	Hole mobility [cm^2 V^−1 s^−1]	V T [V]	I on/Ioff	L and W [μm]	Source–drain electrode	Gate electrode	Insulator	Ref.
[(OctO)8Pc]Tb(Pc)/LB	6.4 × 10^−4		5.5 × 10^4	240 and 28 600	Au	Si	SiO2	433
[(OctO)8Pc]Lu(Pc)/LB	1.7 × 10^−3		10^2
Lu[(α-Oct)8Pc]2/SC	1.5 × 10^−3	−25	10^3	10 and 2000	Au/Cr	n-Type Si	SiO2/OTS	434
Lu[(α-Oct)8Pc]/SC, TA	8.0 × 10^−3	−12.5	10^5	10 and 2000	Au/Cr	n-Type Si	SiO2/OTS
Gd[(α-C8H17)8Pc]2/SC	7.47 × 10^−5	−28.9	10^3	5 and 20 000	Au	Si	SiO2/OTS	435
Gd[Pc(α-C8H17)8]2/SC	1.96 × 10^−5	−27.6	2 × 10^2	20 and 20 000	Au	Si	SiO2/OTS
[(OctO)8Pc]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.60	—	1.4 × 10^5	240 and 28 600	Au	Si	SiO2	436
[(OctO)8Pc]Ho[(15C5)4Pc]Ho[(15C5)4Pc]/LB	0.40	—	—			Si
[(OctO)8Pc]Lu[(15C5)4Pc]Lu[(15C5)4Pc]/LB	0.24	—	—			Si
[(BuO)8Pc]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.0032	—	9.3 × 10	240 and 28 600	Au	Si	SiO2	437
[(HexO)8Pc]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.014	—	8.2 × 10^3
[(DecO)8Pc]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.21	—	9.7 × 10^4
[(C12H25O)8Pc]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.053	—	3.9 × 10
[(HexO)8Pc]Eu[(DEGOMe)8Pc]Eu[(DEGOMe)8Pc]/LB	0.46	—	1.01 × 10^2	240 and 28 600	Au	Si	SiO2	438
[(OctO)8Pc]Eu[(DEGOMe)8Pc]Eu[(DEGOMe)8Pc]/LB	0.17	—	1.02 × 10^3
[(DecO)8Pc]Eu[(DEGOMe)8Pc]Eu[(DEGOMe)8Pc]/LB	0.096	—	3.85 × 10^4
[(C12H25O)8Pc]Eu[(DEGOMe)8Pc]Eu[(DEGOMe)8Pc]/LB	0.014	—	1.15 × 10^2
[(OctO)8Pc]Eu[(12C4)4Pc]Eu[(12C4)4Pc]/LB	0.03	—	2.32 × 10^3	240 and 28 600	Au	Si	SiO2	439
	0.31	—	8.59 × 10^4				SiO2/HMDS
	0.33	—	7.91 × 10^5				SiO2/OTS
[(OctO)8Pc]Eu[(18C6)4Pc]Eu[(18C6)4Pc]/LB	0.02	—	5.5 × 10^3	240 and 28 600	Au	Si	SiO2
	0.28	—	1.73 × 10^5				SiO2/HMDS
[(Dec)4Por]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.78	−3.84	2.20 × 10^4	240 and 28 600	Au	Si	SiO2/HMDS	440
	0.04	−4.02	4.94 × 10^3				SiO2/OTS
[(p-PentOPh)4Por]Eu[(15C5)4Pc]Eu[(15C5)4Pc]/LB	0.05	−1.19	1.48 × 10^4				SiO2/HMDS
	0.03	−4.34	1.03 × 10^4				SiO2/OTS
Tb2[(BuO)8Pc]3/DC	1–2 × 10^−5	—	—	20 and 150	—	Si	SiO2	441

---

Table 2 Performance of sandwich complexes based ambipolar-type OFET devices

Compounds/deposition method	Hole mobility [cm^2 V^−1 s^−1]	I on/Ioff	Electron mobility [cm^2 V^−1 s^−1]	I on/Ioff	−VT [V]	L and W [μm]	Source–drain electrode	Gate electrode	Insulator	Ref.
Y(Pc)2/DFT	0.034		0.031	—	—	—	Au	—	—	432
La(Pc)2/DFT	0.17		0.088	—	—	—	Au	—	—
[(PhO)8Pc]Ho(Pc)/QLS film	1.7 × 10^−4 in air		—	10^3	−21	240 and 28 600	Au	Si	SiO2	442
Ho[(PhO)8Pc]2/QLS film	—		0.54 in N2	10^4	12	240 and 28 600	Au	Si	SiO2
Eu[(NaphO)8Pc]2/QLS film	0.10 in air	10^4	0.01 in N2	10^2		240 and 28 600	Au	Au	PMMA	38
Eu[(SPh)8Pc]2/QLS film	1.1 × 10^−4 in air	—	0.11 in N2	—	—	—	Au	Si	SiO2/HMDS	443
Ho[Pc(SPh)8]2/QLS film	4.9 × 10^−7 in air		3.0 × 10^−3 in N2
[(PhS)8Pc]Eu(Pc)/QLS film	4.4 × 10^−3
[(PhS)8Pc]Ho(Pc)/QLS film	2.6 × 10^−3
(Pc)Eu[f-Pc2]Eu(Pc)^OBu/QLS film	0.8 in air	—	2.3 in N2	—	−3 + 12	240 and 28 600	Au	Si	SiO2/HMDS	221
(TPP)Eu[(PhO)8Pc]Eu[(PhO)8Pc]/QLS film	0.04 in air	10^5	0.08 in N2	10^5	—	240 and 28 600	Au	Si	SiO2/HMDS	444
Eu2[(p-FPhO)8Pc]3/QLS film	0.24	—	0.042	—	−1.1 + 13	W = 28.6 mm, L = 240 μm, W/L = 119	Au	Si	SiO2/HMDS	44
[(PhO)8Pc]Eu[(PhO)8Pc]Eu(Pc)/QLS film	0.68	—	0.014	—	—	240 and 28 600	Au	Si	SiO2/HMDS	42
Eu2[(PhO)8Pc]3/QLS film	0.041	—	0.0026	—	—
Eu2[(NaphO)8Pc]3/QLS film	2.5 × 10^−6	10^2	2.3 × 10^−6	10^2	−8 + 10	240 and 28 600	Au	Si	SiO2/HMDS	40
Solvent-vapor annealing 60 °C	2.6 × 10^−7	10^2	2.3 × 10^−8	10^2	−1 + 3
Solvent-vapor annealing 80 °C	4.9 × 10^−4	10^3	7.1 × 10^−5	10^2	−20 + 18
Solvent-vapor annealing 100 °C	0.11	10^5	0.06	10^5	−3 + 9
Solvent-vapor annealing 120 °C	1.1 × 10^−2	10^4	9.6 × 10^−3	10^4	−4 + 5
[(PhO)8Pc]Eu[(PhO)8Pc]Eu[(C≡CCOOH)Por]/QLS film	7.0 × 10^−7	10^2	7.5 × 10^−7	10^2	−8 + 8	240 and 28 600	Au	Si	SiO2/HMDS	445
[(PhO)8Pc]Eu[(PhO)8Pc]Eu[(C≡CCOOH)Por]/nanoribbon	0.11	10^6	4 × 10^−4	10^4	−3 + 17
CdEu2[(HexS)8Pc]4/QLS film	5.6 × 10^−3	10	1.2 × 10^−3	10	—	—	—	—	—	206
(Pc)2Eu2[f-Pc2]Eu2(Pc)2^SHex/single crystal microsheets	18	10^3–10^4	0.3	10^3–10^4	—	—	Au	Si	SiO2/HMDS	223
[(Py4Por)Eu[(PhO)8Pc]Eu[(PhO)8Pc]]/QLS film	6.9 × 10^−7	10^2	3.07 × 10^−5	10^2	—	—	Au	Si	SiO2/HMDS	446
[(Py4Por)Eu[(PhO)8Pc]Eu[(PhO)8Pc]]/CdS hybrid QLS film	0.09	10^4	0.01	10^3	—
(TClPP)Y(Pc) and fullerene C60 cocrystals	3.72	10^2	2.22	10^2	—	240 and 28 600	Au	Si	SiO2	447
(TFPP)Eu[(PhO)8Pc]Eu[(PhO)8Pc]/QLS film	6.0 × 10^−5	—	1.4 × 10^−4	—	−20 + 38	240 and 28 600	Au	Si	SiO2	448
(TFPP)Eu[(PhO)8Pc]Eu[(PhO)8Pc]/QLS film/CuPc VCD film:	0.16	—	0.30	—	−13 + 25
[(p-HOPh)4Por]Eu(Pc)Eu(Pc)/QLS film	4.85 × 10^−8	10^2–10^3	2.33 × 10^−8	10^2–10^3	—	240 and 28 600	Au	Si	SiO2/HMDS	449
[(p-HOPh)4Por]Eu(Pc)Eu(Pc)/rGO/QLS film	25.53	10^3–10^4	0.77	10^3–10^4	—
[(p-HOPh)4Por]Eu[(PhO)8Pc]Eu[(PhO)8Pc]/QLS film	2.14 × 10^−6	10^2–10^3	2.33 × 10^−8	10^2–10^3	—
[(p-HOPh)4Por]Eu[(PhO)8Pc]Eu[(PhO)8Pc]/rGO/QLS film	30.9	10^3–10^4	39.6	10^3–10^4	—

---

4.1 p-Type OFETs

In 2005, Su et al. reported LB film-based OFETs by using heteroleptic bis(phthalocyaninato) rare earth complexes [(OctO)₈Pc]M(Pc), M = Tb and Lu as the electroactive components, which showed hole mobilities of 6.4 × 10⁻⁴ and 1.7 × 10⁻³ cm² V⁻¹ s⁻¹, respectively, on a SiO₂ dielectric layer with top contacted Au source–drain electrodes.⁴³³ Chen et al. designed amphiphilic tris(phthalocyaninato) rare earth complexes [(OctO)₈Pc]M[(15C5)₄Pc]M[(15C5)₄Pc], M = Eu, Ho and Lu, and fabricated LB film-based OFETs with SiO₂/Si substrates and top contacted Au source–drain electrodes.⁴³⁶ The devices exhibited highly improved hole mobilities in the range of 0.24–0.60 cm² V⁻¹ s⁻¹ with an on/off current ratio of 10⁵, which was ascribed to the intramolecular π–π stacking among Pc rings and the J-aggregation in the LB films.⁴³⁷

In order to study the substituent effect on the assembling properties and OFET performance of tris(phthalocyaninato) triple-deckers [(C_nH_2n+1O)₈Pc]Eu[(15C5)₄Pc]Eu[(15C5)₄Pc], n = 4, 6, and 8, different lengths of alkoxy chains were connected to one of the outer phthalocyanine ligands.⁴³⁷ The results indicated that the molecular packing and intermolecular interaction in LB films are closely related to the length of hydrophobic side chains, thus leading to the change of hole mobilities within 0.0032–0.60 cm² V⁻¹ s⁻¹.

A similar trend was observed for another series of amphiphilic tris(phthalocyaninato) rare earth complexes [(C_nH_2n+1O)₈Pc]Eu[(DEGOMe)₈Pc]Eu[(DEGOMe)₈Pc], DEGOMe = (OC₂H₄)₂OCH₃, n = 6, 8, 10 and 12.⁴³⁸ In the OFET devices with bottom contacted Au source–drain electrodes, the LB films of those triple-deckers also displayed hole mobilities in the range of 0.014–0.46 cm² V⁻¹ s⁻¹, depending on the length of hydrophobic alkoxy chains.

The hydrophilic sides of tris(phthalocyaninato) rare earth complexes were also modified with different crown ether substituents.⁴³⁹ Replacing the 15-crown-5 rings of [(OctO)₈Pc]M[(15C5)₄Pc]M[(15C5)₄Pc] (Fig. 36a) with 12-crown-4 or 18-crown-6 moieties resulted in different packing modes in LB films. Thus, 12-crown-4 and 15-crown5-substituted complexes formed edge-to-edge connected J-aggregates, providing a charge transfer channel parallel to the Pc rings. In contrast, 18-crown-6-substituted complex formed face-to-face connected H-aggregates in LB films, providing a channel along the long axis of the assemblies (Fig. 36b and c). In addition, the authors found that the surface treatment of the SiO₂/Si substrate by hexamethyldisilazane (HMDS) or octadecyltrichlorosilane (OTS) can significantly improve the quality of the LB film, thus leading to an increased carrier mobility.

Fig. 36 (a) Amphiphilic Eu(III) crown-phthalocyaninates [(OctO)Pc]Eu[(crown)₄Pc]Eu[(crown)₄Pc] and schematic arrangement of complexes [(OctO)₈Pc]Eu[(12C4)₄Pc]Eu[(12C4)₄Pc] – (b) and [(OC₈H₁₇)₈Pc]Eu[(18C6)₄Pc]Eu[(18C6)₄Pc] – (c) in the LB film-based OFET devices. Adapted with permission from ref. 439. Copyright 2009, American Chemical Society.

Mixed-ligand (phthalocyaninato)(porphyrinato) Eu(III) triple-deckers (Por*)Eu[(15C5)₄Pc]Eu[(15C5)₄Pc] where Por* = tetra-meso-decylporphyrinate or tetra-meso-(p-pentyloxyphenyl)porphyrinate were synthesized and tested in LB film-based OFET devices.⁴⁴⁰ These triple-deckers adopt a “face-to-face” connected “edge-on” orientation on HMDS- or OTS-treated SiO₂/Si surfaces, leading to the effective intramolecular π–π stacking in films. As a result, the hole mobility of OFET devices can reach 0.78 cm² V⁻¹ s⁻¹, with an on/off ratio of 2.2 × 10⁴ and a threshold voltage of −3.84 V.

In addition to LB films, drop-casted and spin-coated films of tetrapyrrolic sandwiches have also been applied for OFET devices. By using spin-coated films of a liquid crystalline bis(phthalocyaninato) lutetium complex Lu[(α-Oct)₈Pc]₂, Ray et al. fabricated bottom gated OFET devices, which showed a hole mobility of 1.5 × 10⁻³ cm² V⁻¹ s⁻¹ with an on/off current ratio of 10³ at a threshold voltage of −25 V.⁴³⁴ After thermal annealing, the hole mobility was improved to 8.0 × 10⁻³ cm² V⁻¹ s⁻¹, together with an increased on/off ratio by two orders of magnitude and a reduced threshold voltage to half of that of the as-deposited films. Recently, through the study of spin-coated films of the gadolinium complex Gd[(α-Oct)₈Pc]₂,⁴³⁵ Ray et al. revealed that the OFET performance decreases along with scaling down the channel length of devices. The electronic transport properties of drop-casted films of a triple-decker complex Tb₂[(BuO)₈Pc]₃ were reported by Yamashita et al., which also show p-type mobilities in the range of 1–2 × 10⁻⁵ cm² V⁻¹ s⁻¹.⁴⁴¹

4.2 Ambipolar OFETs

The ambipolar semiconducting properties of unsubstituted bis(phthalocyaninato) rare earth complexes were initially reported by Simon and co-workers in 1990s.⁴⁵⁰ The high intrinsic conductivity of Lu(Pc)₂ and Tm(Pc)₂ was ascribed to the narrow molecular energy gaps associated with their radical nature.⁴²⁹ The ambipolar charge transfer properties of M(Pc)₂, M = Y or La were also studied by density functional theory (DFT) calculations, which suggest that the delocalized radical can induce a high HOMO level and a low LUMO level, leading to a small ionization potential and a large electronic affinity. As a result, small injection barriers for both hole and electron from source–drain electrodes (Au) can be expected for these neutral double-deckers.⁴³² The DFT study also revealed small charge reorganization energy and large transfer integral values in the crystal of M(Pc)₂, thus predicted promising ambipolar charge transfer properties. However, in experimental studies, single-component bis(phthalocyaninato) rare earth complex-based OFET devices usually show p-type transporting properties in air, the corresponding n-type operations only can be detected under an inert or reducing atmosphere.

By connecting electron withdrawing phenoxyl groups onto one or both Pc rings of rare earth double-deckers, heteroleptic [(PhO)₈Pc]Ho(Pc) and homoleptic Ho[(PhO)₈Pc]₂ were obtained, which exhibit p- and n-type transporting properties, respectively, in the quasi-Langmuir–Shäfer (QLS) film-based OFET devices.⁴⁴² Introducing other electron withdrawing groups, such as naphthoxyl³⁸ and phenylthiol,⁴⁴³ to the Pc rings of rare earth double-deckers can lead to transforming their semiconducting nature from p-type to ambipolar type by tuning their HOMO/LUMO energy levels.³⁷ A dimeric phthalocyanine ligand was incorporated into rare earth sandwiches, leading to the formation of a yttrium(III) double-decker dimer (Pc)Y[f-Pc₂]Y(Pc)^OBu (cf. Section 2.6.2),²¹⁶ which exhibits a unique biradical state and can be fabricated into high quality QLS films, thus displaying very high carrier mobilities of 2.3 and 0.8 cm² V⁻¹ s⁻¹ for electrons and holes in air and N₂, respectively (Table 2).²²¹

Molecular orbital (MO) levels of sandwiches can be adjusted by the addition of another tetrapyrrolic ring to double-deckers. In this way, a mixed (phthalocyaninato)(porphyrinato) europium(III) triple-decker complex (TPP)Eu[(OPh)₈Pc]Eu[(OPh)₈Pc] was prepared and fabricated into QLS film-based OFET devices, which displayed balanced hole and electron mobilities in air and N₂, respectively.⁴⁴⁴

Tetrapyrrolic sandwich-based air-stable ambipolar OFET devices were first achieved by using a partially fluorinated tris(phthalocyaninato) europium complex Eu₂[(p-FPhO)₈Pc]₃.⁴⁴ The introduction of electron-withdrawing 4-fluoro-phenoxy groups on the peripheral sites resulted in finely controlled HOMO (−5.27 eV) and LUMO (−4.17 eV) levels; both facilitate charge carrier injection from Au electrodes. In particular, the LUMO level is suitable for electron transport under ambient conditions. The triple-decker molecules also form H-aggregates and adopt an “edge-on” conformation in the QLS films, which can provide intense intermolecular interaction and π-electron delocalization. As a result, the QLS film-based single-component OFET devices display an average hole mobility of 0.24 cm² V⁻¹ s⁻¹ and an average electron mobility of 0.042 cm² V⁻¹ s⁻¹ in air, with relatively low threshold voltages of −1.1 and +13 V for holes and electrons, respectively. By taking a similar optimizing strategy, tris(phthalocyaninato) europium complexes [(PhO)₈Pc]Eu[(PhO)₈Pc]Eu(Pc), Eu₂[(PhO)₈Pc]₃ and Eu₂[(NaphO)₈Pc]₃ were obtained. The HOMO and LUMO levels of these triple-deckers were also successfully tuned into the range for air-stable ambipolar semiconductors by the introduction of electron-withdrawing phenoxy⁴² or naphthoxy subsititutes.⁴⁰

Triple-decker complex Eu₂[(NaphO)₈Pc]₃ was used to fabricate ambipolar OFET device by QLS technique, and nanoparticles of the pristine QLS film could be converted into well-defined microstructures using solvent vapour annealing with o-dichlorobenzene (Fig. 37).⁴⁰ Fragment-like irregular polygons, quasi-four-corner sheets, 2D squares, and clovers were obtained at different temperatures (60, 80, 100 and 120 °C, respectively), providing access to the tuning of OFET device performance due the carrier mobilities depending on morphologies and crystallinities. The best result was achieved for the most crystalline 2D squares-based devices with the carrier mobilities of 0.11 ± 0.03 cm² V⁻¹ s⁻¹ for holes and 0.06 ± 0.02 cm² V⁻¹ s⁻¹ for electrons.

Fig. 37 The carrier mobilities for holes and electrons in pristine QLS film of Eu₂[(NaphO)₈Pc]₃ and microstructures obtained by solvent vapour annealing at different temperatures together with SEM images of the corresponding films. Adapted from ref. 40, Copyright (2018) with permission Elsevier.

A further modification was performed by replacing one of the outer Pc ligands with an amphiphilic low-symmetry porphyrin ring (Fig. 38a).⁴⁴⁵ The resulting triple-decker can form 1-D nanoribbons via a “phase-transfer” process (Fig. 38b), which show improved air-stable ambipolar semiconducting properties with mobilities of 0.11 and 4 × 10⁻⁴ cm² V⁻¹ s⁻¹ for holes and electrons, respectively, 3 to 6 orders higher than those of non-crystalline QLS film (Fig. 38c).

Fig. 38 (a) Amphiphilic mixed-ligand Eu(III) complex and SEM images of films showing crystallinity of films formed by the “phase transfer” process – (b) and non-crystalline films obtained by the quasi-Langmuir–Shäfer process – (c) Adapted with permission from ref. 445. Copyright 2016, American Chemical Society.

Further extension of sandwiches was beneficial for their OFET behaviour. Thus a quadruple-decker complex Eu₂Cd[(HexS)₈Pc]₄ exhibited air-stable ambipolar semiconducting characters due to the extended π-system and narrowed bandgap²⁰⁶ The best result for phthalocyanine-based OFETs was achieved with a dimeric phthalocyanine ligand involved sandwich-type triple-decker complex (Pc)₂Eu₂[f-Pc₂]Eu₂(Pc)₂^SHex.²²³ This special sandwich can be fabricated into single crystal microsheets, which display high hole and electron mobilities of 18 and 0.3 cm² V⁻¹ s⁻¹, respectively, with on/off ratios of 10³–10⁴.

Another approach for ambipolar OFETs is the fabrication of bicomponent films and/or cocrystals with triple-decker complexes as one of the components. For example, bicomponent thin film transistors of (TPyP)Eu[(OPh)₈Pc]Eu[(OPh)₈Pc]/CdS show highly improved charge mobilities for holes (0.09 cm² V⁻¹ s⁻¹) and electrons (0.01 cm² V⁻¹ s⁻¹) relative to those of single-component thin film transistors of (TPyP)Eu[(PhO)₈Pc]Eu[(PhO)₈Pc] for holes (6.9 × 10⁻⁷ cm² V⁻¹ s⁻¹) and electrons (3.07 × 10⁻⁵ cm² V⁻¹ s⁻¹).⁴⁴⁶ OFET devices based on the cocrystals of (TClPP)Y(Pc)Y(Pc) and C₆₀ also display very high mobilities of 3.72 and 2.22 cm² V⁻¹ s⁻¹ for holes and electrons, respectively, which is proposed to be due to similar LUMO energy levels of the two components and the low reorganization energy values for holes and electrons at the interface.⁴⁴⁷ On the other hand, ambipolar OFETs have been achieved with a bilayer heterojunction configuration by using triple-decker complexes as air-stable ambipolar semiconductors. The QLS films of (TFPP)Eu[(PhO)₈Pc]Eu[(PhO)₈Pc] were obtained on vacuum deposited (VCD) CuPc templates, and fabricated into two-component bilayer thin film transistors.⁴⁴⁸ The devices possess charge mobilities of 0.16 and 0.30 cm² V⁻¹ s⁻¹ for holes and electrons, respectively, which are 10³–10⁵ times of those of the single-component based devices. The improvement of the performance can be attributed to the heterojunction effect and the H-aggregation of triple-deckers on CuPc films. Bilayer heterojunction FETs were also fabricated by QLS films of [TP(OH)P]Eu(Pc)Eu(Pc) and [TP(OH)P]Eu[(PhO)₈Pc]Eu[(PhO)₈Pc], as the top layers, and reduced graphene oxide (rGO) film as the sublayers. The obtained devices exhibit very high and balanced carrier mobilities of 30.9 and 39.6 cm² V⁻¹ s⁻¹ for holes and electrons, respectively, again due to the heterojunction effect.⁴⁴⁹

4.3 Supramolecular assembling of semiconductive materials based on sandwich crown-substituted complexes

The relationship between morphology and conductivity of layers formed by tetrapyrrole sandwiches in electronic devices makes it possible to apply supramolecular tools to control these factors. For this purpose, molecular blocks capable of exhibiting various types of intermolecular interactions, such as hydrophobic and stacking interactions, as well as hydrogen bonds, were synthesized and studied. However, the assembly of such molecules is a trial-and-error method, since the morphology and conductivity of the film cannot be predicted a priori. To some extent, this problem can be solved by applying cation-induced assembly of crown-substituted phthalocyanines, where well-established cooperative effects govern the formation of well-defined supramolecular architectures (cf. Section 3.3.1).

Thus, Zvyagina et al. used homoleptic crown-substituted bisphthalocyaninates M[(15C5)₄Pc]₂, M = Ce, Tb, and Lu as potential molecular blocks for cation-induced formation of 1D supramolecular polymers {M[(15C5)₄Pc]₂·4⁺}_n where enhanced conductivity was expected (Fig. 39a).⁴⁵¹ Noteworthy, this possibility was envisaged in 1994 by Toupance et al. in one of keystone papers on the construction of ionoelectronic materials based on crown-substituted phthalocyanines; however, experimental verification of this hypothesis was reported only in 2021.

Fig. 39 (a) Modes of interaction of M[(15C5)₄Pc]₂ with potassium cations – intramolecular for M = Tb and Ce vs. intermolecular for M = Lu. (b) The current−voltage characteristic of the drop-cast films of the assemblies formed by M[(15C5)₄Pc]₂ and KBPh₄. Inset shows SEM image of nanowires formed by Lu(III) complex. (c) Schematic illustration of the mechanism of alignment of the {Lu[(15C5)₄Pc]₂·4KBPh₄}_n nanowires. Adapted with permission from ref. 451. Copyright 2021, American Chemical Society.

Indeed, interaction of Lu(III) complex with KBPh₄ in chloroform–acetonitrile mixture resulted in the intermolecular binding of potassium cations which connect neighbouring double-deckers in a cofacial manner, yielding nanowires with a thickness of 10–50 nm and a length of up to 50 μm. The four-probe technique in a four-stripe layout revealed the superior conductivity of these wires up to 11.4 S cm⁻¹, the highest among them being reported for phthalocyanine assemblies (Fig. 39b). The electric properties of these 1D assemblies made it possible to use technologically feasible methods for their controlled alignment on surfaces under an applied electric field (Fig. 39c).

In striking contrast, cerium- and terbium-based complexes underwent mainly intramolecular cation binding without the formation of conductive nanowires (Fig. 39a and b). Apparently, this is associated with a larger interligand distance in these complexes allowing potassium cations to intercalate between crown-substituted ligands.⁴⁵² The films, formed by the resulting assemblies had 8 orders of magnitude lower conductivity due to the lack of intermolecular interactions.

The way to controllable immobilization of supramolecular assemblies onto the electroactive surfaces was paved by Shokurov et al., who used the octopus-like double-decker anchor, octopus-Pc (Scheme 4) which can form self-assembled monolayer (SAM) on the surface of Au electrodes (Fig. 40).⁷³ Face-on orientation of the anchors allowed for the subsequent binding of crown-phthalocyanine molecules H₂[(15C5)₄Pc]via potassium ion bridges. This chemistry was utilized to form a heterogeneous bilayer, in which a single molecule thick adlayer brought an additional redox-state to the system, thus expanding the multistability of the system as a whole. All four redox states available to this system exhibited characteristic absorbance in the visible range, allowing for the switching to be easy read out using optical density measurements. It is expected that the proposed approach can be used in a wide range of switchible materials – single-molecule magnets, conductive and optical devices, etc.

Fig. 40 Formation of a bilayer composed of H₂[(15C5)₄Pc] molecules appended to octopus-Pc SAM on gold via supramolecular assembly. Adapted with permission from ref. 73. Copyright 2022, Wiley-VCH.

5. Non-linear optical materials based on REE sandwich complexes

Metalloporphyrins (MPor) and metallophthalocyanines (MPcs) are important organic nonlinear optical (NLO) materials due to their two-dimensional π-system.^453,454 Among practically relevant properties optical limiting (OL) is particularly interesting as it can be used to protect eyes from damage caused by laser irradiation.

Upon the formation of sandwich complexes with rare earth metals, the π-conjugation systems can be further extended along the axial direction of tetrapyrrol plane, leading to three-dimensionally conjugated molecules, which is desirable for improving the NLO response by regulating the intramolecular charge transfer transitions and/or constructing dipolar and multipolar systems.³

5.1 Second-order nonlinear optical materials

For materials having second-order NLO responses at the molecular level, noncentrosymmetric molecules with strong intramolecular dipolar charge distribution are required. Heteroleptic double-decker tetrapyrrolic complexes generally meet this requirement. However, reports in this field are still very rare. The sole example of heteroleptic double-decker complexes studied for the second-order NLO properties is a mixed porphyrin and tetraazaporphyrin zirconium(IV) complex (OEP)Zr(OETAP).⁴⁵⁵ A negative hyperpolarizability β [(−83 ± 43) × 10⁻³⁰ esu] was obtained at 1907 nm by the electric field-induced second harmonic (EFISH) measurements.

An interesting trial for second-order NLO materials has been extended to homoleptic double-decker complexes with “cubic” octupolar architectures (cf. Section 2.2.2 for structures of complexes). In this regard, Ayhan et al. developed ABAB homoleptic bis(phthalocyaninato) rare earth complexes M[(SHex)₄Pc]₂, M = Nd, Eu, Dy, Y and Lu,^109,110 which represent a new class of cubic molecules with an octupolar charge distribution. Their second-order NLO properties were studied by Hyper–Rayleigh Scattering (HRS) measurements at 1907 nm in chloroform solution. The first molecular hyperpolarizabilities (β_HLS) were evaluated to be in the range of 3010 × 10⁻³⁰ esu to 5760 × 10⁻³⁰ esu for Y(III) and Lu(III) complexes, respectively, which are among the largest values ever reported for nondipolar molecules. Furthermore, it was found that the β_HLS values of these complexes are depended on the number of 4f electrons and ionic radii of the sandwiched lanthanide(III) ions, thus providing opportunity to regulate the second-order NLO responses. Another ABAB homoleptic bis(phthalocyaninato) rare earth complex Lu[(α-PentO)₄Cl₄Pc]₂, containing alternating pentyloxy and chloro substituents on phthalocyanine ligands, was reported by Cao et al.,¹¹¹ which also showed a high molecular hyperpolarizability (β_HLS = 1240 × 10⁻³⁰ esu at 1907 nm, in chloroform).

By applying a porphyrin ligand with di(phenylamino)phenyl and carboxymethyl–phenylethynyl meso-substituents as typical electron-donating and electron-accepting groups, respectively, Sun et al. synthesized the first octupolar bis(porphyrinato) rare earth complex [HTb(DADAPor)₂] (Scheme 15).¹¹³ Because of the reduced protonated nature, this compound does not have any absorption in the near-IR range beyond 800 nm. Consequently, an off-resonant hyperpolarizability (β_HLS = 1700 × 10⁻³⁰ esu at 1907 nm, in chloroform) was obtained for this compound, which is the largest off-resonant hyperpolarizability for octupolar molecular materials.

Density functional theory (DFT) studies were also conducted to clarify the nature of the second-order NLO properties of octupolar homoleptic bis(phthalocyaninato) rare earth complexes.⁴⁵⁶ A series of cubic molecules based on the bis(phthalocyaninato) yttrium Y(Pc)₂ skeleton, containing two couples of electron-donating and electron-withdrawing groups in an alternate manner (Fig. 41a), were investigated. According to the calculation results: (1) the static hyperpolarizability (β_HLS, 10³–10⁴ a.u.) increases along with increasing electronegativity of the peripheral substituents; (2) a dynamic perturbation (excitation) will induce a resonance effect with a much higher dynamic β_HLS (10⁴–10⁶ a.u.), which is also positively correlated with the electronegativity of peripheral substituents; (3) the molecular conformation affects the second-order NLO properties. The maximal static β_HLS appears at a rotation angle θ of 70°, but not 90°, because Y(Pc)₂ molecule employs a cuboid instead of a standard cube (Fig. 41b and c). These findings are useful for the design of new octupolar NLO materials, given that the theoretically predicted maximal β_HLS values can reach 10⁵ × 10⁻³⁰ esu for such octupolar molecules.

Fig. 41 (a) Molecular structures of the series of bis(phthalocyaninato) yttrium double-decker skeletons used for DFT modelling. (b) Plot of the calculated static β_HRS as a function of the rotation angle θ for the complex Y(Pc4)₂. (c) Cuboid structure of the bis(phthalocyaninato) yttrium double-decker skeleton with a rotation angle of 90°. Adapted with permission from ref. 456. Copyright 2015, Wiley-VCH.

5.2 Third-order nonlinear optical materials

The tetrapyrrol macrocycles are very suitable for the third-order NLO materials due to three special advantages, including (1) the large number of free π electrons playing the role of oscillators driven by an external periodically vibrating electromagnetic field; (2) the extended delocalized π-system acting as a broad oscillating space; and (3) the huge polarizing/hyperpolarizing moments due to the long distance between the peripheral electron-donating and withdrawing substituents. The double-decker and multi-decker tetrapyrrols further improves the total three advantages in comparison with the single-decker ones.

The study of third-order NLO materials based on double-decker phthalocyanines started from 1990s. Shirk and coworkers investigated the third-order NLO performance of double-deckers M(Pc)₂ (M = Sc, Y, Eu, Gd, Lu) using the degenerate four-wave mixing (DFWM) method.⁴⁵⁷ The second-order hyper-polarizabilities (γ) and third-order susceptibility χ⁽³⁾ values of M(Pc)₂ are about an order of magnitude higher than those of typical monomeric metallophthalocyanines. Among these double-deckers, the third-order NLO performance of Sc(Pc)₂, Lu(Pc)₂ and Yb(Pc)₂ are better than others. This therefore suggests that the large increase in the hyperpolarizabilities of bis(phthalocyaninato) complexes over those of the monophthalocyaninato compounds could not be simply attributed to the extended delocalization length from monomers to double-deckers.

In the past 20 years, the Z-scan method was widely used in most third-order NLO experiments. Sheng and coworkers studied a series of homoleptic bis(phthalocyaninato) rare earth complexes M[(OctO)₈Pc]₂, M = Pr, Ce, Sm, and Lu.^458–460 The NLO performance of these compounds are tested using the exciting laser wavelength of 800 nm (Z-scan, 170 ns). The two-photon absorption cross-section σ_2PA values of this series are in the range of (3–254) × 10³ GM, showing their strong nonlinear optical response. More importantly, the authors measured the real part of the second-order molecular hyperpolarizability (γ_R) of the neutral complex Pr[(OctO)₈Pc]₂⁰˙ together with its anion Pr[(OctO)₈Pc]₂⁻ and cation Pr[(OctO)₈Pc]₂⁺. They found that the γ_R values of neutral, anionic and cationic forms are 440 × 10⁻³², 180 × 10⁻³², and 220 × 10⁻³² esu, respectively.⁴⁶⁰ This is a key experimental phenomenon which shows the free single electron in the neutral radical molecule can be used as a good oscillator in the NLO response. When this free single electron is paired in the anionic form or removed in the cationic form, the NLO response will be significantly weakened. Following a similar mentality, Plekhanov et al. examined another complex Lu[(SHex)₈Pc]₂.⁴⁶¹ Using a similar Z-scan test (300 ns) at an exciting laser wavelength of 1550 nm, the complex achieved a remarkable nonlinear refractive index (n₂) value of (300 ± 3) × 10⁻⁵ cm² GW⁻¹. They found that the n₂ for unsubstituted complex Lu(Pc)₂ is only (228 ± 3) × 10⁻⁵ cm² GW⁻¹, which is a quarter lower than that of Lu[(SHex)₈Pc]₂. This comparative investigation revealed that the peripheral electron-withdrawing substituents are beneficial to the NLO response because they could provide more free electrons to the π conjugated system.

Karpo et al. studied the NLO properties of a series of butyl-substituted complexes M(Bu₈Pc)₂, M = Eu, Dy, Er, and Lu in THF solution using the open aperture z-scan technique with 350 ps laser pulses at 532 nm.⁴⁶² All the compounds exhibited reverse saturable absorption (RSA), and application or rate equations formalism evidenced that RSA is mainly contributed by the triplet state absorption. Absorption cross-sections from the ground state (σ₀), excited singlet and triplet (σ₁ and σ₂) were deduced, evidencing that σ₂ ≫ σ₀ > σ₁, the largest ratio σ₂/σ₀ was found for erbium complex.

Sekhosana et al. conducted a series of comparative studies on the NLO properties of double-decker phthalocyanines with electron-withdrawing pyridin-4-yloxy substituents (PyO).^47,463–466 They found that the imaginary part of the third-order susceptibility χ⁽³⁾_I value is 106 × 10⁻¹² esu for Yb[(α-OPy)₄Pc]₂ and 338 × 10⁻¹² esu for Yb[(β-OPy)₄Pc]₂ (Y20) under the Z-scan test (10 ns, in DMSO) at the wavelength of 532 nm. When the –OR substituents are linked onto the macrocycle α-sites, there will be a p–π-super-conjugation between the p_z(O) and π(Pc) orbitals. When the –OR substituents are connected on the β-sites instead of the α-sites, the oscillation space of π electrons will be more effectively extended, which is positive to the periodic oscillation in the nonlinear forced vibration.

Another interesting discovery is that the χ⁽³⁾_I value for the neutral complex Nd[(tBuPhO)₄Pc]⁰˙ is 377 × 10⁻¹² esu, but it gains spectacular increase to 1280 × 10⁻¹² esu for the ionic complex Nd[(tBuPhO)₄Pc]⁻Nd(OAc)₂⁺.⁴⁷ Although the latter complex does not have unpaired π-electrons, a huge positive/negative ion-pair is constructed by the bisphthalocyaninate anion and its counter-ion, rendering it as a (δ⁺)–(δ⁻) pair with a very strong dipole moment, which shows an improved NLO property.

Throughout the investigations on various RE(Pc)₂ molecules, it is easy to find out two rules to get the double-decker tetrapyrrols with high NLO performance: (1) the neutral radical M(Pc)₂⁰˙ is a good NLO material due to the additional free single π electron. When this radical complex is oxidized/reduced and then this free single π electron is removed/paired, it lowers the overall NLO response. However, it still can be improved by ion pair formation.⁴⁷ (2) The peripheral electron-donating substituents, especially the –OR groups, are beneficial to NLO response because they could provide more free electrons to the π conjugated system. If the –OR groups are linking at the β-sites, they will provide an extended oscillating space, and thus an improved NLO response. In comparison with these positions, the efficiency of substitution at α-sites is not obvious.

Sheng and coworkers also tested a series of heteroleptic double-deckers [(OctO)₈Pc]M(Pc), M = Y, Ho, Er, Yb, and found that their σ_2PA is in the range of (126–183) × 10³ GM,⁴⁶⁷ which is still at the same level of M[(OctO)₈Pc]₂, M = Pr, Ce, Sm, and Lu series.^458–460

Jiang and coworkers explored binuclear phthalocyanine-containing sandwich-type rare-earth complexes.⁴⁶⁸ They revealed that the χ⁽³⁾_I value of (Pc)M[f-Pc₂]M(Pc)^SHex (18.6 × 10⁻¹² esu) is higher than that of Pr[(SHex)₈Pc]₂ (15.6 × 10⁻¹² esu), due to the enlarged π-conjugation by incorporating a binuclear phthalocyanine ligand.

Chen et al. conducted another comparative study on a series of porphyrin-appended europium(III) bis(phthalocyaninato) complexes (Fig. 42a).⁴⁶⁹ The χ⁽³⁾_I values for [(α-ZnPor)Pc]Eu(Pc), [(β-ZnPor)Pc]Eu(Pc) and [(β-ZnPor)₂Pc]Eu(Pc) are 111 × 10⁻²², 74 × 10⁻²², and 133 × 10⁻²² m² V⁻² which are in cases smaller than χ⁽³⁾_I for the unsubstituted Eu(Pc)₂ complex (194 × 10⁻²² m² V⁻²). This result shows that the using of larger ligands is not always beneficial to NLO properties. In the porphyrin-appended Pc ligands, there is no unified conjugating system between the Por and Pc macrocycles, which cannot provide a channel for electron oscillation. So the effective electron oscillating space is still located at the Eu(Pc)₂ part.

Fig. 42 Examples of π-extended sandwich complexes used for NLO studies: (a) Eu(III) bisphthalocyaninates with appended porphyrin units;⁴⁶⁹ (b and c) double- and triple-decker complexes with meso-pyrenyl-substituted porphyrin ligands.³¹

In order to improve the third-order NLO responses of rare-earth based tetrapyrrolic sandwiches, multi-deckers were constructed to extend the delocalized π-system and increase the oscillating space. However, the χ⁽³⁾_I value of a homoleptic triple-decker Nd₂[(β-tBuPhO)₄Pc]₃ is found to be 350 × 10⁻¹² esu,⁴⁷⁰ which is slightly lower than that of the corresponding double-decker Yb[(β-tBuPhO)₄Pc]₂ (χ⁽³⁾_I = 420 × 10⁻¹² esu) in the same Z-scan test (10 ns, 0.3 mmol L⁻¹ in DMF).⁴⁶⁵ Similar results were also obtained in the comparison of mixed (phthalocyaninato)(porphyrinato) europium triple- and double-decker complexes, in which γ_R for [(Pyr)₄Por]Eu(Pc)Eu(Pc) is also lower than γ_R for [(Pyr)₄Por]Eu(Pc) (Fig. 42b and c).³¹

However, Oluwole et al. showed that symmetrical Eu₂[(BuO)₈Pc]₃ has a higher χ⁽³⁾_I value in comparison with Eu[(BuO)₈Pc]₂.¹⁰¹ NLO response and OL properties can be improved by lowering the molecular symmetry via the introduction of diethyleneglycol substituents (Scheme 11), which in turn can be used to graft complexes to quantum dot surfaces gaining even more spectacular improvement of OL characteristics.

The third-order NLO properties of multi-decker sandwiches with more than a three layered structure were also examined by using Z-scan technique.^{48,209,471–473} Along with expanding the π–π interaction system in the order of triple-, quadruple-, quintuple-, and sextuple-deckers, as expected, the NLO responses are gradually improved due to the extended oscillating space. Remarkable χ⁽³⁾_I and γ_R values of 4910 × 10⁻¹² and 12 700 × 10⁻³⁰ esu were recorded for a sextuple-decker Nd₃Cd₂[(tBuPhO)₄Pc]₆ in the Z-scan test (10 ns, 532 nm, 3.0 × 10⁻⁴ M in CH₂Cl₂),⁴⁸ suggesting the great potential for application.

While solution measurements provide fundamental NLO characteristics of synthesized complexes, their practical application requires immobilization of molecules on surfaces or incorporation into polymeric matrices for coating fabrication. The latter approach was used by Liu et al. to form optical limiters based on heteroleptic sextuple-decker complex (Pc)Sm(Pc)Cd(Pc*)Cd(Pc*)Cd(Pc)Sm(Pc), where Pc* = (BuO)₈Pc.⁴⁷⁴ Due to RSA mechanism this complex showed excellent OL characteristics in chloroform solution (Fig. 43a), which inspired the authors to probe its practical application. Thus, this complex was dispersed in gel glass formed by methyltriethoxysilane polycondensation (Fig. 43b), affording materials capable of efficient optical limiting – while linear transmittance of a glass sample containing 1 mass% of complex was 84% (Fig. 43c), the minimum normalized transmittances decreased to 10% (Fig. 43d).

Fig. 43 (a) NLO properties of (Pc)Sm(Pc)Cd(Pc*)Cd(Pc*)Cd(Pc)Sm(Pc) in CHCl₃ at a concentration of 2.0 × 10⁻⁵ mol L⁻¹ at different incident intensities. (b) Photographs of the gel glass of the complex with different concentrations, G1–G5 where the number corresponds to weight percentage of the complex. (c) Transmittance spectra of glasses G1–G5. (d) NLO properties of gel glass G1 at different incident intensities. Reproduced from ref. 474 with permission from the Royal Society of Chemistry.

6. Conclusions

The data presented in this review showed that tetrapyrrole sandwich complexes based on rare earth elements are a unique molecular platform for targeted design of functional materials. Over the last 20 years, the field of REE sandwich complexes has made great progress in both basic and applied research. The variety of architectures and properties of sandwiches presented in this review shows fantastic opportunities for their application in advanced technologies, particularly SMM-based quantum information storage and processing, where spins are used as information carriages and quantum effects play an important role.

We can expect that the prospects for the use of REE tetrapyrrolic sandwiches will greatly expand in the coming years both with the further development of their synthetic chemistry and further elaboration of approaches to hybrid materials and thin film-based devices. The latter aspect is particularly important in terms of application, as the requirements in small amounts of complexes needed for the fabrication of such devices partially alleviate the need in large scale synthesis.

Thus, despite the rather large number of papers already published and reviewed, this area of research still has many avenues for further development, so the authors are confident that this review will encourage scientists from various fields of chemistry and materials science to tackle this exciting and promising topic.

List of abbreviations

3nCn, n = 4, 5, 6	3n-crown-n-ether
acac^−	Acetylacetonate
ClN	Chloronaphthalene
c-Pc2	Clamshell-bisphthalocyanine
DClB	Dichlorobenzene
dpm^−	Dipivaloylmethanate
DZPc	Diazepinoporphyrazine
f-Pc2	Fused bisphthalocyanine
Hp	Hemiporphyrazine
LF	Ligand-field
Ln	Lanthanide
Menth	Menthyl
Naph	Naphthyl
Nc	Naphthalocyanine
NLO	Nonlinear optics
OFET	Organic field-effect transistor
P or Por	Porphyrin
Pc	Phthalocyanine
Pz	Porphyrazine
QT	Quantum technology
QTM	Quantum tunneling of magnetization
REE	Rare earth element
SB	Schiff base
SMM	Single-molecule magnet
SOC	Spin–orbit coupling
SWNT	Single-wall nanotube
TAnP	Tetra-p-anisylpoprhyrin
TBP	Tetrabenzoporphyrin
TBTAP	Tetrabenzotriazporphyrin
TClB	1,2,4-Trichlorobenzene
TClPP	Tetra-p-chlorophenylpoprhyrin
TM	Transition metal
TTP	Tetra-p-tolylporphyrin

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

A. G. Martynov (Chapter 2) thanks Russian Science Foundation (18-73-10174-P). Y. Horii thanks the JSPS KAKENHI Grant Number JP21K14645. K. Katoh thanks the JSPS KAKENHI Grant Number JP24750119, JP15K05467. Y. Bian thanks the Financial support from the National Natural Science Foundation of China (22271012) and Beijing Natural Science Foundation (2202028). J. Jiang thanks the Financial support from the National Natural Science Foundation of China (22235001, 22175020), M. Yamashita thanks the JSPS KAKENHI Grant Number JP19H05631, JP20225003, CREST (JPMJCR12L3) from JST, the National Natural Science Foundation of China (NSFC, 22150710513), and the 111 project (B18030) from China.

Notes and references

1. D. K. P. Ng and J. Jiang, Chem. Soc. Rev., 1997, 26, 433–442.
2. J. W. Buchler and D. K. P. Ng, in The Porphyrin Handbook, ed. K. M. Kadish, K. M. Smith and R. Guilard, Academic Press, New York, 2000, vol. 3, pp. 245–294.
3. J. Jiang, K. Kasuga and D. P. Arnold, in Supramolecular Photosensitive and Electroactive Materials, ed. H. S. Nalwa, Elsevier, 2001, pp. 113–210.
4. R. Weiss and J. Fischer, in The Porphyrin Handbook, ed. K. M. Kadish, K. M. Smith and R. Guilard, Academic Press, New York, 2003, vol. 16, pp. 171–246.
5. V. E. Pushkarev, L. G. Tomilova and V. N. Nemykin, Coord. Chem. Rev., 2016, 319, 110–179.
6. V. E. Pushkarev, L. G. Tomilova and Y. V. Tomilov, Russ. Chem. Rev., 2008, 77, 875–907.
7. J. Jiang and D. K. P. Ng, Acc. Chem. Res., 2009, 42, 79–88.
8. M. Bouvet, P. Gaudillat and J.-M. M. Suisse, J. Porphyrins Phthalocyanines, 2013, 17, 628–635.
9. H. Wang, B. W. Wang, Y. Bian, S. Gao and J. Jiang, Coord. Chem. Rev., 2016, 306, 195–216.
10. W. L. Chan, C. Xie, W. S. Lo, J. C. G. Bünzli, W. K. Wong and K. L. Wong, Chem. Soc. Rev., 2021, 50, 12189–12257.
11. X. Tang, Q. Liu, C. Wei, X. Lv, Z. Jin, Y. Chen and J. Jiang, J. Rare Earths, 2021, 39, 113–120.
12. P. A. Barrett, C. E. Dent and R. P. Linstead, J. Chem. Soc., 1936, 1719.
13. I. S. Kirin, P. N. Moskalev and Y. A. Makashev, Russ. J. Inorg. Chem., 1965, 10, 1065–1066.
14. G. A. Corker, B. Grant and N. J. Clecak, J. Electrochem. Soc., 1979, 126, 1339–1343.
15. D. Markovitsi, T.-H. Tran-Thi, R. Even and J. Simon, Chem. Phys. Lett., 1987, 137, 107–112.
16. R. Rousseau, R. Aroca and M. L. Rodríguez-Méndez, J. Mol. Struct., 1995, 356, 49–62.
17. L. G. Tomilova, E. V. Chernykh, N. A. Ovchinnikova, E. V. Bezlepko, V. M. Mizin and E. A. Lukyanets, Opt. Spectrosc., 1991, 70, 775–778.
18. K. Kasuga, M. Ando, H. Morimoto and M. Isa, Chem. Lett., 1986, 1095–1098.
19. M. M’Sadak, J. Roncali and F. Garnier, J. Chim. Phys., 1986, 83, 211–216.
20. K. Kasuga, M. Tsutsui, R. C. Petterson, K. Tatsumi, N. Van Opdenbosch, G. Pepe and E. F. Meyer, J. Am. Chem. Soc., 1980, 102, 4835–4836.
21. S. I. Troyanov, L. A. Lapkina, V. E. Larchenko and A. Y. Tsivadze, Dokl. Chem., 1999, 367, 192–196.
22. A. De Cian, M. Moussavi, J. Fischer and R. Weiss, Inorg. Chem., 1985, 24, 3162–3167.
23. K. Takahashi, Y. Tomita, Y. Hada, K. Tsubota, M. Handa, K. Kasuga, K. Sogabe and T. Tokii, Chem. Lett., 1992, 759–762.
24. K. Takahashi, M. Itoh, Y. Tomita, K. Nojima, K. Kasuga and K. Isa, Chem. Lett., 1993, 1915–1918.
25. L. A. Lapkina, E. Niskanen, H. Rönkkömäki, V. E. Larchenko, K. I. Popov and A. Y. Tsivadze, J. Porphyrins Phthalocyanines, 2000, 4, 588–590.
26. I. V. Nefedova, Y. G. Gorbunova, S. G. Sakharov and A. Y. Tsivadze, Russ. J. Inorg. Chem., 2005, 50, 165–173.
27. J. W. Buchler, H.-G. Kapellmann, M. Knoff, K.-L. Lay and S. Pfeifer, Zeitschrift für Naturforsch. B, 1983, 38, 1339–1345.
28. J. W. Buchler, K. Elsässer, M. Kihn-Botulinski and B. Scharbert, Angew. Chem., Int. Ed. Engl., 1986, 25, 286–287.
29. J. W. Buchler, A. De Cian, J. Fischer, M. Kihn-Botulinski, H. Paulus and R. Weiss, J. Am. Chem. Soc., 1986, 108, 3652–3659.
30. J. W. Buchler, A. De Cian, J. Fischer, M. Kihn-Botulinski and R. Weiss, Inorg. Chem., 1988, 27, 339–345.
31. J. K. Duchowski and D. F. Bocian, J. Am. Chem. Soc., 1990, 112, 8807–8811.
32. H. Heiner, A. Tutaß, M. Göldner, U. Cornelissen and H. Homborg, Z. Anorg. Allg. Chem., 2001, 627, 485–497.
33. L. A. Lapkina, S. G. Sakharov, N. Y. Konstantinov, V. E. Larchenko, Y. G. Gorbunova and A. Y. Tsivadze, Russ. J. Inorg. Chem., 2007, 52, 1758–1768.
34. M. Moussavi, A. De Cian, J. Fischer and R. Weiss, Inorg. Chem., 1986, 25, 2107–2108.
35. P. Turek, P. Petit, J. J. Andre, J. Simon, R. Even, B. Boudjema, G. Guillaud and M. Maitrot, J. Am. Chem. Soc., 1987, 109, 5119–5122.
36. N. Ishikawa, M. Sugita, T. Ishikawa, S.-Y. Koshihara and Y. Kaizu, J. Am. Chem. Soc., 2003, 125, 8694–8695.
37. G. Lu, M. Bai, R. Li, X. Zhang, C. Ma, P.-C. Lo, D. K. P. Ng and J. Jiang, Eur. J. Inorg. Chem., 2006, 3703–3709.
38. X. Kong, Q. Jia, F. Wu and Y. Chen, Dyes Pigm., 2015, 115, 67–72.
39. X. Kong, Z. Dong, Y. Wu, X. Li, Y. Chen and J. Jiang, Chin. J. Chem., 2016, 34, 975–982.
40. X. Kong, G. Lu, X. Zhang, X. Li, Y. Chen and J. Jiang, Dyes Pigm., 2017, 143, 203–210.
41. X. Kong, X. Zhang, D. Gao, D. Qi, Y. Chen and J. Jiang, Chem. Sci., 2015, 6, 1967–1972.
42. D. Li, H. Wang, J. Kan, W. Lu, Y. Chen and J. Jiang, Org. Electron., 2013, 14, 2582–2589.
43. T. V. Dubinina, A. D. Kosov, E. F. Petrusevich, S. S. Maklakov, N. E. Borisova, L. G. Tomilova and N. S. Zefirov, Dalton Trans., 2015, 44, 7973–7981.
44. J. Kan, Y. Chen, D. Qi, Y. Liu and J. Jiang, Adv. Mater., 2012, 24, 1755–1758.
45. K. E. Sekhosana, E. Amuhaya, J. Mack and T. Nyokong, J. Mater. Chem. C, 2014, 2, 5431–5437.
46. K. E. Sekhosana, M. H. Manyeruke and T. Nyokong, J. Mol. Struct., 2016, 1121, 111–118.
47. K. E. Sekhosana, R. Nkhahle and T. Nyokong, ChemistrySelect, 2018, 3, 6671–6682.
48. K. E. Sekhosana and T. Nyokong, Dyes Pigm., 2020, 172, 107836.
49. C. R. Ganivet, B. Ballesteros, G. de la Torre, J. M. Clemente-Juan, E. Coronado and T. Torres, Chem. – Eur. J., 2013, 19, 1457–1465.
50. T. V. Dubinina, K. V. Paramonova, S. A. Trashin, N. E. Borisova, L. G. Tomilova and N. S. Zefirov, Dalton Trans., 2014, 43, 2799–2809.
51. T. V. Dubinina, M. S. Belousov, S. S. Maklakov, V. I. Chernichkin, M. V. Sedova, V. A. Tafeenko, N. E. Borisova and L. G. Tomilova, Dyes Pigm., 2019, 170, 107655.
52. N. Kobayashi and A. B. P. Lever, J. Am. Chem. Soc., 1987, 109, 7433–7441.
53. O. E. Sielcken, M. M. Van Tilborg, M. F. M. Roks, R. Hendriks, W. Drenth and R. J. M. Nolte, J. Am. Chem. Soc., 1987, 109, 4261–4265.
54. T. Toupance, V. Ahsen and J. Simon, J. Am. Chem. Soc., 1994, 116, 5352–5361.
55. T. Toupance, H. Benoit, D. Sarazin and J. Simon, J. Am. Chem. Soc., 1997, 119, 9191–9197.
56. Y. G. Gorbunova, L. A. Lapkina, A. G. Martynov, I. V. Biryukova and A. Y. Tsivadze, Russ. J. Coord. Chem., 2004, 30, 245–251.
57. A. May, P. Majumdar, A. G. Martynov, L. A. Lapkina, S. I. Troyanov, Y. G. Gorbunova, A. Y. Tsivadze, J. Mack and T. Nyokong, J. Porphyrins Phthalocyanines, 2020, 24, 589–601.
58. Y. G. Gorbunova, A. G. Martynov and A. Y. Tsivadze, in Handbook of Porphyrin Science, ed. K. M. Kadish, K. M. Smith and R. Guilard, World Scientific Publishing, 2012, vol. 24, pp. 271–388.
59. K. P. Birin, Y. G. Gorbunova and A. Y. Tsivadze, J. Porphyrins Phthalocyanines, 2006, 10, 931–936.
60. A. G. Martynov, E. A. Safonova, A. Y. Tsivadze and Y. G. Gorbunova, Coord. Chem. Rev., 2019, 387, 325–347.
61. S. L. Selektor, A. V. Shokurov, V. V. Arslanov, Y. G. Gorbunova, K. P. Birin, O. A. Raitman, F. Morote, T. Cohen-Bouhacina, C. Grauby-Heywang and A. Y. Tsivadze, J. Phys. Chem. C, 2014, 118, 4250–4258.
62. A. V. Shokurov, D. S. Kutsybala, A. G. Martynov, A. V. Bakirov, M. A. Shcherbina, S. N. Chvalun, Y. G. Gorbunova, A. Y. Tsivadze, A. V. Zaytseva, D. Novikov, V. V. Arslanov and S. L. Selektor, Langmuir, 2020, 36, 1423–1429.
63. D. S. Kutsybala, A. V. Shokurov, A. G. Martynov, A. V. Yagodin, V. V. Arslanov, Y. G. Gorbunova and S. L. Selektor, Symmetry, 2022, 14, 340.
64. S. P. Babailov, M. A. Polovkova, G. A. Kirakosyan, A. G. Martynov, E. N. Zapolotsky and Y. G. Gorbunova, Sens. Actuators, A, 2021, 331, 112933.
65. S. P. Babailov, M. A. Polovkova, E. N. Zapolotsky, G. A. Kirakosyan, A. G. Martynov and Y. G. Gorbunova, J. Porphyrins Phthalocyanines, 2022, 26, 334–339.
66. R. J. Holmberg, M. A. Polovkova, A. G. Martynov, Y. G. Gorbunova and M. Murugesu, Dalton Trans., 2016, 45, 9320–9327.
67. Y. Horii, S. Kishiue, M. Damjanović, K. Katoh, B. K. Breedlove, M. Enders and M. Yamashita, Chem. – Eur. J., 2018, 24, 4320–4327.
68. A. G. Martynov, A. V. Bykov, Y. G. Gorbunova and A. Y. Tsivadze, Russ. Chem. Bull., 2018, 67, 2195–2200.
69. A. G. Martynov, O. V. Zubareva, Y. G. Gorbunova, S. G. Sakharov, S. E. Nefedov, F. M. Dolgushin and A. Y. Tsivadze, Eur. J. Inorg. Chem., 2007, 4800–4807.
70. A. G. Martynov, E. A. Safonova, Y. G. Gorbunova and A. Y. Tsivadze, Russ. J. Inorg. Chem., 2010, 55, 347–354.
71. A. Y. Tsivadze, A. G. Martynov, M. A. Polovkova and Y. G. Gorbunova, Russ. Chem. Bull., 2011, 60, 2258–2262.
72. M. A. Polovkova, A. G. Martynov, K. P. Birin, S. E. Nefedov, Y. G. Gorbunova and A. Y. Tsivadze, Inorg. Chem., 2016, 55, 9258–9269.
73. A. V. Shokurov, A. V. Yagodin, A. G. Martynov, Y. G. Gorbunova, A. Y. Tsivadze and S. L. Selektor, Small, 2022, 18, 2104306.
74. H. Zhou, K. Wang, D. Qi and J. Jiang, Dalton Trans., 2014, 43, 1699–1705.
75. W. Lv, P. Zhu, Y. Bian, C. Ma, X. Zhang and J. Jiang, Inorg. Chem., 2010, 49, 6628–6635.
76. W. Zheng, B.-B. Wang, J.-C. Lai, C.-Z. Wan, X.-R. Lu, C.-H. Li and X.-Z. You, J. Mater. Chem. C, 2015, 3, 3072–3080.
77. M. Gonidec, F. Luis, À. Vílchez, J. Esquena, D. B. Amabilino and J. Veciana, Angew. Chem., Int. Ed., 2010, 49, 1623–1626.
78. E. A. Kuzmina, T. V. Dubinina and L. G. Tomilova, New J. Chem., 2019, 43, 9314–9327.
79. Y. Chen, W. Cao, K. Wang and J. Jiang, Inorg. Chem., 2015, 54, 9962–9967.
80. Y. Chen, F. Ma, X. Chen, B. Dong, K. Wang, S. Jiang, C. Wang, X. Chen, D. Qi, H. Sun, B. Wang, S. Gao and J. Jiang, Inorg. Chem. Front., 2017, 4, 1465–1471.
81. Y. Chen, F. Ma, X. Chen, B. Dong, K. Wang, S. Jiang, C. Wang, X. Chen, D. Qi, H. Sun, B. Wang, S. Gao and J. Jiang, Inorg. Chem., 2017, 56, 13889–13896.
82. C. Liu, M. Li, Y. Zhang, H. Tian, Y. Chen, H. Wang, J. Dou and J. Jiang, Eur. J. Inorg. Chem., 2019, 2940–2946.
83. Y. Chen, F. Ma, Y. Zhang, L. Zhao, K. Wang, D. Qi, H. Sun and J. Jiang, Inorg. Chem. Front., 2018, 5, 2006–2012.
84. S. Takamatsu, T. Ishikawa, S. Koshihara and N. Ishikawa, Inorg. Chem., 2007, 46, 7250–7252.
85. D. V. Konarev, S. S. Khasanov, M. S. Batov, A. G. Martynov, I. V. Nefedova, Y. G. Gorbunova, A. Otsuka, H. Yamochi, H. Kitagawa and R. N. Lyubovskaya, Inorg. Chem., 2019, 58, 5058–5068.
86. M. Gonidec, D. B. Amabilino and J. Veciana, Dalton Trans., 2012, 41, 13632.
87. M. Gonidec, I. Krivokapic, J. Vidal-Gancedo, E. S. Davies, J. McMaster, S. M. Gorun and J. Veciana, Inorg. Chem., 2013, 52, 4464–4471.
88. A. Şenocak, B. Köksoy, E. Demirbaş and M. Durmuş, J. Phys. Org. Chem., 2019, 32, e3907.
89. F. Bertani, N. Cristiani, M. Mannini, R. Pinalli, R. Sessoli and E. Dalcanale, Eur. J. Org. Chem., 2015, 7036–7042.
90. E. A. Kuzmina, T. V. Dubinina, N. E. Borisova, B. N. Tarasevich, V. I. Krasovskii, I. N. Feofanov, A. V. Dzuban and L. G. Tomilova, Dyes Pigm., 2020, 174, 108075.
91. E. A. Kuzmina, T. V. Dubinina, P. N. Vasilevsky, M. S. Saveliev, A. Y. Gerasimenko, N. E. Borisova and L. G. Tomilova, Dyes Pigm., 2021, 185, 108871.
92. E. A. Safonova, A. G. Martynov, S. E. Nefedov, G. A. Kirakosyan, Y. G. Gorbunova and A. Y. Tsivadze, Inorg. Chem., 2016, 55, 2450–2459.
93. R. Wang, R. Li, Y. Bian, C.-F. Choi, D. K. P. Ng, J. Dou, D. Wang, P. Zhu, C. Ma, R. D. Hartnell, D. P. Arnold and J. Jiang, Chem. – Eur. J., 2005, 11, 7351–7357.
94. Y. Chen, F. Ma, X. Chen, Y. Zhang, H. Wang, K. Wang, D. Qi, H.-L. Sun and J. Jiang, Inorg. Chem., 2019, 58, 2422–2429.
95. Y. Gao, R. Li, S. Dong, Y. Bian and J. Jiang, Dalton Trans., 2010, 39, 1321–1327.
96. M. Damjanović, Y. Horie, T. Morita, Y. Horii, K. Katoh, M. Yamashita and M. Enders, Inorg. Chem., 2015, 54, 11986–11992.
97. Y. Horii, Y. Horie, K. Katoh, B. K. Breedlove and M. Yamashita, Inorg. Chem., 2018, 57, 565–574.
98. J. Mack and N. Kobayashi, Chem. Rev., 2011, 111, 281–321.
99. N. Ishikawa and Y. Kaizu, Chem. Lett., 1998, 183–184.
100. S. Alpugan, Ü. İşci, F. Albrieux, C. Hirel, A. G. Gürek, Y. Bretonnière, V. Ahsen and F. Dumoulin, Chem. Commun., 2014, 50, 7466.
101. D. O. Oluwole, A. V. Yagodin, N. C. Mkhize, K. E. Sekhosana, A. G. Martynov, Y. G. Gorbunova, A. Y. Tsivadze and T. Nyokong, Chem. – Eur. J., 2017, 23, 2820–2830.
102. V. E. Pushkarev, A. Y. Tolbin, N. E. Borisova, S. A. Trashin and L. G. Tomilova, Eur. J. Inorg. Chem., 2010, 5254–5262.
103. A. Pedrini, M. Perfetti, M. Mannini and E. Dalcanale, ACS Omega, 2017, 2, 517–521.
104. V. E. Pushkarev, A. Y. Tolbin, F. E. Zhurkin, N. E. Borisova, S. A. Trashin, L. G. Tomilova and N. S. Zefirov, Chem. – Eur. J., 2012, 18, 9046–9055.
105. Y. Horii, K. Katoh, B. K. Breedlove and M. Yamashita, Chem. Commun., 2017, 53, 8561–8564.
106. Y. S. Korostei, V. G. Tarasova, V. E. Pushkarev, N. E. Borisova, A. K. Vorobiev and L. G. Tomilova, Dyes Pigm., 2018, 159, 573–575.
107. K. P. Birin, A. I. Poddubnaya, E. V. Isanbaeva, Y. G. Gorbunova and A. Y. Tsivadze, J. Porphyrins Phthalocyanines, 2017, 21, 406–415.
108. W. Lu, C. Wang, X. Li, X. Zhan, D. Qi and Y. Bian, Inorg. Chem. Commun., 2017, 81, 18–21.
109. M. M. Ayhan, A. Singh, C. Hirel, A. G. Gürek, V. Ahsen, E. Jeanneau, I. Ledoux-Rak, J. Zyss, C. Andraud and Y. Bretonnière, J. Am. Chem. Soc., 2012, 134, 3655–3658.
110. M. M. Ayhan, A. Singh, E. Jeanneau, V. Ahsen, J. Zyss, I. Ledoux-Rak, A. G. Gürek, C. Hirel, Y. Bretonnière and C. Andraud, Inorg. Chem., 2014, 53, 4359–4370.
111. W. Cao, K. Wang, I. Ledoux-Rak and J. Jiang, Inorg. Chem. Front, 2016, 3, 1146–1151.
112. Q. Zhi, F. Ma, C. Wang, Y. Chen, H. Wang, H. Sun and J. Jiang, Eur. J. Inorg. Chem., 2019, 1329–1334.
113. J. Sun and Y. Han, RSC Adv., 2017, 7, 22855–22859.
114. S. Kyatskaya, J. R. Galán Mascarós, L. Bogani, F. Hennrich, M. Kappes, W. Wernsdorfer and M. Ruben, J. Am. Chem. Soc., 2009, 131, 15143–15151.
115. S. Klyatskaya, A. Eichhöfer and W. Wernsdorfer, Eur. J. Inorg. Chem., 2014, 4179–4185.
116. O. Hampe, S. Klyatskaya, T. Karpuschkin, M. Vonderach, P. Weis, M. Ruben and M. M. Kappes, Int. J. Mass Spectrom., 2012, 325–327, 183–188.
117. M. Urdampilleta, S. Klyatskaya, J.-P. Cleuziou, M. Ruben and W. Wernsdorfer, Nat. Mater., 2011, 10, 502–506.
118. B. Ballesteros, G. de la Torre, A. Shearer, A. Hausmann, M. Â. Ã. Herranz, D. M. Guldi and T. Torres, Chem. – Eur. J., 2010, 16, 114–125.
119. H. Pan, L. Gong, W. Liu, C. Lin, Q. Ma, G. Lu, D. Qi, K. Wang and J. Jiang, Dyes Pigm., 2018, 156, 167–174.
120. K. Wang, H. Pan and J. Jiang, Chem. – Eur. J., 2015, 21, 18461–18465.
121. Y. Li, Y. Bian, M. Yan, P. S. Thapaliya, D. Johns, X. Yan, D. Galipeau and J. Jiang, J. Mater. Chem., 2011, 21, 11131.
122. S. Sakaue, A. Fuyuhiro, T. Fukuda and N. Ishikawa, Chem. Commun., 2012, 48, 5337.
123. K. P. Birin, Y. G. Gorbunova, A. Y. Tsivadze and A. Y. Tsivadze, J. Porphyrins Phthalocyanines, 2009, 13, 283.
124. K. P. Birin, Y. G. Gorbunova and A. Y. Tsivadze, Magn. Reson. Chem., 2010, 48, 505–515.
125. K. P. Birin, Y. G. Gorbunova and A. Y. Tsivadze, Macroheterocycles, 2010, 3, 210–217.
126. K. P. Birin, K. a Kamarova, Y. G. Gorbunova and A. Y. Tsivadze, J. Porphyrins Phthalocyanines, 2013, 17, 1027–1034.
127. K. P. Birin, Y. G. Gorbunova and A. Y. Tsivadze, Dalton Trans., 2011, 40, 11539–11549.
128. K. P. Birin, A. I. Poddubnaya, Y. G. Gorbunova and A. Y. Tsivadze, Macroheterocycles, 2017, 10, 514–519.
129. K. P. Birin, Y. G. Gorbunova and A. Y. Tsivadze, Dalton Trans., 2012, 41, 23–28.
130. H.-G. Jin, X. Jiang, I. A. Kühne, S. Clair, V. Monnier, C. Chendo, G. Novitchi, A. K. Powell, K. M. Kadish and T. S. Balaban, Inorg. Chem., 2017, 56, 4864–4873.
131. D. González-Lucas, S. C. Soobrattee, D. L. Hughes, G. J. Tizzard, S. J. Coles and A. N. Cammidge, Chem. – Eur. J., 2020, 26, 10724–10728.
132. P. Zhu, X. Zhang, H. Wang, Y. Zhang, Y. Bian and J. Jiang, Inorg. Chem., 2012, 51, 5651–5659.
133. X. Wu, W. Lv, Q. Wang, H. Wang, X. Zhang and J. Jiang, Dalton Trans., 2011, 40, 107–113.
134. J. Lu, L. Zhou, Q. Meng, H. Sui, Y. Li and X. Tai, Dyes Pigm., 2015, 113, 138–144.
135. T. Komeda, H. Isshiki, J. Liu, K. Katoh, M. Shirakata, B. K. Breedlove and M. Yamashita, ACS Nano, 2013, 7, 1092–1099.
136. K. Katoh, S. Yamashita, N. Yasuda, Y. Kitagawa, B. K. Breedlove, Y. Nakazawa and M. Yamashita, Angew. Chem., Int. Ed., 2018, 57, 9262–9267.
137. T. V. Dubinina, A. D. Kosov, E. F. Petrusevich, N. E. Borisova, A. L. Trigub, G. V. Mamin, I. F. Gilmutdinov, A. A. Masitov, S. V. Tokarev, V. E. Pushkarev and L. G. Tomilova, Dalton Trans., 2019, 48, 13413–13422.
138. R. Stefak, N. Ratel-Ramond and G. Rapenne, Inorg. Chim. Acta, 2012, 380, 181–186.
139. Y. Zhang, H. Kersell, R. Stefak, J. Echeverria, V. Iancu, U. G. E. Perera, Y. Li, A. Deshpande, K.-F. Braun, C. Joachim, G. Rapenne and S.-W. Hla, Nat. Nanotechnol., 2016, 11, 706–712.
140. N. Giménez-Agulló, C. S. de Pipaón, L. Adriaenssens, M. Filibian, M. Martínez-Belmonte, E. C. Escudero-Adán, P. Carretta, P. Ballester and J. R. Galán-Mascarós, Chem. – Eur. J., 2014, 20, 12817–12825.
141. E. N. Tarakanova, P. A. Tarakanov, V. E. Pushkarev and L. G. Tomilova, J. Porphyrins Phthalocyanines, 2014, 18, 149–154.
142. P. A. Tarakanov, E. N. Tarakanova, P. V. Dorovatovskii, Y. V. Zubavichus, V. N. Khrustalev, S. A. Trashin, K. De Wael, M. E. Neganova, D. V. Mischenko, J. L. Sessler, P. A. Stuzhin, V. E. Pushkarev and L. G. Tomilova, Dalton Trans., 2018, 47, 14169–14173.
143. E. N. Tarakanova, S. A. Trashin, A. O. Simakov, T. Furuyama, A. V. Dzuban, L. N. Inasaridze, P. A. Tarakanov, P. A. Troshin, V. E. Pushkarev, N. Kobayashi and L. G. Tomilova, Dalton Trans., 2016, 45, 12041–12052.
144. E. N. Tarakanova, S. A. Trashin, P. A. Tarakanov, V. E. Pushkarev and L. G. Tomilova, Dyes Pigm., 2015, 117, 61–63.
145. E. N. Tarakanova, O. A. Levitskiy, T. V. Magdesieva, P. A. Tarakanov, V. E. Pushkarev and L. G. Tomilova, New J. Chem., 2015, 39, 5797–5804.
146. E. N. Tarakanova, P. A. Tarakanov, A. O. Simakov, T. Furuyama, N. Kobayashi, D. V. Konev, O. A. Goncharova, S. A. Trashin, K. De Wael, I. V. Sulimenkov, V. V. Filatov, V. I. Kozlovskiy, L. G. Tomilova, P. A. Stuzhin and V. E. Pushkarev, Dalton Trans., 2021, 50, 6245–6255.
147. T. V. Dubinina, D. V. Dyumaeva, S. A. Trashin, M. V. Sedova, A. S. Dudnik, N. E. Borisova, L. G. Tomilova and N. S. Zefirov, Dyes Pigm., 2013, 96, 699–704.
148. J. H. Helberger, A. von Rebay and D. B. Hevér, Justus Liebigs Ann. Chem., 1938, 533, 197–215.
149. H. Y. Xu, C. H. Hu, Q. F. Liu, W. X. Zhao and Y. Liu, Appl. Mech. Mater., 2012, 190–191, 571–574.
150. Q. F. Liu, Adv. Mater. Res., 2013, 750–752, 1816–1821.
151. Q. F. Liu, Adv. Mater. Res., 2013, 690–693, 573–576.
152. N. E. Galanin and G. P. Shaposhnikov, Russ. J. Org. Chem., 2012, 48, 851–857.
153. A. I. Koptyaev, N. E. Galanin and G. P. Shaposhnikov, Russ. J. Org. Chem., 2019, 55, 944–950.
154. A. I. Koptyaev, N. E. Galanin, V. V. Travkin and G. L. Pakhomov, Dyes Pigm., 2021, 186, 108984.
155. A. I. Koptyaev, M. I. Bazanov and N. E. Galanin, Russ. J. Org. Chem., 2020, 56, 788–796.
156. N. E. Galanin, L. A. Yakubov and G. P. Shaposhnikov, Russ. J. Org. Chem., 2011, 47, 941–947.
157. V. E. Pushkarev, V. V. Kalashnikov, S. A. Trashin, N. E. Borisova, L. G. Tomilova and N. S. Zefirov, Dalton Trans., 2013, 42, 12083.
158. V. E. Pushkarev, V. V. Kalashnikov, A. Y. Tolbin, S. A. Trashin, N. E. Borisova, S. V. Simonov, V. B. Rybakov, L. G. Tomilova and N. S. Zefirov, Dalton Trans., 2015, 44, 16553–16564.
159. Z. Gross, N. Galili and I. Saltsman, Angew. Chem., Int. Ed., 1999, 38, 1427–1429.
160. S. Nardis, F. Mandoj, M. Stefanelli and R. Paolesse, Coord. Chem. Rev., 2019, 388, 360–405.
161. H. L. Buckley, M. R. Anstey, D. T. Gryko and J. Arnold, Chem. Commun., 2013, 49, 3104.
162. K. C. Armstrong, S. Hohloch, T. D. Lohrey, R. A. Zarkesh, J. Arnold and M. R. Anstey, Dalton Trans., 2016, 45, 18653–18660.
163. G. Lu, S. Yan, M. Shi, W. Yu, J. Li, W. Zhu, Z. Ou and K. M. Kadish, Chem. Commun., 2015, 51, 2411–2413.
164. G. Lu, C. He, K. Wang, J. Sun, D. Qi, L. Gong, C. Wang, Z. Ou, S. Yan, S. Zeng and W. Zhu, Inorg. Chem., 2017, 56, 11503–11512.
165. G. Lu, J. Li, S. Yan, C. He, M. Shi, W. Zhu, Z. Ou and K. M. Kadish, Dyes Pigm., 2015, 121, 38–45.
166. G. Lu, J. Li, X. Jiang, Z. Ou and K. M. Kadish, Inorg. Chem., 2015, 54, 9211–9222.
167. G. Lu, J. Li, S. Yan, W. Zhu, Z. Ou and K. M. Kadish, Inorg. Chem., 2015, 54, 5795–5805.
168. G. Lu, C. He, Y. Fang, L. Wang and W. Zhu, New J. Chem., 2018, 42, 2498–2503.
169. H. Furuta, T. Asano and T. Ogawa, J. Am. Chem. Soc., 1994, 116, 767–768.
170. H. Furuta, H. Maeda and A. Osuka, Chem. Commun., 2002, 1795–1804.
171. W. Cao, H. Wang, X. Wang, H. K. Lee, D. K. P. Ng and J. Jiang, Inorg. Chem., 2012, 51, 9265–9272.
172. Y. Zhang, W. Cao, K. Wang and J. Jiang, Dalton Trans., 2014, 43, 9152.
173. W. Cao, C. Gao, Y.-Q. Zhang, D. Qi, T. Liu, K. Wang, C. Duan, S. Gao and J. Jiang, Chem. Sci., 2015, 6, 5947–5954.
174. H. Wang, W. Cao, T. Liu, C. Duan and J. Jiang, Chem. – Eur. J., 2013, 19, 2266–2270.
175. Q. Ma, X. Feng, W. Cao, H. Wang and J. Jiang, CrystEngComm, 2013, 15, 10383.
176. H. Wang, C. Liu, T. Liu, S. Zeng, W. Cao, Q. Ma, C. Duan, J. Dou and J. Jiang, Dalton Trans., 2013, 42, 15355.
177. Z. Li, F. Gao, Z. Xiao, G. Ao, X. Wu, Y. Fang, Z. Nie, T. H. Wei, J. Yang, Y. Wang, X. Zhang, J. Zuo and Y. Song, Dyes Pigm., 2015, 119, 70–74.
178. Z. Li, F. Gao, Z. Xiao, X. Wu, J. Zuo and Y. Song, Opt. Laser Technol., 2018, 103, 42–47.
179. F. Gao, X.-M. Zhang, L. Cui, K. Deng, Q.-D. Zeng and J.-L. Zuo, Sci. Rep., 2015, 4, 5928.
180. F. Gao, X. Feng, L. Yang and X. Chen, Dalton Trans., 2016, 45, 7476–7482.
181. A. Santria, A. Fuyuhiro, T. Fukuda and N. Ishikawa, Inorg. Chem., 2017, 56, 10625–10632.
182. K. Kizaki, M. Uehara, A. Fuyuhiro, T. Fukuda and N. Ishikawa, Inorg. Chem., 2018, 57, 668–675.
183. A. Santria, A. Fuyuhiro, T. Fukuda and N. Ishikawa, Dalton Trans., 2019, 48, 7685–7692.
184. A. Santria and N. Ishikawa, Inorg. Chem., 2020, 59, 14326–14336.
185. K. Kizaki, A. Santria and N. Ishikawa, Inorg. Chem., 2021, 60, 2037–2044.
186. R. H. Platel, T. Teixeira Tasso, W. Zhou, T. Furuyama, N. Kobayashi and D. B. Leznoff, Chem. Commun., 2015, 51, 5986–5989.
187. J. A. Elvidge and R. P. Linstead, J. Chem. Soc., 1952, 5008.
188. W. Liu, H. Pan, Z. Wang, K. Wang, D. Qi and J. Jiang, Chem. Commun., 2017, 53, 3765–3768.
189. W. Liu, S. Zeng, X. Chen, H. Pan, D. Qi, K. Wang, J. Dou and J. Jiang, Inorg. Chem., 2018, 57, 12347–12353.
190. F. Gao, M.-X. Yao, Y.-Y. Li, Y.-Z. Li, Y. Song and J.-L. Zuo, Inorg. Chem., 2013, 52, 6407–6416.
191. J. Y. Hu, Y. Ning, Y. S. Meng, J. Zhang, Z. Y. Wu, S. Gao and J. L. Zhang, Chem. Sci., 2017, 8, 2702–2709.
192. G. Q. Jin, Y. Ning, J. X. Geng, Z. F. Jiang, Y. Wang and J. L. Zhang, Inorg. Chem. Front., 2020, 7, 289–299.
193. S. Sarwar, S. Sanz, J. van Leusen, G. S. Nichol, E. K. Brechin and P. Kögerler, Dalton Trans., 2020, 49, 16638–16642.
194. H. Sugimoto, T. Higashi, A. Maeda, M. Mori, H. Masuda and T. Taga, J. Chem. Soc., Chem. Commun., 1983, 1234–1235.
195. H. Sugimoto, T. Higashi, A. Maeda, Y. Hirai, J. Teraoka, M. Mori, H. Masuda and T. Taga, J. Less Common Met., 1985, 112, 387–392.
196. A. Maeda and H. Sugimoto, J. Chem. Soc., Faraday Trans. 2, 1986, 82, 2019.
197. S. Tsuchiya, A. Fuyuhiro, T. Fukuda and N. Ishikawa, J. Porphyrins Phthalocyanines, 2014, 18, 933–936.
198. J.-Y. Ge, H.-Y. Wang, J. Li, J.-Z. Xie, Y. Song and J.-L. Zuo, Dalton Trans., 2017, 46, 3353–3362.
199. J. Ge, Y. Qiu, H. Wang, J. Su, P. Wang and Z. Chen, Chem. – Asian J., 2020, 15, 3013–3019.
200. J.-Y. Ge, L. Cui, J. Li, F. Yu, Y. Song, Y.-Q. Zhang, J.-L. Zuo and M. Kurmoo, Inorg. Chem., 2017, 56, 336–343.
201. J.-Y. Ge, H.-Y. Wang, J. Su, J. Li, B.-L. Wang, Y.-Q. Zhang and J.-L. Zuo, Inorg. Chem., 2018, 57, 1408–1416.
202. T. Fukuda, T. Biyajima and N. Kobayashi, J. Am. Chem. Soc., 2010, 132, 6278–6279.
203. T. Fukuda, K. Hata and N. Ishikawa, J. Am. Chem. Soc., 2012, 134, 14698–14701.
204. T. Fukuda and N. Ishikawa, J. Porphyrins Phthalocyanines, 2014, 18, 615–629.
205. H. Wang, K. Qian, K. Wang, Y. Bian, J. Jiang and S. Gao, Chem. Commun., 2011, 47, 9624–9626.
206. G. Lu, X. Kong, H. Wang, Y. Y. Chen, C. Wei, Y. Y. Chen and J. Jiang, New J. Chem., 2019, 43, 15763–15767.
207. H. Wang, K. Wang, Y. Bian, J. Jiang and N. Kobayashi, Chem. Commun., 2011, 47, 6879.
208. H. Wang, N. Kobayashi and J. Jiang, Chem. – Eur. J., 2012, 18, 1047–1049.
209. H. Wang, D. Qi, Z. Xie, W. Cao, K. Wang, H. Shang and J. Jiang, Chem. Commun., 2013, 49, 889–891.
210. C. Liu, W. Yang, Y. Zhang and J. Jiang, Inorg. Chem., 2020, 59, 17591–17599.
211. K. Katoh, Y. Horii, N. Yasuda, W. Wernsdorfer, K. Toriumi, B. K. Breedlove and M. Yamashita, Dalton Trans., 2012, 41, 13582–13600.
212. Y. Horii, K. Katoh, N. Yasuda, B. K. Breedlove and M. Yamashita, Inorg. Chem., 2015, 54, 3297–3305.
213. Y. Horii, K. Katoh, G. Cosquer, B. K. Breedlove and M. Yamashita, Inorg. Chem., 2016, 55, 11782–11790.
214. Y. Horii, M. Damjanović, M. R. Ajayakumar, K. Katoh, Y. Kitagawa, L. Chibotaru, L. Ungur, M. Mas-Torrent, W. Wernsdorfer, B. K. Breedlove, M. Enders, J. Veciana and M. Yamashita, Chem. – Eur. J., 2020, 26, 8621–8630.
215. N. Kobayashi, H. Lam, W. A. Nevin, P. Janda, C. C. Leznoff, T. Koyama, A. Monden and H. Shirai, J. Am. Chem. Soc., 1994, 116, 879–890.
216. K. Wang, D. Qi, H. Wang, W. Cao, W. Li, T. Liu, C. Duan and J. Jiang, Chem. – Eur. J., 2013, 19, 11162–11166.
217. T. Morita, M. Damjanović, K. Katoh, Y. Kitagawa, N. Yasuda, Y. Lan, W. Wernsdorfer, B. K. Breedlove, M. Enders and M. Yamashita, J. Am. Chem. Soc., 2018, 140, 2995–3007.
218. C. Wei, G. Lu, C. Guo, X. Lv, X. Tang, Q. Liu, X. Cai, Y. Chen and J. Jiang, Org. Electron., 2021, 93, 106151.
219. T. Morita, K. Katoh, B. K. Breedlove and M. Yamashita, Inorg. Chem., 2013, 52, 13555–13561.
220. K. Katoh, T. Morita, N. Yasuda, W. Wernsdorfer, Y. Kitagawa, B. K. Breedlove and M. Yamashita, Chem. – Eur. J., 2018, 24, 15522–15528.
221. G. Lu, K. Wang, X. Kong, H. Pan, J. Zhang, Y. Chen and J. Jiang, ChemElectroChem, 2018, 5, 605–609.
222. X. Cai, C. Wei, J. Dong, Q. Liu, Y. Wu, G. Lu, Y. Chen and J. Jiang, J. Mater. Sci.: Mater. Electron., 2019, 30, 1976–1983.
223. G. Lu, X. Kong, J. Sun, L. Zhang, Y. Chen and J. Jiang, Chem. Commun., 2017, 53, 12754–12757.
224. S. Lee, K. ichi Yamashita, N. Sakata, Y. Hirao, K. Ogawa and T. Ogawa, Chem. – Eur. J., 2019, 25, 3240–3243.
225. W. Wernsdorfer, N. Aliaga-Alcalde, D. N. Hendrickson and G. Christou, Nature, 2002, 416, 406–409.
226. R. Sessoli, D. Gatteschi, A. Caneschi and M. A. Novak, Nature, 1993, 365, 141–143.
227. D. Gatteschi and R. Sessoli, Angew. Chem., Int. Ed., 2003, 42, 268–297.
228. ed. Molecular Nanomagnets and Related Phenomena, S. Gao, Springer Berlin Heidelberg, Berlin, Heidelberg, 2015, vol. 164.
229. J. D. Rinehart and J. R. Long, Chem. Sci., 2011, 2, 2078–2085.
230. D. N. Woodruff, R. E. P. Winpenny and R. A. Layfield, Chem. Rev., 2013, 113, 5110–5148.
231. A. R. Farrell, J. A. Coome, M. R. Probert, A. E. Goeta, J. A. K. Howard, M.-H. Lemée-Cailleau, S. Parsons and M. Murrie, CrystEngComm, 2013, 15, 3423–3429.
232. K. Katoh, Y. Yoshida, M. Yamashita, H. Miyasaka, B. K. Breedlove, T. Kajiwara, S. Takaishi, N. Ishikawa, H. Isshiki, Y. F. Zhang, T. Komeda, M. Yamagishi and J. Takeya, J. Am. Chem. Soc., 2009, 131, 9967–9976.
233. N. Ishikawa, M. Sugita, T. Okubo, N. Tanaka, T. Iino and Y. Kaizu, Inorg. Chem., 2003, 42, 2440–2446.
234. F. S. Guo, B. M. Day, Y. C. Chen, M. L. Tong, A. Mansikkamäki and R. A. Layfield, Science, 2018, 362, 1400–1403.
235. S. G. McAdams, A.-M. Ariciu, A. K. Kostopoulos, J. P. S. Walsh and F. Tuna, Coord. Chem. Rev., 2017, 346, 216–239.
236. M. Feng and M.-L. Tong, Chem. – Eur. J., 2018, 24, 7574–7594.
237. F. Habib and M. Murugesu, Chem. Soc. Rev., 2013, 42, 3278–3288.
238. H. L. C. Feltham and S. Brooker, Coord. Chem. Rev., 2014, 276, 1–33.
239. S. Demir, I. R. Jeon, J. R. Long and T. D. Harris, Coord. Chem. Rev., 2015, 289–290, 149–176.
240. P. Zhang, Y. N. Guo and J. Tang, Coord. Chem. Rev., 2013, 257, 1728–1763.
241. X.-Y. Wang, C. Avendaño and K. R. Dunbar, Chem. Soc. Rev., 2011, 40, 3213–3238.
242. C. Benelli and D. Gatteschi, Introduction to Molecular Magnetism. From Transition Metals to Lanthanides, 2015.
243. Lanthanides and Actinides in Molecular Magnetism, ed. R. A. Layfield and M. Murugesu, 2015.
244. M. Yamashita and K. Katoh, Molecular Magnetic Materials, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2016, pp. 79–101.
245. D. Gatteschi, R. Sessoli and J. Villain, Molecular Nanomagnets, Oxford University Press, 2006.
246. K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341–351.
247. A. Abragam and B. Bleaney, Electron paramagnetic resonance of transition ions, Oxford University Press, 2012.
248. A. Lunghi, F. Totti, R. Sessoli and S. Sanvito, Nat. Commun., 2017, 8, 14620.
249. L. Gu and R. Wu, Phys. Rev. Lett., 2020, 125, 117203.
250. M. Briganti, F. Santanni, L. Tesi, F. Totti, R. Sessoli and A. Lunghi, J. Am. Chem. Soc., 2021, 143, 13633–13645.
251. J. M. Zadrozny, M. Atanasov, A. M. Bryan, C.-Y. Lin, B. D. Rekken, P. P. Power, F. Neese and J. R. Long, Chem. Sci., 2013, 4, 125–138.
252. N. Ishikawa, M. Sugita, T. Ishikawa, S. Y. Koshihara and Y. Kaizu, J. Phys. Chem. B, 2004, 108, 11265–11271.
253. W. Wernsdorfer, M. Murugesu and G. Christou, Phys. Rev. Lett., 2006, 96, 3–6.
254. L. Gunther, Quantum Tunneling of Magnetization—QTM ’94, Springer Netherlands, Dordrecht, 1995, pp. 413–434.
255. J. R. Friedman, M. P. Sarachik, J. Tejada and R. Ziolo, Phys. Rev. Lett., 1996, 76, 3830–3833.
256. S. Bertaina, S. Gambarelli, T. Mitra, B. Tsukerblat, A. Müller and B. Barbara, Nature, 2008, 453, 203–206.
257. E. Burzurí, F. Luis, B. Barbara, R. Ballou, E. Ressouche, O. Montero, J. Campo and S. Maegawa, Phys. Rev. Lett., 2011, 107, 097203.
258. W. Wernsdorfer and R. Sessoli, Science, 1999, 284, 133–135.
259. M. N. Leuenberger and D. Loss, Nature, 2001, 410, 789–793.
260. S. Sproules, Electron Paramagnetic Resonance, 2017, vol. 25, pp. 61–97.
261. M. J. Graham, J. M. Zadrozny, M. S. Fataftah and D. E. Freedman, Chem. Mater., 2017, 29, 1885–1897.
262. A. Gaita-Ariño, F. Luis, S. Hill and E. Coronado, Nat. Chem., 2019, 11, 301–309.
263. M. Atzori and R. Sessoli, J. Am. Chem. Soc., 2019, 141, 11339–11352.
264. J. Bartolomé, F. Luis and J. F. Fernández, Molecular magnets: physics and applications, Springer-Verlag/Sci-Tech/Trade, 2013.
265. F. Aquilante, J. Autschbach, R. K. Carlson, L. F. Chibotaru, M. G. Delcey, L. De Vico, I. F. Galván, N. Ferré, L. M. Frutos, L. Gagliardi, M. Garavelli, A. Giussani, C. E. Hoyer, G. Li Manni, H. Lischka, D. Ma, P. Å. Malmqvist, T. Müller, A. Nenov, M. Olivucci, T. B. Pedersen, D. Peng, F. Plasser, B. Pritchard, M. Reiher, I. Rivalta, I. Schapiro, J. Segarra-Martí, M. Stenrup, D. G. Truhlar, L. Ungur, A. Valentini, S. Vancoillie, V. Veryazov, V. P. Vysotskiy, O. Weingart, F. Zapata and R. Lindh, J. Comput. Chem., 2016, 37, 506–541.
266. E. Coronado, Nat. Rev. Mater., 2020, 5, 87–104.
267. S. Stepanow, J. Honolka, P. Gambardella, L. Vitali, N. Abdurakhmanova, T. C. Tseng, S. Rauschenbach, S. L. Tait, V. Sessi, S. Klyatskaya, M. Ruben and K. Kern, J. Am. Chem. Soc., 2010, 132, 11900–11901.
268. S. Lumetti, A. Candini, C. Godfrin, F. Balestro, W. Wernsdorfer, S. Klyatskaya, M. Ruben and M. Affronte, Dalton Trans., 2016, 45, 16570–16574.
269. M. Urdampilleta, N. V. Nguyen, J. P. Cleuziou, S. Klyatskaya, M. Ruben and W. Wernsdorfer, Int. J. Mol. Sci., 2011, 12, 6656–6667.
270. S. Thiele, F. Balestro, R. Ballou, S. Klyatskaya, M. Ruben and W. Wernsdorfer, Science, 2014, 344, 1135–1138.
271. A. L. Thomas, Phthalocyanine Research and Applications, Taylor & Francis Inc, 1990.
272. N. Ishikawa, J. Jiang, O. Bekaroğlu, J. Liu, P. Lo and D. K. P. Ng, Functional Phthalocyanine Molecular Materials, Springer Berlin Heidelberg, Berlin, Heidelberg, 2010, vol. 135.
273. J. W. Perry, K. Mansour, I. S. Lee, X. Wu, P. V. Bedworth, C. Chen, D. Ng, S. R. Marder, P. Miles, T. Wada, M. Tian and H. Sasabe, Science, 1996, 273, 1533–1536.
274. E. Palomares, M. V. Martı and J. R. Durrant, Chem. Commun., 2004, 2112–2113.
275. B. Lim, G. Y. Margulis, J.-H. Yum, E. L. Unger, B. E. Hardin, M. Grätzel, M. D. McGehee and A. Sellinger, Org. Lett., 2013, 15, 784–787.
276. N. Ishikawa, M. Sugita and W. Wernsdorfer, Angew. Chem., Int. Ed., 2005, 44, 2931–2935.
277. N. Ishikawa, M. Sugita, N. Tanaka, T. Ishikawa, S. Koshihara and Y. Kaizu, Inorg. Chem., 2004, 43, 5498–5500.
278. M. Damjanovic, T. Morita, K. Katoh, M. Yamashita and M. Enders, Chem. – Eur. J., 2015, 21, 14421–14432.
279. K. M. Kadish, K. M. Smith and R. Guilard, The porphyrin handbook, Phthalocyanines: spectroscopic and electrochemical characterization, Academic Press, 2003, vol. 16.
280. R. Sessoli, Nature, 2017, 548, 400–401.
281. L. Escalera-Moreno, J. J. Baldoví, A. Gaita-Ariño and E. Coronado, Chem. Sci., 2018, 9, 3265–3275.
282. F.-S. Guo, B. M. Day, Y.-C. Chen, M.-L. Tong, A. Mansikkamäki and R. A. Layfield, Angew. Chem., Int. Ed., 2017, 56, 11445–11449.
283. C. A. P. Goodwin, F. Ortu, D. Reta, N. F. Chilton and D. P. Mills, Nature, 2017, 548, 439–442.
284. N. Ishikawa, T. Iino and Y. Kaizu, J. Phys. Chem. A, 2002, 106, 9543–9550.
285. J. L. Liu, Y. C. Chen, Y. Z. Zheng, W. Q. Lin, L. Ungur, W. Wernsdorfer, L. F. Chibotaru and M. L. Tong, Chem. Sci., 2013, 4, 3310–3316.
286. J. L. Liu, Y. C. Chen and M. L. Tong, Chem. Soc. Rev., 2018, 47, 2431–2453.
287. L. Sorace, C. Benelli and D. Gatteschi, Chem. Soc. Rev., 2011, 40, 3092–3104.
288. K. W. H. Stevens, Proc. Phys. Soc., London, Sect. A, 1952, 65, 209–215.
289. J. Sievers, Z. Phys. Chem., 1982, 45, 289–296.
290. D. Schmitt, J. Phys., 1986, 47, 677–681.
291. C. Görller-Walran and K. Binnemans, Handbook on the physics and chemistry of rare earths. Volume 23, Elsevier, North-Holland, Amsterdam, 1996.
292. L. F. Chibotaru, in Molecular Nanomagnets and Related Phenomena, ed. S. Gao, Springer Berlin Heidelberg, Berlin, Heidelberg, 2015, pp. 185–229.
293. L. Ungur and L. F. Chibotaru, Chem. – Eur. J., 2017, 23, 3708–3718.
294. A. L. Wysocki and K. Park, Inorg. Chem., 2020, 59, 2771–2780.
295. G. Taran, E. Bonet and W. Wernsdorfer, J. Appl. Phys., 2019, 125, 142903.
296. G. Taran, E. Bonet and W. Wernsdorfer, Phys. Rev. B, 2019, 99, 180408.
297. D. Tanaka, T. Inose, H. Tanaka, S. Lee, N. Ishikawa and T. Ogawa, Chem. Commun., 2012, 48, 7796.
298. N. Ishikawa, T. Iino and Y. Kaizu, J. Am. Chem. Soc., 2002, 124, 11440–11447.
299. F. Ara, H. Oka, Y. Sainoo, K. Katoh, M. Yamashita and T. Komeda, J. Appl. Phys., 2019, 125, 183901.
300. K. Katoh, H. Isshiki, T. Komeda and M. Yamashita, Coord. Chem. Rev., 2011, 255, 2124–2148.
301. K. Katoh, H. Isshiki, T. Komeda and M. Yamashita, Chem. – Asian J., 2012, 7, 1154–1169.
302. T. Komeda, H. Isshiki, J. Liu, Y.-F. Zhang, N. Lorente, K. Katoh, B. K. Breedlove and M. Yamashita, Nat. Commun., 2011, 2, 217.
303. R. Vincent, S. Klyatskaya, M. Ruben, W. Wernsdorfer and F. Balestro, Nature, 2012, 488, 357–360.
304. A. Candini, S. Klyatskaya, M. Ruben, W. Wernsdorfer and M. Affronte, Nano Lett., 2011, 11, 2634–2639.
305. E. Moreno-Pineda, T. Komeda, K. Katoh, M. Yamashita and M. Ruben, Dalton Trans., 2016, 45, 18417–18433.
306. T. Komeda, K. Katoh and M. Yamashita, Single molecule magnet for quantum information process, 2019.
307. T. Inose, D. Tanaka, J. Liu, M. Kajihara, P. Mishra, T. Ogawa and T. Komeda, Nanoscale, 2018, 10, 19409–19417.
308. E. Moreno-Pineda, C. Godfrin, F. Balestro, W. Wernsdorfer and M. Ruben, Chem. Soc. Rev., 2018, 47, 501–513.
309. M. Ganzhorn, S. Klyatskaya, M. Ruben and W. Wernsdorfer, ACS Nano, 2013, 7, 6225–6236.
310. M. Urdampilleta, S. Klayatskaya, M. Ruben and W. Wernsdorfer, ACS Nano, 2015, 9, 4458–4464.
311. E. Moreno-Pineda, M. Damjanović, O. Fuhr, W. Wernsdorfer and M. Ruben, Angew. Chem., Int. Ed., 2017, 56, 9915–9919.
312. I. V. Krainov, J. Klier, A. P. Dmitriev, S. Klyatskaya, M. Ruben, W. Wernsdorfer and I. V. Gornyi, ACS Nano, 2017, 11, 6868–6880.
313. C. Godfrin, S. Thiele, A. Ferhat, S. Klyatskaya, M. Ruben, W. Wernsdorfer and F. Balestro, ACS Nano, 2017, 11, 3984–3989.
314. E. Moreno-Pineda, S. Klyatskaya, P. Du, M. Damjanović, G. Taran, W. Wernsdorfer and M. Ruben, Inorg. Chem., 2018, 57, 9873–9879.
315. C. Godfrin, R. Ballou, E. Bonet, M. Ruben, S. Klyatskaya, W. Wernsdorfer and F. Balestro, npj Quantum Inf., 2018, 4, 53.
316. N. Koike, H. Uekusa, Y. Ohashi, C. Harnoode, F. Kitamura, T. Ohsaka and K. Tokuda, Inorg. Chem., 1996, 35, 5798–5804.
317. K. Katoh, B. K. Breedlove and M. Yamashita, Chem. Sci., 2016, 7, 4329–4340.
318. J. Kan, H. Wang, W. Sun, W. Cao, J. Tao and J. Jiang, Inorg. Chem., 2013, 52, 8505–8510.
319. K. Katoh, T. Kajiwara, M. Nakano, Y. Nakazawa, W. Wernsdorfer, N. Ishikawa, B. K. Breedlove and M. Yamashita, Chem. – Eur. J., 2011, 17, 117–122.
320. H. Wang, K. Wang, J. Tao and J. Jiang, Chem. Commun., 2012, 48, 2973.
321. K. Katoh, N. Yasuda, M. Damjanović, W. Wernsdorfer, B. K. Breedlove and M. Yamashita, Chem. – Eur. J., 2020, 26, 4805–4815.
322. B. Cirera, J. Matarrubia, T. Kaposi, N. Giménez-Agulló, M. Paszkiewicz, F. Klappenberger, R. Otero, J. M. Gallego, P. Ballester, J. V. Barth, R. Miranda, J. R. Galán-Mascarós, W. Auwärter and D. Ecija, Phys. Chem. Chem. Phys., 2017, 19, 8282–8287.
323. T. Orlando, M. Filibian, S. Sanna, N. Giménez-Agullo, C. S. De Pipaón, P. Ballester, J. R. Galán-Mascarós and P. Carretta, J. Phys.: Condens. Matter, 2016, 28, 386002.
324. N. Ishikawa, M. Sugita and W. Wernsdorfer, J. Am. Chem. Soc., 2005, 127, 3650–3651.
325. T. Fukuda, W. Kuroda and N. Ishikawa, Chem. Commun., 2011, 47, 11686–11688.
326. Y. Horii, K. Katoh, K. Sugimoto, R. Nakanishi, B. K. Breedlove and M. Yamashita, Chem. – Eur. J., 2019, 25, 3098–3104.
327. N. Ishikawa, S. Otsuka and Y. Kaizu, Angew. Chem., Int. Ed., 2005, 44, 731–733.
328. H. Wang, T. Liu, K. Wang, C. Duan and J. Jiang, Chem. – Eur. J., 2012, 18, 7691–7694.
329. T. Fukuda, K. Matsumura and N. Ishikawa, J. Phys. Chem. A, 2013, 117, 10447–10454.
330. K. Katoh, R. Asano, A. Miura, Y. Horii, T. Morita, B. K. Breedlove and M. Yamashita, Dalton Trans., 2014, 43, 7716–7725.
331. A. Y. Tolbin, V. E. Pushkarev, L. G. Tomilova and N. S. Zefirov, Mendeleev Commun., 2009, 19, 78–80.
332. N. Ishikawa and Y. Kaizu, Chem. Phys. Lett., 1993, 203, 472–476.
333. A. G. Martynov, M. A. Polovkova, G. S. Berezhnoy, A. A. Sinelshchikova, F. M. Dolgushin, K. P. Birin, G. A. Kirakosyan, Y. G. Gorbunova and A. Y. Tsivadze, Inorg. Chem., 2020, 59, 9424–9433.
334. A. G. Martynov, M. A. Polovkova, G. S. Berezhnoy, A. A. Sinelshchikova, V. N. Khrustalev, K. P. Birin, G. A. Kirakosyan, Y. G. Gorbunova and A. Y. Tsivadze, Inorg. Chem., 2021, 60, 9110–9121.
335. Y. Chen, C. Liu, F. Ma, D. Qi, Q. Liu, H.-L. Sun and J. Jiang, Chem. – Eur. J., 2018, 24, 8066–8070.
336. H. Zhang, R. Wang, P. Zhu, Z. Lai, J. Han, C.-F. Choi, D. K. P. Ng, X. Cui, C. Ma and J. Jiang, Inorg. Chem., 2004, 43, 4740–4742.
337. J. D. Rinehart, M. Fang, W. J. Evans and J. R. Long, J. Am. Chem. Soc., 2011, 133, 14236–14239.
338. J. D. Rinehart, M. Fang, W. J. Evans and J. R. Long, Nat. Chem., 2011, 3, 538–542.
339. K. Katoh, K. Kagesawa and M. Yamashita, Materials and Energy, World Scientific Publishing, 2018, vol. 10, pp. 271–344.
340. K. Katoh, T. Komeda and M. Yamashita, Chem. Rec., 2016, 16, 987–1016.
341. K. Katoh, Y. Aizawa, T. Morita, B. K. Breedlove and M. Yamashita, Chem. – Eur. J., 2017, 23, 15377–15386.
342. T. Sato, S. Matsuzawa, K. Katoh, B. K. Breedlove and M. Yamashita, Magnetochemistry, 2019, 5, 65.
343. D. Pinkowicz, R. Podgajny and B. Sieklucka, Molecular Magnetic Materials, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2016, pp. 279–300.
344. L. Bogani and W. Wernsdorfer, Nat. Mater., 2008, 7, 179–186.
345. L. Margheriti, D. Chiappe, M. Mannini, P. E. Car, P. Sainctavit, M. A. Arrio, F. B. De Mongeot, J. C. Cezar, F. M. Piras, A. Magnani, E. Otero, A. Caneschi and R. Sessoli, Adv. Mater., 2010, 22, 5488–5493.
346. A. Hofmann and Z. Salman, J. Phys.: Conf. Ser., 2014, 551, 012055.
347. T. Yamabayashi, K. Katoh, B. K. Breedlove and M. Yamashita, Molecules, 2017, 22, 1–11.
348. T. Sato, B. K. Breedlove, M. Yamashita and K. Katoh, Angew. Chem., Int. Ed., 2021, 60, 21179–21183.
349. M. S. Haghighi, M. Rath, H. W. Rotter and H. Humborg, Z. Anorg. Allg. Chem., 1993, 619, 1887–1896.
350. G. Ostendorp, H. W. Rotter and H. Homborg, Zeitschrift für Naturforsch. B, 1996, 51, 567–573.
351. G. Ostendorp, H. W. Rotter and H. Homborg, Z. Anorg. Allg. Chem., 1996, 622, 235–244.
352. J. Janczak, R. Kubiak and A. Jezierski, Inorg. Chem., 1999, 38, 2043–2049.
353. R. Kubiak and J. Janczak, Inorg. Chim. Acta, 2010, 363, 2461–2466.
354. M. A. Lebedeva, T. W. Chamberlain and A. N. Khlobystov, Chem. Rev., 2015, 115, 11301–11351.
355. P. D. W. Boyd and C. A. Reed, Acc. Chem. Res., 2005, 38, 235–242.
356. D. Sun, F. S. Tham, C. A. Reed, L. Chaker and P. D. W. Boyd, J. Am. Chem. Soc., 2002, 124, 6604–6612.
357. A. L. Litvinov, D. V. Konarev, A. Y. Kovalevsky, I. S. Neretin, P. Coppens and R. N. Lyubovskaya, Cryst. Growth Des., 2005, 5, 1807–1819.
358. M. Suzuki, Z. Slanina, N. Mizorogi, X. Lu, S. Nagase, M. M. Olmstead, A. L. Balch and T. Akasaka, J. Am. Chem. Soc., 2012, 134, 18772–18778.
359. T. Ishii, N. Aizawa, R. Kanehama, M. Yamashita, K. ichi Sugiura and H. Miyasaka, Coord. Chem. Rev., 2002, 226, 113–124.
360. H. Wang, K. Qian, D. Qi, W. Cao, K. Wang, S. Gao and J. Jiang, Chem. Sci., 2014, 5, 3214–3220.
361. H. Iwami, J. Xing, R. Nakanishi, Y. Horii, K. Katoh, B. K. Breedlove, K. Kawachi, Y. Kasama, E. Kwon and M. Yamashita, Chem. Commun., 2020, 56, 12785–12788.
362. N. Ishikawa, Y. Mizuno, S. Takamatsu, T. Ishikawa and S. Koshihara, Inorg. Chem., 2008, 47, 10217–10219.
363. S. Takamatsu and N. Ishikawa, Polyhedron, 2007, 26, 1859–1862.
364. R. Pederson, A. L. Wysocki, N. Mayhall and K. Park, J. Phys. Chem. A, 2019, 123, 6996–7006.
365. Y. Horii, M. Damjanović, K. Katoh and M. Yamashita, Dalton Trans., 2021, 50, 9719–9724.
366. M. Gonidec, E. S. Davies, J. McMaster, D. B. Amabilino and J. Veciana, J. Am. Chem. Soc., 2010, 132, 1756–1757.
367. T. Inose, D. Tanaka, H. Tanaka, O. Ivasenko, T. Nagata, Y. Ohta, S. De Feyter, N. Ishikawa and T. Ogawa, Chem. – Eur. J., 2014, 20, 11362–11369.
368. D. Tanaka, N. Sumitani, T. Inose, H. Tanaka, N. Ishikawa and T. Ogawa, Chem. Lett., 2015, 44, 668–670.
369. N. Sun, H. Wang, T. Liu, D. Qi and J. Jiang, Dalton Trans., 2019, 48, 1586–1590.
370. L. Ruan, J. Tong, F. Luo, G. Qin, F. Tian, X. Jiao and X. Zhang, Inorg. Chem., 2021, 125, 10165–10172.
371. M. Dailey and C. Besson, CrystEngComm, 2021, 23, 7151–7161.
372. L. Vitali, S. Fabris, A. M. Conte, S. Brink, M. Ruben, S. Baroni and K. Kern, Nano Lett., 2008, 8, 3364–3368.
373. Z. Deng, S. Rauschenbach, S. Stepanow, S. Klyatskaya, M. Ruben and K. Kern, Phys. Scr., 2015, 90, 98003.
374. B. Warner, F. El Hallak, N. Atodiresei, P. Seibt, H. Prüser, V. Caciuc, M. Waters, A. J. Fisher, S. Blügel, J. van Slageren and C. F. Hirjibehedin, Nat. Commun., 2016, 7, 12785.
375. F. Ara, Z. K. Qi, J. Hou, T. Komeda, K. Katoh and M. Yamashita, Dalton Trans., 2016, 45, 16644–16652.
376. Y. S. Fu, J. Schwöbel, S. W. Hla, A. Dilullo, G. Hoffmann, S. Klyatskaya, M. Ruben and R. Wiesendanger, Nano Lett., 2012, 12, 3931–3935.
377. C. Wäckerlin, F. Donati, A. Singha, R. Baltic, S. Rusponi, K. Diller, F. Patthey, M. Pivetta, Y. Lan, S. Klyatskaya, M. Ruben, H. Brune and J. Dreiser, Adv. Mater., 2016, 28, 5195–5199.
378. Y. Zhang, Y. Wang, P. Liao, K. Wang, Z. Huang, J. Liu, Q. Chen, J. Jiang and K. Wu, ACS Nano, 2018, 12, 2991–2997.
379. J. Schwöbel, Y. Fu, J. Brede, A. Dilullo, G. Hoffmann, S. Klyatskaya, M. Ruben and R. Wiesendanger, Nat. Commun., 2012, 3, 953.
380. L. Malavolti, L. Poggini, L. Margheriti, L. Chiappe, P. Graziosi, B. Cortigiani, V. Lanzilotto, F. Buatier de Mongeot, P. Ohresser, E. Otero, F. Choueikani, P. Sainctavit, I. Bergenti, V. A. Dediu, M. Mannini and R. Sessoli, Chem. Commun., 2013, 49, 11506–11508.
381. A. Lodi Rizzini, C. Krull, T. Balashov, J. J. Kavich, A. Mugarza, P. S. Miedema, P. K. Thakur, V. Sessi, S. Klyatskaya, M. Ruben, S. Stepanow and P. Gambardella, Phys. Rev. Lett., 2011, 107, 177205.
382. C. Nistor, C. Krull, A. Mugarza, S. Stepanow, C. Stamm, M. Soares, S. Klyatskaya, M. Ruben and P. Gambardella, Phys. Rev. B: Condens. Matter Mater. Phys., 2015, 92, 184402.
383. A. Lodi Rizzini, C. Krull, T. Balashov, A. Mugarza, C. Nistor, F. Yakhou, V. Sessi, S. Klyatskaya, M. Ruben, S. Stepanow and P. Gambardella, Nano Lett., 2012, 12, 5703–5707.
384. A. Candini, D. Klar, S. Marocchi, V. Corradini, R. Biagi, V. De Renzi, U. Del Pennino, F. Troiani, V. Bellini, S. Klyatskaya, M. Ruben, K. Kummer, N. B. Brookes, H. Huang, A. Soncini, H. Wende and M. Affronte, Sci. Rep., 2016, 6, 1–8.
385. J. Gómez-Segura, I. Díez-Pérez, N. Ishikawa, M. Nakano, J. Veciana and D. Ruiz-Molina, Chem. Commun., 2006, 2866–2868.
386. M. Gonidec, R. Biagi, V. Corradini, F. Moro, V. De Renzi, U. Del Pennino, D. Summa, L. Muccioli, C. Zannoni, D. B. Amabilino and J. Veciana, J. Am. Chem. Soc., 2011, 133, 6603–6612.
387. D. Klar, A. Candini, L. Joly, S. Klyatskaya, B. Krumme, P. Ohresser, J.-P. Kappler, M. Ruben and H. Wende, Dalton Trans., 2014, 43, 10686–10689.
388. L. Smykalla, P. Shukrynau and M. Hietschold, J. Phys. Chem. C, 2012, 116, 8008–8013.
389. G. Serrano, E. Velez-Fort, I. Cimatti, B. Cortigiani, L. Malavolti, D. Betto, A. Ouerghi, N. B. Brookes, M. Mannini and R. Sessoli, Nanoscale, 2018, 10, 2715–2720.
390. S. Marocchi, A. Candini, D. Klar, W. Van Den Heuvel, H. Huang, F. Troiani, V. Corradini, R. Biagi, V. De Renzi, S. Klyatskaya, K. Kummer, N. B. Brookes, M. Ruben, H. Wende, U. Del Pennino, A. Soncini, M. Affronte and V. Bellini, ACS Nano, 2016, 10, 9353–9360.
391. M. Mannini, F. Bertani, C. Tudisco, L. Malavolti, L. Poggini, K. Misztal, D. Menozzi, A. Motta, E. Otero, P. Ohresser, P. Sainctavit, G. G. Condorelli, E. Dalcanale and R. Sessoli, Nat. Commun., 2014, 5, 1–8.
392. H. Oka, K. Katoh, Y. Okada, D. Oka, T. Hitosugi, M. Yamashita and T. Fukumura, Chem. Lett., 2021, 50, 1489–1492.
393. M. Serri, M. Mannini, L. Poggini, E. Vélez-Fort, B. Cortigiani, P. Sainctavit, D. Rovai, A. Caneschi and R. Sessoli, Nano Lett., 2017, 17, 1899–1905.
394. F. Pineider, E. Pedrueza-Villalmanzo, M. Serri, A. M. Adamu, E. Smetanina, V. Bonanni, G. Campo, L. Poggini, M. Mannini, C. De Julián Fernández, C. Sangregorio, M. Gurioli, A. Dmitriev and R. Sessoli, Mater. Horiz., 2019, 6, 1148–1155.
395. A. Dahal and M. Batzill, Nanoscale, 2014, 6, 2548–2562.
396. M. Cinchetti, V. A. Dediu and L. E. Hueso, Nat. Mater., 2017, 16, 507–515.
397. A. Zhao, Q. Li, L. Chen, H. Xiang, W. Wang, S. Pan, B. Wang, X. Xiao, J. Yang, J. G. Hou and Q. Zhu, Science, 2005, 309, 1542–1544.
398. V. Madhavan, W. Chen, T. Jamneala, M. F. Crommie and N. S. Wingreen, Science, 1998, 280, 567–569.
399. T. Komeda, H. Isshiki, J. Liu, K. Katoh and M. Yamashita, ACS Nano, 2014, 8, 4866–4875.
400. R. Robles, N. Lorente, H. Isshiki, J. Liu, K. Katoh, B. K. Breedlove, M. Yamashita and T. Komeda, Nano Lett., 2012, 12, 3609–3612.
401. J. Hellerstedt, A. Cahlík, M. Švec, B. De La Torre, M. Moro-Lagares, T. Chutora, B. Papoušková, G. Zoppellaro, P. Mutombo, M. Ruben, R. Zbořil and P. Jelinek, Nanoscale, 2018, 10, 15553–15563.
402. S. Fahrendorf, N. Atodiresei, C. Besson, V. Caciuc, F. Matthes, S. Blügel, P. Kögerler, D. E. Bürgler and C. M. Schneider, Nat. Commun., 2013, 4, 2425.
403. J. Granet, M. Sicot, B. Kierren, S. Lamare, F. Chérioux, F. Baudelet, Y. Fagot-Revurat, L. Moreau and D. Malterre, Nanoscale, 2018, 10, 9123–9132.
404. H. Isago and M. Shimoda, Chem. Lett., 1992, 147–150.
405. A. Mugarza, C. Krull, R. Robles, S. Stepanow, G. Ceballos and P. Gambardella, Nat. Commun., 2011, 2, 490.
406. N. Tsukahara, S. Shiraki, S. Itou, N. Ohta, N. Takagi and M. Kawai, Phys. Rev. Lett., 2011, 106, 187201.
407. M. Studniarek, C. Wäckerlin, A. Singha, R. Baltic, K. Diller, F. Donati, S. Rusponi, H. Brune, Y. Lan, S. Klyatskaya, M. Ruben, A. P. Seitsonen and J. Dreiser, Adv. Sci., 2019, 6, 1901736.
408. S. Iijima, Nature, 1991, 354, 56–58.
409. M. Menon and D. Srivastava, Phys. Rev. Lett., 1997, 79, 4453–4456.
410. S. J. Tans, A. R. M. Verschueren and C. Dekker, Nature, 1998, 393, 49–52.
411. V. I. Preprint, S. W. Emmons and A. M. Leroi, Nature, 1999, 402, 253–254.
412. W. Lu and C. M. Lieber, Nat. Mater., 2007, 6, 841–850.
413. P. M. Ajayan, T. W. Ebbesen, T. Ichihashi, S. Iijima, K. Tanigaki and H. Hiura, Nature, 1993, 362, 522–525.
414. E. Dujardin, T. W. Ebbesen, H. Hiura and K. Tanigaki, Science, 1994, 265, 1850–1852.
415. B. W. Smith, M. Monthioux and D. E. Luzzi, Nature, 1998, 396, 323–324.
416. T. Pichler, H. Kuzmany, H. Kataura and Y. Achiba, Phys. Rev. Lett., 2001, 87, 267401.
417. S. Bandow, M. Takizawa, K. Hirahara, M. Yudasaka and S. Iijima, Chem. Phys. Lett., 2001, 337, 48–54.
418. R. Kitaura, N. Imazu, K. Kobayashi and H. Shinohara, Nano Lett., 2008, 8, 693–699.
419. M. del Carmen Giménez-López, F. Moro, A. La Torre, C. J. Gómez-García, P. D. Brown, J. van Slageren and A. N. Khlobystov, Nat. Commun., 2011, 2, 407.
420. R. Nakanishi, M. Yatoo, K. Katoh, B. Breedlove and M. Yamashita, Materials, 2016, 10, 7.
421. R. Nakanishi, J. Satoh, K. Katoh, H. Zhang, B. K. Breedlove, M. Nishijima, Y. Nakanishi, H. Omachi, H. Shinohara and M. Yamashita, J. Am. Chem. Soc., 2018, 140, 10955–10959.
422. K. Katoh, J. Sato, R. Nakanishi, F. Ara, T. Komeda, Y. Kuwahara, T. Saito, B. K. Breedlove and M. Yamashita, J. Mater. Chem. C, 2021, 9, 10697–10704.
423. A. I. A. Abd El-Mageed and T. Ogawa, RSC Adv., 2019, 9, 28135–28145.
424. A. Tsumura, H. Koezuka and T. Ando, Appl. Phys. Lett., 1986, 49, 1210–1212.
425. C. A. Di, G. Yu, Y. Liu and D. Zhu, J. Phys. Chem. B, 2007, 111, 14083–14096.
426. C. Wang, H. Dong, W. Hu, Y. Liu and D. Zhu, Chem. Rev., 2012, 112, 2208–2267.
427. M. H. Hoang, Y. Kim, M. Kim, K. H. Kim, T. W. Lee, D. N. Nguyen, S. J. Kim, K. Lee, S. J. Lee and D. H. Choi, Adv. Mater., 2012, 24, 5363–5367.
428. Y. Zhang, X. Cai, Y. Bian and J. Jiang, in Functional Phthalocyanine Molecular Materials, ed. J. Jiang, Springer Berlin Heidelberg, Berlin, Heidelberg, 2010, vol. 135, pp. 275–322.
429. G. Guillaud, M. Al Sadoun, M. Maitrot, J. Simon and M. Bouvet, Chem. Phys. Lett., 1990, 167, 503–506.
430. A. V. Shokurov, D. S. Kutsybala, A. G. Martynov, O. A. A. Raitman, V. V. V. Arslanov, Y. G. G. Gorbunova, A. Y. Tsivadze and S. L. Selektor, Thin Solid Films, 2019, 692, 137591.
431. F. S. Kim, E. Ahmed, S. Subramaniyan and S. A. Jenekhe, ACS Appl. Mater. Interfaces, 2010, 2, 2974–2977.
432. Y. Zhang, X. Cai, D. Qi, Y. Bian and J. Jiang, J. Phys. Chem. C, 2008, 112, 14579–14588.
433. W. Su, J. Jiang, K. Xiao, Y. Chen, Q. Zhao, G. Yu and Y. Liu, Langmuir, 2005, 21, 6527–6531.
434. N. B. Chaure, J. L. Sosa-Sanchez, A. N. Cammidge, M. J. Cook and A. K. Ray, Org. Electron., 2010, 11, 434–438.
435. S. Barard, D. Mukherjee, S. Sarkar, T. Kreouzis, I. Chambrier, A. N. Cammidge and A. K. Ray, J. Mater. Sci.: Mater. Electron., 2020, 31, 265–273.
436. Y. Chen, W. Su, M. Bai, J. Jiang, X. Li, Y. Liu, L. Wang and S. Wang, J. Am. Chem. Soc., 2005, 127, 15700–15701.
437. Y. Chen, R. Li, R. Wang, P. Ma, S. Dong, Y. Gao, X. Li and J. Jiang, Langmuir, 2007, 23, 12549–12554.
438. R. Li, P. Ma, S. Dong, X. Zhang, Y. Chen, X. Li and J. Jiang, Inorg. Chem., 2007, 46, 11397–11404.
439. Y. Gao, P. Ma, Y. Chen, Y. Zhang, Y. Bian, X. Li, J. Jiang and C. Ma, Inorg. Chem., 2009, 48, 45–54.
440. P. Ma, Y. Chen, N. Sheng, Y. Bian and J. Jiang, Eur. J. Inorg. Chem., 2009, 954–960.
441. K. Katoh, K. Yamamoto, T. Kajiwara, J. Takeya, B. K. Breedlove and M. Yamashita, J. Phys.: Conf. Ser., 2011, 303, 012035.
442. Y. Chen, D. Li, N. Yuan, J. Gao, R. Gu, G. Lu and M. Bouvet, J. Mater. Chem., 2012, 22, 22142–22149.
443. Y. Chen, X. Kong, G. Lu, D. Qi, Y. Wu, X. Li, M. Bouvet, D. Sun and J. Jiang, Mater. Chem. Front., 2018, 2, 1009–1016.
444. X. Zhang and Y. Chen, Inorg. Chem. Commun., 2014, 39, 79–82.
445. G. Lu, X. Kong, P. Ma, K. Wang, Y. Chen and J. Jiang, ACS Appl. Mater. Interfaces, 2016, 8, 6174–6182.
446. X. Zhang, L. Liu, J. Xiao, Z. Sun and P. Li, J. Mater. Res. Technol., 2020, 9, 13682–13691.
447. C. Wang, D. Qi, G. Lu, H. Wang, Y. Chen and J. Jiang, Inorg. Chem. Front., 2019, 6, 3345–3349.
448. D. Gao, X. Zhang, X. Kong, Y. Chen and J. Jiang, ACS Appl. Mater. Interfaces, 2015, 7, 2486–2493.
449. X. Kong, G. Lu, X. Wang, S. Zhao, D. Sun, X. Li, Y. Chen and J. Jiang, J. Mater. Sci.: Mater. Electron., 2019, 30, 12437–12446.
450. R. Madru, G. Guillaud, M. Al Sadoun, M. Maitrot, J. J. André, J. Simon and R. Even, Chem. Phys. Lett., 1988, 145, 343–346.
451. A. I. Zvyagina, A. E. Aleksandrov, A. G. Martynov, A. R. Tameev, A. E. Baranchikov, A. A. Ezhov, Y. G. Gorbunova and M. A. Kalinina, Inorg. Chem., 2021, 60, 15509–15518.
452. A. G. Martynov, Y. G. Gorbunova and A. Y. Tsivadze, Prot. Met. Phys. Chem. Surfaces, 2011, 47, 465–470.
453. M. O. Senge, M. Fazekas, E. G. A. Notaras, W. J. Blau, M. Zawadzka, O. B. Locos and E. M. N. Mhuircheartaigh, Adv. Mater., 2007, 19, 2737–2774.
454. G. de la Torre, P. Vazquez, F. Agullo-Lopez and T. Torres, Chem. Rev., 2004, 104, 3723–3750.
455. J. P. Collman, J. L. Kendall, J. L. Chen, T. A. Eberspacher and C. R. Moylan, Inorg. Chem., 1997, 36, 5603–5608.
456. D. Qi and J. Jiang, ChemPhysChem, 2015, 16, 1889–1897.
457. J. S. Shirk, J. R. Lindle, F. J. Bartoli, Z. H. Kafafi, A. W. Snow and M. E. Boyle, J. Nonlinear Opt. Phys. Mater., 1992, 01, 699–726.
458. J. L. Wu, B. Gu, N. Sheng, D. Liu and Y. Cui, Appl. Phys. Lett., 2014, 105, 171113.
459. N. Sheng, D. Liu, B. Gu, J. He and Y. Cui, Dyes Pigm., 2015, 122, 346–350.
460. B. Ren, N. Sheng, B. Gu, Y. Wan, G. Rui, C. Lv and Y. Cui, Dyes Pigm., 2017, 139, 788–794.
461. A. I. Plekhanov, T. V. Basova, R. G. Parkhomenko and A. G. Gürek, Opt. Mater., 2017, 64, 13–17.
462. A. B. Karpo, V. E. Pushkarev, V. I. Krasovskii and L. G. Tomilova, Chem. Phys. Lett., 2012, 554, 155–158.
463. K. E. Sekhosana and T. Nyokong, Opt. Mater., 2014, 37, 139–146.
464. K. E. Sekhosana and T. Nyokong, Opt. Mater., 2015, 47, 211–218.
465. K. E. Sekhosana and T. Nyokong, Inorg. Chim. Acta, 2016, 450, 87–91.
466. K. E. Sekhosana, M. Shumba and T. Nyokong, Polyhedron, 2017, 138, 154–160.
467. N. Sheng, B. Gu, B. Ren, Y. Wang, L. Han, J. Wang, H. Cao, M. Guan, X. Zhai and J. Sha, Dyes Pigm., 2017, 136, 553–558.
468. C. Huang, K. Wang, J. Sun and J. Jiang, Eur. J. Inorg. Chem., 2014, 1546–1551.
469. L. Chen, R. Hu, J. Xu, S. Wang, X. Li, S. Li and G. Yang, Spectrochim. Acta, Part A, 2013, 105, 577–581.
470. K. E. Sekhosana, E. Amuhaya, S. Khene and T. Nyokong, Inorg. Chim. Acta, 2015, 426, 221–226.
471. H. Shang, H. Wang, K. Wang, J. Kan, W. Cao and J. Jiang, Dalton Trans., 2013, 42, 1109–1115.
472. K. E. Sekhosana and T. Nyokong, RSC Adv., 2019, 9, 16223–16234.
473. X. Chen, D. Qi, C. Liu, H. Wang, Z. Xie, T. W. Chen, S. M. Chen, T. W. Tseng and J. Jiang, RSC Adv., 2019, 10, 317–322.
474. C. Liu, W. Yang, J. Wang, X. Ding, H. Ren, Y. Chen, Z. Xie, T. Sun and J. Jiang, Dalton Trans., 2021, 50, 13661–13665.

---