Eugenio
Coronado
* and
José R.
Galán-Mascarós
Instituto de Ciencia Molecular, Universidad de Valencia, Dr. Moliner 50, E-46100, Burjasot, Spain. E-mail: eugenio.coronado@uv.es
First published on 24th November 2004
One of the most important trends in the field of organic conductors aims towards the design of hybrid materials, following the strategy of combining in a single compound organic cationic radicals able to give rise to conducting architectures with functional molecular anions able to add a second physical property of interest. Ferromagnetic metals, magnetic superconductors, chiral conductors and switchable conductors are some of the most recent and spectacular materials obtained from this approach. The design and physical studies of some relevant examples are discussed and highlighted.
Eugenio Coronado | Eugenio Coronado, born in Valencia in 1959, has been Professor of Inorganic Chemistry at the University of Valencia since 1993 and Director of the University's Institute of Molecular Science (ICMol) since its foundation in 2000. He is author of numerous contributions in the chemistry and physics of functional molecular materials, with particular emphasis on magnetic and electrical properties. |
José R. Galán-Mascarós | José R. Galán-Mascarós received his PhD in Chemistry from the University of Valencia (1999), and worked as a postdoctoral researcher at Texas A&M University (1999–2001). From 2002 he is a research fellow (Ramón y Cajal program) at the University of Valencia, with his main interest in the study of multifuncional molecular materials from an organic–inorganic hybrid approach. |
Fig. 1 Building blocks for molecular conductors and superconductors: tetrathiafulvalene (a, TTF) and derivatives, bis(ethylenedithio)TTF (b, ET), tetramethyltetraselenafulvalene (c, TSF), bis(ethylenediseleno)TTF (d, BEST), bis(ethylenethio)TTF (e, BET), bis(ethylenedithio)TSF (f, BETS). |
Although with many intrinsic benefits, molecular superconductors cannot compete with inorganic superconductors in terms of processing, stability and specially critical temperatures. The common low dimensionality of these radical salts, usually 1D or 2D systems, is a clear drawback. Still, the two-network hybrid structure of these salts can also be an advantage towards other aims as, for example, in the preparation of dual-function materials when using as counter-ions functional anions that are able to exhibit a second physical (or chemical) property of interest. Therefore, this hybrid approach opens the possibility of designing dual-function and, in general, multifunctional materials, which is one of the main targets in contemporary materials science. This class of materials is designed to exhibit either coexistence or coupling between the properties. Coupling between properties is of course desired in many cases as it may bring about synergistic effects between them. In such a situation one can envision that acting upon one property via external stimuli the other property can be modified and vice versa, in the search for switchable materials. But the opposite situation, with the two properties being independent due to the lack of electronic interaction between the two networks, is also a very important case of study since it may lead to combination of properties that are difficult or impossible to coexist in naturally occurring materials, like, for example, ferromagnetism and superconductivity.6
The clear challenge here is the preparation of materials exhibiting an appropriate architecture to provide the desired combination or coupling of properties. Obviously the molecular approach offers unparalleled opportunities in this area since it enables chemical design of the molecular building blocks in order to make them suitable for self-assembly to form supramolecular structures with the desired interactions and properties.
In this article we will briefly examine the different strategies developed for the design of hybrid molecular conductors exhibiting, in addition to the electrical conductivity, a second physical property. Most of the research in this area has been focused in the search for materials combining conducting and magnetic properties. In this context, ferromagnetic molecular conductors represent the most recent and significant advance. These materials will be described in the first part. In the second part the interest of designing hybrid molecular materials combining conductivity with optical activity will be presented. This development is a very recent one but is very promising as it provides the opportunity of tuning the electrical properties of the material via external physical stimuli (electrical and magnetic fields). In this case the chiral structure of the organic network is the basis of this effect. To finish, we present in the third part another possible approach to switch the electrical properties of these molecular conductors. This is based on the mutual influence between the two molecular networks. Thus, changes in the conductivities of the organic network may arise from structural changes produced in the inorganic network upon application of an external stimulus (light, for example).
The general strategy towards molecular magnetic conductors consists of the preparation of radical salts of TTF derivatives using anionic paramagnetic complexes of transition metals as counter-ions. The first candidates used for this purpose where simple paramagnetic anions (Fig. 2) such as MX4n−, analogous structurally to the simple diamagnetic anions that gave rise to the first molecular superconductors based on TTF derivatives, such as ClO4−. The higher charge/size ratio of these paramagnetic analogs is one of the reasons why most of these salts behave as semiconductors or insulators, inducing also new packing arrays, with the presence of highly oxidized species of the organic radicals. Nevertheless, paramagnetic organic metals7,8 and even superconductors9,10 have been obtained with this strategy.
Fig. 2 Building blocks for molecular magnetic conductors and superconductors. Inorganic paramagnetic anions (top left to right): [MX4]−, [M(CN)6]3−, [M(ox)3]3−, (bottom left and right) [M(H2O)(PW11O39)]5− and [M(H2O)(P2W17O61)]8−. |
Exploring the influence of charge and size, polyoxometalates with paramagnetic centers have also been used for this purpose. Although the high charges and sizes of these anions were supposed to have a strong polarizing effect on the organic layers, from the combination with ET, paramagnetic semiconductors and metals have been obtained with Keggin11 [Fig. 2 (bottom left)] and Dawson–Wells12 [Fig. 2 (bottom right)] anions, respectively.
The first paramagnetic superconductors were discovered in 1995 in the family of hybrid crystals formed by layers of ET separated by trisoxalatometalate [MIII(ox)3]3− anions and solvent molecules.9,13 For these compounds the onset of superconductivity appears at temperatures between 4 and 8.3 K, and this phenomenon is very sensitive to the solvent used. This observation illustrates how in molecular systems almost negligible structural changes can have dramatic consequences on the physical properties. In this case, the experimental data clearly indicate that both networks are independent.
The second family of paramagnetic superconductors was reported in 1996, in salts of the selenium-containing BETS with anions of the type [MX4]− (X = Cl, Br).10 In this case the two networks and also the anions are much closer to each other. Thus, π–d and d–d interactions, although weak, have important effects on the physical properties of these materials. The most striking result has been the appearance of superconductivity induced by high external magnetic fields.14 This result apparently is against the physical laws since a high enough applied magnetic field destroys the superconducting state. In this case, however, the reverse effect is observed due to the presence of localized magnetic moments in the material. A plausible explanation to such behavior is related to the compensation between the magnetic field seen by the electronic spin of the superconductor and the exchange magnetic field due to the magnetic ions.15
Discrete paramagnetic anions, as described above, can only afford magnetic ordering at very low temperatures, and usually of antiferromagnetic nature. In order to achieve ferromagnetic ordering, a convenient strategy consists of using a polymeric anionic network instead of discrete molecules. The presence of a ferromagnetic layer interleaved with conducting layers of TTF derivatives could give rise to novel magneto-resistance properties, since the ferromagnetic layer will create an internal magnetic field that can be quite strong, and that will obviously affect the conduction electrons. On the other hand, if in this approach the two networks are independent, this may lead to the preparation of a ferromagnetic superconductor, although superconductivity and ferromagentism are thought to be mutually exclusive.
This strategy was successfully developed using bimetallic oxalate-based layered complexes (Fig. 3) as the magnetic component. This route is much more demanding, from both the synthetic and crystal growth points of view. It requires the formation of the polymeric inorganic network via self assembly of the precursors [Cr(ox)3]3− and MnII, induced by the presence of oxidized ETs which are generated electrochemically. Crystals with alternating layers of the [MnCr(ox)3]− anionic network and of ET were successfully obtained, which behave as ferromagnets below 5.5 K and are metallic down to at least 0.2 K [Fig. 4(a)].16 These two electronic sublattices, responsible for the ferromagnetism and the metallic conductivity, are quasi-independent. The low dimensionality is another added originality for such a system. In fact, the magneto-resistance behavior shows a large anisotropy. When the external field is applied perpendicular to the layers (in the direction of the easy axis of magnetization) well below the ferromagnetic ordering temperature (0.7 K) a negative magneto-resistance is observed up to 2 T, then the resistance reaches a minimum and starts to increase upon further increase of the magnetic field [Fig. 4(a)]. On the other hand, when the magnetic field is applied parallel to the layers (θ = 0°) the magneto-resistance response is weaker and opposite in sign, showing below the magnetic ordering temperature a sharp peak in the resistance for very weak fields, and a very small decrease in resistance for higher fields. The presence of this peak below Tc indicates that the conducting layer feels the presence of the internal magnetic field created by the ferromagnetic layer and demonstrates the magnetic interaction between the two sublattices.
Fig. 3 Multilayered structure of the ferromagnetic metal (ET)3[MnCr(ox)3] (a), and top view of the organic (β-phase) (b) and inorganic layers (c). |
Fig. 4 (a) Magneto-resistance measurements of the ferromagnetic metal (ET)3[MnCr(ox)3] at 0.7 K for the applied field perpendicular (θ = 90°) and parallel (θ = 0°) to the ferromanetic layers. (b) Thermal variation of the ac magnetic susceptibility at 332 Hz for (ET)3[CoCr(ox)3] showing the appearance of magnetic ordering. |
Taking advantage of the possibilities offered by the molecular chemistry in terms of molecular design, the family of ferromagnetic molecular metals has been extended, from the original (ET)3[MnCr(ox)3] salt, to other TTF derivatives and other bimetallic oxalato layers. The main results are summarized in Table 1.
MIII | D | x | Packing | MII | T c/K | σ RT/S cm−1 |
---|---|---|---|---|---|---|
Cr | ET | ca. 3 | β | Mn | 5.5 (ferro) | 250 (metal) |
Co | 9.2 (ferro) | 1 (metal/semi) | ||||
Cr | BEST | ca. 2 | ? | Mn | 5.6 (ferro) | 10−6 (insulator) |
Co | 10.8 (ferro) | 10−6 (insulator) | ||||
Cr | BETS | ca. 3 | α | Mn | 5.2 (ferro) | 1 (metal/semi) |
Cr | BET | ca. 3 | ? | Mn | 5.6 (ferro) | 4 (metal/semi) |
Co | 13.0 (ferro) | 21 (metal/semi) | ||||
Rh | ET | 2.53 | β | Mn | paramagnetic | 13 (metal/semi) |
Substitution of MnII by CoII in the anionic network has yielded an analogous isostructural material with anionic layers [CoCr(ox)3]− alternating with the ET layers. This change has improved the critical temperature for ferromagnetic ordering up to 9.2 K [Fig. 4(b)],17 while maintaining the metallic conductivity. Substitution of CrIII by RhIII has yielded an analogous isostructural material with anionic layers [MnRh(ox)3]−.18 This compound is particularly relevant for several reasons. On one hand, the presence of the heavier RhIII units has allowed for a much better and complete structural characterization. The initial salts were found to present an apparent structural disorder, with the inorganic layers not well defined in the structural model proposed. This material has shown the real incommensurate nature of these salts, with two different unit cells for the organic and inorganic subsystems (Fig. 5), which do not match perfectly, with a non-stoichiometric formula, as (ET)2.53[MnRh(ox)3]·CH2Cl2. Since in the other cases the reflections from the inorganic subsystem could not be found, the accurate formula cannot be determined but should be formulated as (ET)x[MnCr(ox)3] (2.6 < x < 3). From salt to salt, x will also vary slightly depending on the relative sizes of the building units. On the other hand, the presence of the diamagnetic RhIII units breaks the magnetic ordering, maintaining a paramagnetic regime for the magnetically diluted MnII ions in the inorganic network at all temperatures. This situation is interesting because it allows us to study the conducting properties without the presence of an internal magnetic field, maintaining essentially the same crystal structure. The first studies on this direction have confirmed that the negative magneto-resistance does not appear in any case; therefore it must be related to the magnetic ordering. Still, this RhIII derivative, although metallic, does not show any transition into a superconducting state. Instead, a decrease in the conductivity has been observed at low temperatures, probably due to a weak charge localization.
Fig. 5 Representation of the incommensurate structure of the salt (ET)2.53[MnRr(ox)3]. |
This finding suggests that the lack of superconductivity found for (ET)x[MnCr(ox)3] is not only related to the appearance of a ferromagnetic state, but is intrinsic to the organic layer and may be related to structural reasons. The incommensurate nature of this material could not allow for such a transition to appear, since no long-range structural order is present between the layers. In addition, the ethylenic groups of the ET molecules appear structurally disordered over the two possible boat conformations, a feature that usually has a tremendous effect upon the conducting properties.19 A possible approach to overcome this type of structural disorder consists of using a chiral analog as an organic donor to the ET molecule. Still, this possibility is very demanding from the synthetic point of view.
Another possibility in the attempt to improve the conducting properties consists of using selenium derivatives such as BETS as organic donor, since BETS has been shown to provide stronger intermolecular contacts through the selenium atoms (Fig. 1f). The analogous multilayer material can be obtained for this donor with the [MnCr(ox)3]− anions. Unfortunately, in this salt the BETS packing changes from the β-type adopted by the ET derivative to an α-type (Fig. 6) which is less favourable for electron delocalizaton. In fact, the salt (BETS)x[MnCr(ox)3]·CH2Cl2 (x ≈ 3) is also metallic at room temperature, but with a smaller conductivity (σ = 1 S cm−1) and a resistance that starts to increase below 150 K due to a weak localization of charges (Fig. 7).20
Fig. 6 Multilayered structure of the ferromagnetic metal (BETS)x[MnCr(ox)3] (a), and top view of the organic network showing the α-packing (b). |
Fig. 7 Thermal dependence of the electrical conductivity for the ferromagnetic metal (BETS)x[MnCr(ox)3]. |
Other TTF derivatives, such as BEDO, BEST and BET have been also used for the preparation of analogous ferromagnetic salts, but no crystals of such compounds are available to date, which makes it impossible to extract correlations between the physical properties and structure (Table 1). From measurements on powder samples and pressed pellets, these compounds show analogous behavior, with metallic regimes and ferromagnetic transitions from 5 up to 12.8 K. The only exception being the BEST derivatives, which behave as insulators, showing stoichiometries very close to (BEST)2[MIICr(ox)3].
In conclusion, the molecular approach has proven to be very useful for the design and preparation of materials combining magnetism and conductivity. It is remarkable how it allows for the preparation of a whole series of layered structures with alternating conducting and magnetic networks, with the possibility of tuning the macroscopic properties by the appropriate selection of the building blocks. Following this strategy one can change the properties at will from insulating to metallic, and from paramagnetic to ferromagnetic, maintaining the overall structural features in such a way that the correlations between the two properties can be properly examined and determined.
Chirality is a property of molecules. Serendipitously, several chiral inorganic conductors are known, such as tellurium or β-manganese,22 but there cannot be an established strategy to deliberately induce chirality in such systems. Some examples of molecular conductors and superconductors formed by achiral building blocks have been reported to crystallize in chiral space groups without any further synthetic control.23 But the best approach to designing a chiral conductor takes advantage of the chirality of the molecules.
Although conducting polymers can be made chiral using organic chiral molecules as constituent subunits,24 few examples of these materials are known among the crystalline molecular conductors25 where chirality has been introduced by the use of chiral TTF derivatives. This approach has some disadvantages, such as the limited availability of these synthetically-demanding derivatives, and also the general trend that symmetric TTF derivatives give better and more regular packing, which clearly favours the electrical properties. The complementary strategy, i.e. to use chiral charge-compensating counter-ions in the preparation of the radical salts, offers, however, a wider variety of possibilities. This will allow the use of readily-available organic donors, and opens up the way for the design and preparation of many different systems. The use of chiral anions has already been loosely explored,26 using organic anions with chiral alkyl chains, but such results regarding chirality are unexplored, and the effects upon the electrical conductivity are only related to the disorder present in such systems.
We have recently proposed the use of inorganic stable chiral complexes as a better approach, as in the compound [ET]3[Sb2(L-tart)2]·CH3CN (Fig. 8, L-tart = (2R,3R)-(+)-tartrate),27 which combines the ET radical with the chiral dimer [Sb2(L-tart)2]2−.28 This material is formed by alternating layers of anions and of ET radical cations. The latter arrange following a typical packing of the so-called α-phases, with three different types of ET molecule, A, B and C. Two of them appear to be almost completely oxidized with a charge close to +1 (A and B) while the third one (C) is close to neutrality according to their intramolecular bonding distances. These molecules stack forming regular chains along the c axis following an …ABCABC… sequence, with long intrachain distances (ca. 4.2 Å) shorter than interchain ones (ca. 3.4 Å) and with an orthogonal 21 axis defining the symmetry of the conducting layer.
Fig. 8 Multilayered structure of the chiral hybrid conductor compound (ET)3[Sb2(L-tart)2]·CH3CN. |
The transport properties in the range 2–300 K classify this material as a semicondutor, with a relatively high conductivity at room temperature (2 S cm−1). The semiconducting behavior and high resistance at low temperatures has made the magneto-resistance measurements inconclusive regarding the possible presence of electrical magneto-chiral anisotropy. But this strategy can now be extended to other anions and TTF derivatives, in the search for metallic behavior. It is important to note the possibility of also using paramagnetic chiral anions, where the presence of local magnetic moments in the materials could give rise to stronger correlations between magnetic fields and electrical conductivity, as observed in other examples described before, and increasing the complexity of the materials that could exhibit three different physical properties.
One possibility is the use of photochromic species, such as nitroprusside and other mononitrosyl metallic complexes, that can access extremely long-living electronic excited states by irradiation with light.29 This kind of excitation produces important geometrical changes, affecting the bonding isomerism of the NO group bound to the metal, and therefore such a process could have an effect upon the conducting properties if local or long-range structural changes occur.
In the search for such materials several salts have been prepared to date with different TTF derivatives.30–33 Although in all these compounds the metastable state can be accessed by light irradiation, the low penetration of light into the conducting system and the low efficiency of the process (usually in regular transparent salts the fraction of anions involved in the transition is between 30 and 50%) means that only a very low percentage of anions in the solid will reach the excited state, and thus no effects on the conductivity of the materials have been observed. Two examples, however, are particularly appealing, namely the salts (TTF)7[Fe(CN)5NO]2 and ET4K[Fe(CN)5NO]2.
The salt (TTF)7[Fe(CN)5NO]2,34 as with most of the other examples, presents a layered structure with alternating layers of TTF radicals, although the TTF molecule usually favours 1D stacking. In this case the 2D layers are built from centrosymmetric hexamers and monomers orthogonally packed to form a novel type of κ pattern (Fig. 9). The anionic layer also possesses some unique features, e.g. the NO groups in the nitroprusside anions (which usually appear crystallographycally disordered due to their random distribution over at least two out of the six octahedral positions) are perfectly localized in this case, and penetrate the organic layer. This is due to the presence of some interactions between the TTF molecules and the NO groups. Indeed, DSC measurements in irradiated polycrystalline samples have suggested the existence of a metastable state for the nitroprusside anions at temperatures as high as 240 K, well over the usual temperatures shown in other nitroprusside salts.35 This effect was found to be sample dependent, what can be related to the small fraction of nitroprusside complexes being excited, depending on the penetration depth of the irradiation in each case.
Fig. 9 Multilayered structure (top) and view of the organic layer (bottom) of the salt (TTF)7[Fe(CN)5NO]2. |
The ET4K[Fe(CN)5NO]236,37 salt is also an interesting case, with a structure that presents short contacts between anions and cations, and behaving as a quasi-2D metal down to very low temperatures. EPR measurements under light irradiation showed the appearance of light-induced electron localization in the conduction band. This localized state has been attributed to the appearance of a fluctuation potential (Anderson's localization) in the ET layer related to the transition of part of the nitroprusside anions to the metastable state upon laser irradiation. However, this hypothesis could not be unambiguously confirmed
Another possibility regarding switchable molecular building blocks is that of metallic complexes exhibiting spin-crossover (SCO) transition.38 Most of the examples that show this behavior are Fe(II) complexes that, having a diamagnetic low spin configuration as the ground state, can access the low-lying excited state of the high spin configuration by means of temperature, pressure or light irradiation. In this case the change is magnetic, since paramagnetic centers can be ‘generated’ and also structural, since the two configurations are considerably different in size, with changes of over 0.2 Å for the metal-to-ligand distances of the coordination sphere. The possibility of combining such species with TTF derivatives into conducting systems is very attractive, but until now no succesful examples have been reported. This is mostly due to the fact that the systems that show efficient spin-transition are either neutral or positively charged. The development of anionic complexes will be key for such a combination, and some efforts are being made on this regard.
Owing to these synthetic difficulties, a more likely combination would be that of mixing readily-available cationic spin-crossover systems with anionic precursors giving rise to molecular conductors such as [Ni(dmit)2]x− and its analogs.39 This strategy has already yielded some interesting results, as the salt [Fe(sal2-trien)][Ni(dmit)2]3, prepared from electrolysed solutions of {[Fe(sal2-trien)][Ni(dmit)2]} (Fig. 10).40 This material exhibits a structural arrangement typical of [Ni(dmit)2]δ− fractional oxidation state compounds, namely layers of [Ni(dmit)2]δ− units separated from each other by sheets of disordered [Fe(sal2-trien)]+ cations. The layers of Ni(dmit)2 units enable the occurrence of electronic transport. At 300 K the electrical conductivity of this compound is 0.20 S cm−1, exhibiting semiconductor behavior. The magnetic behavior indicates that the molecular arrangement of this compound favours the LS of Fe(III). However, there is experimental evidence for cooperative SCO in a second phase, not fully characterized yet, which co-crystallizes in minor amounts (≤10%) together with [Fe(sal2-trien)][Ni(dmit)2]3.41
Fig. 10 Scheme of the molecular components of the {[Fe(sal2-trien)][Ni(dmit)2]} salt (top) and thermal dependence of the magnetic susceptibility (bottom) for {[Fe(sal2-trien)][Ni(dmit)2]} showing the thermal hysteresis loop. |
It can be argued in general that in the case of light irradiation for switchable materials, the effects in conductors will be weak, since the conducting electrons will absorb most of the incident light, but in all cases non-negligible. Nevertheless, the reponse upon changes in pressure, temperature and other external stimuli, such as magnetic fields, should also be remarkable.
It is important to stress that even when the possibilities are wide open, so far only a few examples of dual-function molecular conductors have been studied. Until now remarkable findings have been made in the field of magnetic conductors, but only prospective work has been done in other areas, where the first steps are being currently established. A relevant example is that of introducing chirality by the use of chiral building blocks, an approach that is valid not only for conductors but also for other types of molecular materials, such as the molecule-based magnets.42 The introduction of switchable species into molecular materials could also lead to unprecedented synergetic systems where the physical properties can be controlled and/or tuned by external stimuli, in the way for novel approaches to read/write devices and sensors.
Other combinations are possible, of course, and continuous efforts are being devoted in different directions. Particularly appealing are those combinations that include ‘non-electronic’ properties such as, for example, materials combining electronic and ionic transports. Some of them based on ET,43 and also on other organic radicals such as TCNQ,44 have recently been reported, with the use of organic crown ether ligands incorporated into the structure of conducting networks, creating channels able to conduct protons and other small cations. The ET-based ionic and electronic conductors are related to the first family of paramagnetic superconductors,9 where 18-crown-6 is incorporated into the inorganic layers. Thus, the structure is no longer truly hexagonal, but the 18-crown-6 molecules stack in the solid state creating channels that allow for proton mobility. In addition to the ionic conductivity the compounds are metallic down to 180 K.
Finally, it should be noted that the hybrid approach to multifunctionality is not restricted to crystalline materials. Organized thin films and nano objects are also under investigation. Some promising results have already been reported in hybrid Langmuir–Blodgett films wherein monolayers of localized spins alternate with monolayers of delocalized electrons.45
This journal is © The Royal Society of Chemistry 2005 |