2.1. Design and Characterization
Advanced methods of crystal engineering are the key to the synthesis of molecular magnets. Cyanido complexes are typical starting materials in the synthesis of modern molecular magnetic materials. A short contact between paramagnetic centers is required to observe the magnetic order. The Me-CN-Me’ distance (where Me is the d- or f-electron) depends only on the radius involved in the metal bridge, since the linear cyanido bridge does not allow for less than 5 Å of Me-Me contact. This is a serious problem in the synthesis of ferromagnetic materials, and the only possibility to reduce the distance between paramagnetic centers is to use metal atoms with a high coordination number and the formation of as many cyanido bridges as possible to create the Me–Me side-on interactions. Therefore, it is believed that complexes based on mixed ligands, including cyano-organic ones, are not good precursors of magnetic materials. We decided to take a different approach, so we wondered whether complexes containing cyanido and, e.g., bipyridyl ligands, are suitable for the synthesis of molecular ferromagnets. What will be the influence of ligands other than cyanido on the arrangement of the structure of the tested compounds? Will the presence of a bulky ligand containing the ring system cause the emergence of new intermolecular interactions that will force an unusual arrangement of cyanido bridges and thus cause paramagnetic centers to come closer together?
In the manuscript, to show the possibilities of our method, we used the worst starting materials, in terms of ferromagnetic properties, that we could find, i.e., we used diamagnetic Zn
2+ and Cd
2+ as cations and [W(CN)
6(bpy)]
2−/− as anions. Thus, only one metallic center, in our case, W(V) with the
d1 configuration, could be responsible for the ferromagnetic interactions. However, due to the large W-W distance [W-CN-(Zn/Cd)-NC-W distance] of more than 8 Å, in particular, 8.168 or 8.159 Å for Zn and Cd, respectively (see cif files), at first look, such interactions are excluded. We also used an anion complex, synthesized by one of us in 1988, containing mixed ligands of the anionic formula [W(CN)
6(bpy)] with 2,2′-bipyridine as a ligand blocking the formation of a complex 3D system with cyanido bridge interactions [
14]. This ligand should introduce steric hindrances to the surroundings of the metallic center of W(V), thus reducing the possibility of magnetic interactions. The use of cations with the
d10 configuration also excludes the problem of high/low spin change in polymer products. Thus, we did our best to prevent the synthesis of ferromagnetic materials, following the typical thinking on molecular magnets.
We present here the structure of three salts with Zn2+ and Cd2+ cations and anions [W(CN)6(bpy)]2−/− used in the synthesis of soft ferromagnetic material. We also present preliminary evidence of its high-temperature ferromagnetism and a possible explanation of ferromagnetic interactions. Our patented method requires several critical conditions to achieve ferromagnetic interactions. First of all, the most obvious is the use of paramagnetic centers. In the presented case, we used the anion [W(CN)6(bpy)]− with the d1 configuration. This anion has a typical magnetic moment associated with one unpaired electron (approx. 1.73 μB). The second condition is the use of a cyanido complex, as cyanido ligands serve as strong bridges between the cations and the anions that make up the polymer structures, which is typical in all papers dealing with molecular magnets. The third condition is the use of cations that coordinate with the nitrogen end of the cyanido ligands and form a polymeric, insoluble precipitate in a fast reaction (Zn2+ or Cd2+ cations, in our case; in the literature, typically, paramagnetic cations are used to observe metal–metal interactions in the Me-CN-Me chain). The following three conditions, we discovered, are the most important. These are: (1) the applications of the diamagnetic analog of the paramagnetic center (in the present case, it is an anion [W(CN)6(bpy)]2−); (2) the diamagnetic anion must also form an insoluble precipitate with the cation used, across the cyanido bridges; and (3) the product must have lower solubility than its paramagnetic counterpart. All these conditions are met by the two systems described here. The structures of diamagnetic and paramagnetic precursors are the most important.
2.2. The structures of Precursors
Due to requirements of low solubility of the formed complexes, after several months of growth under special conditions, we were able to isolate single crystals of three salts, which are precursors of two final products showing magnetic properties. These are for W(V):[Zn(bpy)Cl][W(CN)
6(bpy)] (
1) and [Cd(bpy)(H
2O)(NO
3)][W(CN)
6(bpy)] (
2) and one for W(IV), [Cd(bpy)(H
2O)][W(CN)
6(bpy)]·2H
2O (
3). Unfortunately, despite over a year of attempts to obtain crystals of the Zn analog from W(IV), it was not possible to obtain crystals of sizes suitable for diffraction measurements on a single crystal. The product can be precipitated as a very fine orange powder (
4). The structures of both W(V) complexes are very similar and show 1D molecular chains, presented in
Figure 1 for the Cd analog (extended data, as well as the structure of the Zn analog, can be found in
supplementary materials). The distance between the paramagnetic centers (8.168 Å for
1 and 8.159 Å for
2) is long enough to prevent spin–spin interactions between tungsten atoms, and both salts show typical magnetic moments of ca. 1.73 μ
B. The coordination environment around the cations adopts the geometry of a slightly distorted trigonal bipyramid (
1) or a prism (
2), while the complex anion in all the described compounds adopts the geometry of the dodecahedron. Importantly, the anions in structures
1 and
2 only use two cyanido ligands at positions 2 and 4 as cation-bridging ligands. In addition, half of the cation charge is neutralized by a cation-labile anion (Cl
− at
1 or NO
3− at
2). The W-W distance is shorter than expected for W-CN-Cd-NC-W, since the N-Cd-N angle is 80.18°. One-dimensional W-Cd or W-Zn chains are separated by bpy molecules, as shown in
Figure 1c. These chains are linked by the π–π stacking interactions, which results in a separation of adjacent W atoms by at least 9 Å (chains separated by bpy ligands in the complex W) or by more than 12.965 Å when the W atoms are separated from the Cd atoms by the bpy ligand (
2).
The structure of the diamagnetic salt
3 of W(IV) with the d
2 configuration shown in
Figure 2 is much more complicated because three cyanido ligands of the [W(CN)
6(bpy)]
2− anion are involved in the formation of cation-anionic bridges. The structure of the described compound is of the 2D type, with the layers separated by molecules of water of crystallization. The coordination environment of the Cd
2+ cation adopts the geometry of an octahedron with a coordinated one-labile water molecule. The shortest W-W distance in
3 is 8.509 Å.
2.3. The Ferromagnetic Transformation
We present here the procedure of ferromagnetic materials’ synthesis based on zinc and cadmium salts—
5 with Zn
2+ and
6 with Cd
2+ cations. The most important fact is that the creation of the final product begins with the formation of an appropriate mixture of cyanido precursors, complexes of W(V) and W(IV) in the form of their PPh
4+ salts, well soluble in a H
2O-MeCN mixture. To this mixture, zinc or cadmium nitrates water solution is added. Due to much lower solubility of the W(IV) salts, they start to precipitate, at first forming diamagnetic zinc or cadmium salts of the [W(CN)
6(bpy)]
2− ion {due to very low solubility of zinc salt, we obtained too small crystals for X-ray measurements, but the cadmium salt had a structure described earlier (complex
3)}. As the salts precipitate, the W(IV) anion concentration decreases, and the W(V) cadmium or zinc salts start to co-precipitate, but as cations, in the earlier precipitated W(IV) salts crystals, remaining coordinatively unsaturated. The W(V) anion tends to coordinate with cations (zinc or cadmium) on the W(IV) substrate. This results in forcing the W(V) zinc or cadmium salts to mimic the structure of the W(IV) substrate, at least to some extent. This process is responsible for the closer W-W distances and formation of ferromagnetic interactions between W(V) centers in the material formed. With the restricted method that we discovered, we were able to increase the magnetic moment of this complex, without changes in the substrates used, from typical molar content in the product for a d
1 system (from expected 1.45 μ
B for 80% of W(V)) to ca. 12.1 μ
B for
5 and 11.1 μ
B for
6 per tungsten atom, without change in the substrates used. In terms of magnetic properties, the increase in gram susceptibility was from 1.67·10
−6 to 1.03·10
−4 for
5 and from 1.67·10
−6 to 8.69·10
−5 for
6. We were unable to determine the Curie temperature, as it was found to be higher than the complex decomposition temperature; still, the product retained its magnetic moment unchanged (within experimental error) up to 150 °C. In dc magnetic measurements, we did not exceed 300 K, as water of hydration was released (see TG curves in
Supplementary Materials Figures S1 and S2). This could have contaminated the measuring chamber.
It must be stressed that the high magnetic moment depends strongly on the synthetic conditions, and small changes can result in a dramatic decrease in the magnetic moment of the samples. It is obvious that the small dimension of former crystals of the W(IV) precursor is crucial for its high surface area, and thus, the presence of numerous places for W(V) deposition. We found that ferromagnetic materials 5 and 6 are very sensitive to the solvent presence, and even acetone added to the ferromagnetic product decreases the magnetic moment dramatically, even up to 1.45 μB per tungsten atom, thus to a value expected for the paramagnetic sample. After solvent removal, the magnetic moment increases, and the ferromagnetic properties return to their original value.
Figure 3a,b presents the IR spectra, in the ν
CN range, of
1–
4 precursors compared to ferromagnetic samples
5 and
6. As in general, the ν
CN bands are very sensitive to symmetry, the oxidation state of metal and the types of bridges formed, bands in this region serve as a possible source of ferromagnetic material structure. As can be seen, the spectra of ferromagnetic salts are the sum of bands of precursors. This may indicate that ferromagnetic materials preserve the types of interactions and structures observed in the precursors. As the W(V) ν
CN bands are of very low intensity compared to the W(IV) ones, the bands of this last precursor are mainly observed.
We also present the powder diffractograms of the precursors and ferromagnetics
5 and
6 (
Figure 3c,d). It can be seen that ferromagnetic materials
5 and
6 are crystalline (the background is caused by the apieson used to immobilize the samples). The substantial structural changes in the ferromagnetic materials compared to their precursors can be observed. W(V) salts are formed on W(IV) crystals, preserving, at the beginning of crystallization, the Me
2+-Me
2+ distances; thus, the substantial changes in powder diffractograms are expected. However, as W(V) is present in 4:1 excess over the W(IV) complex, gradually, the more distant from the crystal surface structure of W(V) salts of Me
2+ observed in
1 or
2 begins to recover, as seen in
Figure 3.