Article Cite This: Cryst. Growth Des. 2018, 18, 6020−6027
pubs.acs.org/crystal
Synthesis and Magnetic Properties of Solid Solutions of Ferromagnetic Layered Coordination Polymers Prepared from Solution and the Solid State Tristan Neumann,† Michał Rams,‡ Carsten Wellm,† and Christian Näther*,† †
Institut für Anorganische Chemie, Christian-Albrechts-Universität zu Kiel, Max-Eyth-Straße 2, 24118 Kiel, Germany Institute of Physics, Jagiellonian University, Łojasiewicza 11, 30-348 Kraków, Poland
‡
Crystal Growth & Design 2018.18:6020-6027. Downloaded from pubs.acs.org by UNIV OF SUNDERLAND on 10/17/18. For personal use only.
S Supporting Information *
ABSTRACT: Reaction of mixtures of Co(NCS)2 and Ni(NCS)2 with ethylisonicotinate in EtOH in a ratio of 1:2 leads to the formation of solid solutions (mixed crystals) of the layered compound [CoxNi1−x(NCS)2(ethylisonicotinate)2]n that are always contaminated with an additional crystalline phase. Pure samples of the desired layer compound are accessible by using an excess of the metal salts, but in this case the Co content in the solid solutions, as determined by atomic absorption spectroscopy and energy dispersive X-ray spectroscopy, is always lower than that used in the synthesis. Therefore, solid solutions of CoxNi1−x(NCS)2(ethylisonicotinate)4 were synthesized that upon heating transform into solid solutions of the desired layered compound, and in this case the Co/Ni ratio exactly corresponds to that used in the synthesis. The formation of solid solutions of all compounds is already indicated by comparison of the experimental powder X-ray diffraction pattern of the solid solutions with that of physical mixtures with the same Co/Ni ratio. Magnetic and specific heat measurements reveal that the solid solutions show ferromagnetic ordering, as it is the case for the homometallic compounds reported recently, and that the magnetic ordering temperature increases linearly with increasing Ni content. AC measurements show that some Co/Ni inhomogeneity is present, which is more pronounced in the solid solutions prepared via the solid state than in those prepared from solution.
■
cations are linked by the anionic ligands into chains.32−35 However, the homometallic compounds show dominating ferromagnetic interactions, and upon cooling ferromagnetic ordering is observed with very different critical temperatures for the Co and the Ni compound.36,37 For the solid solutions also, ferromagnetic ordering is observed with no sign for spinglass behavior, and the critical temperatures increase linearly with increasing Ni content.31 This result is somehow surprising because it indicates that no disorder at the metal cation positions is present. However, we also have found that homogeneous samples cannot be obtained by simple mixing of all reactants in a one-pot reaction as used for the synthesis of the homometallic counterparts, because of the different solubilities of the Co and the Ni compound as determined by atomic absorption spectroscopy (AAS). In this case, Ni-rich particles precipitate in the beginning under kinetic control that become Co-rich with further reaction time. Therefore, for these samples both the magnetic as well as the specific heat measurements clearly show very broad ferromagnetic transitions, because each particle exhibits its own critical
INTRODUCTION The synthesis of magnetic coordination compounds with varying magnetic properties is still a growing field in coordination chemistry. This includes compounds that show three-dimensional (3D) magnetic ordering but also lowdimensional materials like, e.g., single chain magnets (SCMs) or single molecule magnets (SMMs).1−20 One important part in this field consists of the development of strategies for a modification of such compounds to tune their magnetic properties in more detail. In this context, the synthesis of solid solutions (mixed crystals) of different metal cations might be a promising tool, but compared to the overall literature this strategy is rarely used.21−30 One reason for this finding might be the fact that sometimes homogeneous samples are difficult to prepare and that possible disorder of the cations might lead to a more complicated magnetic behavior, even if this might not be necessarily the case. In this context, we have reported on solid solutions of coordination polymers with the general composition [CoxNi1−x(NCS)2(4-acetylpyridine)2]n, in which the metal cations are linked by pairs and single anionic ligands into layers.31 It is noted that in thiocyanate chemistry this arrangement is rarely observed, because in most cases onedimensional (1D) compounds are obtained, in which the © 2018 American Chemical Society
Received: June 6, 2018 Revised: August 30, 2018 Published: September 5, 2018 6020
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
Synthesis of 2-Ni. Crystalline powders were obtained by stirring a mixture of Ni(NCS)2 (175 mg, 1 mmol) and ethylisonicotinate (149 μL,1 mmol) in ethanol (5.0 mL) for 2 days. Synthesis of 2-CoxNi1‑x. Crystalline powders were obtained by completely dissolving the desired ratio Co(NCS)2 and Ni(NCS)2 in ethanol. Afterward, ethylisonicotinate was added in equimolar amounts (1:1) to the clear solution, and the reaction mixture was stirred for 2 days. Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDX). These measurements were performed using a Philips ESEM XL 3 with an EDAX New XL-30 Detector, at an acceleration voltage of 20 kV. Atomic Absorptions Spectroscopy (AAS). The AAS experiments were performed with a PerkinElmer Aanalyst 300. Each sample was dissolved in water with 2.5 mL of HNO3 for 100 mL of analyte. IR Spectroscopy. The IR data were measured using a Bruker ALPHA-P ATP IR spectrometer with WINFIRST control software from ATI Mattson. Differential Scanning Calorimetry (DSC). The DSC experiments were performed using a DSC 1 star system with STARe Excellence software from Mettler-Toledo AG under dynamic nitrogen atmosphere. The instrument was calibrated using standard reference materials. Magnetic Measurements. Magnetic measurements were performed using a Physical Property Measurement System (PPMS) from Quantum Design and QD MPMS-5XL. The data were corrected for core diamagnetism. Specific Heat Measurements. Specific heat was measured by the relaxation technique using a Quantum Design PPMS. Powder samples were pressed into pellets. Apiezon N grease was used to ensure thermal contact of the samples with the calorimeter. The heat capacity of the grease was measured before each run and subtracted. Powder X-ray Diffraction (PXRD). The PXRD measurements were performed by using a Stoe Transmission Powder Diffraction System (STADI P) with CuKα radiation that was equipped with a linear position-sensitive MYTHEN 1K detector from STOE & CIE.
temperature. If such mixtures are stirred for longer time homogeneous, samples were obtained, because the solubility of the solid solutions is lower than that of the homometallic compounds. Alternatively one can precipitate from clear solutions of Co(NCS)2 and Ni(NCS)2 by adding the coligand in exactly that ratio that is found in the final compound, and in this case homogeneous samples with sharp transitions and defined critical temperatures were obtained.31 To investigate if this strategy can be generally used for tuning of the magnetic properties of such compounds, we tried to prepare solid solutions for similar compounds with ethylisonicotinate as the coligand that show the same layer topology as those with 4-acetylpyridine (Figure S1) and for which also ferromagnetic ordering is observed.38 Unfortunately, in this case the reaction of the metal salts with ethylisonicotinate in a ratio given by the molecular formula (1:2) leads to the formation of samples that are always contaminated with the coligand-rich compound M(NCS)2(ethylisonicotinate)4 (M = Co; 1-Co, Ni; 1-Ni). These compounds are in equilibria with the desired layer compounds [M(NCS)2(ethylisonicotinate)2]n (M = Co; 2-Co, Ni; 2-Ni) with bridging thiocyanate anions and consist of discrete complexes, in which the thiocyanate ligands are only terminal N-bonded to the metal cations. This is frequently observed for compounds with thio- or selenocyanate anions and indicates that this coordination is energetically favored, presumably because of the low chalcophilicity of Co and Ni.39,40 However, pure samples of of [CoxNi1−x(NCS)2(ethylisonicotinate)2]n (2-CoxNi1‑x) can be prepared using an excess of the metal salts, but in this case the Co/Ni ratio does not correspond to that used in the synthesis and consequently is difficult to control. Therefore, we tried to prepare solid solutions of the discrete complexes with the composition CoxNi1−x(NCS)2(ethylisonicotinate)4 (1-CoxNi1‑x) to transform them into solid solutions of the desired layer compounds 2-CoxNi1‑x by thermal annealing, a strategy that was never used before for this purpose.41−43 Here we report on these investigations.
■
■
RESULTS AND DISCUSSION Synthesis of Solid Solutions from the Liquid State. Because the layered compounds 2-Co and 2-Ni are isotypic and contain metal cations neighbouring in the Periodic Table, very small differences in their PXRD pattern are expected, which is already obvious from their calculated pattern and their unit cell parameters (Table S1 and Figure S2 in Supporting Information). However, to check if some reasonable splitting is observed in the powder pattern, which might allow us to prove solid solution formation, the layered compounds 2-Co and 2Ni were synthesized to prepare physical mixtures with the desired Co/Ni ratio (Figures S3−S6). If the powder patterns of 2-Co and 2-Ni are compared with that of physical mixtures with a Co/Ni ratio of 25:75, 50:50, and 75:25 using α-quartz as a standard, only very small differences are observed (Figure S7). However, the (1̅12) reflections at Bragg angles of 2-theta 14.5° for 2-Co and 14.7° for 2-Ni are reasonably resolved. As an additional tool, we also measured IR spectra for the physical mixtures to investigate if two bands for the CN stretching vibrations are observed, but no significant differences were detected (Figure S8). In first experiments, we tried to prepare solid solutions using a similar procedure as that reported for the synthesis of solid solutions with 4-acetylpyridine.31 Therefore, Co(NCS)2 and Ni(NCS)2 were dissolved in EtOH in a ratio of 50:50, and afterward 2 equivalents of ethylisonicotinate was added. The precipitate that formed was stirred for 2 days, filtered off, and investigated by PXRD, which clearly shows that all samples were contaminated with the discrete complexes M(NCS)2-
EXPERIMENTAL SECTION
Reagents. Co(NCS)2, Ba(NCS)2·3H2O, and ethylisonicotinate were obtained from Alfa Aesar, and Ni(SO4)2·6H2O was purchased from Merck. All chemicals and solvents were used without further purification. Ni(NCS)2 was prepared by a reaction of equimolar amounts of NiSO4·6H2O and Ba(NCS)2·3H2O in water. The resulting precipitate of BaSO4 was filtered off, and the solvent was removed completely using a rotary evaporator leading to a green residue of Ni(NCS)2. The purity was checked by powder X-ray diffraction (PXRD). The purity of all compounds was checked by PXRD and by IR spectroscopy. Synthesis of 1-Co. Crystalline powders were obtained by stirring a mixture of Co(NCS)2 (175 mg, 1 mmol) and ethylisonicotinate (894 μL, 6 mmol) in ethanol (5.0 mL) for 2 days. Synthesis of 1-Ni. Crystalline powders were obtained by stirring a mixture of Ni(NCS)2 (175 mg, 1 mmol) and ethylisonicotinate (894 μL, 6 mmol) in ethanol (5.0 mL) for 2 days. Synthesis of 1-CoxNi1‑x. Crystalline powders were obtained by completely dissolving the desired ratio Co(NCS)2 and Ni(NCS)2 in ethanol. Afterward, ethylisonicotinate was added in a small excess (1:6) to the clear solution, and the reaction mixture was stirred for 2 days. Synthesis of 2-Co. Crystalline powders were obtained by stirring a mixture of Co(NCS)2 (175 mg, 1 mmol) and ethylisonicotinate (149 μL, 1 mmol) in ethanol (5.0 mL) for 2 days. The purity was checked by PXRD and by IR spectroscopy. 6021
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
(ethylisonicotinate)4 (M = Co, Ni) and terminal N-bonded anionic ligands (Figure S9). This is frequently observed in thiocyanate chemistry, because the coligand-rich compounds with only terminal N-bonded thiocyanate anions are more stable because Co(II) and Ni(II) cations are not very chalcophilic and therefore do not prefer to coordinate to sulfur. To shift the reaction in the direction of the desired crystalline phase, the metal salts were reacted with ethylisonicotinate in a ratio of 1:1. In this case, the powder pattern clearly show that pure products have formed without any sign for the presence of the discrete complexes (Figure S10). Surprisingly, investigations using AAS reveal that the Co/Ni ratio does not correspond to that used in the synthesis and that the Co content is always much lower (Table S2). This was completely different for the corresponding compounds with 4acetylpyridine, where always the correct ratios were found even if the solubility of the pure cobalt (6.18 g/L) and nickel (1.15 g/L)44 compound in ethanol is significantly different. To check if for the compounds with ethylisonicotinate much larger differences in the solubility are observed, the solubilities of 2Co and 2-Ni were determined in EtOH by AAS (Table S3), which shows that they are much larger compared to that of the compounds with 4-acetylpyridine (2-Co = 38.7 g/L and 2-Ni = 4.5 g/L) but that the overall difference is similar. However, a simple explanation is not at hand, but to check if the deviating stoichiometry (excess of metal salt) might be responsible for this phenomena we tried to synthesize solid solutions of the 4acetylpyridine compound using a ratio between Co(NCS)2, Ni(NCS)2, and 4-acetylpyridine of 0.5:0.5:1. It is noted that in the previous article we used the exact stoichiometric ratio (0.5:0.5:2), which leads to solid solutions with the same Co/ Ni ratio as that used in the synthesis. However, in this case AAS proves that even in this sample the Co content is much lower (Co/Ni ratio = 33:67) than that used in the synthesis and roughly in the same range as observed for the ethylisonicotinate solid solutions. This shows that the synthesis of solid solutions of the desired composition depends on several factors, with the ratio of the reactants as an essential one. To prepare samples with the desired Ni ratio of 75:25, 50:50, and 25:75, we always used an excess of cobalt in the synthesis, and in this case the cobalt content is higher but still does not correspond to the desired values. However, for these batches the Co/Ni ratio was measured for selected crystals and the bulk by EDX and all values are closely distributed indicating that homogeneous samples were obtained (Table S4). The formation of homogeneous crystals is also indicated by comparison of their PXRD pattern with that of physical mixtures, because in the former only one peak is observed for the (1̅12) reflection, whereas two peaks are observed for the physical mixtures (Figure 1). Moreover, the position of the Bragg peaks is shifted to higher values with increasing Ni content and the full width at half-maximum (fwhm) is similar to that of the physical mixtures, suggesting that homogeneous samples were obtained, which is in agreement with the results of the EDX measurements. It is noted that SEM images of these samples and of 2-Co and 2-Ni show a similar particle size and therefore should not influence the width of the reflections (Figure S11). Synthesis of Solid Solutions from the Solid State. Because the Co/Ni ratio always deviated from that used in the synthesis, and thus cannot be exactly predetermined, we tried to prepare solid solutions by an alternative approach that we
Figure 1. Experimental PXRD pattern of 2-Co, 2-Ni, and 2-CoxNi1‑x (x = 0.35, 0.55, and 0.71; black) and of physical mixtures of 2-Co and 2-Ni with roughly the same Co/Ni ratio (red). The position of the 1̅12 reflection of 2-Co and 2-Ni is indicated by a vertical line.
frequently use for the synthesis of thio- and selenocyanate coordination polymers with bridging anionic ligands, when no pure samples are available. This approach is based on the thermal decomposition of discrete complexes of the general composition M(NCS)2(L)4 (M = divalent 3d metal cation; L = coligand) in which the metal cations are octahedral coordinated by two terminal N-bonded anionic ligands and four neutral coligands. Upon heating the coligands are usually removed in discrete steps, which enforces the formation of the coligand deficient compounds with the desired bridging coordination.39,41−43 The synthesis and crystal structures of the discrete complexes with the composition M(NCS)2(ethylisonicotinate)4 (M = Co; 1-Co and M = Ni; 1-Ni) are already reported in the literature.45,46 Both compounds are also isotypic with very similar lattice parameters and consequently a very similar PXRD pattern (Table S5 and Figure S12). In the beginning, the homometallic compounds were synthesized (Figures S13−S16) to prepare physical mixtures that were investigated by PXRD (Figure S17). As it is the case for 2-Co and 2-Ni both powder patterns are very similar, but these crystalline phases can be distinguished using, e.g., the (2̅11) and (130) reflection at Bragg angles of 2-theta = 18.0° and 20.2° respectively. Just as for the compounds with a bridging coordination (see above), the IR spectra do not allow us to prove the formation of solid solutions (mixed crystals) (Figure S18). For the synthesis of the solid solutions, a clear solution of Co(NCS)2 and Ni(NCS)2 in ratios of 25:75, 50:50, and 75:25 was reacted with a slight excess of ethylisonicotinate (metal salt to coligand: 1:6) to prepare pure samples, and the precipitates formed within 2 days were filtered off and investigated by PXRD (Figure 2). Comparison of these powder patterns with that of 1-Co and 1-Ni as well as physical mixtures with the same Co/Ni ratio clearly indicate that solid solutions have formed (Figure 2). The width of the reflections of the homometallic counterparts are comparable to that of the solid solutions, and all samples show a similar particle size (Figure S20). In this case the Co/ Ni ratio, as determined by AAS (Table S6), corresponds exactly to that used in the synthesis. This is further confirmed by EDX measurements, which additionally indicate that homogeneous samples were obtained (Table S7). We also determined the unit cell volume as a function of the Co/Ni ratio by Pawley fits, leading to a linear trend, which proves that Vegard’s law is valid (Figure S21). In this context it is noted 6022
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
Figure 3. DSC curve of 1-CoxNi1‑x with x = 0.5 (black curve) and of a physical mixture of 1-Co and 1-Ni with a Co/Ni ratio of 50:50 (red curve). For the DSC curves of x = 0.25 and x = 0.75, see Figures S23−S25 in the Supporting Information.
Figure 2. Experimental PXRD pattern of 1-Co, 1-Ni, and 1-CoxNi1‑x (x = 0.75, 0.25, and 0.75; black) and of physical mixtures of 1-Co and 1-Ni with the same Co/Ni ratio (red). The position of the 130 reflection of 1-Co and 1-Ni is indicated by a vertical line. For the full range of the pattern, see Figure S19.
that similar experiments for the layered compounds do not lead to satisfying results because the values for the unit cell volumes scatter. It must be kept in mind that for these compounds the difference between the unit cell parameters is extremely small, and errors of a few cubic Angstrom do not lead to a linear behavior. Moreover, the crystal structure of the pure Ni compound was recently determined by single crystal X-ray diffraction, which reveal disorder of the pyridine substituent. This disorder might change from batch to batch and also will influence the unit cell volumes to some extent. However, on the basis of all experiments it can be assumed that even for the layered compounds Vegard’s law is valid. Measurements using thermogravimetry and differential thermoanalysis (TG-DTA) of 1-Co, 1-Ni, and the solid solutions show two mass steps that are accompanied by endothermic events in the DTA curve. The mass loss in the first step is in reasonable agreement with that calculated for the removal of half of the ethylisonicotinate ligands, but the mass steps are poorly resolved, and thus no differences between the solid solutions and the physical mixtures are observed (Figure S22). To improve the resolution, measurements using differential scanning calorimetry (DSC) were performed, and in this case one endothermic event is observed for the solid solutions, whereas two peaks are observed for the physical mixtures, which can be traced back to the different decomposition temperatures of 1-Co and 1-Ni (Figure 3 and Figure S23−S25). PXRD investigations of the residues obtained after the first TG step shows that the solid solutions 2-CoxNi1‑x (x = 0.25, 0.5, and 0.75) were obtained as pure phases and that there is no significant difference to samples obtained from solution, which is also indicated by SEM investigations (Figures S26 and S27). Investigations on the Stability of the Solid Solutions. Concerning the synthesis of solid solutions, an important question is related to their stability, namely, whether they are formed by kinetic or thermodynamic control. To investigate this question, a 1:1 mixture of 1-Co and 1-Ni or 2-Co and 2Ni was suspended in ethanol with an excess of solid and stirred for 3 weeks, and the identity of the solid was investigated each week by PXRD (Figure 4).
Figure 4. Experimental PXRD pattern of a physical mixture (red) of 1-Co and 1-Ni (top) and of 2-Co and 2-Ni (bottom) in a ratio of 50:50 and of the residues obtained stirring this mixture in ethanol for 20 days (black) together with the calculated pattern for the homometallic compounds. In the bottom of each figure, the calculated peak positions are given to show how many reflections are involved. For the full range, see Figure S28.
This clearly shows that within about 20 days the reflections of the individual compounds disappeared, and for those reflections that are separated in the mixture, only one reflection is observed in the final product, which proves solid solution formation (Figure 4). The cobalt content was afterward determined in the solid solutions leading to x = 0.49 for 1CoxNi1‑x and to 0.36 for 2-CoxNi1‑x. Therefore, at room6023
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
temperature the solid solutions are thermodynamically stable compared to the homometallic compounds. Magnetic Measurements. To finally prove the formation of solid solutions and to investigate their influence on the magnetic properties, several samples were investigated by magnetic and specific heat measurements, and these results were compared with those for the homometallic compounds 2Co and 2-Ni. The room-temperature value of the χT product ranges between 1.23 cm3 K mol−1 for 2-Ni and 3.18 cm3 K mol−1 for 2-Co and increases linearly with increasing Co content (Figures S29 and S30). Susceptibility measurements performed at 100 Oe at low temperature reveal that for all samples the susceptibility increases rapidly below 10 K (Figure 5: top). For 2-Ni and Ni-
Figure 6. Ferromagnetic critical temperature determined from magnetic susceptibility (top) and from specific heat (bottom) for samples of 2-CoxNi1‑x prepared from solution (black) or by annealing (red).
Moreover, all dχ/dT peaks for 2-CoxNi1‑x are broader than that for 2-Ni, indicating a range of critical temperatures for each solid solution composition, which was not observed for the corresponding 4-acetylpyridine solid solutions.44 This seems to be more pronounced for samples prepared by thermal annealing, and the width of distribution of Tc was marked in Figure 6 for them. Specific Heat Measurements. The specific heat was measured for all samples at zero field, with the aim to observe the magnetic ordering. The lattice contribution to the specific heat was estimated, assuming that it is identical for all 2CoxNi1‑x compounds, which is justified by almost identical molar masses of Co and Ni. A sum of Debye and Einstein models was fitted using the merged data of 2-Co (8−15 K) and 2-Ni (20−50 K), i.e., using the temperature ranges well above respective Tc. The values of characteristic phonon energies obtained are θD = 87(2), θE = 171(3) K, and the amplitudes of both contributions are aD = 2.8(1), aE = 6.00(7) J mol−1 K−1 (Figure S33). The magnetic contribution obtained by subtracting the same lattice contribution for the whole 2CoxNi1‑x series is shown in Figure 7. The overall magnetic behavior is very well reflected as Cmagn(T) maxima are observed, indicating a second-order phase transition. In contrast to the susceptibility measurements, the heat-capacity was measured down to 0.4 K, and in this case also for 2-Co the critical temperature can be determined. The values for the critical temperatures determined from the heat-capacity measurements are very similar to values obtained from magnetic measurements. If Tc is plotted as a function of the Co/Ni ratio, a linear trend is also observed (Figure 6, bottom), and there are no significant differences of determined Tc between samples prepared by thermal decomposition and by crystallization from solution. The C(T) peaks for 2-Co and 2-Ni have the λ-shape pointing for a second-order transition. For the solid solutions,
Figure 5. Field cooled magnetic susceptibility χ measured at field of 100 Oe as a function of temperature (top) and first derivative dχ/dT (bottom) for 2-CoxNi1‑x prepared from solution (s) or from annealing (a) together with that of 2-Co and 2-Ni.
rich samples, near saturation is observed, which originates from their ferromagnetic behavior. The increase in susceptibility moves to lower temperatures with increasing Co content, but even for Ni-deficient compounds one can anticipate that saturation is reached at very low temperatures. This is different for 2-Co because this compound orders at too low temperatures. The ferromagnetic behavior is also reflected by the field dependence of the magnetization measured at 1.8 K (Figure S31). For all samples near saturation is observed, which is typical for a ferromagnet, and the saturation magnetization increases with increasing Co content. The continuous change is also visible in magnetization hysteresis loop (Figure S32). Summarizing, all these measurements show that the ferromagnetic behavior is retained in the solid solutions and that the overall magnetic behavior changes continuously as a function of the actual Co/Ni ratio. The critical temperatures Tc of the ferromagnetic transitions were determined calculating the derivative dχ/dT (Figure 5, bottom) to find the temperature of the maximal slope of χ(T). This shows that the critical temperature increases with increasing Ni content almost linearly (Figure 6, top). 6024
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
Figure 7. Magnetic contribution to the specific heat measured at zero field for 2-CoxNi1‑x prepared from solution (s) or by thermal annealing (a) together with that of 2-Co and 2-Ni. Lines are to guide the eye.
Figure 8. Real part of AC magnetic susceptibility for 2-CoxNi1‑x prepared from solution x = 0.55(s); by thermal annealing x = 0.5(a). Lines are to guide the eye.
maxima (parameter φ) of samples prepared by thermal decomposition points to larger local disorder than in samples prepared from solution, even if the overall Co/Ni ratio for different crystallites of the batches prepared by both routes seems to be not scattering over a larger range. It means that it scatters differently within the clusters in each crystal, and this might be traced back to the formation of the solid solutions in the solid state. Such a reaction must surely proceed via nucleation and growth of a new phase for which the diffusion of the reactants might be important. Under the assumption that this is different for each reactant, it would explain why clusters are formed in which the Co/Ni ratio varies over a wider range. Interestingly, this would also mean that during thermal decomposition something like a “phase separation” occurs. This hypothesis is difficult to prove and is only valid if one assumes that the disorder in the precursor solid solutions obtained from solution is comparable to that in 2-CoxNi1‑x, but one can investigate if the larger local disorder is generally observed in solid solutions prepared by this route. Therefore, the solid state synthesis of solid solutions of the 4acetylpyridine compounds might be helpful, which was formerly not needed, because for samples prepared in solution the Co/Ni ratio always corresponds to that used in the synthesis. In this context, it would also be of interest why no local disorder is observed for the acetylpyridine solid solutions. This might depend on small variations of the synthetic procedure, which would be difficult to control, but the most pronounced difference is the fact that in the present work the metal to ligand ratio must be changed, which leads to samples in which the Co content is much lower than that used in the synthesis. Therefore, investigations on the local disorder of solid solutions with 4-acetylpyridine prepared by a different metal to ligand ratio would also be of high interest. Finally, it is noted that in principle variations of the Co:Ni ratio should also be detectable by PXRD, but to prove such small variations is beyond the accuracy of this method.
the peaks are still present and the C/T(T) drops to zero at low T, which means the ordered state is reached. However, these peaks are significantly broader than for 2-Ni, which means that inhomogeneities of the Ni and Co distribution in the whole sample reach a level that influences the homogeneity of the critical temperature. For samples obtained by thermal annealing, the C(T) peaks are even broader than for samples obtained from solution. AC Measurements. Finally, it is noted that for the corresponding compounds with 4-acetylpyridine AC measurements were performed that show maxima that do not depend on the frequencies.44 Therefore, spin-glass behavior, e.g., due to disordering of the cations was excluded, which was in agreement with the occurrence of a second order phase transition observed in the heat-capacity measurements. This also suggests that the magnetic exchange between Co and Ni cations is ferromagnetic and all this should also be the case for the present compounds 2-CoxNi1‑x because a similar behavior is observed in the heat-capacity measurements. However, for two samples of similar composition prepared by thermal annealing and from solution, AC measurements were performed, which show a very small shift of the AC maxima Tm for samples prepared from solution, which is more pronounced in samples prepared by thermal annealing (Figure 8). The relative temperature shift of the χ′ susceptibility peak on a decade of frequency φ = ΔTm/[TmΔ(log f)] is 0.0080 for x = 0.5(a) sample and 0.0027 for x = 0.55(s) sample. Such small values indicate the formation of a cluster spin-glass.47 As a test, the AC measurement with similar precision was made for 2-Ni, and no Tm shift is observed (Figure S34). The frequency dependence of χac points to some importance of local disorder caused by Co−Ni distribution inhomogeneity’s on the microscopic scale. The regions with a slightly higher Ni content and a bit higher Tc than average create ferromagnetic clusters. The relaxation of such clusters in the vicinity of Tc is observed in the AC susceptibility, and their freezing leads to a ferromagnetic cluster glass. For such a system, the specific heat has a broadened peak, the C(T) dependence below Tc is convex, and the field-cooled susceptibility is monotonous, as it is for 2-CoxNi1‑x. In the limit of vanishing disorder, a canonical ferromagnet would be observed, but there is a smooth evolution between a ferromagnet and cluster spin-glass, and as a simplification the family of 2-CoxNi1‑x can be described as ferromagnets. However, the much larger shift of the AC
■
CONCLUSION This work is part of an overall project on the synthesis of solid solutions to modify the magnetic properties of coordination polymers in more detail. As a first step solid solutions of a ferromagnetic layered compound with ethylisonicotinate were prepared, and it was shown that the critical temperature linearly increases with increasing Ni content, as already 6025
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
(2) Sun, H. L.; Wang, Z. M.; Gao, S. Strategies towards single-chain magnets. Coord. Chem. Rev. 2010, 254, 1081−1100. (3) Bogani, L.; Vindigni, A.; Sessoli, R.; Gatteschi, D. Single chain magnets: Where to from here? J. Mater. Chem. 2008, 18, 4750−4758. (4) Coulon, C.; Miyasaka, H.; Clérac, R. Single-Chain Magnets:Theoretical Approach and Experimental Systems. Struct. Bonding (Berlin) 2006, 122, 163−206. (5) Coulon, C.; Pianet, V.; Urdampilleta, M.; Clérac, R. Single-Chain Magnets and Related Systems; Springer: Berlin Heidelberg, 2015; Vol. 164, pp 143−184. (6) Dhers, S.; Feltham, H. L. C.; Brooker, S. A toolbox of building blocks, linkers and crystallisation methods used to generate singlechain magnets. Coord. Chem. Rev. 2015, 296, 24−44. (7) Craig, G. A.; Murrie, M. 3d single-ion magnets. Chem. Soc. Rev. 2015, 44, 2135−2147. (8) Bar, A. K.; Pichon, C.; Sutter, J.-P. Magnetic anisotropy in twoto eight-coordinated transition−metal complexes: Recent developments in molecular magnetism. Coord. Chem. Rev. 2016, 308 (Part 2), 346−380. (9) Pedersen, K. S.; Sigrist, M.; Sørensen, M. A.; Barra, A.-L.; Weyhermüller, T.; Piligkos, S.; Thuesen, C. A.; Vinum, M. G.; Mutka, H.; Weihe, H.; Clérac, R.; Bendix, J. [ReF6]2−: A Robust Module for the Design of Molecule-Based Magnetic Materials. Angew. Chem. 2014, 126, 1375−1378. (10) Bogani, L.; Wernsdorfer, W. Molecular spintronics using singlemolecule magnets. Nat. Mater. 2008, 7, 179−186. (11) Sottini, S.; Poneti, G.; Ciattini, S.; Levesanos, N.; Ferentinos, E.; Krzystek, J.; Sorace, L.; Kyritsis, P. Magnetic Anisotropy of Tetrahedral Co(II) Single-Ion Magnets: Solid-State Effects. Inorg. Chem. 2016, 55, 9537−9548. (12) Murrie, M. Cobalt(II) single-molecule magnets. Chem. Soc. Rev. 2010, 39, 1986−1995. (13) Rams, M.; Böhme, M.; Kataev, V.; Krupskaya, Y.; Büchner, B.; Plass, W.; Neumann, T.; Tomkowicz, Z.; Näther, C. Static and dynamic magnetic properties of the ferromagnetic coordination polymer [Co(NCS)2(py)2]n. Phys. Chem. Chem. Phys. 2017, 19, 24534−24544. (14) Weng, D.-F.; Wang, Z.-M.; Gao, S. Framework-structured weak ferromagnets. Chem. Soc. Rev. 2011, 40, 3157−3181. (15) Frost, J. M.; Harriman, K. L. M.; Murugesu, M. The rise of 3-d single-ion magnets in molecular magnetism: towards materials from molecules? Chem. Sci. 2016, 7, 2470−2491. (16) Rams, M.; Tomkowicz, Z.; Böhme, M.; Plass, W.; Suckert, S.; Werner, J.; Jess, I.; Näther, C. Influence of the Metal Coordination and the Co-ligand on the Relaxation Properties of 1D Co(NCS)2 Coordination Polymers. Phys. Chem. Chem. Phys. 2017, 19, 3232− 3243. (17) Liu, K.; Shi, W.; Cheng, P. Toward heterometallic singlemolecule magnets: Synthetic strategy, structures and properties of 3d−4f discrete complexes. Coord. Chem. Rev. 2015, 289−290, 74− 122. (18) Caneschi, A.; Gatteschi, D.; Totti, F. Molecular magnets and surfaces: A promising marriage. A DFT insight. Coord. Chem. Rev. 2015, 289−290, 357−378. (19) Wang, X. Y.; Avendano, C.; Dunbar, K. R. Molecular magnetic materials based on 4d and 5d transition metals. Chem. Soc. Rev. 2011, 40, 3213−3238. (20) Martínez-Lillo, J.; Faus, J.; Lloret, F.; Julve, M. Towards multifunctional magnetic systems through molecular-programmed self assembly of Re(IV) metalloligands. Coord. Chem. Rev. 2015, 289− 290, 215−237. (21) Chorazy, S.; Stanek, J. J.; Nogas, W.; Majcher, A. M.; Rams, M.; Koziel, M.; Juszynska-Galazka, E.; Nakabayashi, K.; Ohkoshi, S.; Sieklucka, B.; Podgajny, R. Tuning of Charge Transfer Assisted Phase Transition and Slow Magnetic Relaxation Functionalities in {Fe9‑xCox[W(CN)8]6} (x = 0−9) Molecular Solid Solution. J. Am. Chem. Soc. 2016, 138, 1635−46.
observed for similar solid solutions with 4-acetylpyridine by our group. However, important chemical and physical differences become obvious that show that solid solutions of simple coordination compounds can be more complicated and cannot simply be achieved by mixing all reactants in a ratio given by the desired composition. Moreover, much effort is needed to prove if “macroscopic” homogeneous solid solutions were obtained, for which the ratio of the metal cations does not scatter very much over the sample. In the case where the metal cations are neighbors in the Periodic Table, the simultaneously investigation of physical mixtures of the individual components can be helpful. But even if samples are obtained that are homogeneous on a larger scale, it is of extraordinary importance to get some information on their homogeneity on a microscale, which can differ if different synthetic routes were used as shown in this work. For the overall goaltuning of the critical temperature by solid solution formationit appears to be unimportant, because always a linear trend is observed, but on closer inspection, one can detect micro inhomogeneities, which in fact are expected for such compounds. This automatically leads to broader magnetic transitions, which might not be disturbing in this case but which can provide important information on synthetic aspects of coordination chemistry. The synthesis of coordination compounds like, e.g., solid solutions, might be performed by different synthetic routes, but which route will be the best depends on the requirements a compound should fulfill.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.cgd.8b00865. Crystal data, PXRD patterns, IR spectra, AAS measurements, EDX measurements, DSC curves, magnetic susceptibility plots, magnetization curves, hysteresis curves, estimations of lattice contribution to specific heat (PDF).
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Christian Näther: 0000-0001-8741-6508 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This project was supported by the Deutsche Forschungsgemeinschaft (Project No. Na 720/6-1) and the State of Schleswig-Holstein and by National Science Centre, Poland (Project No. 2017/25/B/ST3/00856). We thank Prof. Dr. Wolfgang Bensch for access to his experimental facilities. Part of the experiments was carried out with equipment financed by the European Regional Development Fund, within the Polish Innovation Economy Operational Program (Contract No. POIG.02.01.00-12-023/08).
■
REFERENCES
(1) Miyasaka, H.; Julve, M.; Yamashita, M.; Clérac, R. Slow Dynamics of the Magnetization in One-Dimensional Coordination Polymers: Single-Chain Magnets. Inorg. Chem. 2009, 48, 3420−3437. 6026
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027
Crystal Growth & Design
Article
and Cooperative Magnetic Properties. Z. Anorg. Allg. Chem. 2013, 639, 2696−2714. (40) Wöhlert, S.; Ruschewitz, U.; Näther, C. Metamagnetism and Slow Relaxation of the Magnetization in the 2D Coordination Polymer: [Co(NCSe)2(1,2-bis(4-pyridyl)ethylene)]n. Cryst. Growth Des. 2012, 12, 2715−2718. (41) Suckert, S.; Rams, M.; Germann, L.; Cegiełka, D. M.; Dinnebier, R. E.; Näther, C. Thermal transformation of a 0D thiocyanate precursor into a ferromagnetic 3D coordination network via a layered intermediate. Cryst. Growth Des. 2017, 17, 3997−4005. (42) Neumann, T.; Ceglarska, M.; Rams, M.; Germann, L. S.; Dinnebier, R. E.; Suckert, S.; Jess, I.; Näther, C. Structures, thermodynamic relations and magnetism of novel stable and metastable Ni(NCS)2 coordination polymers. Inorg. Chem. 2018, 57, 3305−3314. (43) Wöhlert, S.; Runčevski, T.; Dinnebier, R. E.; Ebbinghaus, S. G.; Näther, C. Synthesis, Structures, Polymorphism, and Magnetic Properties of Transition Metal Thiocyanato Coordination Compounds. Cryst. Growth Des. 2014, 14, 1902−1913. (44) Wellm, C.; Rams, M.; Neumann, T.; Ceglarska, M.; Näther, C. Tuning of the Critical Temperature in Magnetic 2D Coordination Polymers by Mixed Crystal Formation. Cryst. Growth Des. 2018, 18, 3117. (45) Feng, X.-L.; Zhang, Y.-P. Tetrakis(ethylpyridine-4-carboxylatekN)bis(thiocyanato-kN)cobalt(II). Acta Crystallogr., Sect. E: Struct. Rep. Online 2012, 68, m786. (46) Soliman, S. M.; Elzawy, Z. B.; Abu-Youssef, M. A. M.; Albering, J.; Gatterer, K.; Ohrstrom, L.; Kettle, S. F. A. Towards the chemical control of molecular packing: syntheses and crystal structures of three trans-[NiL4(NCS)2] complexes. Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater. 2014, 70, 115−125. (47) Mydosh, J. A. Spin Glasses: An Experimental Introduction; Francis Ltd: London, 1993.
(22) Wang, Y. Q.; Yue, Q.; Gao, E. Q. Effects of Metal Blending in Random Bimetallic Single-Chain Magnets: Synergetic, Antagonistic, or Innocent. Chem. - Eur. J. 2017, 23, 896−904. (23) Chakraborty, P.; Enachescu, C.; Humair, A.; Egger, L.; Delgado, T.; Tissot, A.; Guenee, L.; Besnard, C.; Bronisz, R.; Hauser, A. Light-induced spin-state switching in the mixed crystal series of the 2D coordination network {[Zn1‑xFex(bbtr)3](BF4)2}infinity: optical spectroscopy and cooperative effects. Dalton Trans. 2014, 43, 17786−17796. (24) Erickson, P. K.; Miller, J. S. Thin film Co[TCNE]2 and VyCo1−y[TCNE]2 magnetic materials. J. Magn. Magn. Mater. 2012, 324, 2218−2223. (25) Kong, X. J.; Guo, C.; Liu, G. X.; Wang, Y.; Nishihara, S. Synthesis, crystal structure, and magnetic properties of a 3d-3d mixed heterometallic coordination polymer. Russ. J. Coord. Chem. 2012, 38, 134−139. (26) Li, B.; Zhang, X.; Tian, J.; Zhang, J. The magneto-structural correlation of two novel 1D antiferromagnetic chains with different magnetic behaviors. Polyhedron 2011, 30, 3100−3105. (27) Pokhodnya, K. I.; Vickers, E. B.; Bonner, M.; Epstein, A. J.; Miller, J. S. Solid Solution VxFe1‑x[TCNE]2·zCH2Cl2 Room-Temperature Magnets. Chem. Mater. 2004, 16, 3218−3223. (28) Vickers, E. B.; Senesi, A.; Miller, J. S. Ni[TCNE]2·zCH2Cl2 (Tc=13 K) and VxNi1−x[TCNE]y·zCH2Cl2 solid solution room temperature magnets. Inorg. Chim. Acta 2004, 357, 3889−3894. (29) Fursova, E.; Shvedenkov, Y.; Romanenko, G.; Ikorskii, V.; Ovcharenko, V. Solid solutions of heterospin molecular magnets. Polyhedron 2001, 20, 1229−1234. (30) Pokhodnya, K. I.; Burtman, V.; Epstein, A. J.; Raebiger, J. W.; Miller, J. S. Control of Coercivity in Organic-Based Solid Solution VxCo1−x[TCNE]2·zCH2Cl2 Room Temperature Magnets. Adv. Mater. 2003, 15, 1211−1214. (31) Wellm, C.; Rams, M.; Neumann, T.; Ceglarska, M.; Näther, C. Tuning of the Critical Temperature in Magnetic 2D Coordination Polymers by Mixed Crystal Formation. Cryst. Growth Des. 2018, 18, 3117−3123. (32) Palion-Gazda, J.; Machura, B.; Lloret, F.; Julve, M. Ferromagnetic Coupling Through the End-to-End Thiocyanate Bridge in Cobalt(II) and Nickel(II) Chains. Cryst. Growth Des. 2015, 15, 2380−2388. (33) Wöhlert, S.; Wriedt, M.; Fic, T.; Tomkowicz, Z.; Haase, W.; Näther, C. Synthesis, Crystal Structure and Magnetic Properties of the Coordination Polymer [Fe(NCS)2(1,2-bis(4-pyridyl)-ethylene)]n Showing a Two Step Metamagnetic Transition. Inorg. Chem. 2013, 52, 1061−1068. (34) Tahli, A.; Maclaren, J. K.; Boldog, I.; Janiak, C. Synthesis and crystal structure determination of 0D-, 1D- and 3D-metal compounds of 4-(pyrid-4-yl)-1,2,4-triazole with zinc(II) and cadmium(II). Inorg. Chim. Acta 2011, 374, 506−513. (35) Shurdha, E.; Lapidus, S. H.; Stephens, P. W.; Moore, C. E.; Rheingold, A. L.; Miller, J. S. Extended Network Thiocyanate- and Tetracyanoethanide-Based First-Row Transition Metal Complexes. Inorg. Chem. 2012, 51, 9655−9665. (36) Werner, J.; Rams, M.; Tomkowicz, Z.; Runčevski, T.; Dinnebier, R. E.; Suckert, S.; Näther, C. Thermodynamically Metastable Thiocyanato Coordination Polymer that shows Slow Relaxations of the Magnetization. Inorg. Chem. 2015, 54, 2893−2901. (37) Werner, J.; Runčevski, T.; Dinnebier, R.; Ebbinghaus, S. G.; Suckert, S.; Näther, C. Thiocyanato Coordination Polymers with Isomeric Coordination Networks − Synthesis, Structures, and Magnetic Properties. Eur. J. Inorg. Chem. 2015, 2015, 3236−3245. (38) Suckert, S.; Rams, M.; Böhme, M.; Germann, L. S.; Dinnebier, R. E.; Plass, W.; Werner, J.; Näther, C. Synthesis, structures, magnetic and theoretical investigations of Co and Ni coordination polymers with layered thiocyanate networks. Dalton Trans. 2016, 45, 18190− 18201. (39) Näther, C.; Wöhlert, S.; Boeckmann, J.; Wriedt, M.; Jeß, I. A Rational Route to Coordination Polymers with Condensed Networks 6027
DOI: 10.1021/acs.cgd.8b00865 Cryst. Growth Des. 2018, 18, 6020−6027