Magneto-structural Correlations in Ni2+-Halide···Halide-Ni2+ Chains

Abstract

We present the magnetic properties of a new family of S = 1 molecule-based magnets, NiF₂(3,5-lut)₄·2H₂O and NiX₂(3,5-lut)₄, where X = HF₂, Cl, Br, or I (lut = lutidine C₇H₉N). Upon creation of isolated Ni-X···X-Ni and Ni-F-H-F···F-H-F-Ni chains separated by bulky and nonbridging lutidine ligands, the effect that halogen substitution has on the magnetic properties of transition-metal-ion complexes can be investigated directly and in isolation from competing processes such as Jahn-Teller distortions. We find that substitution of the larger halide ions turns on increasingly strong antiferromagnetic interactions between adjacent Ni²⁺ ions via a novel through-space two-halide exchange. In this process, the X···X bond lengths in the Br and I materials are more than double the van der Waals radius of X yet can still mediate significant magnetic interactions. We also find that a simple model based on elongation/compression of the Ni²⁺ octahedra cannot explain the observed single-ion anisotropy in mixed-ligand compounds. We offer an alternative that takes into account the difference in the electronegativity of axial and equatorial ligands.

Figure 1

Phase diagram of the S = 1 antiferromagnetic chain recreated from refs (17) and (18). Compounds for which J_⊥ values are not quantitatively established are represented by arrows along the x-axis. From the results in ref (19), it is known that the J_⊥/J ratio for NiI₂(3,5-lut)₄ is extremely small. We expect it will be the same for the isostructural NiBr₂(3,5-lut)₄. Reproduced with permission from ref (17). Copyright 2014 American Physical Society. Reproduced with permission from ref (18). Copyright 2014 World Scientific Publishing.

Figure 2

Low-temperature structure of NiX₂(3,5-lut)₄. (a) Layout of the local environment around each Ni²⁺ ion (silver). (b) Ni–HF₂···HF₂–Ni chain showing that the F–H–F molecule axis is oriented along the c direction in Ni(HF₂)₂(3,5-lut)₄. (c) Ni–I···I–Ni chains in NiI₂(3,5-lut)₄ (isostructural with X = Cl and Br), where Ni²⁺ ions in adjacent chains are offset in the c direction. (d) Lutidine molecules keep Ni–X···X–Ni chains well separated in the a–b plane. Lutidine hydrogen atoms have been omitted for the sake of clarity.

Figure 3

150 K structure of NiF₂(3,5-lut)₄·2H₂O. (a) Ni–F···F–Ni chain with lutidine paddle wheels rotated by 45° on adjacent sites. Water molecules located between F^– ions along Ni–F···F–Ni bond pathways are positionally disordered (water hydrogen atoms have been omitted for the sake of clarity). (b) View down the Ni²⁺ chains highlighting the rotation of the lutidine paddle wheels along the c-axis. Lutidine hydrogen atoms in panels a and b have been omitted for the sake of clarity.

Figure 4

Magnetometry data for single crystals of NiF₂(3,5-lut)₄·2H₂O. (a) Zero-field-cooled susceptibility (χ_mol) measured at μ₀H = 0.1 T for the field applied along the z-axis (orange triangles) and within the x–y plane (purple circles). Solid lines are fits to the models described in the text. (b) Pulsed-field magnetization (line) calibrated using similar temperature dc-field SQUID (circles) measurements with the field applied within the x–y plane. (c) Pulsed-field magnetization (line) calibrated using g_z obtained from modeling χ(T) with the field applied parallel to the z-axis. SQUID data (triangles) are also shown. The differences between the data sets are caused by a magnetocaloric cooling in the pulsed-field measurements. The inset shows a peak in the pulsed-field dM/dH data (line) centered at μ₀H_c is consistent with the feature in the 1.8 K SQUID dM/dH data (triangles).

Figure 5

Magnetometry data for an aligned single crystal of Ni(HF₂)₂(3,5-lut)₄. (a) Zero-field-cooled susceptibility plotted vs temperature for the field applied parallel (blue triangles) and perpendicular (black circles) to the crystallographic c-axis. Solid lines correspond to a fit to the model described in the text. The inset is a representative sketch of an S = 1 energy-level diagram, showing the labeling scheme for transitions observed in ESR measurements of Ni(HF₂)₂(3,5-lut)₄ and NiBr₂(3,5-lut)₄. (b) Pulsed-field magnetization (solid black line) calibrated to similar temperature DC-field SQUID data (black circles) with the field perpendicular to the c-axis. SQUID magnetometry data with the field parallel to the c-axis (blue triangles) are also shown.

Figure 6

(a) ESR spectra of Ni(HF₂)₂(3,5-lut)₄ recorded at 20 K and frequencies of 203.2, 321.6, and 406.4 GHz. Large resonances are observed in all three spectra, which were fitted using a model with single-ion anisotropy D. The obtained parameters were then simulated (Sim) in good agreement with the data. (b) Field frequency positions of the observed resonances (circles) and the result of the fit (lines) with the following extracted parameters: g_xy = 2.22(1), g_z = 2.16(2), and D = +11.0(1) K. (c) Temperature dependence of the 321.6 GHz Ni(HF₂)₂(3,5-lut)₄ ESR spectra.

Figure 7

Illustration of magnetization ellipsoids for Ni atoms in a single crystal of Ni(HF₂)₂(3,5-lut)₄ as determined by refinement of polarized neutron diffraction data (see the text): (a) high-temperature, high-field data and (b) low-temperature, high-field data. The ellipsoids in panel b confirm the xy orientation of the Ni²⁺ moment in the paramagnetic state attributed to D > 0.

Figure 8

(a) Temperature-dependent susceptibility measurements of polycrystalline NiCl₂(3,5-lut)₄ performed at an applied field of 0.1 T and fitted to a model including a single-ion anisotropy contribution (see the text). The inset of panel b shows inverse susceptibility data of NiCl₂(3,5-lut)₄ fitted to a Curie–Weiss model and a small temperature-independent contribution in the range 36–300 K. (b) Magnetization and differential susceptibility (inset) of polycrystalline NiCl₂(3,5-lut)₄. A feature indicating a ground-state energy-level crossing is apparent in the dM/dH data at μ₀H_c = 6.6(5) T. The sharp rise below 0.3 T is due to an artifact of the measurement.

Figure 9

(a) Temperature-dependent susceptibility measurements of powdered NiBr₂(3,5-lut)₄ suspended in Vaseline performed at an applied field of 0.1 T. The data are fitted to a model containing both single-ion and magnetic exchange terms [green (see the text)]. The inset shows inverse susceptibility data fitted to a Curie model in the range of 50–300 K. (b) Pulsed-field magnetization and differential susceptibility (inset) of NiBr₂(3,5-lut)₄. Critical fields are observed in the dM/dH data indicating saturation of moments for the field parallel to the easy plane at μ₀H_sat^xy = 2.3(2) T and the easy axis at μ₀H_sat^z = 11.1(5) T.

Figure 10

(a) ESR spectra of polycrystalline NiBr₂(3,5-lut)₄ collected in derivative mode at low temperatures and high frequencies. Red arrows show the position of the γ_z transition, and blue arrows show the positions of the γ_xy (at high field) and β_xy (at low field) transitions. Gray arrows and ∧ mark the off-axis and double-quantum resonances, respectively. (b) Plot of frequency vs field showing the peak positions (circles). The lines are the expected locations of resonances from the fit to the experimental data described in the text, with the dark gray line and circles corresponding to the off-axis resonances and the brown circles referring to the double-quantum resonances. (c) Temperature dependence of the 260 GHz spectra. The orange line is a simulation at 5 K using the parameters obtained from the fit in panel b.

Figure 11

Bond length and bonded atom electronegativity dependence of D for NiX₂(3,5-lut)₄ (X = HF₂, Cl, Br, or I), NiF₂(3,5-lut)₄·2H₂O, NiCl₂(pyz)₂,^, NiF₂(pyz)₂·3H₂O, and [Ni(pyz)₂(HF₂)₂](SbF₆). The compounds presented in this paper are labeled with X_lut, with their NiX₂(pyz)₂ relatives labeled with X_pyz. The larger red circles are strongly easy-plane, and the smaller bluer circles correspond to an easy-axis D.