Dynamic spin interchange in a tridentate Fe(iii) Schiff-base compound

Dynamic spin interchange where crystals explode with preservation of magnetic memory is observed for a mononuclear hysteretic Fe(iii) Schiff-base compound.


Thermogravimetry experiments
TGA experiments in the range of 295 K to 425 K were carried out at a heating rate of 5 K min -1 on a Perkin Elmer TGA7 apparatus. The balance chamber was kept under a positive flow of nitrogen (Air Liquide N45) of 38 cm 3 min -1 . The sample purge gas was also nitrogen (Air Liquide N45) at a flow rate of 22.5 cm 3 min -1 . The samples with masses of ~6 mg were placed in an open platinum crucible. The mass scale of the instrument was calibrated with a standard 100 mg weight and the temperature calibration was based on the measurement of the of the Curie points (T C ) of alumel alloy (Perkin-Elmer, T C = 427.35K) and nickel (Perkin-Elmer, mass fraction 0.9999, T C = 628.45 K) standard reference materials.
The results of the TG experiments are shown in Figure S4 and Table S1 where m is the initial mass of sample; ∆m is the overall mass loss; and T on , T m , and T end are the temperatures corresponding to the onset, inflection point and end of the mass loss process, respectively. The mean values of the obtained data are: T on =342.7±3.0 K, T m =342.9±3.0 K, T end =343.7±2.8 K, and ∆m = 0.038±0.010 mg. The mean mass loss in percentage is 100∆m/m = 0.57±0.15%. The uncertainties quoted for all quantities are twice the standard error of the mean. 1

Mössbauer spectroscopy measurements
The Mössbauer spectra were recorded in transmission mode at room temperature and at lower temperatures using a conventional constant-acceleration spectrometer and a 50 mCi 57 Co source in a Rh matrix. The low temperature measurements were performed using a liquid helium flow cryostat with a temperature stability of 0.5 K. The velocity scale was calibrated using an α-Fe foil. The spectra were fitted to Lorentzian lines using the WinNormos software program, and the isomer shifts reported are relative to metallic α-Fe at room temperature.
To understand the origin of the doublet asymmetry, spectra were also recorded in the magic angle configuration (54.7° between the α-ray direction and the normal to the absorber). As can be seen from Figure S5, the spectra collected in the "magic angle" geometry at 4.2 K still exhibits the asymmetric doublet, ruling out the texture effect as a possible cause for the doublet lines asymmetry. 2,3 Figure S5 -57 Fe Mössbauer spectrum for [Fe(5-Br-salEen) 2 ]ClO 4 at 4.2 K, measured at the "magic angle" (54.7º) in relation to the  radiation direction.
The spectra measured at different temperatures were also analysed considering two singlet lines. For all temperatures, the same line intensity but different widths were always obtained for the two singlets. Besides, from the guidelines of Figure S6, it is apparent that the asymmetry of the lines is lower at 290 K. These observations led to the conclusion that the Goldankii-Karyagin effect (based on the anisotropy of the Debye-Waller factor for a nucleus in a site with symmetry lower than cubic, predicting different intensities for the doublet lines and that the asymmetry increases with increasing temperature), 4,5 could not explain the lines asymmetry for this complex.  Figure S7 shows the Mössbauer spectrum at 4.2 K obtained from fittings based on one, two and four quadrupole doublets, assuming equal widths and intensities for both doublet lines. The fits considering either one or two sites did not reproduce correctly the experimental spectrum and the one with four sites gave rise to unreasonably different isomer shift and quadrupole splitting values. Figure S8 displays the same spectrum fitted with one quadrupole doublet, leaving the relative lines width (parameter W 21 ) as an adjustable parameter. Both figures and the fitting parameter values presented in Table S3 show that the best fit is the one displayed in Figure S8, indicating that the quadrupole doublet asymmetry should be related to a relaxation phenomenon. This behavior, found in other LS Fe(III) compounds, [6][7][8][9] has been attributed to relatively long paramagnetic relaxation times of the iron when compared to the 57 Fe nuclear Larmor precession time.   Figure S9 -57 Fe Mössbauer spectra measured at 290 K: a) before heating the sample; b) after heating the sample up to 370 K. The hyperfine fitting parameters are shown in Table S3.
As can be seen through Figure S9 and fitting parameters in Table S3, the 290 K spectrum obtained after heating the sample up to 370 K, although very broad, allows to deduce that the major contribution still comes from LS Fe ions. In fact, the  value increases from 0.14 mm s -1 (at 290 K before heating) to 0.24 mm s -1 (at 290 K after heating) indicating a higher contribution of HS states, and the new E Q value (2.0 mm s -1 ), although below the one obtained for the 290 K spectrum before heating (2.6 mm s -1 ) is still well above the E Q values for HS Fe(III) in similar complexes (< 1 mm s -1 , work to be published). Therefore, both hyperfine parameters of the spectrum at 290 K after heating the sample up to 370 K indicate the predominance of LS states and reveal a spin flipping rate higher than 1/ N , where  N is the mean life time of the 57 Fe first excited nuclear state (140 ns), not allowing to quantify the amount of LS and HS contributions.       117.0 Symmetry transformations used to generate equivalent atoms: #1 -x+1,y,-z+1/2 #2 -x+3/2,-y+3/2,z-1/2 #3 -x+3/2,y-1/2,z SI15 Figure S10 -Strictly equivalent views of the ligand in the complex cation for the state i (125 K), on the left, and state iii (300 K), on the right, illustrating conformational differences.