Peripheral Substitution: An Easy Way to Tuning the Magnetic Behavior of Tetrakis(phthalocyaninato) Dysprosium(III) SMMs

Two tetrakis(phthalocyaninato) dysprosium(III)-cadmium(II) single-molecule magnets (SMMs) with different extent of phthalocyanine peripheral substitution and therefore different coordination geometry for the Dy ions were revealed to exhibit different SMM behavior, providing an easy way to tuning and controlling the molecular structure and in turn the magnetic properties of tetrakis(tetrapyrrole) lanthanide SMMs through simple tetrapyrrole peripheral substitution.

I n addition to their interesting organic semiconducting properties [1][2][3][4][5] , the magnetic properties of sandwich-type multiple(phthalocyaninato) lanthanide complexes have aroused increasing research interests in recent years due to their significant importance in high-density storage devices, spintronics, and quantum computations [6][7][8][9] . However, most efforts paid in this direction seem to be focused on the double-decker SMMs with the aim of clarifying the structure-property relationship because of their easy synthesis and well defined molecular structure [10][11][12][13][14][15][16][17][18] . In close relationship with molecular magnet engineering, very lately the bis(phthalocyaninato) lanthanide double-decker as intrinsic SMM scaffold has been fabricated into multiple-decker, in particular quadruple-decker sandwich compounds, and their molecular magnetic properties have also been investigated in a preliminary manner [19][20][21] 8 ]} (M 1 -Cd-M 2 5 Tb-Cd-Tb, Tb-Cd-Y, Y-Cd-Tb) SMMs were designed and synthesized for the purpose of clarifying the effect of long range f-f interaction between the two lanthanide ions separated by a diamagnetic ion on the magnetic properties 20 . In addition, another series of homo-and hetero-nuclear quadruple-decker phthalocyanine analogues (Pc)M(Pc)Cd(Pc)M(Pc) and (Pc)Y(Pc)Cd(Pc)M(Pc) (M 5 Tb, Dy, Er) have also been prepared with their magnetic properties comparatively studied, confirming the long-range interactions between two f-electronic centers 22 . However, the effect of coordination geometry for the lanthanide ions sandwiched between phthalocyanines on the magnetic properties of this novel series of particular quadruple-decker SMMs still remains unexplored due to the lack of effective method in tuning and controlling the quadruple-decker molecular structure, to the best of our knowledge.

Results and Discussion
In the present paper, with the aim of tuning and controlling the molecular structure (actually the coordination geometry of the dysprosium spin carrier), two new tetrakis(phthalocyaninato) metal quadruple-decker complexes {(Pc)Dy[Pc(OC 5 8 ]}, respectively, Figure 1. Single crystal X-ray diffraction analysis clearly reveals the different skew angle (defined as the rotation angle of one ring away from the eclipsed conformation of the two rings) for the bis(phthalocyaninato) dysprosium unit in these two quadrupledecker compounds. This in turn results in their obvious different SMM behavior according to magnetic mea-surements, not only revealing the structure-magnetic property relationship but more importantly providing an easy but effective way towards effectively tuning the SMM behavior of tetrakis(phthalocyaninato) lanthanide quadruple-deckers through simple peripheral substitution.
On the basis of theoretical rationalization, tetrakis ( Single crystals of both quadruple-decker compounds suitable for X-ray diffraction analysis were obtained by diffusing methanol into the solution of corresponding compound in CHCl 3 . Compound 1 crystallizes in the monoclinic system with a C2/c, while 2 in the triclinic system with a P-1 space group. The sandwich nature of both tetrakis(phthalocyaninato) dysprosium(III)-cadmium(II) com-pounds with symmetric quadruple-decker molecular structure was confirmed by single crystal X-ray diffraction analysis, Table S3 (Supplemental Information). As shown in Figure    The static magnetic properties of the two tetrakis(phthalocyaninato) dysprosium quadruple-deckers have been investigated. The temperature dependence of the magnetic susceptibility x M T for 1 and 2 is shown in Figure S3 (Supplemental Information). The values of x M T at 300 K are 28.32 for 1 and 28.04 cm 3 K mol 21 for 2, respectively, both of which are close to 28.34 cm 3 K mol 21 that is expected for two Dy(III) ions [ 6 H 15/2 , S 5 5/2, L 5 5, g 5 4/3] 29-31 . When the temperature gets lowered, the x M T values decrease slowly until about 60 K, then decrease quickly to a minimum value of 23.32 and 22.02 cm 3 K mol 21 at 2 K. Such kind of magnetic behavior for both compounds should mainly originate from the crystal-field effect including thermal depopulation of the dysprosium(III) Stark sublevels and the presence of antiferromagnetic dipole-dipole interaction between the two adjacent double-decker subunits.
According to Curie-Weiss law, fitting the experimental data from 2 to 300 K gives the Curie constant (C), 28.38 and 28.74 cm 3 K mol 21 for 1 and 2, respectively, and Weiss constant h of 25.89 (1) and 26.56 K (2). Such a fact that the field dependence magnetization M (H/T) data at low temperature are far from the saturation magnetization value of 10 m B expected for even one Dy(III) ion (J 5 15/2, g 5 4/3), in combination with the non-superimposition character of the isothermal field dependence M vs H/T curves for 1 and 2, Figure  S4 (Supplemental Information), discloses the presence of the crystalfield effect and the magnetic anisotropy for the Dy(III) ion in the quadruple-decker compounds, suggesting their potential SMM nature [29][30][31][32][33][34][35][36][37][38] .
For the purpose to further reveal the magnetic relaxation of these compounds, the dynamics of magnetization was studied on multicrystalline powder samples of 1 and 2 in a 3.0 Oe ac field oscillating at 1.0-780 Hz. Figure 3 displays the plots of x9 vs. T x0 vs. T in a zero dc field for 1 (A, B) and 2 (C, D). As can be seen, the frequency dependent in-phase (x9) and out-of-phase signals (x0) show the slow relaxation of magnetization for the two compounds, confirming their SMM nature. Nevertheless, for 1, the x0 signal starts to show clear frequency-dependent peak at the frequency as even low as 10 Hz, while 2 shows x0 peak at 320 Hz, revealing a faster relaxation due to the quantum tunneling of magnetization (QTM) than 1 because of the larger deviation of the coordination geometry for the Dy spin carriers from the ideal square antiprism molecular symmetry as indicated by the skew angle of 23.46u for 2, in comparison with the skew angle of 41.42u for 1. Based on a thermally activated mechanism, t 5 t 0 exp(U eff /kT) and t 5 1/(2pn), the Arrhenius law fitting for the picked peaks in the x0 vs T curves in zero field for these two compounds was then carried out, revealing a linear relationship between ln(t) and 1/T in the temperature range of 2.5-6.0 K for 1 and 2.5-3.2 K for 2. This in turn results in an energy barrier with U eff 5 16.42 cm 21 (23.65 K) and pre-exponential factor t 0 5 3.84 3 10 26 s with R 5 0.996 for 1, Figure S5 (Supplemental Information), and U eff 5 12.00 cm 21 (17.28 K) and t 0 5 8.83 3 10 27 s with R 5 0.999 for 2. Obviously, the energy barrier of 1 is larger than that of 2, revealing again the effect of the deviation of the coordination geometry for the dysprosium spin carriers from the ideal SAP polyhedron on the magnetic properties of sandwich-type quadrupledecker complexes 28 . As detailed above, unlike in the solution state, 1 and 2 in the single crystal state possess two similar magic angle but significantly different skew angle for the (Pc9)Dy(Pc9) units, which then results in the different SAP environment. This in turn is responsible for the different energy barrier between these two compounds. Nevertheless, graphical representation of x0 versus x9 (Cole-Cole plot) at 2.0, 3.0, 5.0 K for both 1 and 2 give one semicircle, suggesting the existence of one magnetic relaxation processes, Figure S6 (Supplemental Information). Fitting of the experimental data according to the modified Debye function equation 39 gives the following sets of parameters with a 5 0.19-0.29 for 1 and a 5 0.24-0.25 for 2.
In addition, the dynamic susceptibility was also measured in a static magnetic field H 5 2000 Oe to suppress the QTM for both compounds. As exhibited in Figure 4, very clear peak is observed in the x0 vs T curves of 1 even at the frequency as low as 1.0 Hz, while the peaks for 2 are able to be observed only above relatively high frequency by 100 Hz. Anyway, these results indicate a typical slowing down of the relaxation mechanism. Nevertheless, the ac susceptibility data for these two compounds show an overall reduction in height due to the saturation effect. The corresponding Arrhenius law fitting for the x0 vs T data under an external 2000 Oe field gives effective energy barrier U eff 5 27.35 cm 21 (39.38 K) and t 0 5 7.91 In summary, two new sandwich-type tetrakis(phthalocyaninato) dysprosium complexes with different extent of peripheral substitution on the phthalocyanine ligands have been prepared and structurally characterized. Comparative studies in their magnetic properties reveal the close relationship between the coordination geometry of the dysprosium spin carrier and the SMM behavior. This result is surely helpful for the design and synthesis of novel sandwich-type tetrakis(tetrapyrrole) lanthanide SMMs with their molecular structure and in turn magnetic properties optimized through simple tetrapyrrole peripheral substitution.

Methods
General remarks. 1,2,4-Trichlorobenzene (TCB) and dichloromethane were freshly distilled from CaH 2 under nitrogen. Column chromatography was carried out on silica gel columns (Merck, Kieselgel 60, 70-230 mesh) with the indicated eluents. All other reagents and solvents were used as received.  1 H NMR spectra were recorded on a Bruker DPX 400 spectrometer in CDCl 3 . Spectrum was referenced internally using the residual solvent resonances (d 5 7.26 for 1 H NMR). MALDI-TOF mass spectra were taken on a Bruker BIFLEX III ultrahighresolution Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer with alpha-cyano-4-hydroxycinnamic acid as matrix. Elemental analyses were performed on an Elementar Vavio El III.
Single crystal X-ray diffraction determination. Data were collected on a Oxford Diffraction Gemini E diffractometer with Mo Ka radiation (l 5 0.7107 Å ) at 120 K. Final unit cell parameters were derived by global refinements of reflections obtained from integration of all the frame data. The collected frames were integrated by using the preliminary cell-orientation matrix. CrysAlisPro Agilent Technologies software was used for collecting frames of data, indexing reflections, and determination of lattice constants and SCALE3 ABSPACK for absorption correction. The structures were solved by the direct method (SHELXS-97) and refined by full-matrix least-squares (SHELXL-97) on F 2 . Anisotropic thermal parameters were used for the nonhydrogen atoms and isotropic parameters for the hydrogen atoms. Hydrogen atoms were added geometrically and refined using a riding model. Crystallographic data and other pertinent information for the complex are summarized in Table S3 (Supplemental Information). Selected bond distances and bond angles with their estimated standard deviation are listed in Table S4 (Supplemental Information). CCDC 926978 for 1 and CCDC 994848 for 2 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.