A new member of a class of rod-like Mn12 single molecule magnets using 2-(pyridine-2-ly)propan-2-ol

This paper reports on the synthesis, structure and magnetic properties of a new type of rod like Mn12 metal cluster, [Mn12O7(OH)2(OMe)2(dmhmp)4(O2CPh)11(H2O)] (6) where the ligand (dmhmpH) is 2-(pyridine-2-yl)propan-2-ol. Compound (6) was obtained by reacting MnCl2·4H2O with dmhmpH in the presence of benzoic salt and Et3N. The resulting crystalline material is assigned to the triclinic space group P1̄. Although compound (6) displays ferromagnetic and antiferromagnetic competition behavior, this does not prevent the molecule from functioning as a single-molecule magnet (SMM). The SMM behavior is evidenced by observing frequency dependent out-of-phase ac signals as well as magnetization hysteresis loops at low temperatures in a micro-SQUID study. A brief comparison between all rod-like Mn12 materials is also given in the manuscript.


Introduction
Because of their natural abundance and the fact that they are involved in numerous biologically related processes, 1 st row transition metal ions have attracted the interest of the scientic community. 1 For example, manganese plays a key role in photo system II (PSII). 2 Oxo-bridged manganese clusters are of great interest because of their reactivity, redox properties, ferroelectricity as well as their magnetochemistry. 3 In applications in magnetochemistry, the most important branch is singlemolecule magnets (SMMs). 4 Single molecule magnets (SMMs) are metal clusters that simultaneously have large molecular spins and signicant negative magneto-anisotropy. Upon combining these two factors, an energy barrier for spin ip is created. The manganese ion, especially Mn(III), due to its high spin number S ¼ 2, and the negative anisotropy parameter resulting from the elongation of Jahn-Teller distortion, make them ideal candidates for constructing SMMs. 5 Since the discovery of the rst SMMs [Mn 12 O 12 (O 2 CR) 16 (H 2 O) x ] (x ¼ 3 or 4) in 1993, 6 a dynasty of SMMs has arisen, not only concerning their magnetism but also metal-oxo chemistry.
MeOH that is rich in MeOH, otherwise desired compound 6 is not produced in high yields.
The ORTEP plot for compound 6 is shown in Fig. 1. The metal-ligand bond lengths are listed in Table 1. On calculating the charge by their structural data, and combining these with the BVS result, 10 unlike compound 1-3 where the cation has a charge of +34, compound 6, with a charge +33, is more similar to compound 5. Although, unlike compound 1-3 which possess inversion symmetry, all of these complexes have similar structural characteristics with a central [Mn 4 O 6 ] unit of uncompleted dicubane connected with two [Mn 4 O] tetrahedral units on both two ends of the rod. Compounds 6 and 5 have no intramolecular symmetry. All of compounds 1-3 contain two Mn(II) (Mn(6) and Mn(6 0 ) in ref. 8) ions located on the two tips on the end of the rod. In compounds 5 and 6, there are three Mn(II) ions, one of which being located on the tip of one end, the other two being located on a lateral of the tetrahedral on the other end but does not occupy the tip position. In a comparison of compounds 5 and 6, their ligands are obviously different. Compound 5 has twelve PhCOO À units, while compound 6 has eleven PhCOO À units. A detailed study of the structure of 6 revealed that the whole metal-oxo core is enclosed in a package formed from four dmhmp À ligands and eleven PhCOO À units, where each dmhmp À binds a Mn(III) ion. The PhCOO À units have three different binding modes: one monodentate h 1 type coordinated with one manganese ion, another type m 2 that bridges two metal ions in a h 1 :h 1 fashion. The nal m 3 bridges three metal ions in a h 1 :h 2 manner. (For detailed information on the binding modes, see Fig. S1. †)

Magnetic susceptibility study of compound 6
Variable temperature dc magnetic susceptibility measurements were carried on a polycrystalline sample of compound 6 restrained by paralm under a 1000 G magnetic eld in the temperature range of 2.0-300 K. The c M T vs. T data are shown in Fig. 2. It can be seen that the c M T value is 27.3 cm 3 K mol À1 at 300 K, which is signicantly smaller than the spin-only value: 40.13 cm 3 K mol À1 for nine uncoupled Mn(III) and three Mn(II) ions. This phenomenon indicates that antiferromagnetic coupling plays an important role at room temperature. The c M T values then steadily decrease to 20.0 cm 3 K mol À1 as the temperature is cooled down to 50 K and then is nearly constant down to 4 K followed by a drop to 16.8 cm 3 K mol À1 at 2 K. The abnormal feature of the c M T values indicate a ferromagnetic and antiferromagnetic competition behavior in compound 6. Nevertheless, the c M T value down to 2.0 K roughly gives a S ¼ 11/2 ground state with a g value of 1.95. The g value deviating from 2.0 can be attributed to intermolecular antiferromagnetic coupling and/or zero-eld splitting.
The spin number as well as the zero-eld splitting parameters were then obtained by variable elds, variable temperatures reduced magnetization measurements. Based on the known information regarding compound 6, it is a ferromagnetic and antiferromagnetic competition case in which low lying high spin excited states may approach to the ground state in energy. To avoid the interference of high spin excited states and to obtain the parameter of the low spin ground state, a lower magnetic eld should be applied. 11 Therefore, in this experiment, magnetic elds in the range from 1000 G to 9000 G were applied in the temperature range of 2-4 K. The experimental results are illustrated in Fig. 3, where the solid red lines are the ttings based on eqn (1).
It can be seen that the eqn (1) gives a reasonably good tting and reasonable spin parameters are obtained which gives S ¼ 11/2, g ¼ 2.0, D ¼ À0.34 K. Based on this set of parameters, we postulate that compound 6 can function as a single-molecule magnet (SMM).

a.c. magnetic susceptibility
To further conrm that compound 6 is, in fact, a SMM, a.c. magnetic susceptibility measurements were carried. In the rst attempt, we applied typical measurements on the sample which applies to an a.c. magnetic eld 3.5 G in the temperature range 1.9-9.0 K with the frequencies below 1000 Hz. Although frequency-dependent out-of-phase signals were seen, none of the signal showed peak feature above 1.9 K. We therefore examined a wider scope of oscillating frequencies from 10 to 10 000 Hz in the temperature range 1.9-6.1 K. Fig. 4 presents the results of temperature dependent and frequency dependent (log u) scans, which show two peaks above 1.9 K. The energy barrier for spin ipping was estimated by the Arrhenius formula: 12 ln(s) ¼ ln(s 0 ) + U eff /kT (2) which is shown in the inset of Fig. 4 to give a value of U eff ¼ 9.2 K. This result is somewhat lower than the theoretical value D(S 2 À 0.25) ¼ 10.2 K. This implies that somewhat quantum tunnelling of magnetization (QTM) 13 occurs at this moment.
Armed with this data, it can be concluded that compound 6 is a SMM. Further detail a.c. magnetic susceptibility examinations regarding magnetization relaxation process was analysed by the Debye model of Argand's plot (c 0 M vs. c 00 M plot) which is also known as Cole-Cole plot. 14 The Debye model is described as eqn (3) and (4): (3) where c 0 M and c 00 M are the in-phase and out-of-phase a.c. magnetic susceptibilities. Both are the functions of scanning frequencies u. The parameters c S represents the adiabatic susceptibility, c T the isothermal susceptibility and s the magnetization relaxation time. The value for parameter a varies in the range 0-1 and provide an estimation of the distribution of relaxations. An Argand's plot for compound 6 at 1.9 K is shown in Fig. 5. The parameters in eqn (3) and (4) obtained from the best tting are given as c T ¼ 11.64 cm 3 mol À1 , c S ¼ 4.09 cm 3 mol À1 , a ¼ 0.35, and s ¼ 2.03 Â 10 À3 s. From this result, we can conclude a distribution of single relaxation for compound 6.
This was also observed in many other manganese based single molecule magnets (SMMs).

Magnetization hysteresis studies
Magnetization hysteresis measurements were then carried on a single crystal of 6 in a micro-SQUID array with the magnetic eld aligned with the easy axis of the crystal. Fig. 6(a) presents the hysteresis loops observed in the temperature range 0.025-0.9 K. Obviously, the hysteresis loops can be seen at temperatures below 0.9 K providing an indication of magnet type behavior. Upon cooling, the magnitude for the coercivity of each loop increases until the lowest reachable temperature 25 mK, which reects the fact that slower magnetization relaxation occurs at lower temperatures. Since the magnetization relaxation process appears to be dominated by quantum tunnelling of the magnetization (QTM) at 25 mK, to explore the nature of the QTM, we examined the hysteresis loops when the temperature is maintained at 25 mK but varying the scan rate in the range of 0.004 T s À1 to 0.128 T s À1 . Several points are worth mentioning in Fig. 6 (b). First, all the loops show a giant transition at zero magnetic eld that can be attributed to the QTM process between the ground spin states. The fact that this occurs at zero eld indicates that there is no signicant intermolecular   interaction to shi the degeneracy of the ground states of spin multiplicity. In another words, no signicant intermolecular exchange bias 15 occurs in this crystal system. Secondly, the step occurring at zero eld increase in size when the scanning rate is slowed. This seems to be the result of a longer time interval the molecule system stays in the QTM window the larger is the QTM amplitude and therefore bigger transition steps are seen. Compared with most typical manganese SMMs [Mn 12 O 12 (O 2 -CR) 16 (H 2 O) 3$4 ], 16 the step feature for compound 6 is signicantly larger, we therefore conclude that compound 6 belongs to the category of SMMs with a fast QTM. Of course, compound 6 possesses the molecular spin S ¼ 11/2, a half integer spin that might hinder the QTM process. Nevertheless, a combination of a hyperne interaction of 55 Mn (I ¼ 5/2, 100%) as well as the lack of symmetry axis of C 3 and/or above which introduces a transverse eld, can still trigger a fast QTM in compound 6. 17

Comparisons with other rod-like Mn 12 complexes
According to the literature, four rod-like Mn 12 complexes have been reported in which two are SMMs. Combining with the compound 6, all ve compounds were synthesized using a hydroxyl-alkyl-pyridine type of ligand. Changing the alkyl group slightly causes signicant differences in the magnetic properties of these complexes. In compound 1, pure hmp ligands with alkyl ¼ methylene leads to a SMM with S ¼ 7(or 6), all of the ac frequencies above 10 Hz have their out-of-phase signals above 1.9 K. Compounds 2 and 3, both contain hep ligands with alkyl ¼ ethyl groups. No SMMs behavior was observed in this series compounds which have molecular spin S ¼ 0. Compound 5 and the current work 6 contain dmhmp with alkyl ¼ isopropyl groups. Although both exhibit SMM behavior, their spin parameters are somewhat different. Compound 5 has molecular spin S ¼ 13/2 and all of its ac out-of-phase signals show only a tail above 1.9 K. However, promising magnetization hysteresis loops have been reported at temperatures below 0.7 K. Thus, compound 5 is a SMM. The current compound 6 also contains alkyl ¼ isopropyl. However, its peripheral ligands are arranged slightly different from those for compound 5 with two coordinated hydroxide ion and a dangling benzoic acid ligand. In this compound, frequency-dependent out-of-phase signals are seen, and two peaks above 1.9 K are seen with frequencies in the interval 1000-10 000 Hz. Solid magnetization hysteresis loops are seen below 0.9 K. The obtained molecular spin for compound 6 was S ¼ 11/2. Hysteresis measurements indicates that this molecular system is a fast QTM of SMM. Details regarding the comparison of these ve compounds are given in Table 2 with their plots in the ESI (Fig. S2 †).
In conclusion, the Mn 12 rod-like complex reported in this study exhibits competing ferromagnetic and antiferromagnetic exchange interactions. As expected because Mn(II) ions are present in its structure these interactions are very weak leading to the existence of several low-lying excited states, as was also evidenced from variable eld-variable temperature magnetization studies. Minor structural changes in the structure of such complexes can lead to small changes of the exchange interactions between the Mn ions, thus leading to a different spin ground state value. This explains why the S values of compounds 5 and 6 are different.

Experimental
All chemicals used in this study were of commercial grade and were used without further purication. The ligand dmhmpH was synthesized according to a literature report. 18 [Mn 12 (dmhmp) In a 100 mL beaker, the ligand dmhmpH 0.13 g (1 mmol) and triethylamine Et 3 N 0.14 mL (1 mmol) were dissolved in a mixed solvent MeCN/MeOH (1/29 mL). MnCl 2 $4H 2 O 0.2 g (1 mmol) and NaO 2 Ph 0.29 g (2 mmol) were added and the mixture was then stirred overnight. Aer ltration, diethylether was allowed to diffuse into the ltrate. Dark black crystals suitable for X-ray crystallographic analysis was obtained aer two weeks. Selected IR data (cm X-ray crystallography X-ray analysis data for 6 was collected on a Bruker D8 Venture diffractometer with Mo Ka radiation (l ¼ 0.71073Å). The temperature was cooled to 200(2) K. The corresponding data was collected by Bruker APEX3 program, and simplied by Bruker SAINT program. Data were corrected by the empirical absorption SADABS program. 19 The SHELXTL program on a PC computer was used for the structure analysis. The structure was solved by direct method of the Shelxs program 20 and renement was done by the Shelxl program 21 based on the full-matrix least squares on F 2 values. Non-hydrogen atoms were rened anisotropically. Hydrogen atoms were xed at the calculated positions and rened using a riding mode. Table 3 summarizes the structure renement parameters for complex 6.
Physical property measurements NMR spectra were collected on a Bruker AV-300 spectrometer. Infrared spectra were obtained using KBr pellets on a Perki-nElmer 1600 spectrometer in the 450-4000 cm À1 range. DC magnetic measurements were collected using a Quantum Design MPMS7 system. Samples were restrained with paralm to prevent torqueing. The magnetic background caused by the gel cap and the paralm were subtracted by blank measurements. The diamagnetism correction was estimated from Pascal's constant. 22 Elemental analyses (C, H, N) and dc magnetic susceptibility measurements were carried by the National Taiwan University Instrument Centre, College of Science. AC magnetic susceptibility data were collected using the Quantum Design MPMS XL7 system of the National Chiao Tung University. Oriented single crystal magnetization hysteresis loops were measured by employing a micro-SQUID array that was described elsewhere. 23 A single crystal was put onto the array and external eld is oriented parallel to the crystal easy axis.

Conclusions
Many factors in a synthetic procedure can inuence the geometries of the resulting products. For example, when hmpH ligands were used, a more symmetric geometry of compound 1 was obtained. SMM type behavior was observed in this compound. While hepH was employed, two Mn 12 clusters (2 and 3) with S ¼ 0 were obtained and none were found to be SMMs. Upon switching the backbone of the hmpH ligand to isopropyl (dmhmpH), the situation becomes more sophisticated. When MnCl 2 and tBuCO 2 À are used as the reactants in a MeCN rich solution, a Mn 7 cluster (4) was isolated which is not a SMM. When Mn(O 2 CPh) 2 was used in a CH 2 Cl 2 rich solution a more distorted rod like Mn 12 (5) was isolated which was shown to be a SMM. Attempts to explore other geometries of manganese clusters by utilizing MnCl 2 and PhCO 2 À as the reactants in a MeOH rich solution, a new type Mn 12 (6) was obtained. According to the work herein, we conclude that this compound is a SMM. Obviously, a number of factors, including peripheral ligands and solvent systems play important roles in determining the geometry of the manganese cluster and so do their magnetic properties. Although without the ability to carefully control the parameters that affect the nature of the product, this study clearly shows that approaches are available to signicantly tune the magnetic properties of the metal clusters.

Conflicts of interest
There are no conicts to declare.