Evidence for Stable High-Temperature Ferromagnetism in Fluorine-Treated C60

It is shown by magnetic �eld dependent ac susceptibility, magnetic force microscopy, and ferromagnetic resonance that exposure of C60 to �uorine at 160 C produces a stable ferromagnetic material with a Curie temperature well above room temperature. e exposure to �uorine is accomplished by decomposing a �uorine-rich polymer, tri�uorochloroethylene [F2C–CFCl]n, which has C60 imbedded in it. Based on previous experimental observations and molecular orbital calculations, it is suggested that the ferromagnetism is arising from crystals of C60–F.


Introduction
ere has been much interest in the material science community in synthesizing molecular-based ferromagnetic materials because of the potential to chemically engineer their properties and the possible ease of production. e C 60 molecule has played an important role in this possibility because of a number of reports of ferromagnetism in it. In 1991 a complex of C 60 and the strong electron donor molecule C 2 N 4 (CH 3 ) 8 (TDAE) were found to be ferromagnetic at 16.1 K [1]. Subsequently ferromagnetism having a Curie temperature of 500 K was reported in a two-dimensional polymeric form of C 60 produced by high pressure (6 GPa) and high temperature in the vicinity of 1000 K [2][3][4]. It is known that C 60 when subjected to UV light forms oligomers most of which are dimers [5]. It has been reported that when the photolysis is done in the presence of oxygen the material is ferromagnetic well above room temperature [6,7]. ere are a number of reports of ferromagnetism in halogenated C 60 . C 60 subjected to a heat treatment in the presence of iodine is shown to be ferromagnetic having a Curie temperature of 60 K [8,9]. Doping with a mix of iodine and bromine produced a material which was ferromagnetic below 30 K [10,11].
In this work it is shown by magnetic �eld dependent AC susceptibility, magnetic force microscopy, and ferromagnetic resonance that C 60 treated with �uorine is ferromagnetic well above room temperature. e C 60 was exposed to �uorine by embedding it in a �uorine-rich polymer, polytri�uorochloroethylene, [F 2 C-CFCl] , (PTFCE) and decomposing the polymer-C 60 mix at high temperature to produce �uorine.

Experimental
e paramagnetic and ferromagnetic resonance measurements were made using a Varian E-9 spectrometer operating at 9.2 GHz with 100 KHz modulation. e temperature of the sample was controlled by �owing heated or cold nitrogen gas through a double-walled quartz tube, which is part of an ADP Heli-Tran system. is system is inserted through the center of the microwave cavity. e magnetization was obtained by measuring the magnetic �eld dependence of the ac susceptibility at 350 kHz using a method similar to that described by Clover and Wolf [12]. e system consists of an HP 204C LC oscillator modi�ed to have an external coil. e sample is contained in the coil, which is in a cryogenic dewar between the poles of a magnet. e change in the frequency of the oscillator, which is proportional to the change in susceptibility, is measured as a function of dc magnetic �eld strength using a HP 5314 frequency counter. e relative susceptibility as a function of dc magnetic �eld is measured by taking the difference in frequency between zero �eld and a given applied �eld. is method of measuring the susceptibility is quite sensitive but not widely used. Magnetic force microscope images were obtained using a Veeco Nano scope IV equipped with a magnetic tip. Raman measurements were made using a J. Y. Horiba confocal micro-Raman system employing a 25 mW He-Ne laser having a wavelength of 632.8 nm and focused to a spot of a 15-micron radius.
Crystalline C 60 , having 99.99% purity, was obtained from the Aldrich Chemical Company. Polytri�uorochloroethylene (PTFCE) was obtained from the Halocarbon Products Corporation. In this paper we will report clear evidence for ferromagnetism in a mixture of C 60 and PTFCE subjected to heat treatment. In any report of ferromagnetism in an organic material, the purity of the starting materials is a critical issue, and we have paid considerable attention to sample analysis. It is particularly important to make sure no magnetic impurities are present in the starting materials. e starting C 60 was analyzed by induction coil plasma mass spectrometry (ICP-MS). e results indicated that signals from all metals were less than 1 part per billion (PPB). Electron paramagnetic resonance (EPR) of the C 60 showed no evidence of the presence of magnetic elements or complexes. EPR is sensitive to 1 part in 10 10 . A very narrow line at 000 was detected which has previously been identi�ed as the C 60 anion [13]. e PTFCE was also subjected to detailed analysis. Energy dispersive X-ray spectroscopy of PTFCE showed only the presence of carbon, �uorine, and chlorine, the constituents of the polymer. No other elements were detected. No magnetic species were detected by EPR. ICP-MS showed no magnetic metals above PPB in the PTFCE.
In a typical synthesis, 0.084 grams of PTFCE were dissolved in acetone and 0.040 grams of C 60 added to the solution. e solution, while subjected to sonication, was allowed to slowly evaporate. e resulting residue was dried for some hours at 50 ∘ C to remove any entrapped acetone. e composite was heated to 160 ∘ C for two minutes and then rapidly quenched to room temperature.

PTFCE as a Source of Fluorine.
PTFCE was heated to 160 ∘ C for one hour and the Raman spectra recorded before and aer the heating. No difference was observed between the frequencies of the Raman spectra indicating the material is not decomposing in the condensed phase.
However, the material sublimes at this temperature. In a second experiment the polymer was heated to 160 ∘ C for one hour in a small beaker which was on a hot plate. e beaker had a chilled slide on the top which enabled collection of the condensed vapor. e Raman spectra of the material condensed on the slide is shown at the bottom of Figure  1. On the top are the spectra in the same frequency region obtained from the polymer before heating. e two lines at 1198 cm −1 and 1298 cm −1 , which are due to C-F vibrations, are completely gone in the spectra of the vapor indicating single �uorines are removed from the carbon atoms of the polymer in the vapor phase providing a source of atomic �uorine rather than diatomic �uorine. Figure 2 is a plot of the ac susceptibility versus dc magnetic �eld at 300 K for the heattreated C 60 -TFPCE. e susceptibility has been normalized to the measured value at 3 Kilo Gauss. By comparing with a measurement on a sample of known magnetization, it is estimated the magnetization at 3000 Gauss is 0.07 emu/gm. No �eld-dependent magnetization is observed in the separated starting materials subjected to the same treatment. Figure 3 shows the details at lower �elds for increasing and decreasing dc magnetic �eld to and from 3000 Gauss showing a small hysteresis. Figure 4 presents the temperature dependence of the magnetization above room temperature measured in a 3000 Gauss �eld and normalized to its value at 300 K. � �t of this data to the Bloch equation,

Magnetization Measurements.
yields a value of of 4.1246 × 10 −5 . e temperature at which ( ) is zero, the Curie temperature, is estimated to be 837 K, from this �t. is is likely an overestimate because of deviations from the Bloch law near due to critical �uctuations. �owever, it is close to estimates of the Curie temperature obtained in C 60 photolyzed in the presence of oxygen.

Magnetic Force Microscope Measurement.
e PTFCE-C 60 material which was heat treated was pressed �at on a slide in order to have a smooth surface. Figure 5 shows a magnetic force microscope image of the material. e brighter regions indicate areas of ferromagnetism.

Ferromagnetic Resonance
Measurements. e �uorinetreated material was heated above its melting point and a 4000 G magnetic �eld applied. e material was then cooled below the melting temperature in the magnetic �eld. is aligns and locks in the direction of maximum magnetization parallel to the direction of the applied magnetic �eld. Figure  6 shows the ferromagnetic resonance spectra for the sample oriented perpendicular and parallel to the direction of the cooling �eld. For a particle having axial symmetry, the dependence of the �eld position of the FMR signal is given by [14], where = 4 / is the anisotropy �eld, is the magnitude of the anisotropy constant, and is the magnetization. e angle, , is between the direction of maximum magnetization and the applied dc �eld, is the magnetic �eld at the center of the FMR signal, and 0 determines the value. Because is lower than the parallel orientation compared to the perpendicular orientation, it can be concluded that is positive. Fitting the data in Figure 6 to (2) allows determination of 0 and . e values for 0 and are 2730 G and 93 G, respectively, at 300 K. e value is 2.287. One of the characteristics of an FMR signal as opposed to an EPR signal is a strong temperature dependence of the line width and �eld position of the resonance. Figure 7 is a plot of the temperature dependence of the �eld position for the sample oriented perpendicular to the direction of the cooling �eld showing a pronounced decrease in the �eld position with decreasing temperature. Figure 8 plots the line width as a function of decreasing temperature showing a marked broadening as the temperature is lowered. e data in Figures  7 and 8 con�rm that the signal is an FMR signal.

A Possible Model for the Origin of Ferromagnetism.
ere has been a previous report of ferromagnetism in C 60 subjected to a different �uorine treatment than that used here [15]. It was observed in C 60 ultrasonically dispersed in dimethylformamide (DMF) solution of polyvinyl di�uoride (PVDF), [H 2 C-CF 2 ] . However, the observed magnetism was not stable having a half-life of 30 hours at room temperature. Further the results were only reproducible 1 out of 15 times. In our observation the results are stable over years and reproducible once the correct synthesis conditions are determined. However, an interesting observation in the PVDF solution made material is that ionization time of �ight mass spectrometry showed the formation of C 60 oligomers having �uorine atoms bonded to each C 60 . Previous density functional molecular orbital calculations of the minimum energy structure of the F-C 60 -C 60 -F dimer indicated that the triplet state has a lower energy than the singlet state by 0.55 eV [16,17]. is suggests that the �uorinated dimer could be a possible source of the unpaired spin necessary to form a ferromagnetic state. However a calculation of the bond dissociation energy (BDE) to dissociate the dimer into 2(F-C 60 ) indicates the dimer would not be stable above 400 K in disagreement with experimental observations.
As discussed in the introduction there have been a number of reports of ferromagnetism at lower temperatures in C 60 subjected to heat treatment in the presence of I, IBr, and H [8][9][10][11]. X-ray diffraction measurements of these materials indicated the ferromagnetism was arising from a cubic phase of C 60 where the C 60 was functionalized with halogens or hydrogens. is is a likely possibility for the structure of the ferromagnetic phase of the �uorinated C 60 observed here. A calculation of the BDE to remove F from C 60 gives a value of 4.36 eV indicating that C 60 -F is stable above 400 K further supporting this possibility [17].
ere has been considerable work done on the �uorination of C 60 which has been discussed in a number of reviews [18,19]. �enerally the �uorination is achieved by exposing C 60 to F 2 gas at high temperature, which of course results in C 60 F where is even. An even would have no net spin. e method of synthesis used here is likely exposing C 60 to atomic �uorine produced by the removal of F from the TFPCE polymer. is would result in being odd giving the entity a net unpaired spin.

Conclusion
Field dependent ac susceptibility, ferromagnetic resonance, and magnetic force microscopy clearly show that exposing C 60 to �uorine at high temperature produces a ferromagnetic phase having a Curie temperature well above room temperature. e results cannot be explained by the presence of magnetic impurities such as Fe, Ni, or Co or compounds of them because ICP-MS and EPR of the starting materials indicate that such materials are present at less than one part per billion. While the structure of the ferromagnetic phase is not determined, some possibilities can be considered. Dimers of C 60 have been suggested to be the origin of the unpaired spin. DFT calculations of the BDE of F-C 60 = C 60 -F indicate it unlikely to be stable above 400 K. On the other hand C 60 -F is predicted to be stable at this temperature. is suggests the structure of the ferromagnetic phase may be cubic lattice of C 60 -F, as observed for ferromagnetism in C 60 -I, C 60 -IBr, and C 60 -H. However further work is needed to con�rm this possibility.