Report

Magnetic Ordering in an Organic Polymer

See allHide authors and affiliations

Science  16 Nov 2001:
Vol. 294, Issue 5546, pp. 1503-1505
DOI: 10.1126/science.1065477

Abstract

We describe preparation and magnetic properties of an organic π-conjugated polymer with very large magnetic moment and magnetic order at low temperatures. The polymer is designed with a large density of cross-links and alternating connectivity of radical modules with unequal spin quantum numbers (S), macrocyclicS = 2 and, cross-linking S = ½ modules, which permits large net S values for either ferromagnetic or antiferromagnetic exchange couplings between the modules. In the highly cross-linked polymer, an effective magnetic moment corresponding to an average S of about 5000 and slow reorientation of the magnetization by a small magnetic field (less than or equal to 1 oersted) below a temperature of about 10 kelvin are found. Qualitatively, this magnetic behavior is comparable to that of insulating spin glasses and blocked superparamagnets.

Recent investigations into the properties of organic magnets have generally concentrated on materials, such as crystalline solids of small molecule radicals or charge transfer salts, in which the exchange interaction involves s- and p-orbitals (1–5). An alternative approach to organic magnets may be based on π-conjugated polymers, as envisioned by Mataga in 1968 (6). Because exchange interactions between electron spin through the π-conjugated system can be made relatively strong, compared to the through-space interactions in molecular solids of organic radicals, this macromolecular approach has a potential for obtaining interesting magnetic properties at relatively high temperatures, even room temperature (7). Although significant progress has been made in the preparation of π-conjugated oligomers and polymers with large values of spin quantum number S, organic polymer magnets have remained elusive (8–15). We report the observation of magnetic properties comparable to that of insulating spin glasses and blocked superparamagnets in an organic π-conjugated polymer.

Our magnetic polymer, polymer 1, is designed with a large density of cross-links and alternating connectivity of radical modules with unequal spin quantum numbers, i.e., macrocyclic S= 2 and cross-linking S = ½ modules (13). This connectivity permits large net Svalues for either ferromagnetic or antiferromagnetic exchange couplings between the modules (Fig. 1A). The cross-linking and connectivity is set in the synthesis of network polyether 2, precursor to polymer 1, based upon Pd-catalyzed Negishi coupling of two tetrafunctionalized macrocyclic monomers 3 and 4 (Fig. 1B) (13,16).

Figure 1

(A) Polymer 1 with ferromagnetic or antiferromagnetic coupling between the macrocyclicS = 2 and cross-linking S = ½ modules. (B) Synthesis of polyether 2 and polymer1. Reagents: (a) t-BuLi, THF, 198 K (2 hours), 253 K (15 min); (b) ZnCl2, from 198 K to ambient temperature; (c) Pd(PPh3)4 (3 mol % per CC bond), THF, 373 K (10 min to 5 hours); (d) Na/K, 15-crown-5, THF-d8, 283 K (several days); and (e) I2, 167 to 170 K.

Representative data for condensations leading to organic insoluble polyether 2 are shown in Table 1 (17, 18). Dissolved metal (Na/K/15-crown-5) reduction of polyether 2, as a gel with perdeuterated tetrahydrofuran (THF-d8), followed by the iodine oxidation of the corresponding carbopolyanion, gave polymer1 (Fig. 1 and Table 1) (19).

Table 1

Synthesis of polyether 2 and generation of polymer 1.

View this table:

The samples of polymer 1 obtained from polyethers2 after long polymerization times (Table 1, run 3) have the most interesting magnetic behavior. Magnetic data of the thermally decomposed polymer 1 (essentially diamagnetic) and metal analyses both preclude any significant interference from magnetic metals on the reported magnetic behavior for polymer 1(19). The magnetic field (H = 0 to 50 kOe) dependence of magnetization (M), measured at several temperatures (T = 1.8 to 20 K), shows an extraordinarily fast rise at low H; however, complete saturation is not attained even at H = 50 kOe (Fig. 2A) (20). Numerical fits of the M versus H/T data (T = 1.8, 2.5, 3.5 K) to a linear combination of one Langevin and three Brillouin functions give average spin quantum numbers (S) and magnetizations at saturation (M sat) that are temperature-dependent (19, 21). Again, because ideal paramagnet (Curie-like) behavior is not found, the maximum value of average S ≈ 5000 at 3.5 K, obtained from theM versus H/T plot, provides only a very approximate estimate for an effective ferro- or ferrimagnetic correlation of 104 electron spin. Values ofM sat = 0.5 to 0.6 μB and their increase with increasing temperature are both compatible with the presence of weak antiferromagnetic and ferromagnetic interactions between the S = 2 and S = ½ modules in polymer 1 (Fig. 1).

Figure 2

(A) Magnetic field (H) dependence of the magnetization (M) of polymer 1at T = 3.5 K, plotted as M/M satversus H/T, where M sat = 0.54 μB is M at saturation. The solid line corresponds to the least-squares fit using linear combination of Langevin and Brillouin functions corresponding to averageS = 5400 (19). (B) Plot of χT versus T (dc susceptibility, χ =M/H).

The plots of χT versus T (dc susceptibility, χ = M/H) show a rapid rise below about 10 K. The values of χT are highly field- dependent, as expected for very large magnetic moments (or values of S). For the smallest applied field, H ≈ 0.5 Oe, χTreaches maximum of about 1000 electromagnetic units (emu) K mol−1 at 3.6 K (Fig. 2B). Very approximately, this corresponds to an effective magnetic moment μeff ≈ 8000 μB or an average S ≈ 4000, comparable to that found in the M versus H/Tplots. These values may be viewed as lower bound estimates, because polymer 1 does not follow the Curie law and quantitative conversion of polyether 2 to polymer 1 is assumed.

More detailed studies of the temperature dependence of M at low applied magnetic fields (H = 0.5 to 1.0 Oe) reveal that the zero field–cooled (ZFC) and the field-cooled (FC) magnetizations diverge below ∼10 K, indicating the slow relaxation (blocking) of the magnetization (Fig. 3). This behavior is highly sensitive to the applied magnetic fields (H); in larger magnetic fields, e.g., H = 5 Oe, the difference between the ZFC and FC magnetizations is undetectable. Even at a relatively low magnetic field,H = 5 Oe, the large values of μeff ≈ 8000 give μeff H/k B≈ 2.7 K, which is within the range of blocking temperatures.

Figure 3

Temperature dependence of the magnetization of polymer 1 measured by increasing the temperature in the indicated field. For H = 0.5 Oe measurements, the sample was cooled either in the zero-field (ZFC) or in the 0.5 Oe field (FC) prior to the measurement. (Inset) Difference between the FC and ZFC magnetization.

The relaxation of the magnetization for polymer 1 at low temperatures is best observed in the temperature and frequency dependence of the ac susceptibility. The temperature dependence of χ′ (in-phase component) shows a steep rise around 10 K, with a broad peak in the 1.9 to 3 K region (Fig. 4A) (19). The peak maximum in χ′ at T m′ shifts to lower temperatures, and it gains intensity with decreasing frequency. The out-of-phase component, χ", is detectable at about 10 K and shows a frequency-dependent maximum in the 1.8 to 2.0 K region (Fig. 4B). The peak maximum in χ" at T m" shifts to lower temperatures and becomes less intense with decreasing frequency; at lower frequencies, e.g., νac < 250 Hz, only continuous increase of χ" is found down to 1.7 K, the lower limit for the temperature control in our superconducting quantum interference device (SQUID) magnetometer. This behavior is consistent with blocking of the magnetic moments in polymer 1 at low temperatures on the time scale of the ac experiment (22). The value of ΔT m′/[T m′Δ(log ω)] ≈ 0.09 to 0.08, indicating the change inT m′ per decade of angular frequency, ω = 2πνac, suggests that the relaxation behavior falls between that in typical superparamagnets {ΔT m′/[T m′Δ(log ω)] ≈ 0.2} and insulating spin glasses {ΔT m′/[T m′Δ(log ω)] ≈ 0.06} (23). The smaller frequency shifts, compared to superparamagnets, may indicate some interaction between the moments (24). For a superparamagnet, the maximaT m", in the out-of-phase component of the susceptibility, allow for derivation of the activation barrierE A associated with moment blocking, according to the Arrhenius law, τ = τ0exp(E A/k B T). In this equation, τ = 1/2πνac is the relaxation time of the magnetization. With the very few T m" data points and narrow frequency range, the plot of ln(τ) versus 1/T gives a straight line (R = 0.996) with an order of magnitude estimates for the barrierE A/k B = 15 K and the microscopic limiting relaxation time τ0 = 9 × 10−8 s. BothE A/k B and τ0 are comparable to those found in slowly relaxing molecular cluster–based superparamagnets (e.g., in Fe8,E A/k B = 22.2 K and τ0 = 1.9 × 10−7 s) (25).

Figure 4

Temperature dependence of the ac susceptibility χ′ (A) and χ" (B) of polymer 1 measured in the zero applied field. The ac driving field is 0.1 Oe and the frequencies are 1000, 800, 500, 250, 100, 50, 25, 10, 5, 2, 0.5, and 0.1 Hz. [Diameter of the symbols for χ" in the inset of (B) represents approximately one standard deviation for three measurements.]

Analogous to ZFC/FC magnetizations, both χ′ and χ", as measured at the frequency of 1000 Hz and the ac driving field of 0.1 Oe, are highly sensitive to the applied H. For H ≈ 0.5 and 1 Oe, T m′ shifts from 2.50 K (at zero-field) to 2.10 and 1.75 K, respectively; the intensity is significantly lowered (at 1 Oe, 60% of the zero-field value). TheT m" shifts to ≤1.7 K even at 0.5 Oe and the intensity decrease of χ" is even more pronounced than for χ′ (at 1 Oe, 25% of the zero-field value at 1.7 K).

Polyethers 2 obtained after short polymerization times (Table 1, runs 1 and 2) give polymers 1 with relatively lower values of average S = 600 to 1500 (seven samples). No peaks in ac susceptibility are detected; however, a small and frequency-dependent χ" < 1 emu mol−1 is observed at low temperatures, suggesting an onset of magnetic blocking.

In conclusion, our experimental data show that organic polymer magnets can be prepared. In polymer 1, both blocking of magnetization and very large magnetic moments are found below a temperature of about 10 K. Overall, the magnetic behavior falls between insulating spin glasses and blocked superparamagnets, but closer to spin glasses.

  • * To whom correspondence should be addressed. E-mail: arajca1{at}unl.edu

REFERENCES AND NOTES

Stay Connected to Science

Navigate This Article