Observation of an ultrahigh-temperature ferromagnetic-like transition in iron-contaminated multiwalled carbon nanotube mats

We report magnetic measurements up to 1200 K on iron-contaminated multiwalled carbon nanotube mats with a Quantum Design vibrating sample magnetometer. Extensive magnetic data consistently show a ferrromagnetic transition at about 1000 K and a ferromagnetic-like transition at about 1275 K. The ferromagnetic transition at about 1000 K is associated with an Fe impurity phase and its saturation magnetization is in quantitative agreement with the Fe concentration measured by an inductively coupled plasma mass spectrometer. On the other hand, the saturation magnetization for the ferromagnetic-like phase (at 1275 K) is about 4 orders of magnitude larger than that expected from the measured concentration of Co or CoFe. We show that this ultrahigh-temperature ferromagnetic-like (UHTFL) transition is not consistent with ferromagnetism of any Fe-carbon phases, carbon-based phases, or magnetic impurities. Alternatively, the observed magnetic behavior of the UHTFL phase is phenomenologically explained in terms of the paramagnetic Meissner effect (orbital ferromagnetism) due to the existence of $\pi$ Josephson junctions in a granular superconductor.

Figure 1

(Color online) (a) Temperature dependence of the susceptibility in a field of 2 Oe for a virgin MWNT mat sample of RS0657 where the metal-based magnetic impurity concentrations (ppm in weight) are $\text{Fe}=6940$, $\text{Co}=36.1$, and $\text{Ni}=20.5$. (b) Temperature dependence of the susceptibility in a field of 5 Oe for a virgin MWNT mat sample of RS0656 where the metal-based magnetic impurity concentrations (ppm in weight) are $\text{Fe}=5340$, $\text{Co}=0.5$, and $\text{Ni}=13.7$. The susceptibility at 320 K appears to be proportional to the Fe concentration, while the susceptibility at 1100 K is independent of the Co concentration.

Figure 2

(Color online) (a) Magnetization vs magnetic field at 320 and 1100 K for sample RS0657 where the Co concentration is 36.1 ppm. (b) Magnetization vs magnetic field at 320 and 1100 K for sample RS0656 where the Co concentration is 0.5 ppm. The saturation magnetization at 1100 K is independent of the Co concentration.

Figure 3

(Color online) (a) Temperature dependence of the magnetization in a field of 20 kOe for a virgin MWNT mat sample of RS0657. The temperature dependence of the magnetization in 20 kOe should be similar to that of the saturation magnetization $\left(M_s\right)$. The solid line is a fit using the curve of $M_s\left(T\right)/M_s\left(0\right)$ versus $T/T_C$ for Ni, appropriately scaled to $T_C=1275\text{K}$. (b) Temperature dependence of the magnetization in a field of 20 kOe for a thermally annealed sample of RS0657. This sample was annealed in air at $480°\text{C}$ for about 5 min.

Figure 4

Temperature dependence of the magnetization in a field of 20 kOe for the residual of sample RS0657, which was obtained by burning off carbon-based materials in air at $550°\text{C}$ for about 10 min. The specific magnetization is calculated using the mass of the pristine MWNT mat sample.

Figure 5

Temperature dependence of the susceptibility for aligned MWNTs in a magnetic field parallel to the tube’s axis. The data for physically separated (PS) tubes are from Ref. 23 and the data for physically coupled bundles are from Ref. 24. These nanotubes were produced using a catalyst-free arc-discharge method.