Limits on Intrinsic Magnetism in Graphene

We have studied magnetization of graphene nanocrystals obtained by sonic exfoliation of graphite. No ferromagnetism is detected at any temperature down to 2 K. Neither do we find strong paramagnetism expected due to the massive amount of edge defects. Rather, graphene is strongly diamagnetic, similar to graphite. Our nanocrystals exhibit only a weak paramagnetic contribution noticeable below 50 K. The measurements yield a single species of defects responsible for the paramagnetism, with approximately one magnetic moment per typical graphene crystallite.

Figure 1

Graphene laminates. (a) Typical SEM micrograph. Inset: photo of the whole sample. (b) Histogram shows the size distribution for 300 crystallites found within the $\sim 1\mu {\mathrm{m}}^2$ area imaged in (a). Crystal sizes were determined as geometrical averages. The inset zooms into the central region of (a). Edges of some of the smallest crystals are outlined for clarity.

Figure 2

Magnetic response of graphene. (a) Magnetic moment $M$ as a function of parallel $H$ at different $T$: (from top to bottom) 2, 3, 4, 5, 10, 15, 20, 50, and 300 K. (b) $M(T)$ in parallel $H$ for the sample in (a). Symbols are the measurements; the curve is the Curie law calculated self-consistently (see text) with an account taken of a constant diamagnetic background $\approx -0.008\mathrm{emu}/\mathrm{g}$ in this particular $H$. Inset: Excess moment $\Delta M$ after subtracting the diamagnetic background (measured at room $T$) as a function of perpendicular $H$.

Figure 3

Analysis of graphene’s weak paramagnetism: (a) Magnetization curves of Fig. 2a plotted as a function of reduced field $H/T$ with the diamagnetic background subtracted. (b) Fits of the data in (a) using the Brillouin function with different values of $J$. For clarity, only data at 2 K are used here. Inset: Zoom of the low-$H$ part of the graph, which is most sensitive to $J$.