Bias-controlled ultrafast demagnetization in magnetic tunnel junctions

We report on the possibility of controlling the maximum demagnetization induced by a femtosecond laser pulse in a CoFeB-based magnetic tunnel junction, by tuning the external bias applied to the junction. We explain this effect in terms of laser-induced spin-polarized currents which can be controlled by tuning the conductivity of the junction. This demonstrates the contribution of spin-polarized currents for laser-induced demagnetization and may pave the way to increase the speed of spintronic devices, using ultrashort laser pulses.

Figure 1

(a) Static Kerr ellipticity $\epsilon$ for an exchange-biased tunneling junction and electrical resistance $R$ as a function of the external magnetic field. The arrows indicate the relative magnetic coupling of the CoFeB layers as a function of the external magnetic fields. (b) TMR value as a function of the external bias voltage applied to the structure. Inset: 3D representation of the layer and contact geometry. The color coding is reported in the Supplemental Material [16]. (c) Typical bias-dependent hysteresis loop of the magnetic tunnel junctions; no apparent dependence is measurable within the experimental noise. The curves are measured on a device with diameter $d=125\mu$m.

Figure 2

(a) Ultrafast variations in the probe beam ellipticity, measured with longitudinal MOKE as a function of the pump-probe delay and the bias voltage. The curves address the changes in the top layer. The fluence impinging on the sample is 4.4 mJ/cm${}^2$. Inset: ultrafast change in the sample reflectivity as a function of the bias voltage; no apparent variation is measurable. (b) Maximum change in ellipticity as a function of the bias and of the relative orientations of the layer magnetizations [see Eqs. (1a) and (1b)]. (c) Maximum demagnetization modulation $\Delta$ achieved in different devices as a function of the corresponding TMR value, at two different fluencies: 4.4 and 9 mJ/cm${}^2$. The curves in (a) and (b) are measured on a device with diameter $d=125\mu$m.

Figure 3

Representative Kerr ellipticity $\epsilon$ for an exchange-biased tunneling junction in either the absence or the presence of optical excitation. The second case is measured at $t=300$ fs after the overlap, at the maximum demagnetization. The reduction in the ellipticity contribution from the top layer is apparent. A small reduction in the pinning strength is also visible in the ellipticity contribution from the pinned layer. This change is smaller than 3% at an impinging fluence of 4.4 mJ/cm${}^2$. The curves are measured on a device with diameter $d=125\mu$m.