Strong linewidth variation for spin-torque nano-oscillators as a function of in-plane magnetic field angle

We measure the microwave signals produced by spin-torque-driven magnetization dynamics excited by direct currents in patterned magnetic multilayer devices at room temperature as a function of the angle of a magnetic field applied in the sample plane. We find strong variations in the frequency linewidth of the signals, with a decrease by more than a factor of 20 as the field is rotated from the magnetic easy axis to the in-plane hard axis. Based on micromagnetic simulations, we identify these variations as due to a transition from spatially incoherent to coherent precession.

Figure 1

(Color online) (a) Differential resistance as a function of current at room temperature for a nanopillar spin valve device with an exchange-biased fixed layer. Inset: top-view electron micrograph of device shape. [(b) and (c)] Resistance as a function of magnetic field for the same device for fields (b) along the easy-axis direction and (c) along the in-plane hard axis. [(d)–(f)] Differential resistance of a nanopillar spin valve device with a thick-fixed layer as a function of (d) current and [(e) and (f)] magnetic field at room temperature.

Figure 2

(Color online) Power spectral density of spin-torque-driven oscillations at room temperature, for field angles ${\theta }_H=0°$, $45°$, and $90°$ (as labeled), for (a) an exchange-biased-fixed-layer sample and (b) a thick-fixed-layer sample. Insets: sample structures. [(c) and (d)] Variation of the linewidth as a function of ${\theta }_H$ for both types of samples.

Figure 3

(Color online) Analysis of the spin-torque-driven microwave signals for the sample with the exchange-biased fixed layer as a function of current and field angle at room temperature and $H=1000\text{Oe}$. (a) Linewidth. (b) Peak frequency. (c) Integrated power within the precessional peak divided by $I^2$. (d) Power spectral density plotted on a logarithmic scale, as a function of ${\theta }_H$, for $I=4\text{mA}$.

Figure 4

(Color online) Analysis of the spin-torque-driven microwave signals for the sample with the thick-fixed layer at room temperature and $H=800\text{Oe}$. (a) Linewidth. (b) Peak frequency. (c) Integrated power divided by $I^2$. (d) Power spectral density plotted on a logarithmic scale, as a function of ${\theta }_H$, for $I=5\text{mA}$.

Figure 5

(Color online) Micromagnetic simulations of the exchange-biased sample. (a) Resistance as a function of time. $R_{\text{P}}$ is the resistance for parallel free and fixed magnetic layers and $R_{\text{AP}}$ corresponds to antiparallel layers. (b) Corresponding power spectral densities. The curves for ${\theta }_H=45°$ and $90°$ are offset by 30 and 60 pW/GHz. The curves for ${\theta }_H=0°$ and $45°$ are scaled by a factor of 4. (c) Calculated linewidth as a function of ${\theta }_H$ for $H=1000\text{Oe}$ and $I=3.5\text{mA}$. [(d) and (e)] Snapshots of the magnetization distribution in the magnetic free layer for (d) ${\theta }_H=0°$ and (e) ${\theta }_H=90°$ at times corresponding to maxima in the precession cycle.