Spin polarization measurements in ferromagnetic SrRuO3 using point-contact Andreev reflection technique

Using the point-contact Andreev reflection (PCAR) technique, we have performed spin polarization measurements in a polycrystalline SrRuO3 to study the difference of the spin polarization between the ballistic and diffusive transport samples. PCAR spectra were measured with Nb tips for T = 4.2 K with changing the contact area mechanically. The results were well fitted by the modified Blonder-Tinkham-Klapwijk (mBTK) model. We estimate the spin polarization P ≈ 0.59, which is slightly larger than that of a single crystal film with P ≈ 0.51 due to the change of transport property.


Introduction
After the discovery of the giant magnetoresistance (GMR), intensive efforts have been devoted to fabricate spintronics devices. Since magnetoresistance is sensitive to the spin polarization of ferromagnetic electrode, studying the spin-polarized transport properties of ferromagnetic materials is crucial for the development of spintronics devices. Spin polarization Pn is defined by difference between minority-spin and majority-spin bands at the Fermi level . Hence, it is possible to define Pn by the following equation [1], where ↑ ( ) and ↓ ( ) are the density of states (DOS) of majority and minority spins at the Fermi level, and ↑ and ↓ are the Fermi velocity, respectively. Spin-resolved photoemission experiments measure the spin polarization P0, which is determined only by the difference of DOS between the majority and minority spins at the Fermi level. On the other hand, transport experiments measure a different spin polarization including the Fermi velocities. When the contact size d at the surface is smaller than mean free path L (d <<L), electron transport is ballistic. In this case, the DOS is weighted linearly with vF ,and P1 (n = 1) is measured. In diffusive electron transport (L < d), the weighting is quadratic in vF (n = 2) and P2 is measured. Therefore, the spin polarization Pn depends on the transport properties of samples; if ↑ is larger than ↓ , P2 is larger than P1.
Owing to the structural simplicity and remarkable chemical stability, SrRuO3 is a good candidate to investigate the transport spin polarization Pn including the Fermi velocities.
SrRuO3 has a ferromagnetic transition temperature at Tc = 160 K and a saturation magnetization of 1.6 μB [2]. To measure the spin polarization, the point contact Andreev reflection (PCAR) technique has been used because the I -V characteristics due to Andreev reflection at a ferromagnet-superconductor interface changes significantly depending on the spin polarization of the ferromagnet.
In this paper, we report the spin polarization of a polycrystalline SrRuO3 in the diffusive transport region with changing the contact area mechanically using PCAR technique where Nb is utilized for a probe tip. By fitting the experimental results to the modified Blonder-Tinkham-Klapwijk (mBTK) model [3], the spin polarization is estimated to be ≈ 0.59, which is slightly larger than that of a single crystal film with ≈ 0.51 [4].

Experiments
The polycrystalline sample of SrRuO3 is synthesized by the conventional solid-state reaction method. The powders of SrCO3 and RuO2 with stoichiometric compositions were carefully mixed and calcined in air at 1100℃ for 24 hours. After careful mixing of the calcined samples, they were shaped into pellets and then sintered at 1100 °C for 24 hours [5]. obtained from fitting the differential conductance dI/dV to the mBTK function. Fig. 1 shows the temperature dependence of electrical resistivity ρ for the polycrystalline

Results and discussion
SrRuO3. An anomaly due to the ferromagnetic transition is visible at T ~ 160 K. The residual resistivity of ~ 420 Ωcm is about 8 times larger than that in a single crystal, suggesting the existence of many grain boundaries.   2 illustrates the temperature dependence of the differential conductance spectra G(V), which is normalized by the conductance at the normal region (Gn). Here, the contact resistance at the interface is controlled to be the same for all the measurements. When the temperature is higher than the Nb superconducting transition temperature TC = 9.5 K, the spectra shows no anomaly. In contrast, as the temperature is decreased, a dip at around the zero bias voltage is We measure the evolution of the differential conductance spectra with increasing the resistivity at the interface, which corresponds to decrease of the contact size. It is important that the measurements are performed without breaking the contact after preparing it, which suppresses the scattering of data. We depict the representative spectra at T = 4.2 K for various resistances in Fig. 3(a)-3(d), where the fitting curve by the mBTK function and the parameters for the fitting are also shown. The superconducting gap of bulk Nb is 1.5 meV, which is larger than that estimated from the present experiments. The gap is suppressed by magnetic scattering at the superconductor/ferromagnet interface. Moreover, the scattering parameter Z increases with decreasing the resistance R at the contact as shown in the inset of Fig. 4. This implies that the effect of ferromagnetic scattering, which suppresses the superconducting current, is increased with increasing the contact size. We plot Pn versus Z in Fig. 4, indicating that Pn is suppressed with increasing Z [6]. From the results, the value of spin polarization at = 0 is evaluated to be Pn ≈ 0.59. The spin polarization of a single-crystal thin film, which is measured in the ballistic transport region, is reported to be P1 = 0.51 at T = 4.2 K [4]. The spin polarization in the present experiments is slightly larger than the single-crystal thin film. The origin is likely due to the difference of transport properties between the samples as follows.
The electrical resistivity of the present sample is ρ ~ 420 Ωcm at T = 4.2 K as shown in Fig.  1. We can estimate the mean free path from the electrical resistivity ρ to be L ~ 0.8 and 0.45 nm for minority and majority carriers, respectively [7]. Furthermore, the contact resistance Rn is approximated by the following equation, where ρ is electrical resistivity [8]. When the contact is assumed to have a circular shape, the contact diameter of this sample is larger than several tens of nanometers, e.g., d ~ 36 nm at R = 60 Ω, representing that the measurements are done in the diffusive transport (L < d) region. Thus, the spin polarization in the present experiments should be larger than that in the singlecrystal as mentioned in Sec. 1 and we can conclude that the effect of the Fermi velocity in Pn is different between the ballistic and diffusive transports for the Nb-SrRuO3 interface. The similar results are found in the spin polarization of La0.7Sr0.3MnO3 [9].