|Home | About | Journals | Submit | Contact Us | Français|
We report the synthesis of single crystalline Co2Si nanowires, and electrical transport studies of single Co2Si nanowire devices at low temperature. The butterfly-shaped magnetoresistance shows interesting ferromagnetic features including negative magnetoresistance, hysteretic switch fields and step-wise drops in magnetoresistance. The non-smooth step-wise magnetoresistance response is attributed to magnetic domain wall pinning and de-pinning motion in the Co2Si nanowires probably at crystal defects or morphology defects. The temperature dependence of the domain wall de-pinning field is observed and is described by a model based on thermally assisted domain wall de-pinning over a single energy barrier.
Magnetic domain wall (DW) motion in ferromagnetic material has attracted extensive interest due to its significance in fundamental investigations and potential for practical applications.1,2 In particular, the current induced DW motion by spin torque and spin valve effects may be applied to the next generation of memory devices and logic circuits.3 Semiconducting and ferromagnetic nanowires have attracted considerable interest as potential building blocks for the next generation of electronic and spintronic devices due to their low dimensionality and unique electromagnetic properties.4–7 DW motion have been observed in polycrystalline nanowires obtained from lithography fabrication or templated growth. Transition metal silicide single crystal nanowires, such as CoSi, MnSi and Fe1−xCoxSi, have recently been successfully synthesized and demonstrated to be ferromagnetic.8–14 However, there is yet no study reporting the DW motion in these synthetic transition metal silicide nanowires.
Here we report experimental investigation of DW motion in single Co2Si nanowires via electrical transport characterization. The butterfly-shaped magnetoresistance (MR) of the Co2Si devices at low temperature of 1.5 K displays rich features, including negative MR, hysteretic MR line shape and step-wise MR changes. The most important feature of our data, the step-wise MR changes, reveals the existence of multiple magnetic domains as well as magnetization moment reversal of these domains. Our data indicates that DW motion of pining and de-pinning occurs in our Co2Si devices during the magnetic field sweep measurements. The temperature dependence of the DW de-pinning field is fitted to a model of DW hopping over a single energy barrier, the magnitude of which is estimated to be 1.3 eV. Our work, for the first time, reports transport charaterization of multiple domains and DW motion in synthetic transition metal silicide materials, which are of significant interest for both fundamental understanding and practical applications of silicide materials.
Synthesis of the Co2Si nanowires is performed using a chemical vapor deposition (CVD) method.9 The silicon substrates are cleaned with buffered hydrofluoric acid (BOE) and placed in the center of a horizontal 1-inch diameter quartz tube furnace. Anhydrous CoCl2 powder (Sigma-Aldrich, 99.9%) is placed in an alumina boat upstream from the substrates. The temperature at the centre of the furnace is set to 800 °C and the CoCl2 powder is heated at around 500 °C. The precursor vapor is carried by Ar gas with a flow of 200 standard cubic centimeters per minute (sccm). The reaction is held for about 30 minutes, and then the furnace is cooled down to room temperature. A “fluffy” black material composed of Co2Si nanowires is clearly visible on the substrates.
Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) studies are carried out to examine the morphology and the crystalline structure of the Co2Si nanowires. The SEM and TEM images show that the Co2Si nanowires typically have diameters of about 30–80 nm and lengths over 10 μm (Fig. 1a, b). The selected area electron diffraction (SAED) pattern of the nanowire shown in Figure 1b can be indexed as Co2Si zone axis  with a crystal structure belonging to space group Pnma (62) with orthorhombic structure and lattice constants a = 4.918, b = 3.738 and c = 7.109 Å (JCPDS Card no. 65–1506) (Fig. 1b inset). Together, Figure 1b and inset reveal that the long axis of the Co2Si nanowire is oriented along  direction with a highly crystalline structure. The high-resolution TEM further confirms the single crystalline nature of these Co2Si nanowires (Fig. 1c).
To investigate the magnetoelectrical transport properties, single nanowire devices are fabricated on a silicon substrate with a 300 nm silicon oxide layer (Fig. 2a). The Co2Si nanowires are transferred from the growth substrate to the silicon/silicon oxide substrate using a contact printing dry transfer process.15,16 Contact electrodes are defined using a standard electron beam lithography technique followed by high vacuum electron beam metal evaporation (70 nm Ti/50 nm Au). Before metal deposition, the samples are typically dipped in buffer oxide etchent (BOE) for 2 seconds to remove the surface oxide layer to ensure a low contact resistance between the nanowires and electrodes. Figure 2b is a representative SEM image of an individual Co2Si device. Electrical measurements are carried out at a temperature down to 1.5 K in a pumped liquid helium Oxford cryostat with variable temperature control.
Using the four-probe measurement method, we first plot the MR of the Co2Si nanowire device with the wire axis parallel to magnetic field with 1 μA current bias (Fig. 2c upper panel). The MR ratio reaches −6% within a 0.5 T applied magnetic field. When the parallel magnetic field sweeps starting from −0.5 T, the MR smoothly increases along the red curve and reaches the maximum value at switch field Hs = 0.08 T. Beyond this maximum point, the MR decreases with a few small steps followed by a couple of abrupt drops before resuming a continuous, smooth decline. Similarly, with the downward sweep of magnetic field from 0.5 T, the MR (black curve) begins with a smooth increase before switching to a stepped decrease at a switch field Hs = −0.08 T. The MR curves from upward and downward magnetic sweeps are highly symmetric to each other, and the whole MR loop is highly reproducible. The butterfly-shaped MR exhibits three major features: negative MR, hysteretic MR behavior and stepped MR drops, which correlate to ferromagnetic properties and DW motion in Co2Si nanowires. In contrast, the MR measurements with the wire axis perpendicular to the applied magnetic field changes much smoother than the former case, and the MR ratio is only −3% (Fig. 2d upper panel).
Co2Si nanowires have previously been reported to be ferromagnetic by measuring the temperature dependence of the magnetization, SQUID magnetization hysteresis and MR hysteresis.9 The hysteresis of MR is generally shown by the fact that the MR curves of upward and downward magnetic field sweeps do not overlap with each other. That is, the maximum resistance from opposite magnetic field sweeps occurs at different values of non-zero switch field, which generally have the same magnitude. Similar to the hysteretic magnetization curve of ferromagnetic materials in a magnetic field, the hysteretic MR also reflects the ferromagnetic coercivity of these nanowires.
In previous studies of Co and Ni nanowires, the MR is positive with a parallel magnetic field and is negative in a perpendicular magnetic field,17,18 indicating an anisotropic magnetoresistance effect (AMR) dominates the system. In contrast, the MR is negative in the Co2Si nanowires in both field directions as plotted in the top panels of Figures 2c and 2d, and as observed in MR of other transition metal silicide nanowires.8,9 This observation implies that AMR does not dominate the MR of the transition metal silicide material. Below the Curie temperature of ferromagnetic materials, exchange splitting of the bands gives rise to a reduction of s-d scattering among majority electrons, resulting in an increased mobility and a reduced resistance.19 With an external magnetic field, the increased degree of spin order contributes to the isotropic negative MR. The similar negative MR observed in polycrystalline ceramic and thin-film manganites is interpreted by the alignment of the ferromagnetic moments of contiguous ferromagnetic grains.20,21 There are also reports that, the antiferromagnetic boundaries lead to an infinitely large negative MR in magnetite films,22,23 which possibly does not apply to our case.
The most interesting feature of our results is the stepped change in the MR curve as plotted in Figure 2c upper panel. Previous studies of ferromagnetic nanowires presented a similar MR behavior, which is explained by DW motion, or the Barkhausen effect.17 During the magnetization and demagnetization process, the magnetic moment of ferromagnetic materials are generally polarized or flipped in the form of multiple domains instead of a single domain due to crystalline defects or shape defects. High-resolution TEM studies show that crystalline defects can often be seen in some of our Co2Si nanowires (Supplementary Fig. S1). In order to present the MR steps, we plot the derivative of MR with respect to magnetic field in Figure 2c lower panel, where dR/dH peaks represent MR steps. When the magnetic field is swept beyond the switch field, the DWs are nucleated, and then successively propagate at magnetic field of dR/dH peaks, and finally are de-pinned at the last dR/dH peak. During the DW propagation process, the height of MR jumps is a measure of the volume of the reversed domain.17 The two major dR/dH peaks at H = 0.157 T and 0.175T on magnetic field upwards sweep and at H = −0.161 T and −0.170 T on magnetic field downwards sweep, corresponds to two major de-pinning sites, which may be related to crystalline defects. However, we do not observe obvious dR/dH peaks in perpendicular magnetic field sweeps as plotted in Figure 2d lower panel, indicating that MR changes fairly smoothly here, in contrast to the MR steps seen in the parallel magnetic field measurements. The smoothness difference of the MR in two magnetic field directions is likely a result of the easy axis of the Co2Si nanowire aligning with the long axis of the nanowire.
The low Curie temperature of 160 K for Co2Si nanowire,9,24 complicates the use of magnetic force microscopy (MFM) to directly observe multiple domains in our Co2Si nanowires at low temperature. To further investigate the DW motion characteristics, we have carried out two-probe measurement of the differential conductance change vs. magenetic field sweep (G-H) in the initial magnetic field sweep and reversible half way magnetic field sweep. Although the two-probed differential conductance measurements do not exclude contact resistance, the G-H measurements make no difference in revealing DW motion to R-H measurements. In Figure 3a, we plot the initial G-H sweep as the black curve and the whole range sweep as the red curve on a similar device. When the magnetic field sweep starts from 0 T, the G-H curve increases with many bumps and kinks. After magnetic field fully polarizes the magnetization of the Co2Si nanowire at H = −0.5 T, the G-H curve follows the red loop in Figure 3a with field sweep. The non-smooth G-H curve of the initial sweep from H = 0 T to −0.5 T denotes that the Co2Si nanowire is already in multiple domains regime before the magnetic field is applied, and that DW motion also takes place during the initial magnetic field sweep. In Figure 3b, differential conductance decreases smoothly as magnetic field sweeps from 0.5 T to a value (H = −0.12 T) approaching the switch field, and then smoothly increases by following the previous route as the magnetic field sweeps back to 0.5 T. The smooth G-H loop between maximum magnetic field and switch field, suggests that magnetization momentums have not flipped until the magnetic field reaches the switch field, and that there is no DW motion yet in the system within that magnetic field range.25 All of these evidences prove that DW motion and multiple domains exist in our Co2Si nanowires, which has not been reported in similar silicide nanowire materials.
DW de-pinning is generally described conceptually by hopping over a single energy barrier of the height E0.26 When a DW is pinned at a defect location and the magnetic field slowly increases up to a critical value, the DW escapes the energy barrier with thermal assistance, resulting in intrinsic uncertainty in the precise value of the de-pinning field. Since the DW de-pining field can differ slightly under identical experimental conditions, we have repeated the MR measurement a few times to determine the mean value of the DW de-pinning field. In general, when temperature is increased, higher thermal energy is expected to activate the DW hop over the energy barrier at a lower magnetic field.
In order to investigate the temperature dependence of the DW de-pinning field in our Co2Si device, we perform G-H measurements at temperatures from 5 K to 45 K with 10 K intervals as plotted in Figure 4a from top to bottom. As expected, the last major DW de-pinnig field pointed by the black arrow, decreases as temperature is increased. At base temperature, 1.5 K, the G-H curve in Figure 4b upper panel shows similar behavior with the devices in Figure 2 and Figure 3. There are two major dG/dH peaks at H = 0.196 T and 0.247 T in the upward magnetic field sweep (red curve) in Figure 4b lower panel, corresponding to two jumps in the G-H curve. The last dG/dH peak, corresponding to the last DW de-pinning site, behaves robust and is legible through the temperature range. Therefore we focus on the last DW de-pinning field Hd as a function of temperature (Fig. 4c), where the average value of Hd is taken from five G-H measurements repeated 5 times at each temperature.
In order to estimate the magnitude of the single energy barrier, we use the Kurkijarvi model, which was originally developed for phase slip in Josephson junctions27 and is here applied to describe the temperature dependence of the de-pinning field of a DW over an energy barrier. In this model, the temperature dependence of the de-pinning field (Hd) mean value is calculated to be
where kB is Boltzmann’s constant, ν = 2 mT/s is the magnetic field sweep rate and Γ0 is the attempt frequency at zero temperature. The value of Γ0 is expected to be in the range between 108 and 1012 Hz, and we assume Γ0 = 1010 Hz here28. The last DW de-pinning field at base temperature 1.5 K, as pointed by the black arrow in Figure 4b, is assumed to be approximated to the zero temperature de-pinning field H0. The normalized values of Hd/H0 vary between 0.8 and 1.0 over the temperature range measured, so we substitute a constant value of 0.9 for Hd/H0 to neglect the magnetic field-dependent part in the denominator of the logarithmic term. With this formula (1), we fit the temperature dependence of the last DW de-pinning field in Figure 4c, and obtain the only fitting parameter E0 = 1.3 eV, which is comparable with values observed on permalloy wires with a symmetric neck or a one-sided notch.28,29 Here, we also check the influence of the values of Γ0 and Hd/H0 on the energy barrier E0 obtained by the fitting. When the values of Γ0 and Hd/H0 vary within the relevant range, the value of E0 only drifts less than 20%. Therefore, the fitting method with fixed values of Γ0 and Hd/H0 is reasonable.
In conclusion, we have measured MR of synthetic Co2Si nanowire devices, and observed butterfly-shaped curves with characteristic features of negative MR, hysteresis loops and abrupt steps which relate to the ferromagnetic properties of Co2Si nanowires and local crystalline defects. This is the first observation of DW motion of pinning and de-pinning in transition metal silicide nanowire devices. The temperature dependence of the DW de-pinning fields can be described by a single energy barrier model, which is used to estimate its magnitude.
We acknowledge technical support from the Center for Quantum Research and the Nanoelectronics Research Facility at UCLA. We thank Dr. Jing Shi, Dr. Ward Beyermann, Dr. Alexandros Shailos, Xinfei Liu for helpful discussions. X.D. acknowledges support by NSF CAREER award 0956171 and the NIH Director’s New Innovator Award Program, part of the NIH Roadmap for Medical Research, through Grant 1DP2OD004342-01.