Introduction

Recent promising achievements in spintronics specially in magnetic thin films conjugated to two-dimensional (2D) materials has made this topic interesting for fundamental studies to explore their important role in the future spintronic-based memory and computing devices1,2,3,4,5. The core deriving fundamental phenomenon in such structures is the spin–orbit interaction (SOI)6,7. To benefit from SOI in spintronic devices, materials with high spin–orbit coupling (SOC), mostly heavy metals like Pt and Ta8 are used in devices in contact to magnetic thin films. Also, due to the recent developments in the field of 2D materials, special focus is put into implementing 2D materials with their intriguing properties instead of those heavy metals9,10 with high SOC. Many studies have demonstrated the successful usage of transition metal dichalcogenides (TMDCs) in contact to ferromagnetic thin films to enhance the SOI, induce surface anisotropy, etc.11,12. We have recently demonstrated that the magnetic anisotropy can be tuned by MoS2 on the surface of Py thin films13 and also predicted interfacial anisotropy can be changed in Co/black-phosphorene14. Here, we alternatively demonstrate the magnetic properties of Ni80Fe20 change by embodiment of MoS2 thin flakes. This shows the whole single ferromagnetic thin film to possess SOC-induced intrinsic magnetic properties.

Fabrication of thin films for spintronics devices based on physical techniques such as sputtering and thermal evaporation have shown the best performance15,16. Besides, electrodeposition method has established to be very promising in producing spin valves with very high number of layer repetitions (above 100 repeated layers17) and also functional nanowires for spin caloritronic devices18,19,20. Although it should be mentioned that electrodeposition lacks the ability to provide ultra-thin films without voids or making multilayers of diverse types of materials in a single growth21. The implementation of 2D materials in contact to ferromagnetic thin films has been challenging22,23 and such structures are made by transferring the as-made 2D layers on the ferromagnetic layers24. In addition to their multi-step fabrication method, the materials contacts are poor that hitherto limits their reproducibility and scalability25. Therefore, developing new fabrication method for making heterostructure of 2D materials/ferromagnetic layers is demanded to achieve higher yield and functionality.

In this work, we use electrodeposition method for fabrication of Py magnetic films and present the co-electrodeposition of MoS2 thin flakes with ionic elements of the solution. The Raman spectroscopy indicates successful embodiment of thin MoS2 flakes inside the grown magnetic film. The magnetic properties of the layer with MoS2 flakes show prominent differences with bare ferromagnetic layer including higher magnetic coercivity and damping parameter which are directly related to the enhancement in SOC of the medium. Our results indicate that our fabrication method has resulted in a good proximity between the MoS2 and the magnetic material for inducing SOC in the ferromagnet. Our method has the possibility of being used for growth of gradients or multilayers of the investigated material through control of the growth conditions like applied growth voltage/current.

Experimental section

Exfoliation of MoS2

Exfoliation of MoS2 was done for a 1 g MoS2 powder (Aldrich, 99%, < 2 μm) in 100 ml distilled water, equivalent to 10 g l−1 concentrations of MoS2 (Fig. 1a). The MoS2 powder was exfoliated for 4 h using a sonication probe unit equipped with a long step horn tip. In order to avoid excessive heating, the probe of sonic was set to operate 0.7 s and rest for 0.3 s and also an ice-water bath was used during the exfoliation. The resulting solution was centrifuged for 30 min at 200 rpm to remove non-exfoliated particles.

Figure 1
figure 1

(a) Schematic of the MoS2 exfoliation by probe sonication (above) and the schematic of electrodeposition of samples (below). (b) The UV–Visible absorption of the exfoliated MoS2 and (c) the Raman spectrum of Py and MoS2@Py samples.

Full size image

Electrodeposition

Electrodeposition was done from two different solutions. First solution for electrodeposition of the Py sample had the composition of 0.4 M NiSO4.6H2O, 0.04 M FeSO4.7H2O, and 0.4 M H3BO3 (originally prepared from Merk) in distilled water. 200 ml of this solution was used for the electrodeposition. For the electrodeposition of the Py conjugated MoS2 sample (MoS2@Py), second solution was prepared as the following. First a 100 ml of the same solution with doubled molarities was prepared and then a 100 ml of the exfoliated MoS2 solution was added to it. This way the final molarities of Ni and Fe ions is similar in both the solutions. Si substrates (single side polished surface) were cut by 1.5 × 1.5 cm slices. In order to clean surface of the Si from native oxide, they were dipped in 10% HF (hydrochloric acid) solution for 45 s, then washed by ethanol, acetone and distilled water respectively, and dried by air pump. Then the Si substrates were immediately transferred to the electrodeposition cell to prevent further surface oxidation. A two-electrode cell configuration with a DC current source was used for the electrodepositions. A 2 × 1 cm platinized Si was used as the anode and Si substrate as the cathode. The Py and MoS2@Py samples were electrodeposited by applying a direct voltage of 10 V at room temperature during 120 and 150 s, respectively.

Characterization

UV–Visible (Perkin Elmer, Lambda25) and Raman spectroscopy (Teksan) were carried out at room temperature. Surface was probed via atomic force microscopy (AFM, nanosurf) measurement. Energy dispersive X-ray spectroscopy (EDX) and elemental mapping were measured by through field emission scanning electron microscope (FESEM, Hitachi). Magnetic hysteresis loops were measured by longitudinal magneto-optical Kerr effect (MOKE), with a 632 nm laser light (a home-made setup). Ferromagnetic resonance (FMR) measurements were performed by a home-made field modulation lock-in technique at the frequency range of 2–20 GHz (for details see supporting information).

Results

Schematic of the exfoliation condition is depicted in Fig. 1a where the force of water molecules results in exfoliation of the MoS2 powder into thin flakes. To characterize the quality of the exfoliated MoS2 in water, UV–Visible absorption measurement was used. Result of this measurements can be seen in Fig. 1b. The A and B peaks at 559 and 663 nm respectively are the characteristic of few layer MoS2 dispersions. After solution preparation and electrodeposition of the layers (Fig. 1a below) Raman characterization is used to see if the MoS2 flakes are imbedded in the body of the Py layer. Figure 1c presents the Raman spectrum for Py and MoS2@Py samples. Raman peaks at 379 and 403 cm−1 clearly show the presence of MoS2 flakes in the electrodeposited layer26. The bare Py sample shows no peak in its Raman spectrum because it has a metallic nature.

The surface topography of the Py and MoS2@Py samples has been observed with AFM and presented in Fig. 2a and b, respectively. The AFM images show that both samples have a similar surface structure with an increased mean surface roughness for the MoS2@Py sample to 45 nm from the 20 nm mean surface roughness of the Py sample. Also, FESEM characterization of the Py and MoS2@Py samples has been performed and results are presented in Fig. 2c and d, respectively. It can be seen that both the samples have a granular structure with an increased grain size for the MoS2@Py sample. Also cross sectional FESEM images of the samples are presented in supporting information which show thicknesses of ~ 50 and ~ 100 nm for Py and MoS2@Py samples (~ 10% error), respectively. The observed higher thickness of the MoS2@Py samples is related to the partial space occupation by MoS2 and also the slightly higher electrodeposition time of this sample. Distribution of MoS2 in Py layer has been evidenced by EDX measurement. Figure 2e–h represents the EDX mapping of Ni, Fe, Mo and S elements where the uniform color distribution shows the uniform embodiment of MoS2. Also, the atomic ratio of Ni:Fe is 4:1.

Figure 2
figure 2

AFM measurement of (a) Py and (b) MoS2@Py samples. The FESEM images of (c) Py and (d) MoS2@Py samples. (e) and (f) are the EDX mapping of Py sample and (g) and (h) the EDX mapping of the MoS2@Py sample.

Full size image

In continue, magnetic properties of the samples are investigated through the MOKE and FMR measurements. Longitudinal MOKE measurement results are presented in Fig. 3a showing that both the samples have an in-plane magnetic anisotropy. Two prominent differences are appeared in the MOKE signal of the samples. One is the much higher coercivity (Hc) of the MoS2@Py sample which is depicted in Fig. 3b. The Hc for the Py sample is ~ 10 Oe and has increased to ~ 30 Oe for the MoS2@Py sample which is equivalent to a 300% increase. In the case of films, the magnetic anisotropy of ferromagnetic layers has been demonstrated to change by proximity to TMDC layered materials due to the d-d hybridization at the interface11,13,14. In the case of MoS2@Py sample, all interfacial directions between Py and MoS2 is possible which overall has resulted in the observed in-plane coercivity change. Generally, magnetic anisotropy is highly dependent on the SOC27,28,29 and by addition of MoS2 as a material with high SOC to the layer, changes in the magnetic anisotropy is expected. One should note that increase of the in-plane coercivity can result from the increase of thickness because of the emerging out of plane magnetic anisotropy component at higher thicknesses30,31,32,33,34. To see if the observed increase of Hc in our samples is due to the relatively higher thickness of the MoS2@Py sample we performed MOKE measurements for different thicknesses of Py layer and only a slight change in the Hc was observed (for the details see supporting information). Therefore, we conclude that the observed changes in the magnetic properties of Py is due to its proximity to MoS2. Also, the Hc of the samples can be affected by the grain size35 and the AFM and SEM images indicate a comparably bigger grain size for the MoS2@Py sample. But the totally different MOKE signal quality of the MoS2@Py sample, including the slope of the plot, indicates that presence of MoS2 is playing a crucial role in this increased Hc.

Figure 3
figure 3

(a) MOKE signal measured for the samples and (b) comparison of the Hc and the MOKE signal intensity of the samples. (c) the FMR signal observed at f = 14 GHz, (b) the fitted f and Hres curve based on Kittel formula and (c) the data of FWHM versus f and their fitting based on the damping relation.

Full size image

The other observation from the MOKE signal is the much higher signal intensity of the MoS2@Py sample relative to Py sample which is increased by 275%. Two mechanism can play role in the observed MOKE signal increase for the MoS2@Py sample: (1) increase of the saturation magnetization (Ms) and (2) increase of light interaction with matter36,37. Here the increase of the MOKE signal cannot be due to the increase of MS, as the MS decreases for the MoS2@Py sample (See FMR section). Therefore, increase of the MOKE signal can be related to the increased interaction between light and the MoS2@Py sample. For the case of multilayered ferromagnet/TMDC heterostructures it has been reported that proper thickness, refraction index and incident angle, can form a cavity38,39. Here we achieved enhancement of the MOKE signal via composing MoS2 flakes with the magnetic layer. Enhancement of light-mater interaction also has been achieved via electrophoretic deposition of MoS2 nanostructures40. Simulation of the MOKE signal for the composed MoS2@Py sample are encouraged to obtain the optimized composing structure. Also, we do not ignore the possibility of the larger grain size in the MoS2@Py being responsible partially for the observed MOKE signal intensity increase. For determining the exact contribution of each parameter further experiments should be designed.

We have also measured the FMR characteristics of the samples to see how MoS2 can affect the magnetization dynamics of the Py layer. The spin dynamic response of the samples investigated at different constant microwave frequencies ranging from 2 to 20 GHz and by sweeping H from 0 to 4500 Oe. The observed FMR spectra were fitted with the derivative of Lorentzian function to determine the resonance field (Hr) and FWHM (∆H) at each frequency. Figure 3c presents the measured FMR signal of the samples (dots) and their fit (solid line) at f = 14 GHz. The frequency dependence of Hr for the Py and MoS2@Py samples can be seen in Fig. 3d (dots). The solid lines in this figure are the fitted data by Kittel’s equation. Dependence of the FMR frequency on the external magnetic field for thin films assuming with infinite dimensions which are saturated in the plane can be described by Kittel formula41,

$$f_{r} = \frac{{{\upmu }_{0} \gamma }}{2\pi }\sqrt {\left( {H + H_{k} + M_{s} } \right)\left( {H + H_{k} } \right)}$$

where \({\upmu }_{0}\) is the permeability of the free space, \(\gamma\) is the gyromagnetic ratio (28 GHz/Tesla), H is the external magnetic field and \(H_{k}\) is the uniaxial anisotropy field which is negligible for Py films with low thickness. The resulted MS from this fit are 10,448 and 9391 Oe for the Py and MoS2@Py samples, respectively. A decrease of about 10% in the Ms of the MoS2@Py sample is observed which can be understood via the nonmagnetic character of the MoS2 that results in a lower magnetization per unit volume. Also, both samples do not show magnetic anisotropy. By using the damping relation42, as the following:

$$\Delta H = \Delta H_{0} + \frac{4\pi \alpha f}{\gamma }$$

and fitting the data we can calculate the damping parameter of the samples. Here ΔH0 is the inhomogeneous broadening, and \(\alpha\) is the Gilbert damping parameter. The fitting of the data based on this equation can be seen in Fig. 3e. The value of ΔH0 for Py and MoS2@Py samples are 114 and 231 Oe respectively. All the obtained parameters from the FMR data are presented in Table 1.

Table 1 The α, Ms, \(\gamma\), Hk, ΔH0 obtained from the fitting of FMR data.
Full size table

In the obtained results, we see a doubling of the damping coefficient in the MoS2@Py sample compared to the Py sample, which has increased from 0.018 to 0.023. There are many reports that show the coating of a nonmagnetic layer on a ferromagnetic layer can lead to the enhancement of the damping parameter43,44,45. Several mechanisms have been proposed for the enhancement of damping parameter in such bilayers. SOC and interfacial d-d hybridization cause the enhancement of the intrinsic damping, while extrinsic enhancement of the damping can arise from two-magnon scattering processes, due to roughness and defects at the interface region46,47,48. In the case of our sample, embodiment of MoS2 in the Py layer results in the increased interfaces between Py and MoS2 and therefore both intrinsic and extrinsic contribution can contribute in the observed increase of the damping parameter. Moreover, the coupling between a FM layer and an adjacent NM layer can enhance the effective damping of the magnetization precession via spin-pumping effect44,49. For example, many groups have recently demonstrated the generation of spin–orbit torque in devices made with the Py/WTe250, CoFeB/MoS2 or WSe251, and Py/MoS252 and the reciprocal process (voltage generation from spin-pumping) in MoS2/Al/Co heterostructures53. In the case of our MoS2@Py sample we observe the phenomena via enhanced damping parameter in a relatively thick magnetic layer thanks to the embodiment of the MoS2 flakes inside the layer which gives a high contact surface area between MoS2 and Py.

Conclusions

MoS2 was successfully embodied within the structure of the Py magnetic thin film by electrodeposition method. The layer with MoS2 flakes shows a higher magnetic coercivity and Gilbert damping parameter, indicating the proper bonding between the MoS2 and the magnetic material. In addition, the cavity of light in the MoS2@Py sample resulted in a three-fold increase of the MOKE signal which opens a pathway for the research on the optimization of MOKE sensors and also fundamental studies in the field. Due to the capability of applying our method for a large set of ferromagnetic/TMDC materials, there is a great potential for further development of functional spintronic and magnonic devices.