Effect of Bias Voltage and Deposition Rate on the Structure and Coercivity of NiFe Films

Abstract

The effect of the bias voltage U_b and the deposition rate ${v}$ on the structure, grain size D, and coercivity H_c of NiFe films with a thickness d from 30 to 980 nm, grown on Si/SiO₂ substrates by DC magnetron sputtering, has been studied. In the case of U_b = 0, a decrease in ${v}$ from values ${v}$ ≈ 27 to ≈7 nm/min is accompanied by an increase in the values of the critical film thickness d_cr from d_cr ≈ 220 nm to d_cr ≈ 270 nm. In this case, H_c in films with d < d_cr is characterized by the dependence H_c ~ D⁶ and varies from ~1 to ~20 Oe. For U_b = –100 V, the effect of the deposition rate on the coercivity is much more noticeable. At ${v}$ = 7 and 14 nm/min, the films exhibit soft magnetic properties (H_c ≈ 0.15–1.4 Oe) and the absence of d_cr for the entire range of studied thicknesses. The films obtained at ${v}$ = 21 and 27 nm/min pass into the “supercritical” state at d ≥ d_cr ≈ 520 nm, and in the range d < d_cr they are characterized by the dependence H_c ~ D³ and an increase in the coercivity from ~0.35 to ~10 Oe.

1 INTRODUCTION

Magnetically soft NiFe films, possessing a low magnitude of coercivity, high magnetic susceptibility, low magnetostriction and high magnetoresistance, are widely used as magnetically sensitive material in sensor devices, in devices for storing and recording information [1–3]. However, with an increase in the thickness of the NiFe film, they lose their soft magnetic properties, which manifests itself in an increase in the value of the coercivity H_c to its maximum value at a film thickness d equal to a certain “critical” d_cr. The existence of a “critical” thickness leads not only to an increase in the coercivity, but also to the appearance of a stripe domain structure [4], rotating [5] and perpendicular [6] magnetic anisotropy. The existence of a “critical” thickness and perpendicular magnetic anisotropy is of interest as a way to increase the information recording density, but it limits the use of NiFe films as a magnetically sensitive layer. The dependence of the H_c value on the film thickness can be used in the design of various spintronic devices in which the use of ferromagnetic layers with different coercivity is a necessary condition. The listed properties of NiFe films stimulate the search for ways to control the d_cr value.

The reasons leading to the appearance of d_cr include magnetic crystallographic anisotropy, magnetostriction, anisotropy of the film microstructure shape, and an increase in the grain size D [4, 7–10]. Moreover, the microstructural texture of the films is largely determined by the technological parameters of film deposition and the properties of the substrate. For this reason, the study of the mechanisms of formation of the microstructural texture of permalloy films providing an increase in d_cr values is given some attention. Thus, in [11, 12], it was shown that an improvement in the texture of films due to the deposition of an additional layer of nonmagnetic material leads to a decrease in H_c. In [13], it is described how an improvement in the texture of FeNi films due to a decrease in the working gas pressure from 1.8 × 10^–2 to 2.8 × 10^–3 mbar leads to a decrease in H_c and an increase in d_cr from 50 to 220 nm. An increase in the substrate temperature from room temperature to 100°C leads to an increase in the critical thickness from ~83 to ~166 nm [4], which is explained by a decrease in the anisotropy constant. Considerably less attention has been paid to the joint effect of the substrate bias potential U_b and the film deposition rate ${v}$. The possibility of obtaining magnetically soft FeNi films with a thickness of 1 μm using an RF bias voltage U_b from –120 to –160 V with high-rate deposition (>150 nm/min) in a setup with two magnetrons was shown in [14]. The results obtained are associated with a decrease in internal stresses and an improvement in the uniformity of the films. In [15], the effect of the deposition rate on the properties of permalloy films with a thickness of d = 100 nm, grown in the absence of a bias potential (U_b = 0), was studied. At ${v}$ = 10, 18, and 28 nm/min, the H_c value was 1.3, 1.1, and 1.4 Oe, respectively. It is of interest to study the mechanism of formation of soft magnetic films under conditions of simultaneous changes in the parameters of U_b and ${v}$.

It should also be noted that the study of the nature of the dependence of H_c of films of subcritical thicknesses (d < d_cr) on the grain size D allows one to associate the mechanism of formation of coercivity [6, 16, 17] either with the dominance of magnetocrystalline anisotropy of crystallites (H_c ∝ D⁶), or with a predominance of coherent uniaxial anisotropy over random magnetocrystalline (H_c ∝ D³). Earlier, only the (H_c ∝ D⁶) dependence was observed for permalloy films [10]. It is of interest to search for the deposition conditions under which the H_c(D) dependence in permalloy films can change according to the D³ law.

With this in mind, the aim of this work was to study the effect of the parameters of U_b and ${v}$ on the microstructural texture, coercivity, and the character of the dependence of H_c on the grain size for permalloy films with a thickness from 30 to 980 nm grown by DC magnetron sputtering on Si/SiO₂ substrates.

2 EXPERIMENT

The deposition of NiFe films was carried out at an ambient temperature in a VUP-5M vacuum setup with a base pressure of 6 × 10^–4 Pa using a planar DC magnetron sputtering system. Targets Ni₈₁Fe₁₉ with a purity of 99.95% (K. Lesker) and argon of extra clean grade (99.998%) were used for deposition. The pressure of working argon gas during sputtering was ~0.2 Pa. The substrate was placed over the center of the target at a distance of ~75 mm. Silicon wafers Si(100) of 100KDB12(100) grade with a thermally oxidized SiO₂ layer of 300 nm in thickness and root mean square surface roughness σ ≈ 0.3 nm were used as substrates. The substrates were subjected to preliminary ultrasonic cleaning in acetone at a temperature of T ≈ 315 K and thermally annealed in vacuum at T ≈ 650 K for 30 min. The thickness of the films was estimated from the time of deposition and the pre-determined deposition rate.

The crystal structure of the films was studied by X‑ray diffraction on a DRON-4 diffractometer with focusing on a plane sample in the Bragg–Brentano geometry (scheme θ–2θ Cu–K_α radiation, λ = 0.15418 nm). The microstructure of the films was studied by scanning electron microscopy (SEM) (Auriga, Carl Zeiss). The thickness of the films was determined by profilometry (Dectak 150, Veeco). The coercivity was evaluated from the hysteresis loops obtained by vibromagnetometry in the magnetization geometry tangential to the film surface at room temperature. Several series of samples obtained at U_b = 0 and –100 V and deposition rates ${v}$ = 7, 14, 21, and 27 nm/min were under study. The deposition rate was increased by increasing the discharge current from 85 to 200 mA, with a corresponding change in the discharge voltage from 345 to 380 V.

3 REZULTS AND DISCUSSION

Figure 1 shows the dependence of the coercivity field H_c on the thickness d for films grown at different deposition rates and U_b = 0 and –100 V. It can be seen that for films grown at U_b = 0, a change in the deposition rate from 7 to 14 nm/min results in a decrease in the “critical” thickness from ~270 to ~220 nm, with a corresponding change in the maximum value of H_c from 20.2 to 21.6 Oe. The minimum value H_c ≈ 1 Oe is observed in both cases. The results obtained agree with the results of [4], where the decrease in d_cr is associated with the deformation caused by magnetostriction and the columnar microstructure of the films. A further increase in the deposition rate up to 27 nm/min does not lead to a noticeable change in the “critical” film thickness. Applying a negative bias voltage U_b = ‒100 V to the substrate during deposition leads to the disappearance of the “critical” thickness in films grown at ${v}$ = 7 and 14 nm/min with thicknesses up to 980 nm. In this case, the value of H_c varies in the range of 0.55–0.80 Oe for ${v}$ = 7 nm/min and 0.15–1.4 Oe for ${v}$ = 14 nm/min. A further increase in the deposition rate to 21 nm/min leads to the appearance of a “critical” thickness d_cr ≈ 520 nm, which increases to ∼540 nm with an increase in ${v}$ up to 27 nm/min. At ${v}$ = 21 nm/min, the maximum value of H_c is equal to 10.2 Oe and decreases to 0.35 Oe. At ${v}$ = 27 nm/min, H_c varies from 9.3 to 1.3 Oe.

Fig. 1.

Dependence of the coercivity H_c on the thickness d of films grown at different U_b and growth rates ${v}$. ${v}$ = (1) 7, (2) 14, (3) 21, and (4)  27 nm/min.

Before proceeding to a discussion of the results, it should be noted that NiFe films deposited at U_b = 0 have a polycrystalline structure with a weakly pronounced NiFe (111) texture. To characterize the dependence of the texture of the films on the deposition parameters, let us turn to Fig. 2, which shows the dependences of the texture coefficient T_hkl on the parameters ${v}$ and U, determined by the formula [18]:

$${{T}_{{hkl}}} = \frac{{{{I}_{{hkl}}}{\text{/}}I_{{hkl}}^{0}}}{{1{\text{/}}hn\sum {{{I}_{{hkl}}}{\text{/}}I_{{hkl}}^{0}} }},$$

where I_hkl is the measured intensity of the diffraction line hkl, $I_{{hkl}}^{0}$ is the line intensity for NiFe powder from JCPDS file, and n is the number of reflexes used. In this case, the increase in T_hkl values reflects the enhancement of the (hkl) texture.

Fig. 2.

Dependence of the texture coefficient T₁₁₁ on the deposition rate ${v}$ and bias voltage U_b.

Figure 2 shows that, at a bias voltage U_b =–100 V, an enhancement of the NiFe (111) texture is observed as compared to the case of films grown on a “grounded” substrate (U_b = 0). An increase in the deposition rate leads to an increase in the texture of the films grown at U = –100 V and to a non-monotonous decrease in the case of U_b = 0 [19, 20].

The choice of U_b = –100 V in the studies performed is explained by the fact that a further increase in U_b with an insignificant improvement in the texture of the films leads to a noticeable decrease in the growth rate of the film due to its sputtering by positive argon ions. An increase in the substrate temperature observed in this case does not lead to a noticeable change in the structure of the films [20].

Figure 3 shows the images of the cross-section of the films deposited at different ${v}$ and U_b.

Fig. 3.

Cross sections of films of (a) “supercritical” and (b) “subcritical” thicknesses d grown at bias voltages (a) U_b = 0 and (b) ‒100 V and different deposition rates ${v}$.

An increase in the deposition rate leads to an increase in the number of atoms of the sputtered target and their energy. At the used working gas pressure of ∼0.2 Pa, the mean free path of the sputtered target particles exceeded the distance between the target and the substrate and, therefore, the energy of particles reaching the substrate did not change when passing from the target to the substrate. Thus, an increase in the deposition rate was accompanied by an increase in the migration ability of adatoms and the formation of a more uniform microstructure of the films. At the same time, bombardment of the growing film with positive Ar ions prevented the formation of the columnar microstructure (Fig. 3b).

Figure 4 shows the results of X-ray structural analysis of the films obtained at ${v}$ = 14 and 27 nm/min. The interplanar distance a_[hkl] along the crystallographic axis [hkl] was determined by the Bragg–Wolfe formula 2a_[hkl] sin θ = nλ, where θ is the diffraction angle, λ = 0.15418 nm, and n = 1. The grain size was determined by the Scherrer formula L = λ/β cos θ, where β is the broadening of the diffraction line. Figure 4a shows the dependence of the internal deformations ε = Δa/a of the samples under study on the thickness.

Fig. 4.

Dependences of (a) deformation ε and (b) grain size D on the thickness of films d grown at U_b = 0 and –100 V and rates: (1) ${v}$ = 14 and (2) 27 nm/min.

First of all, it is necessary to note the difference in the deformations of the films grown at U_b = 0 and ‒100 V. It can be seen that the films grown at U_b = 0 are characterized by compressive deformation (ε < 0), while the films grown at U_b = –100 V experience tensile deformation (ε > 0).

Soft magnetic films grown at U_b = 0 and ${v}$ = 14 nm/min in the “subcritical” state (d_cr ≤ 220 nm) have ε ≈ –5 × 10^–4, the value of which increases to ε ≈ –10 × 10^–4 upon transition to the “supercritical” state. For soft magnetic films grown at U_b = –100 V and ${v}$ = 27 nm/min, ε decreases from 86 × 10^–4 to 35 × 10^‒4 units with an increase in the film thickness to the “critical” one, d_cr ≈ 520 nm. A further increase in the film thickness up to 810 nm leads to a decrease in ε to 12 × 10^–4 units. For films grown at ${v}$ = 14 nm/min and U_b = –100 V, the value of ε decreases from 156 × 10^–4 to 20 × 10^–4 with an increase in the film thickness from 30 to 980 nm. In this case, no transition to the “supercritical” state is observed.

Thus, it can be argued that for films grown at U_b = –100 V and v from 7 to 27 nm/min, internal deformations do not have a noticeable effect on the transition of the film to the “supercritical” state. The results obtained agree with [14], in which for films with a thickness of 1 μm deposited at U_b = –100 V, the disappearance of the “supercritical” state is associated with the absence of deformations and the transition of the columnar microstructure to a homogeneous one.

Since the grain size has a significant effect on the value of the coercivity [6], let us consider how the grain size changes with increasing film thickness (Fig. 4b). Regardless of U_b and ${v}$, the grain size grows with the increase in the film thickness; however, the grain size of films grown at U_b = –100 V increases more slowly and, in accordance with the chaotic anisotropy model [7], leads to a shift of d_cr into the region of greater thicknesses. It can be seen that the films grown at U_b = 0 pass into the “supercritical” state at a film thickness of 220–270 nm and a grain size D_cr ≈ 7 nm. For films deposited at U_b = –100 V and ${v}$ = 21 and 27 nm/min, the “critical” thickness is ∼520 nm and D_c ≈ 11 nm.

Figure 5 shows the dependences of the coercivity H_c on the grain size D for films obtained at deposition rates 14 and 27 nm/min at U_b = 0 and –100 V. The calculated dependences H_c ∝ D⁶ and H_c ∝ D³ (dashed line) are also shown there.

Fig. 5.

Dependences of the coercivity H_c on the grain size D for films obtained at U_b = 0 and –100 V and rates (a) ${v}$ = 14 and (b) 27 nm/min.

It can be seen that at ${v}$ = 14 nm/min and U_b = 0 (Fig. 5a), the dependence of the coercivity on the grain size is H_c ∝ D⁶ at D ≤ 7.5 nm, and H_c ∝ 1/D at D ≥ 7.5 nm, which corresponds to nanocrystalline materials [8]. Applying a negative bias voltage U_b = ‒100 V leads to an enhancement of the film texture (Fig. 2) and a change in the dependence of H_c on D. At U_b = –100 V, H_c varies from 0.15 to 1.4 Oe over the entire range of grain sizes.

An increase in the film growth rate to 27 nm/min at U_b = 0 does not change the nanocrystalline structure of the film, and the H_c ∝ D⁶ dependence is retained (Fig. 5b). At U_b = –100 V, the power-law dependence H_c(D) changes from H_c ∝ D⁶ to H_c ∝ D³, which does not contradict the known data [16, 17] and is associated with an insignificant improvement in the texture of the nanocrystalline film, which leads to the appearance of uniform magnetic anisotropy.

4 CONCLUSIONS

The possibility of obtaining magnetically soft NiFe films up to 980 nm in thickness by applying a negative voltage to the substrate and selecting the deposition rate is experimentally shown. In the absence of a bias voltage, the coercivity H_c changes ∝ D⁶, the “critical” thickness is d_cr ≈ 230 nm and the maximum value of H_c changes from 17 to 25 Oe depending on the deposition rate. The application of a negative bias voltage leads to bombardment of the growing film with argon ions, which improves the texture of the films, reduces the grain size, complicates the formation of a columnar microstructure and leads to a change in the dependence of H_c on D. At U_b = –100 V and ${v}$ = 21 and 27 nm/min, H_c changes ∝ D³, the maximum value of H_c decreases to ~8 Oe, and d_cr increases to ~520 nm. A decrease in ${v}$ to 7 and 14 nm/min leads to a further decrease in the dependence of H_c on D. With an increase in the film thickness up to 980 nm, the grain size increases from 7 to 14 nm, H_c changes from 0.15 to 1.4 Oe, and “critical” thickness is absent. It should be noted that the H_c ∝ D³ dependence in NiFe films is observed for the first time.