Observation of room temperature ferromagnetism and exchange bias in a ⁵⁵Mn⁺ ion-implanted unintentionally doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ single crystal

Highlights

• Exchange bias effect and room temperature ferromagnetism are observed in $\beta$-${\text{Ga}}_2{\text{O}}_3$:Mn.

• The single domain and core-shell models are used to study the exchange bias system.

• A new explanation is proposed for the origin of the observed ferromagnetism.

Abstract

In this manuscript, the structural and magnetic properties of a ⁵⁵Mn⁺ ion-implanted unintentionally doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ single crystal are investigated. The implanted ${\mathrm{Mn}}^{+}$ ions cause crystalline degeneration due to the formation of defects in the surface of the substrate. Room temperature ferromagnetism (FM) is observed in the samples with implantation doses of 5 $\times$ ${10}^{14}{\mathrm{cm}}^{-2}$, 1 $\times$ ${10}^{15}{\mathrm{cm}}^{-2}$ and 5 $\times$ ${10}^{15}{\mathrm{cm}}^{-2}$. The zero-field-cooled (ZFC)/field-cooled (FC) measurement of magnetization reveals that the ion implantation process induces an antiferromagnetic (AFM) phase. The temperature evolution of the remanence asymmetry ($M_E$) demonstrates the typical behavior of the exchange bias effect due to the coexistence of FM and AFM phases. Based on the distribution of ${\mathrm{Mn}}^{+}$ ions and defects and the relationship between the remanence asymmetry ($M_E$) and the exchange bias field ($H_E$), the single domain and core-shell models are used to investigate the exchange bias system. The formation mechanism of the ferromagnetic order is briefly discussed, which is rather different from the one corresponding to the bound magnetic polaron (BMP) model.

Introduction

As one of the main approaches bridging semiconductors and ferromagnets, dilute magnetic semiconductors (DMSs), which occupy both charge and spin degrees of freedom, have attracted great interest for their potential applications in the realm of spintronics, such as spin injection [1], [2], magnetic multilayers [3], [4] and other quantum structures [5], [6]. The transition metal (TM)-doped oxide semiconductor is regarded as one branch of DMS study [7], [8], [9], [10], [11]. Currently, some efforts have been made to investigate the magnetic properties of TM-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ materials through both theoretical calculations and experiments. The results of these studies suggest that some of the TM-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ are ferromagnetic at room temperature. For instance, Hojo et al. found room-temperature ferromagnetic order in Fe-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ prepared by a solid-state reaction and suggested that the inhomogeneous ${\mathrm{Fe}}^{3+}$ ions were responsible for the observed ferromagnetism [12]. Room temperature ferromagnetism was also observed in ${\text{Ga}}_2{\text{O}}_3$:Fe epitaxial thin films grown on $\alpha$-Al₂O₃ (0001) substrates by the LMBE technique [13]. By characterizing the structure and the magnetic anisotropy, the authors concluded that the room temperature ferromagnetism in ${\text{Ga}}_2{\text{O}}_3$:Fe epitaxial thin films was intrinsic.

Probably inspired by the successful preparation of high-quality ferromagnetic Mn-doped GaAs and GaN samples, some efforts have been made on the investigation of Mn-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$. However, similar to Fe doping, Mn-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ samples prepared by various methods exhibit different magnetic properties. There was no ferromagnetic signal in a poly-crystalline ${\text{Ga}}_{2-x}$${\text{Mn}}_{2x}$O₃ sample prepared by standard solid-state reaction method [14]. An extrinsic trace ferromagnetic signal was found at low temperature in ${\text{Ga}}_{2-x}$${\text{Mn}}_{2x}$O₃ single crystal grown by the floating zone technique [15]. In contrast, predominant ferromagnetism with a Curie temperature above 300 K was recently observed in Mn-doped $\beta$-${\mathrm{Ga}}_2{\mathrm{O}}_3$ epitaxial thin films without secondary phases. Further research indicated the enhancement of magnetization with the increase of the Mn concentration, and a maximal saturation magnetization of 33.1 emu/${\mathrm{cm}}^3$ with Mn doping concentration of 53.10 at.$\%$ [16].

The discrepancy of experimental results in Mn-doped $\beta$-${\mathrm{Ga}}_2{\mathrm{O}}_3$ suggests that the origin of the ferromagnetism in bulk materials and epitaxial films may be different. The trace ferromagnetic signal observed at low temperature in ${\text{Ga}}_{2-x}$${\text{Mn}}_{2x}$O₃ single crystal grown by the floating zone technique can be attributed to the ferromagnetic secondary phases, such as MnGa and Mn₃O₄. However, the structural analysis ruled out the formation of MnGa or Mn₃O₄ in the epitaxial thin films. Moreover, the Curie temperature of the ferromagnetic epitaxial film far exceeds its bulk counterpart. To explain the magnetic property of the epitaxial thin films, Guo et al. [16] suggested that defects introduced in the film/substrate interface were involved in the ferromagnetic coupling. They use bound magnetic polaron (BMP) model [17] to elucidate the ferromagnetic interaction between vacancies and transition metal ions. Furthermore, a recent study theoretically revealed that, in the absence of transition metal doping, a potential long-range ferromagnetic order [18] is formed by the coupling of gallium vacancies. It is worth noting that the lowest Mn concentration in Ref. [16] is 6.06 at.$\%$; if defect-related ferromagnetism is considered, the lowest TM doping concentration may be extended to a smaller range in some dilute magnetic oxide films, as proposed by Coey et al. [17].

Besides, for Mn-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$, first-principles calculations within the generalized gradient approximation (GGA) method reveal that the ground state of the magnetic interaction changes with the substitution sites as well as with the distance of Mn-Mn cations [19]. Therefore, the inhomogeneous distribution of Mn ions in the doped thin films probably leads to the coexistence of ferromagnetic and antiferromagnetic phases, resulting in the exchange bias effect at the AFM/FM interface, which is not touched by the formal relevant investigations.

As a material engineering process far from thermal equilibrium, ion implantation has been used to tailor DMSs with unique magnetic properties [20], [21], [22], [23], [24]. In the implantation process, the concentration and distribution of magnetic ions are controlled by the dose, incident energy, and angle. By controlling the implantation condition, a low dopant concentration of TM ions towards the dilute limit can be achieved. Furthermore, for a low-energy implantation experiment, the incident ions will be distributed over the near-surface region of the target materials while introducing defects, which allows us to study the magnetic properties of the samples when TM ions and defects coexist.

Nevertheless, few reports have been concerned with the magnetic property of TM-doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ obtained by ion implantation. In this paper, unintentionally doped $\beta$-${\text{Ga}}_2{\text{O}}_3$ single crystals are implanted with low-energy ⁵⁵Mn⁺ ions to prepare ferromagnetic specimens. The structure, distribution of manganese ions and magnetic properties of the ion-implanted samples are investigated. It is found that the magnetic properties of the Mn-implanted samples change with the implantation dose. Room temperature ferromagnetism (FM) is observed in the samples with implantation doses of 5 $\times$ ${10}^{14}{\mathrm{cm}}^{-2}$, 1 $\times$ ${10}^{15}{\mathrm{cm}}^{-2}$ and 5 $\times$ ${10}^{15}{\mathrm{cm}}^{-2}$. The temperature-dependent magnetization-magnetic field curves demonstrate the coexistence of ferromagnetism and antiferromagnetism in the implanted samples. A brief discussion of the exchange bias effect is presented.