The successful fabrication of (Ga,Mn)As films has attracted a lot of attention in the research of diluted magnetic semiconductors (DMSs).^[1–5] However, the substitution of Mn for Ga induces both spin and charge dopings, which makes it difficult to control the spin and carrier densities separately. On the other hand, the maximum T_C of (Ga,Mn)As is only 200 K,^[6] which is still below the room temperature that is the requirement for practical applications. Recently, a series of new bulk form DMSs with decoupled charge and spin dopings have been reported.^[7–9] According to the crystal structure, they are classified as 111-type Li(Zn,Mn)As,^[10] Li(Zn,Mn)P,^[11] 122-type (Ba,K)(Zn,Mn)₂As₂,^[12] (Ca,Na)(Zn,Mn)₂As₂,^[13] and 1111-type (La,Ba)(Zn,Mn)AsO,^[14] (La,Sr)(Zn,Mn)AsO,^[15] La(Zn,Mn,Cu)SbO,^[16] etc. In these new DMSs, spins are introduced by the substitution of Mn²⁺ for Zn²⁺, and carriers can be introduced at a different site, which enables the investigation of how the carrier or spin density affects T_C. Further more, muon spin rotation (μSR) measurements have demonstrated that these DMSs have the same ferromagnetic mechanism as that of (Ga,Mn)As.^[17,18] Investigation of these bulk form DMSs will therefore help to understand the common ferromagnetism in all diluted magnetic systems.

Among these novel DMSs, 122-type DMSs have been extensively studied. The Curie temperature T_C of 122-type (Ba,K)(Zn,Mn)₂As₂ has reached ∼ 230 K.^[19] Very recently, a 122-type DMS Ba(Zn,Co)₂As₂ with n-type carriers and T_C ∼ 45 K has been reported.^[20] The 122-type series of DMSs do not have the same crystal structure. For example, (Ca,Na)(Zn,Mn)₂ As₂^[13] has a hexagonal CaAl₂Si₂ structure, while (Ba,K)(Zn,Mn)₂As₂ has a tetragonal ThCr₂Si₂ structure,^[12] and (Ba,K)(Cu,Mn)₂Se₂ has an orthorhombic BaCu₂S₂ structure.^[21] In this paper, we report the successful synthesis of a new 122 DMS (Ca,Na)(Zn,Mn)₂Sb₂ through the doping of Na and Mn into a hexagonal crystal CaZn₂Sb₂. CaZn₂Sb₂ has been known as nonmagnetic with small energy gap.^[22] We find that co-doping Na and Mn into CaZn₂Sb₂ induces a ferromagnetic ordering with a maximum 10 K. Our work represents a new member of DMS families.

Polycrystalline samples of (Ca_1-xNa_x)(Zn_2-xMn_x)Sb₂ were synthesised through solid-state reaction method. High purity elements of Ca, Na, Zn, Mn, and Sb were mixed, pressed into pellets, and heated at 200°C for 10 h in evacuated silica tubes. The mixture was then heated at 750°C for 40 h before cooling to room temperature at a rate of 10°C/h. The crystal structure was characterized via powder x-ray diffraction (XRD) at room temperature with a powder x-ray diffractometer (Model EMPYREAN) using monochromatic Cu K_α1 radiation. The DC magnetization measurements were conducted on a Quantum Design magnetic property measurement system. The resistivity was measured via the typical four-probe method, and the magneto-resistance was measured by a physical property measurement system.

In Fig. 1(a), we show the x-ray diffraction patterns for (Ca_1-xNa_x)(Zn_2-xMn_x)Sb₂ (x = 0.08, 0.10, 0.12, 0.16, 0.20). The XRD patterns of the samples can be well indexed by a CaAl₂Si₂-type hexagonal crystal structure which is shown in the inset of Fig. 1(b). Some traces of ZnSb and Sb impurities were observed in the samples and marked as triangles and stars. They are both paramagnetic and have no influences on the ferromagnetic behaviors discussed in the following. In Fig. 1(b), we show the Rietveld refinement for (Ca_0.88Na_0.12)(Zn_1.88Mn_0.12)Sb₂ by GSAS-II.^[23] The resultant weighted reliable factor is 9.8%, indicating that this sample has a good quality. From the Rietveld refinement structural analysis, we obtain the lattice parameter c and plot it in Fig. 1(c). The lattice parameter c monotonically increases from 7.478 Å for x = 0.08 to 7.499 Å for x = 0.20, indicating the successful doping of Na and Mn. The synthesis of the sample with higher doping level of was not successful and some secondary phase started to appear, which suggests that the chemical solubility in (Ca_1-xNa_x)(Zn_2-xMn_x)Sb₂ is 20%.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (a) The XRD patterns for (Ca1–xNax)(Zn2–xMnx)Sb2 (x = 0.08, 0.10, 0.12, 0.16, 0.20). Traces of impurities ZnSb (*) and Sb (▲) are marked. (b) The Rietveld refinement of (Ca0.88Na0.12)(Zn1.88Mn0.12)Sb2. Inset is the crystal structure of CaZn2Sb2. (c) The lattice parameter c of (Ca1–xNax)(Zn2–xMnx)Sb2.

In Fig. 2, we show the temperature dependent magnetization of Ca(Zn_1.9Mn_0.1)Sb₂ measured under field-cooled conditions with an external field B_ext∼ 100 Oe. No magnetic ordering is observed below 150 K, indicating that doping Mn only will not induce any type of magnetic ordering. We plot 1/(M-M₀) versus temperature as the inset of Fig. 2. It is a straight line. This result is consistent with that of previously reported 1111-type and 122-type DMSs.^[12,14] That is, Mn substitution for Zn without carriers will not induce any type of magnetic ordering.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. The temperature dependence of the DC magnetization for Ca(Zn1.9Mn0.1)Sb2 measured under field-cooling (FC) with external field of 100 Oe. Inset: the plot of 1/(M-M0) versus temperature.

In Fig. 3(a), we show the temperature dependent magnetization of (Ca_1-xNa_x)(Zn_2-xMn_x)Sb₂ (x = 0.08,0.10,0.12,0.16,0.20) measured under field-cooled conditions with an external field B_ext ∼ 1000 Oe. We can see that below T ∼ 20 K, the magnetization starts to increase abruptly. The saturation moments at the base temperature of 2 K decrease from 1.06 μ_B/Mn for x = 0.08 to 0.56 μ_B/Mn for x = 0.20. This is caused by the increase of the nearest-neighbour (NN) Mn-Mn antiferromagnetic interaction. The probability of finding two Mn atoms at NN Zn site is for x = 0.08 and increases to 29.16% for x = 0.20. We plot the first derivative of magnetization versus temperature in Fig. 3(b). The Curie temperature is defined as the minimum value from the first derivative of magnetization versus temperature. T_C is ∼ 5.0 K for x = 0.08, ∼ 6.0 K for x = 0.10 ∼ 7.0 K for x = 0.12, ∼ 8.5 K for x = 0.16, and finally increases to ∼ 10.0 K for x = 0.20. The maximum Curie temperature for (Ca,Na)(Zn,Mn)₂Sb₂ is three times lower than ∼ 33 K for (Ca,Na)(Zn,Mn)₂As₂.^[13] This may be because the lattice constants of (Ca,Na)(Zn,Mn)₂Sb₂ are larger than those of (Ca,Na)(Zn,Mn)₂As₂.^[13] Similar trend has also been observed in (Sr,Na)(Zn,Mn)₂As₂, which has the hexagonal CaAl₂Si₂ structure and of which the lattice constant is between those of (Ca,Na)(Zn,Mn)₂As₂ and (Ca,Na)(Zn,Mn)₂Sb₂, has the maximum Curie temperature of ∼ 24 K.^[24] We use the Curie–Weiss formula to fit the curve above , where θ is the Weiss temperature, C is a constant indicating the effective magnetic moment, and M₀ is a temperature-independent contribution. We plot the reverse of versus temperature in Fig. 3(c), and the linear fitting to data at high temperature intercept with x-axis is the Weiss temperature. The Weiss temperature θ is ∼ 7.0 K for x = 0.08 and then increases to ∼ 10.2 K for x = 0.20. We calculate the effective magnetic moment with . The resultant M_eff is ∼ 4.49μ_B/Mn, 4.41 μ_B/Mn, 4.67 μ_B/Mn, 4.03 μ_B/Mn, and 4.55 μ_B/Mn for x = 0.08, 0.10, 0.12, 0.16, and 0.20, respectively. These values are close to 5.9 μ_B expected for Mn²⁺, indicating that the Mn atoms in this material have a high spin state S = 5/2. The fitting data for T_C, θ, and M_eff are tabulated in Table 1.

Fig. 3.

Figure Option ViewDownloadNew Window

Fig. 3. (a) The temperature dependence of DC magnetization for (Ca1-xNax)(Zn2-xMnx)Sb2 (x = 0.08, 0.10, 0.12, 0.16, 0.20) measured under field-cooling with external field of 1000 Oe. (b) The derivative of moment versus temperature for (Cax1-xNax)(Znx2-xMnx)Sb2 (x=0.08, 0.10, 0.12, 0.16, 0.20). The arrow marks the Curie temperature of x = 0.08. (c) The plot of 1/(M-M0) versus temperature for (Cax1-xNax)(Znx2-x Mnx)Sb2 (x = 0.08, 0.10, 0.12, 0.16, 0.20). The arrow marks the Weiss temperature of x = 0.20. (d) Iso-thermal magnetization for (Cax1-xNax)(Znx2-xMnx)Sb2 (x = 0.08, 0.10, 0.12, 0.16, 0.20) at 2 K.

Table 1.

Table 1.

Table 1. The Curie temperature TC, Weiss temperature θ, effective moment Meff, and coercive field HC for (Ca1-xNax)(Zn2-x Mnx)Sb2. . x TC/K θ/K Meff/(μB/Mn) HC/Oe 0.08 5.0 7.0 4.49 3 0.10 6.0 8.8 4.41 14 0.12 7.0 9.7 4.67 71 0.16 8.5 10.1 4.03 98 0.20 10.0 10.2 4.55 245

Table 1. The Curie temperature TC, Weiss temperature θ, effective moment Meff, and coercive field HC for (Ca1-xNax)(Zn2-x Mnx)Sb2. .

In Fig. 3(d), we show the iso-thermal magnetization at 2 K for (Ca_1-xNa_x)(Zn_2-xMn_x)Sb₂ (x = 0.08, 0.10, 0.12, 0.16, 0.20). Clear hysteresis loops can be observed in all samples, indicating that ferromagnetic ordering exists for all doping levels at 2 K. The coercive field H_C is ∼ 3 Oe for x = 0.08, ∼ 14 Oe for x = 0.10, ∼ 71 Oe for x = 0.12, ∼ 98 Oe for x = 0.16, and finally increases to ∼ 245 Oe for x = 0.20. The increasing of the Curie temperature and coercive field with higher doping level is likely arising from the RKKY-like interaction of the Mn spins which are mediated by the hole carriers. Similar trend has been observed in other DMS materials that have been investigated by nuclear magnetic resonance and muon spin rotation experiments.^[25,26] The H_C values are also tabulated in Table 1. The value is smaller than ∼ 400 Oe for (Ca,Na)(Zn,Mn)₂As₂.^[13]

In Fig. 4(a), we show the electrical resistivity of (Ca_1-xNa_x)(Zn_2-xMn_x)₂Sb₂ (x = 0.08, 0.10, 0.12, 0.16, 0.20) down to 10 K. All samples display a metallic behavior. Besides, the resistivity increases monotonically from ∼ 0.002 for x = 0.08 to ∼ 0.008 mm for x = 0.20, indicating that doping more Mn atoms increases the probabilities of carriers been scattered by magnetic fluctuations. This behavior has also been observed in a 122-type DMS (Sr,Na)(Cd,Mn)₂As₂, which has the same CaAl₂Si₂ structure.^[27]

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (a) Resistivity for (Ca1-xNax)(Zn2-xMnx)Sb2 (x = 0.08, 0.10, 0.12, 0.16, 0.20). (b) The magneto-resistance for (Ca0.9Na0.1)(Zn1.9Mn0.1)Sb2 under the applied external magnetic fields of 0 T, 3 T, 5 T, 9 T. (c) The field dependence of MR for (Ca0.9Na0.1)(Zn1.9Mn0.1)Sb2 from −9 T to 9 T at 2 K. The MR is defined as .

In Fig. 4(b), we show the magneto-resistance (MR) of (Ca_0.9Na_0.1)(Zn_1.9Mn_0.1)Sb₂ measured under the applied external fields of 0 T, 3 T, 5 T, 9 T. The field dependent resistivity decreases below T_C ∼ 6 K, indicating a negative magneto-resistance effect. This is due to the suppression of scattering of carriers by magnetic fluctuations under the external field. We plot the field dependence of MR for (Ca_0.9 Na_0.1)(Zn_1.9 Mn 0.1)Sb₂ from −9 T to 9 T at 2 K in Fig. 4(c). The MR is defined by the expression and reaches 12% at 9 T. This value is comparable to for (Ca,Na)(Zn,Mn)₂As₂ at 7 T.[]

To summarize, we turn nonmagnetic CaZn₂Sb₂ into ferromagnetic ordering by doping Na and Mn into CaZn₂Sb₂. Na provides carriers and Mn provides local moments, respectively. A new DMS material (Ca,Na)(Zn,Mn)₂Sb₂ with Curie temperature T_C∼ 10 K has been successfully synthesized. Iso-thermal magnetization measurements show that the coercive field is ∼ 245 Oe. Further more, this material shows a negative magneto-resistance with MR . This new ferromagnetic semiconductor stands as a new member in novel DMS families with decoupled charge and spin dopings, and provides a new choice to develop junctions with the same CaAl₂Si₂ structure of iron-based superconductors and other antiferromagnets.

Note added During the preparation of this manuscript, we became aware of the same material (Ca,Na)(Zn,Mn)₂Sb₂ has just been published in Physical Review Materials by Yu et al.^[28] The results in our independent work are quantitatively consistent to theirs.