betaFeedback

WorksPhysics and Astronomy

Superexchange Interaction in Insulating EuZn2_{2}P2_{2}

Abstract

We report magnetic and transport properties of single-crystalline EuZn2_{2}P2_{2}, which has trigonal CaAl2_2Si2_2-type crystal structure and orders antiferromagnetically at ≈\approx23~K. Easy abab-plane magneto-crystalline anisotropy was confirmed from the magnetization isotherms, measured with a magnetic field applied along different crystallographic directions (abab-plane and cc-axis). Positive Curie-Weiss temperature indicates dominating ferromagnetic correlations. Electrical resistivity displays insulating behavior with a band-gap of ≈ \approx\,0.177~eV, which decreases to ≈ \approx\,0.13~eV upon application of a high magnetic field. We explained the intriguing presence of magnetic interactions in an intermetallic insulator by the mechanism of extended superexchange, with phosphorus as an anion mediator, which is further supported by our analysis of the charge and spin density distributions. We constructed the effective Heisenberg model, with exchange parameters derived from the ab initio DFT calculations, and employed it in Monte-Carlo simulations, which correctly reproduced the experimental value of Néel temperature.

Introduction

Interplay between magnetism and band topology generates alluring magnetic and electronic properties, such as superconductivity [1], non-trivial electronic states, and anomalous Hall effect [2, 3] in Zintl-phase materials. Most interestingly, magnetically coupled topological charge carriers can be modulated by changing the spin structure with an application of external magnetic field. Eu-based Zintl-phase materials are in focus for exhibiting quantum phenomena as well as non-trivial topological properties, with potential applications in thermoelectric and spintronic devices [2, 3, 4, 5]. EuT₂X₂ (TT = transition metal and XX = pnictogen) compounds have been argued to have various magnetic ground states. For example, a collinear antiferromagnetic (AFM) structure has been reported for EuCd₂As₂, EuZn₂As₂, and EuMg₂Bi₂ [6, 7, 8], whereas a helical AFM structure has been predicted for EuIn₂As₂ [2], which possesses robust non-trivial topological features. Field evolution of the magnetic structure has been reported for EuCd₂As₂ and EuMg₂Bi₂ - diffraction studies (resonant X-ray and neutron, respectively) have shown that below the Néel temperature, TNT_{\mathrm{N}}, the Eu²⁺ moments are ferromagnetically aligned within the abab-plane and antiferromagnetically coupled with those in the adjacent abab-plane, forming an A-type AFM structure [6, 9]. Applied magnetic field breaks the collinearity of AFM structure, turns spin arrangement into noncollinear one, and finally in polarized paramagnetic state [6, 7, 10]. It is worth to mention here, that EuMg₂Bi₂, EuCd₂As₂, EuZn₂As₂, EuZn₂Sb₂, EuCd₂Sb₂, and EuZn₂P₂ all have the same trigonal crystal structure (space group Pˉ3m1\bar{P}3m1) [6, 7, 10, 11, 12]. Notwithstanding, large variations in the magnetoconductivity were observed, manifesting Dirac state in EuCd₂As₂ [6, 10], semiconducting/insulating state in EuZn₂P₂ [12] or a flat band in EuZn₂As₂ [7].

Coexistence of insulating behavior and the AFM order makes EuZn₂P₂ a very interesting system. Magnetic and electronic transport properties of that compound have recently been reported, and the authors claimed that the compound is an A-type antiferromagnet with TNT_{\mathrm{N}} (≈ 23 K) and only 23\frac{2}{3} of the magnetic entropy expected for Eu²⁺ ion (Rln⁡8\mathcal{R}\ln8) recovered above TNT_{\mathrm{N}} [12]. An insulating behavior of the compound (with a band gap of 0.11 eV) was observed from electrical transport as well as DFT calculations. In the same article, it has been reported that the magnetic easy-axis is the crystallographic cc-axis [12]. In contrast, most of the materials that crystallize in a structure similar to EuZn₂P₂ have the easy axis in the abab-plane and the hard axis along the cc-axis [6, 7, 10, 11].

In this context, we revisited the magnetization, electrical resistivity and specific heat of high-quality single crystal to verify the easy- and hard-axes and the reason for unrecovered magnetic entropy. Notably, we established that the underlying magnetic interaction is governed by the exchange mechanism but not arising from the dipolar interactions as proposed in the Ref. [12]. To support our argument, we performed an abab initio DFT calculation, and extracted the anisotropic exchange parameters. Using the exchange parameters, we simulated the specific heat and magnetization as a function of temperature using Monte Carlo method with the effective Heisenberg model.

Primary results presented in this paper are organized in the following manner. A detailed description of the used methods can be found in Sec. II. The details of the crystal structure and electronic band structure obtained from the DFT calculations are presented in Sec. III A. In Sec. III B, we have characterized the magnetic properties of the compound along the principal crystallographic directions. The specific heat study and electrical resistivity are described in Sec. III D and Sec. III C, respectively. The Monte Carlo (MC) simulation of the magnetic ordering and specific heat using the exchange parameters obtained from DFT calculations, are shown in Sec. III E. To understand more the magnetic interactions among Eu atoms, a comprehensive magnetocaloric effect study is shown in the Supplemental Material (SM) [13]. Finally, we conclude our findings in Sec. IV.

Methods

Experimental

Single crystals of EuZn2_2P2_2 were grown from Sn flux. Starting materials including Eu, Zn, red P, and Sn were mixed in a molar ratio 1:2:2:45. The mixture was sealed in a quartz ampule and heated up very slowly (about 5 days) to 1100°C, kept at that temperature for 24 hours and then cooled to 800°C (at 2°C/h rate), when the flux was removed by centrifuging.

Composition of obtained single crystals was examined with a scanning electron microscope (SEM) (FEI Technologies) with an energy dispersive X-ray spectrometer (EDS) (Genesis XM4). Single crystals’ quality and their orientation were identified using the Laue back-scattering diffraction (Laue-COS, Proto Manufacturing). We took special care in the orientation of the crystal and precisely assigned the crystallographic directions. Magnetization measurements were carried out on the oriented single crystals using a SQUID magnetometer (MPMS-XL, Quantum Design). The electrical resistivity measurements were performed on a PPMS platform (Quantum Design) by a conventional 4-probe method. A rectangular-prism-shaped sample (1.36 × 1 × 0.8 mm3^3) was cut from an oriented single-crystal and electrical contacts were made with 50 μ\mum gold wire and silver paint. Measurements were performed in the magnetic field applied parallel to the cc-axis, while an electric current of 3 mA was within the abab-plane. We restricted our electrical transport measurements to temperatures above 120 K, due to the instrumental limitations of measuring high resistance.

Computational details

The first-principles DFT calculations were performed using the projector augmented-wave (PAW) potentials [14] implemented in the Vienna Ab initio Simulation Package (VASP) code [15, 16, 17]. The calculations were performed within the generalized gradient approximation (GGA) in the Perdew, Burke, and Ernzerhof (PBE) parameterization [18].

In this study, we used the experimental lattice parameters while the atoms positions were optimized. The magnetic unit cell containing two formula units was optimized, with the 12 × 12 × 4 kk-point Γ\Gamma-centered grid in the Monkhorst–Pack scheme [19]. For converging the optimization loop, we took the energy difference of 10−610^{-6} eV and 10−810^{-8} eV for ionic and electronic degrees of freedom, respectively. The calculations were performed with the energy cut-off set to 400 eV. In calculations, the Eu 4f4f electrons were treated as valence states. Correlation effects were introduced within DFT+U, proposed by Dudarev et al. [20]. Similar to previous study [12], we assumed U=5U = 5 eV as a realistic value for the Eu 4f4f states.

Results and discussions

Crystal structure and electronic band structure

EuZn2_2P2_2 crystallizes in CaAl2_2Si2_2-type structure (P3ˉm1P\bar{3}m1 space group, No. 164), with a=4.087a = 4.087 Å, and c=7.010c = 7.010 Å [21]. The atoms are located at the high symmetry Wyckoff positions: Eu 1a (0,0,0), Zn 2d (2/3,1/3, zZnz_{\mathrm{Zn}}), and P 2d (1/3,2/3, zPz_{\mathrm{P}}). After optimization, we found free parameters as zZn=0.3683z_{\mathrm{Zn}} = 0.3683 and zP=0.2704z_{\mathrm{P}} = 0.2704, which are close to reported in previous study (zZn=0.3695z_{\mathrm{Zn}} = 0.3695 and zP=0.2692z_{\mathrm{P}} = 0.2692) [12].

Similarly to EuCd2_2As2_2 [6], at low temperatures the system realizes A-type AFM structure with moments in abab-planes Figure 1(a). From the DFT calculations, the Eu magnetic moment is nearly 7 μB\mu_{\mathrm{B}} (inferred by used UU value), which corresponds to J=S=7/2J = S = 7/2 of Eu2+^{2+} ion. The electronic band structure, shown in Figure 1(b), is characterized by the indirect gap, realized between Γ\Gamma and M points. Similar behavior was earlier reported for EuMg2_2Sb2_2 [22]. Values of the indirect and direct gaps strongly depend on the treatment of 4f4f electrons, as well as a correlation parameter UU (see Tab. S1 in the SM [13]), due to the band structure renormalization, similar to that reported for the rare-earth nickelates [23]. The band gap reported experimentally is around 0.1 eV (cf. Electrical transport and [12]), consistent with DFT calculations.

Image not extracted

Figure 1. The crystal structure of EuZn2_2P2_2 with P3ˉm1P\bar{3}m1 symmetry and A-type AFM order with in-plane magnetic moments (a). Electronic band structure for Eu 4f4f electrons treated as valence states (b) along high symmetry direction of the Brillouin zone (c). The orange and blue lines present results in the absence and presence of the spin–orbit coupling, respectively.

Role of ff electrons and correlation. We compare the electronic band structure for Eu 4f4f electrons treated as a core state and valence state. In both cases, the electronic band structure exhibits similar features (cf. Figure S3 and Figure S4 in the SM [13]). However, treating 4f4f electrons as a core state leads to the strong overestimation of band gaps (cf. Tab. S1 in the SM [13]). Including the Hubbard-type local interaction UU in the DFT calculations (DFT+U) shifted the 4f4f valence state to lower energies (see Fig. S4 in the SM [13]). We wish to mention here, that the Eu 4f4f states for assumed U=5U = 5 eV are located around the binding energy 1.5 eV, and are well visible as the flat bands Figure 1(b). Additionally, from the comparison band structures calculated without and with spin-orbit coupling, we can conclude that this coupling strongly affects only the Eu 4f4f states. This is reflected in the strong decoupling of the flat band around binding energy 1.5 eV (cf. orange and blue lines on Figure 1(b), related to the results in the absence and presence of the spin–orbit coupling, respectively).

Figure S3. The electronic band structure of EuZn2P2\mathrm{EuZn}_{2}\mathrm{P}_{2} with Eu 4f4f electrons treated as core states. The orange and blue lines present results in the absence and presence of the spin–orbit coupling, respectively.

Figure S4. Influence of the Hubbard UU on the electronic band structure. Results for UU equal 3 eV, 5 eV, and 7 eV (from top to bottom). The orange and blue lines present results in the absence and presence of the spin–orbit coupling, respectively.

Magnetization

Figure 2(a) shows the variation of static magnetic susceptibility (χ=M/H\chi= M/H) with temperature (2–30 K) in a magnetic field of 1 kOe applied parallel and perpendicular to the crystallographic cc-direction. The compound orders antiferromagnetically below TN=23.5T_{\mathrm{N}} = 23.5 K, which is consistent with that reported earlier [12]. The χ(T)\chi(T) measured in 5 kOe (H⊥cH \perp c) is fitted with the Curie–Weiss law, in the temperature range 30–300 K:

Magnetic susceptibility, inverse susceptibility, and magnetization isotherms

Figure 2. (a) Magnetic susceptibility measured in magnetic field of 1 kOe, applied parallel and perpendicular to the cc-axis. (b) Inverse susceptibility vs. TT, (in H=5H = 5 kOe), red line represents fit with the Eq. (1). (c) Magnetization isotherms, measured at T=2T = 2 K in magnetic field applied parallel and perpendicular to the cc-axis. The horizontal line corresponds to the 7μB7\mu_{\mathrm{B}} of Eu2+^{2+} ion.

χ=CT−θp(1)\chi= \frac{C}{T-\theta_p} \tag*{(1)}

where CC is the Curie constant and θp\theta_p is the paramagnetic Curie–Weiss temperature [as shown with red line in Figure 2(b)]. The fit of χ−1(T)\chi^{-1}(T) with Eq. (1) provided an effective magnetic moment, μeff=8.19 μB\mu_{\mathrm{eff}} = 8.19\,\mu_{\mathrm{B}}, and θp=27.45\theta_p = 27.45 K. The value of μeff\mu_{\mathrm{eff}} is close to the theoretical effective magnetic moment of Eu2+^{2+} ion [gJJ(J+1)=7.94 μBg_J\sqrt{J(J+1)} = 7.94\,\mu_{\mathrm{B}}]. The large positive value of θp\theta_p suggests the dominating presence of ferromagnetic (FM) interactions.

Figure 2(c) shows the magnetic field dependent magnetization M(H)M(H) measured at 2 K for both H⊥cH \perp c and H∥cH \parallel c directions. For H∥cH \parallel c, magnetization increases continuously with increasing magnetic field up to Hsat∥≈24H_{\mathrm{sat}}^{\parallel} \approx24 kOe, where a saturation is observed. The saturated moment is 6.8 μB\mu_{\mathrm{B}}, which is very close to the theoretical value for spin-only magnetic moment of Eu2+^{2+} ion (i.e. 7 μB\mu_{\mathrm{B}}). For H⊥cH \perp c the saturated moment is the same, although the saturating field is significantly smaller: Hsat⊥≈14H_{\mathrm{sat}}^{\perp} \approx14 kOe.

To characterize the underlying magnetic interactions in the EuZn2_2P2_2, χ(T)\chi(T) in different magnetic fields were collected as shown in Figure S6, Figure S6) in the SM [13]. In both H∥cH \parallel c and H⊥cH \perp c, TNT_{\mathrm{N}} shifts towards lower temperatures and transition gradually broadens with the increasing magnetic field. Figure S7, Figure S7) in the SM [13], shows the magnetic isotherms, M(H)M(H) measured in the temperature range 5–30 K in the field of both orientations. A gradual change from the AFM-typical saturation to the Brillouin-type behavior is observed while increasing the temperature in the ordered region.

Magnetic susceptibility measured in different magnetic fields applied perpendicular and parallel to the c-axis

Figure S6. Magnetic susceptibility measured in different magnetic fields applied (a) perpendicular and (b) parallel to the cc-axis.

Magnetic isotherms at different temperatures in magnetic fields applied perpendicular and parallel to the c-axis

Figure S7. Magnetic isotherms at different temperatures in magnetic fields applied (a) perpendicular and (b) parallel, to the cc-axis.

Next, we calculated magneto-crystalline anisotropy energy (MAE), presented in Figure 3. In the calculations, we assumed that the Eu magnetic moment with A-type AFM order is locked in the acac plane, while angle ϕ\phi is relative to aa direction. Figure 3 shows the minimum of the MAE when magnetic moments are parallel to the aa direction (ϕ=0\phi= 0). The state with magnetic moments parallel to the cc direction (ϕ=π/2\phi= \pi/2) has larger energy. These two configurations are separated by the energy difference ≈0.8\approx0.8 meV per Eu atom. Our results strongly suggest that the abab is the easy plane, which differs from the previous report [12]. This finding is in agreement with very recent study of EuZn2_2P2_2, which also claims that the hard axis is parallel to cc crystallographic direction [24].

Figure 3. Magneto-crystalline anisotropy energy (MAE) of Eu ions obtained with DFT calculations. The orientation of the magnetic moment is denoted by the angle relative to the aa direction within the acac plane (inset). Minimum at ϕ=0\phi= 0 corresponds to the magnetic moments parallel to aa direction.

Specific heat study

Figure 4 shows the temperature dependence of heat capacity (CpC_p) measured in zero magnetic fields, which reveals a sharp transition close to TNT_{\mathrm{N}}. Above 50 K, CpC_p is fitted with a combination of Debye and Einstein terms, and the result is used to extract the magnetic entropy of the compound in the entire temperature range. The Debye model comprises acoustic phonons, whereas the Einstein

Temperature dependence of the specific heat

Figure 4. Temperature dependence of the specific heat of EuZn2_2P2_2 (left axis). The solid red curve is the Debye-Einstein fit with Eq. (2). The green curve represents the residuals of that fit. The blue line (right axis) shows the corresponding magnetic entropy.

model tackles optical ones. The model we adopted[25]:

Cph(T)=(1−m)9Rn(TθD)3∫0θD/Tx4ex dx(ex−1)2+m3Rn(θET)2eθE/T(eθE/T−1)2(2)C_{\mathrm{ph}}(T) = (1-m)9Rn\left(\frac{T}{\theta_{\mathrm{D}}}\right)^3 \int_{0}^{\theta_{\mathrm{D}}/T}\frac{x^4 e^x\,\mathrm{d}x}{(e^x-1)^2} + m3Rn\left(\frac{\theta_{\mathrm{E}}}{T}\right)^2\frac{e^{\theta_{\mathrm{E}}/T}}{(e^{\theta_{\mathrm{E}}/T}-1)^2} \tag*{(2)}

where mm is the relative weight of the Einstein term, nn, RR, θD\theta_{\mathrm{D}} and θE\theta_{\mathrm{E}} are: the number of atoms per formula unit, universal gas constant, Debye temperature, and Einstein temperature, respectively. Fit of the Eq. (2) provided the values of mm, θD\theta_{\mathrm{D}} and θE\theta_{\mathrm{E}}: 0.44(3), 479.88(2) K and 135.54(1) K, respectively. θD>θE\theta_D > \theta_E indicates the dominance of acoustic phonons at higher temperatures. The magnetic entropy (SmagS_{\mathrm{mag}}) was calculated as Smag=∫2 K300 K1T(Cp−Cph)dTS_{\mathrm{mag}} = \int_{2\ \mathrm{K}}^{300\ \mathrm{K}} \frac{1}{T}(C_p-C_{\mathrm{ph}})\mathrm{d}T. It increases with increasing TT above 2 K, and reaches a maximum of 17 Jmol−1^{-1}K−1^{-1} above TNT_N (see Figure 4), which is very close to theoretical entropy Rln⁡(2S+1)=17.3R\ln(2S+1)=17.3 Jmol−1^{-1}K−1^{-1} for Eu2+^{2+} (S=7/2S=7/2). The earlier report on the same compound had shown that the saturation magnetic entropy attained a maximum of 23Rln⁡8\frac{2}{3}R\ln8, which is much smaller than expected for the Eu2+^{2+} [12]. The analysis in that paper has employed two-terms Debye model without including the Einstein model. We fitted the same model to our specific heat data (see Fig. S9 in the SM [13]), but it did not retrieve the full magnetic entropy. In contrast, our fit of Eq. (2) (including both the Debye and Einstein model) shows that the saturation entropy is very close to the Rln⁡8R\ln8.

Above TNT_N there are short-range spin fluctuations present, and hence we restricted the fitting range to 50–300 K. The same model of specific heat was adapted for several EuT2_2X2_2-type compounds [8, 22].

Electrical transport

The temperature dependence of the electrical resistivity (ρ\rho) measured with the electric current flowing in the hexagonal abab-plane (j⊥cj\perp c-axis) at different magnetic fields is shown in Figure 5(a). At 300 K, the ρ\rho is 3.35 Ω\Omegacm, with decreasing temperature it increases, and below 128 K the compound becomes an insulator. In magnetic fields up to 140 kOe, the curves appear similar and the magnitude of ρ\rho decreases, but the system remains insulating. The inset of Figure 5(a) shows field-dependent resistivity ρ(H)\rho(H), at 125 K, which tends to decrease due to reducing band gap. To estimate the band gap (EgE_g) the Arrhenius model was applied by fitting the resistivity data with an equation:

Resistivity versus temperature in different magnetic fields, inverse-temperature resistivity plot, and band gap versus magnetic field

Figure 5. Resistivity vs. temperature in different magnetic fields transverse to the crystallographic cc-axis and with current flowing within abab-plane. Inset: Magnetic field dependence of the resistivity at 125 K. (b) ln⁡ρ\ln\rho vs. 1/T1/T plot, for zero applied field, and 150–300 K range of TT. The red curve represents the fit with Eq. (3). (c) Band gap as a function of the magnetic field.

ρ∝eEg/2kBT.(3)\rho\propto e^{E_g/2k_{\mathrm{B}}T}. \tag*{(3)}

Dependence of the ln⁡ρ\ln\rho on the inverse temperature is shown in Figure 5(b), which is linear, and yielded Eg=0.177E_g=0.177 eV when fitted with Eq. (3). The band gap decreases monotonically with the application of a magnetic field [Figure 5(c)] down to 0.13 eV in 140 kOe. The small difference between the experimentally observed and the theoretically predicted band gap may be due to the shift of Fermi level induced by the defect level present in the charge gap [12].

Effective Heisenberg model and mechanism of magnetic interaction

To study the magnetic ordering of EuZn2_2P2_2 theoretically, we formulated the effective Heisenberg model [26],

Heff=−∑⟨i,j⟩Jijeiej,(4)\mathcal{H}_{\mathrm{eff}}=-\sum_{\langle i,j\rangle}J_{ij}\mathbf{e}_i\mathbf{e}_j, \tag*{(4)}

where ei\mathbf{e}_i is the unit vector in the direction of the magnetic moment at site ii, and JijJ_{ij} are the direction-dependent exchange parameters. In this notation, the value of on-site atomic magnetic moment SiS_i is included into the exchange parameters JijJ_{ij} and the summation runs only once over each pair of ii and jj.

The model concerns only the magnetic moments of Eu, which form two-dimensional triangular layers stacked over each other without any interlayer offset [see Figure 6(a)]. From the DFT calculations, we found that the triangular layers possess the ferromagnetic coupling J∥=1.23J_{\parallel} = 1.23 meV, while the AFM interlayer coupling J⊥=−0.35J_{\perp} = -0.35 meV (detailed procedure can be found in the SM [13]). The obtained values of the exchange parameters are comparable with those reported in the literature [27]. The magnitude of ferromagnetic (intralayer) coupling is bigger than the AFM (interlayer) coupling (∣J∥∣>∣J⊥∣|J_{\parallel}| > |J_{\perp}|). This relation between exchange coupling parameters is expected and related to the distances between Eu atoms (shorter in abab plane than along cc direction).

Figure 6. (a) Coupling exchange parameters found from ab initio calculations. (b) The average magnetization of a single Eu layer and (c) specific heat obtained from the Monte Carlo simulations. Red and blue lines, correspond to the results obtained for 8×8×168 \times8 \times16 and 16×16×3216 \times16 \times32 spin blocks, respectively.

To correlate the experimental data with the effective Heisenberg model, we have used two different spin blocks of size 8×8×168 \times8 \times16 and 16×16×3216 \times16 \times32. The temperature dependence of the specific heat and average magnetization of a single layer is simulated with Monte Carlo method using SpinMC.jl package [28]. Results of our Monte Carlo simulations are presented in Figure 6. The magnetic order parameter (defined as an average magnetization of the triangular Eu layer), falls abruptly around 21 K [Figure 6(b)]. Around the same temperature, the specific heat also possesses a clear peak [Figure 6(c)], which directly indicates a magnetic phase transition from AFM to PM phase. Theoretically obtained Néel temperature TN≈21T_N \approx21 K is in excellent agreement with experimental one. Presented results weakly depend on the size of the investigated system [cf. red and blue lines on Figure 6(b, Figure 6c)]. The presence of two exchange coupling parameters, with opposite sign, are also reflected in the magnetocaloric effect, which is shown in the SM [13].

Magnetic interaction mechanism

We propose an extended superexchange mechanism [29], related to superposed electronic states of cation in a specific path, as a source of the magnetic interaction in EuZn2_2P2_2. In the case of typical superexchange interaction, transition metal cations interact indirectly through intervening nonmetal anions. The superexchange interaction, explained by Goodenough–Kanamori–Anderson rule [30], strongly depends on the cation-anion-cation bond angles. It was argued that the 180° bond stabilizes an AFM ground state, while a 90° bond is responsible for the FM interactions. Nevertheless, the mentioned mechanism is not limited by only said two bond angles, rather allowed to be extended to any bond angles. The extended version of the superexchange interaction was previously used to explain the magnetic ground state of several compounds, viz. CrOX (XX = Cl, Br) [29, 31], or RbLnSe2_2 (LnLn = Ce, Pr, Nd, Gd). [32]

In the case of EuZn2_2P2_2, the magnetic interaction is mediated by P atoms, which is allowed through two different paths [see Figure 7(a)]. In the abab plane, the magnetic interaction is realized by “path A” (blue dashed line), containing one P atom. Here, the Eu-P-Eu bond angle (≈85.52∘\approx85.52^\circ) is close to the ideal 90° value, which favors the ferromagnetic coupling. On the other hand, the magnetic interaction between two Eu atoms along the cc-direction is mediated by two P atoms located between two Eu-layers and marked by “path B” (green dashed line), in Figure 7(a). In this case, the Eu-P-P bond angle (≈105.90∘\approx105.90^\circ) is relatively large and corresponds to the AFM exchange coupling. Very similar bond lengths, (and hence the bond angles) of the Zintl phases simultaneously possessing insulating, and AFM properties are discussed in Ref. [12], although without any exchange interactions. The authors of that paper phenomenologically approximated the strength ratio of the AFM and FM interactions as 1:-4. We accounted for the Heisenberg exchange interactions J⊥J_{\perp} and J∥J_{\parallel}, and showed that J∥/J⊥J_{\parallel}/J_{\perp} is very close to -3.5. Thus, we propose that the extended superexchange mechanism can be used to explain the AFM ordering in insulating Zintl phases, that possess strong intralayer FM and weak interlayer AFM couplings. Moreover, such extended superexchange mechanism should be predominant in this class of materials.

Figure 7. (a) Bond lengths and bond angles in EuZn2_2P2_2. Blue and green dashed lines show paths participating in the extended superexchange mechanism. Modification of the (b) charge and (c) spin density distribution in EuZn2_2P2_2 obtained from DFT calculations. Yellow and cyan isosurfaces indicate the accumulation and depletion of the charge/spin density, respectively.

To further strengthen our argument, we have calculated the modification of charge as well as spin density distribution of the crystal using the DFT technique. Here, we can define modification of the charge density as [33]:

Δρ=ρEuZn2P2−(ρEu+ρZn2+ρP2),(5)\Delta\rho= \rho_{\mathrm{EuZn_2P_2}} - (\rho_{\mathrm{Eu}} + \rho_{\mathrm{Zn_2}} + \rho_{\mathrm{P_2}}), \tag*{(5)}

where ρEuZn2P2\rho_{\mathrm{EuZn_2P_2}} denotes density distribution for EuZn2_2P2_2, while ρi\rho_i density distribution for separated component forming this compound (i.e., i=Eui = \mathrm{Eu}, Zn, and P). The spin density redistribution can be defined in an analogous way. Figure 7(b) shows a transfer of charge density mostly between Zn and P, which is related to the realization of strong bonding between these atoms. Nevertheless, the electron density around Eu and Zn decreases and increases around P. This suggests the anion-like character of the phosphorus in chemical bonding of EuZn2_2P2_2, which we predicted in the previous paragraph applying of extended superexchange scheme. Interestingly, a contrasting behavior is observed in the case of the spin density rearrangement [see Figure 7(c)]. In this case, we observed a transfer of the spin density only between Eu and P atoms. Indeed, DFT analysis indicates the occurrence of the small magnetic moment on P atoms (with magnitude ≈0.02 μB\approx0.02\,\mu_{\mathrm{B}}). Observed spin density redistribution between Eu and P atoms, suggests that the P atoms play the role of the mediator within the magnetic interaction mechanism. Such rearrangement of the spin density further supports our choice of virtual hopping path (marked with red and green in Figure 7(a)). Additionally, we notice, that the electronic states closest to the “Fermi level” (top of valence band) are realized mostly by the Zn and P orbitals, what is well visible on the electronic density of states projected on atoms (see Figure S5 in SM [13]). This finding is in agreement with earlier investigation of EuZn2_2P2_2 [12]. This further supports out prediction about the role of Zn and P states in the mechanism of magnetic interaction, and

Figure S5. Electronic partial density of states (PDOS). The shaded gray area corresponds to the total DOS for the ff orbitals treated as core states. Red, green, and blue color solid lines correspond to the PDOS of Eu, Zn, and P, respectively. Similarly, dashed black and brown lines, correspond to the total DOS and Eu PDOS, respectively, when ff orbitals are treated as valence states.

forming of the AFM order.

Summary

The temperature and magnetic field-dependent magnetic susceptibility, isothermal magnetization, magnetocaloric effect, heat capacity, and electrical resistivity of EuZn2_2P2_2 single crystal indicate an antiferromagnetic phase transition at TN=23.5T_{\mathrm{N}} = 23.5 K due to the ordering of Eu2+^{2+} spins. M(H)M(H) at T=2T = 2 K, saturates to 6.8 μB6.8\,\mu_{\mathrm{B}} at Hsat⊥=14H_{\mathrm{sat}}^{\perp} = 14 kOe for H⊥c\mathrm{H}\perp c and at the Hsat∥=24H_{\mathrm{sat}}^{\parallel} = 24 kOe for H∥c\mathrm{H}\parallel c directions. The magnetic entropy retrieved from the heat capacity data is equal to 1717 J mol−1^{-1}K−1^{-1}, which closely matches the theoretical one. M(H)M(H) confirms that the easy-axis lies in the abab-plane. Electrical resistivity reveals the insulating nature of EuZn2_2P2_2, with band gap ≈0.17\approx0.17 eV, which survives (only slightly reduced), up to 140 kOe. Based on the magnetization data in function of both temperature and magnetic field, we sketched a magnetic phase diagram of the compound for both H⊥c\mathrm{H}\perp c and H∥c\mathrm{H}\parallel c directions. Figure 8. TNT_{\mathrm{N}} and HsatH_{\mathrm{sat}} are obtained from χ(T)\chi(T) and M(H)M(H) data, respectively.

Magnetic phase diagram of EuZn2P2

Figure 8. Magnetic phase diagram of EuZn2_2P2_2, drawn for magnetic field applied perpendicular and parallel to the cc-axis, using values of Néel temperature (red points) and saturation field (green points).

We proposed an extended superexchange interaction to be responsible for the magnetic ordering. The charge and spin density modifications indicate that phosphorus atoms play the role of anionic mediators, necessary in the superexchange mechanism. We formulated an effective model of the magnetic interaction, based on the ab initio DFT calculations. We found that the interaction in abab plane, realized via the Eu-P-Eu path, with bond angle of ≈86∘\approx86^\circ is moderately ferromagnetic (J∥=1.23J_{\parallel} = 1.23 meV), while that along cc direction (Eu-P-P-Eu path, with bond angle ≈106∘\approx106^\circ) is weakly antiferromagnetic (J⊥=−0.35J_{\perp} = -0.35 meV). The experimentally determined Néel temperature was successfully reproduced by the Monte Carlo simulations with those interaction parameters, which strongly supports the superexchange mechanism we propose. Theoretical calculation of the magneto-crystalline anisotropy energy of the compound. in the acac plane reveals that the easy axis is along the crystallographic aa direction. This finding is in agreement with a recent investigation [24]. Our results shed also new light on the mechanism of magnetic interactions in the Eu based compounds, like EuZn₂P₂.

Acknowledgments

Some figures in this work were rendered using Vesta [34]. We would like to thank F. L. Buessen for sharing his Monte Carlo code. This work was supported by National Science Centre (NCN, Poland) under Projects No. 2021/41/B/ST3/01141 (K.S., P.W., and D.K.) and 2021/43/B/ST3/02166 (A.P.). A.P. is grateful to Laboratoire de Physique des Solides in Orsay (CNRS, University Paris Saclay) for hospitality during a part of the work on this project.

Superexchange Interaction in Insulating EuZn2_2P2_2

Karan Singh; Shovan Dan; A. Ptok; T. A. Zaleski; O. Pavlosiuk; P. Wiśniewski; D. Kaczorowski

In this Supporting Information, we present additional results:

  • § 9 Structure and composition characterization.

  • Figure S1 Laue diffractogram of the EuZn2_2P2_2.

  • Figure S2 EDS spectrum.

  • § 10 Exchange parameters based on DFT calculations.

  • Figure S3 The electronic band structure of EuZn2_2P2_2 with Eu 4f4f electrons treated as core states.

  • Table S1 Band gaps from DFT calculations dependently by UU value.

  • Figure S4 Influence of the Hubbard UU on the electronic band structure (ff electrons treated as valence states).

  • Figure S5 Electronic partial density of states (PDOS).

  • § 11 Magnetic susceptibility and magnetization.

  • Figure S6 Magnetic susceptibility.

  • Figure S7 Magnetic isotherms at different temperatures.

  • § 12 Specific heat.

  • Figure S8 CP/T3C_P/T^3 and its fit using Debye and Einstein contribution.

  • Figure S9 CP/TC_P/T fitted using two Debye model and concerned SmagS_{\mathrm{mag}}.

  • § 13 Magnetocaloric effect.

  • Figure S10 ΔSM(T)\Delta S_M(T) in H∥c\mathbf{H} \parallel c, and H⊥c\mathbf{H} \perp c directions.

Laue diffractogram of the EuZn2P2 crystal

Figure S1. Laue diffractogram of the EuZn2_2P2_2 crystal obtained with incident X-ray beam along [1] direction.

Energy-dispersive X-ray spectrum of the EuZn2P2 crystal

Figure S2. Energy-dispersive X-ray spectrum collected on arbitrary spot on the EuZn2_2P2_2 crystal. Weight and atomic composition are shown in the inset.

Hubbard UUindirect gap (eV)direct gap (eV)
w/o UU (f states in core) – Figure S30.71200.8047
U=3U = 3 eV – Figure S40.17840.3678
U=5U = 5 eV – Figure S40.42510.5933
U=7U = 7 eV – Figure S40.54120.6953

Table S1. Band gaps from DFT calculations

Specific heat divided by temperature cubed and its Debye and Einstein contributions

Figure S8. Cp/T3C_p/T^3 and its fit using Debye and Einstein contribution. Decomposed Debye and Einstein contribution (CD/T3C_D/T^3 and CE/T3C_E/T^3) are also plotted as a function of TT.

Specific heat divided by temperature and magnetic entropy

Figure S9. Left-axis: Cp/TC_p/T fitted to (in the range 40–300 K) the two-Debye model as proposed in Ref. 12. Fitting residuals are plotted with a blue line. Right-axis: the magnetic entropy. The entropy values obtained with the two-Debye-terms model and our with fit including one Debye and one Einstein mode are shown with olive and orange dotted line, respectively. The actual magnetic entropy for Eu2+^{2+} is Rln⁡8=17.3R\ln8 = 17.3 J mol−1^{-1} K−1^{-1}.

Plots of $\Delta S_M$ versus temperature for magnetic fields perpendicular and parallel to the c-axis, with insets showing the low-field regions

Figure S10. ΔSM(T)\Delta S_M(T) calculated in magnetic fields (a) perpendicular and (b) parallel to the c-axis. Insets show the field region 0→250 \to25 kOe for H∥cH\parallel c and 0→100 \to10 kOe for H⊥cH\perp c directions.

Structure and Composition Characterization

To analyze the chemical composition, Energy-dispersive X-ray spectroscopy (EDS) experiment was carried out on a few pieces of the synthesized EuZn2_2P2_2 crystal, which shows that stoichiometry is a good agreement with the nominal composition (see Figure S2). The Laue diffractogram of the oriented EuZn2_2P2_2 crystal is shown in Figure S1, it perfectly corresponds to the trigonal structure with space-group P3ˉm1P\bar{3}m1.

Exchange parameters based on DFT calculations

To find the parameters of exchange between magnetic moments of Eu, we performed DFT calculation within the supercell containing 2×2×12 \times2 \times1 magnetic unit cells. We calculate the ground states (GS) energies for four fixed configurations of magnetic moments:

  • AFM configuration, with the GS energy −211.38165-211.38165 eV,

  • AFM configuration with one inverted magnetic moment, and the GS energy −211.36549-211.36549 eV,

  • FM configuration, with the GS energy −211.37605-211.37605 eV,

  • FM configuration with one inverted magnetic moment, and the GS energy −211.36269-211.36269 eV.

In the simplest case, the GS energy of the mentioned configurations may be used to represent the total energy of the system as a function of the magnetic moments, in the form:

E=E0−∑ijJ∥eiej−∑ijJ⊥eiej(6)E = E_0 - \sum_{ij} J_{\parallel} e_i e_j - \sum_{ij} J_{\perp} e_i e_j \tag*{(6)}

where summation over sites ii and jj is perform within the FM layer (J∥J_{\parallel} term) or between AFM layers (J⊥J_{\perp} term). E0E_0 denotes the energy of the system in the absence of magnetic order (i.e. non-magnetic system). Solving the system of equations for the mentioned configuration, we found E0=−211.349E_0 = -211.349 eV, J∥=1.23J_{\parallel} = 1.23 meV, and J⊥=−0.35J_{\perp} = -0.35 meV.

Magnetic Susceptibility and Magnetization

Specific Heat

For a compound containing Einstein contribution, it is expected to have a hump in the Cp/T3C_p/T^3 vs. TT plot. However, in the present case, the hump occurs close to the magnetic ordering temperature and not properly visible in naked eye. For visualizing the feature, we have decomposed the specific heat data into the Debye and Einstein, as shown in Figure S8 and plotted the Cp/T3C_p/T^3 vs. TT plot, which clearly shows the hump near to the ordering temperature due to Einstein contribution.

We have also shown that the two-Debye model, as proposed in ref: 12, does not able to rescue the full magnetic entropy of Eu2+^{2+} ions (Figure S9).

Magnetocaloric Effect

The magnetocaloric effect (MCE) is an important phenomenon providing an overview of underlying magnetic interactions, apart from its importance for technological applications. For the latter aspect, the absolute value of magnetic entropy change (∣ΔSM∣|\Delta S_M|), and the relative cooling power play crucial roles. However, recently, MCE is being also used as a tool to explain the complex competition of magnetic interactions, eg. in skyrmions [35]. The magnetic entropy gain upon change of magnetic field from H1H_1 to H2H_2, can be obtained from Maxwell’s relation:

ΔSM(T,ΔH)=∫H1H2(∂M∂T)dH.(7)\Delta S_M(T,\Delta H)=\int_{H_1}^{H_2}\left(\frac{\partial M}{\partial T}\right)\mathrm{d}H. \tag*{(7)}

Nevertheless, the magnetic entropy change is positive for FM interactions and negative for AFM interactions, dominating in given field range, and temperature [35].

Figure S10 shows the ΔSM\Delta S_M for both H⊥cH\perp c and H∥cH\parallel c field directions. ∣ΔSM∣|\Delta S_M| attains maxima near TNT_N with ≈14.9\approx14.9 and 1717 J·kg−1^{-1}K−1^{-1}, for magnetic field applied in perpendicular and parallel to cc-axis, respectively. Closer observation of ΔSM\Delta S_M at the low fields (see the insets of Figure S10), indicates a positive contribution in the ΔSM\Delta S_M value, below TNT_N, and thus the domination of FM interactions, in full agreement with our discussion of magnetization.

References

  1. [1]T. F. Fässler and C. Kronseder, BaSn₃: A superconductor at the border of zintl phases and intermetallic compounds. real-space analysis of band structures, Angew. Chem., Int. Ed. Engl. 36, 2683 (1997).DOI
  2. [2]Y. Xu, Z. Song, Z. Wang, H. Weng, and X. Dai, Higher-Order Topology of the Axion Insulator EuIn₂As₂, Phys. Rev. Lett. 122, 256402 (2019).DOI
  3. [3]J. Yan, Z. Z. Jiang, R. C. Xiao, W. J. Lu, W. H. Song, X. B. Zhu, X. Luo, Y. P. Sun, and M. Yamashita, Field-induced topological Hall effect in antiferromagnetic axion insulator candidate EuIn₂As₂, Phys. Rev. Res. 4, 013163 (2022).DOI
  4. [4]C. Chen, W. Xue, S. Li, Z. Zhang, X. Li, X. Wang, Y. Liu, J. Sui, X. Liu, F. Cao, et al., Zintl-phase Eu₂ZnSb₂: A promising thermoelectric material with ultralow thermal conductivity, PNAS 116, 2831 (2019).
  5. [5]Z. Zada, H. Ullah, R. Zada, S. Zada, A. Laref, S. Azam, A. A. Khan, and M. Irfan, Structure stability, half metallic ferromagnetism, magneto-electronic and thermoelectric properties of new zintl XCr₂Bi₂ (X= Ca, Sr) compounds for spintronic and renewable energy applications, Phys. B: Condens. Matter 607, 412866 (2021).
  6. [6]M. C. Rahn, J.-R. Soh, S. Francoual, L. S. I. Veiga, J. Strempfer, J. Mardegan, D. Y. Yan, Y. F. Guo, Y. G. Shi, and A. T. Boothroyd, Coupling of magnetic order and charge transport in the candidate Dirac semimetal EuCd₂As₂, Phys. Rev. B 97, 214422 (2018).DOI
  7. [7]Z.-C. Wang, E. Been, J. Gaudet, G. M. A. Alqasseri, K. Fruhling, X. Yao, U. Stuhr, Q. Zhu, Z. Ren, Y. Cui, C. Jia, B. Moritz, S. Chowdhury, T. Devereaux, and F. Tafti, Anisotropy of the magnetic and transport properties of EuZn₂As₂, Phys. Rev. B 105, 165122 (2022).DOI
  8. [8]S. Pakhira, M. A. Tanatar, and D. C. Johnston, Magnetical, thermal, and electronic-transport properties of EuMg₂Bi₂ single crystals, Phys. Rev. B 101, 214407 (2020).DOI
  9. [9]S. Pakhira, T. Heitmann, S. X. M. Riberolles, B. G. Ueland, R. J. McQueeney, D. C. Johnston, and D. Vaknin, Zero-field magnetic ground state of EuMg₂Bi₂, Phys. Rev. B 103, 024408 (2021).DOI
  10. [10]J.-R. Soh, E. Schierle, D. Y. Yan, H. Su, D. Prabhakaran, E. Weschke, Y. F. Guo, Y. G. Shi, and A. T. Boothroyd, Resonant x-ray scattering study of diffuse magnetic scattering from the topological semimetal EuCd₂As₂ and EuCd₂Sb₂, Phys. Rev. B 102, 014408 (2020).DOI
  11. [11]F. Weber, A. Cosseev, S. Drobnik, A. Faibt, K. Grube, A. Nateprov, C. Pfeifferer, M. Uhlarz, and H. v. Löhneysen, Low-temperature properties and magnetic order of EuZn₂Sb₂, Phys. Rev. B 73, 014427 (2006).DOI
  12. [12]T. Berry, V. J. Stewart, B. W. Y. Redemann, C. Ly-gouras, N. Varnava, D. Vanderbilt, and T. M. McQueen, A-type antiferromagnetic order in the Zintl-phase insulator EuZn₂P₂, Phys. Rev. B 106, 054420 (2022).DOI
  13. [13]Supplemental Material at [URL will be inserted by publisher] contains results presenting: (1) Structure and composition characterization, (2) Exchange parameters from DFT calculations, (3) Magnetic susceptibility and magnetization, (4) Specific heat, (5) Magnetocaloric effect and Ref. [35].
  14. [14]P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B 50, 17953 (1994).pubmed.ncbi.nlm.nih.gov/9976227
  15. [15]G. Kresse and J. Hafner, Ab initio molecular-dynamics simulation of the liquid-metal–amorphous-semiconductor transition in germanium, Phys. Rev. B 49, 14251 (1994).pubmed.ncbi.nlm.nih.gov/10010505
  16. [16]G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B 54, 11169 (1996).pubmed.ncbi.nlm.nih.gov/9984901
  17. [17]G. Kresse and D. Joubert, From ultrasoft pseudopotentials to the projector augmented-wave method, Phys. Rev. B 59, 1758 (1999).DOI
  18. [18]J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized gradient approximation made simple, Phys. Rev. Lett. 77, 3865 (1996).pubmed.ncbi.nlm.nih.gov/10062328
  19. [19]H. J. Monkhorst and J. D. Pack, Special points for Brillouin-zone integrations, Phys. Rev. B 13, 5188 (1976).DOI
  20. [20]S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, and A. P. Sutton, Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study, Phys. Rev. B 57, 1505 (1998).DOI
  21. [21]L. Frik and A. Mewis, Stacking variants of SrPtSb and CaAl₂Si₂ analogous units, Z. anorg. allg. Chem. 625, 126 (1999).
  22. [22]S. Pakhira, F. Islam, E. O’Leary, M. A. Tanatar, T. Heitmann, L.-L. Wang, R. Prozorov, A. Kaminski, D. Vaknin, and D. C. Johnston, A-type antiferromagnetic order in semiconducting EuMg₂Sb₂ single crystals, Phys. Rev. B 106, 024418 (2022).DOI
  23. [23]A. Ptok, S. Basak, P. Piekarz, and A. M. Oleś, Influence of electrons on the electronic band structure of rare-earth nickelates, Condensed Matter 8, 19 (2023).DOI
  24. [24]S. Krebber, M. Kopp, C. Garg, K. Kummer, J. Sichelschmidt, G. Schulz, G. Poelchen, M. Mende, K. Warawa, M. D. Thomson, A. V. Tarasov, D. Y. Usachov, D. V. Vyalikh, H. G. Roskos, J. Müller, C. Krellner, and K. Kliemt, Colossal magnetoresistance in EuZn₂P₂ and its electronic and magnetic structure (2023), arXiv:2302.14539.DOI
  25. [25]E. Gopal, Specific heats at low temperatures (Springer Science & Business Media, 2012).
  26. [26]M. Pajda, J. Kudrnovský, I. Turek, V. Drchal, and P. Bruno, Ab initio calculations of exchange interactions, spin-wave stiffness constants, and curie temperatures of Fe, Co, and Ni, Phys. Rev. B 64, 174402 (2001).DOI
  27. [27]Y. Guo, B. Wang, X. Zhang, S. Yuan, L. Ma, and J. Wang, Magnetic two-dimensional layered crystals meet with ferromagnetic semiconductors, InfoMat 2, 639 (2020).DOI
  28. [28]F. L. Buessen, SpinMC.jl Monte Carlo package, https://github.com/fbuessen/SpinMC.jl.
  29. [29]F. Zhang, Y.-C. Kong, R. Pang, L. Hu, P.-L. Gong, X.-Q. Shi, and Z.-K. Tang, Super-exchange theory for polyvalent anion magnets, New J. Phys. 21, 053033 (2019).
  30. [30]J. B. Goodenough, Magnetism and the Chemical Bond (Interscience Publishers, 1963).DOI
  31. [31]S. W. Jang, D. H. Kiem, J. Lee, Y.-G. Kang, H. Yoon, and M. J. Han, Hund’s physics and the magnetic ground state of CrOX (X = Cl, Br), Phys. Rev. Mater. 5, 034409 (2021).DOI
  32. [32]L. Azzouz, M. Halit, Z. Charif, H. Baaziz, M. Rérat, H. Denawi, and C. F. Matta, Magnetic semiconductor properties of RbLnSe₂ (Ln = Ce, Pr, Nd, Gd): A density functional study, J. Magn. Magn. Mater. 501, 166448 (2020).DOI
  33. [33]A. Potk, K. J. Kapcia, and A. Ciechan, Electronic properties of Bi₂Se₃ doped by 3d transition metal (Mn, Fe, Co, or Ni) ions, J. Phys.: Condens. Matter 33, 065501 (2020).
  34. [34]K. Momma and F. Izumi, VESTA3 for three-dimensional visualization of crystal, volumetric and morphology data, J. Appl. Crystallogr. 44, 1272 (2011).DOI
  35. [35]S. Jamaluddin, S. K. Manna, B. Giri, P. P. Madduri, S. S. Parkin, and A. K. Nayak, Robust antiskyrmion phase in bulk tetragonal Mn–Pt (Pd)–Sn Heusler system probed by magnetic entropy change and AC-susceptibility measurements, Adv. Funct. Mater. 29, 1901776 (2019).DOI

Paper details

vlnc:W_4je9ba9z
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents