betaFeedback

WorksPhysics and Astronomy

Unconventional magnetic glassiness in non-centrosymmetric Sm7_7Pd3_3: Interplay of magnetic frustration, long-range order, and frozen domains

Abstract

We present a comprehensive investigation of the intricate spin dynamics in the non-centrosymmetric compound Sm7_7Pd3_3, revealing the coexistence of spin glass, domain glass, and ferromagnetic (FM) behaviors. Magnetic field-dependent measurements indicate large coercivity, suggesting ferromagnetic domain formation below the Curie temperature (TC∼173T_C \sim 173 K), while temperature-dependent magnetization data point to antiferromagnetic (AFM) coupling, highlighting the competition between FM and AFM interactions. Detailed ac susceptibility, isothermal remanent magnetization, and aging effect measurements demonstrate the presence of two distinct types of glassiness in the sample, and their possible origins are discussed extensively. Magnetization measurements reveal the mixing of the J = 5/2 ground state of Sm3+^{3+} with the excited J = 7/2 multiplet, lying 965 K above. The specific heat data show further crystalline electric field splitting of the J = 5/2 state into a ground-state doublet and a fourfold-degenerate excited state.

Introduction

Describing the magnetic state of a complex compound with multiple magnetically active atomic positions is often a non-trivial task. Such compounds rarely conform to classic definitions of ferromagnetic (parallel moment alignment inside magnetic domains), antiferromagnetic (antiparallel moment alignment resulting in near-zero magnetization), or canonical spin-glass (equal probability of AFM and FM configurations due to geometrical frustration of the spin system) order [1, 2, 3]. Altermagnetism is a telling example of an ordered system where the classic approach is no longer sufficient to define the magnetic state [4, 5, 6], and, likely, it will not be the last one. Often magnetic systems exhibit multiple characteristic behaviors – for example, displaying FM behavior along one crystallographic direction and AFM behavior along another or combine magnetic frustration with well-established long-range order [7, 8, 9]. Acquiring a fundamental understanding of magnetism in such materials presents a substantial challenge, particularly when conventional approaches, such as magnetic structure determination via neutron diffraction, are not feasible due to the inherently low scattering cross section of the constituent elements and/or the presence of complex dynamic spin fluctuations [10]. Yet to advance the current paradigm of magnetic ordering, we must explore and explain these complex examples with utmost rigor.

The non-centrosymmetric intermetallic compounds with the general formula R7T3\mathrm{R}_7\mathrm{T}_3 (where R represents a rare-earth element and T denotes a transition metal) have attracted significant research interest due to their diverse range of exotic physical phenomena, including unconventional electrical transport, large coercivity with nearly zero magnetization, exchange bias, giant magnetocaloric effect (GMCE), and glassy magnetic states [11, 12, 13].

For example, the lack of inversion symmetry in the crystal structure induces unconventional superconductivity in La7T3\mathrm{La}_7\mathrm{T}_3 (T = Ir, Rh, and Ni) via time-reversal symmetry breaking [11, 12, 14]. All R7T3\mathrm{R}_7\mathrm{T}_3 compounds crystallize in the hexagonal Th7Fe3\mathrm{Th}_7\mathrm{Fe}_3-type structure (P63mcP6_3mc) at room temperature, where the rare-earth atoms occupy three distinct crystallographic sites (two at 6c6c and one at 2b2b) [15]. Each 6c6c site belongs to a distinct atomic plane and one of these planes is shared with 2b2b position. This intricate structural arrangement results in competing magnetic interactions, with small energy differences between various configurations, leading to frustrated ground states. Furthermore, the delicate interplay between the localized 4f4f moments of R atoms and the itinerant dd states of T atoms enhances the complexity of the magnetic and electronic structures of these compounds [16, 17, 18].

For instance, Nd7Rh3\mathrm{Nd}_7\mathrm{Rh}_3 exhibits two antiferromagnetic (AFM) transitions at TN1=32T_{\mathrm{N}1} = 32 K and TN2=10T_{\mathrm{N}2} = 10 K, along with an anomalous field-induced first-order AFM-to-FM transition below TN2T_{\mathrm{N}2}, exhibiting the signature of phase coexistence after the field cycling [13]. The co-existing FM and AFM regions form a phase separated state, which is distinct from a typical spin-glass state and gives rise to an unconventional time-dependent magnetic behavior [13, 19]. Termed “magnetic glass” [13, 19] to differentiate from “spin glass”, such states are common in compounds with strong magnetostructural coupling, including rare earths manganates and kinetically arrested Gd5Ge4\mathrm{Gd}_5\mathrm{Ge}_4 [9, 19, 20, 21]. Similar order-order successive magnetic transitions are observed in Ho7Rh3\mathrm{Ho}_7\mathrm{Rh}_3: a transition from a paramagnetic state to an incommensurate AFM structure at TN=32T_{\mathrm{N}} = 32 K, followed by a spin reorientation transition around 22 K, and finally, the emergence of a ferromagnetic component for T<9T < 9 K [16]. Notably, short-range magnetic correlations persist in Ho7Rh3\mathrm{Ho}_7\mathrm{Rh}_3 well above TNT_{\mathrm{N}} (∼2TN\sim2T_{\mathrm{N}}), underscoring the presence of competing interactions even in the absence of long-range order [16]. A particularly intriguing example is Nd7Pd3\mathrm{Nd}_7\mathrm{Pd}_3, which undergoes a second-order paramagnetic-to-AFM transition at TN=37T_{\mathrm{N}} = 37 K, followed by a first-order AFM-to-FM transition at TC=33T_{\mathrm{C}} = 33 K, accompanied by a structural change from P63mcP6_{3}mc to Cmc21Cmc2_{1}, which is necessary to accommodate the ground state FM structure [17]. The close proximity of TNT_{\mathrm{N}} and TCT_{\mathrm{C}} in this compound underscores the fragility of AFM state [17]. We note that first-order transition at TCT_{\mathrm{C}} is practically anhysteretic with giant magnetocaloric effect and, if not for a high cost of Pd, could be considered as a very promising GMCE material [17, 22].

Among R7R_{7}Pd3_{3} compounds, our previous work highlights the remarkable magnetic behavior of Sm7_{7}Pd3_{3}. Sm7_{7}Pd3_{3} crystallizes in the Th7_{7}Fe3_{3}-type structure at room temperature and orders ferro- or ferrimagnetically below TC≈170T_{\mathrm{C}} \approx170 K [23, 24]. Our previous study establishes that the crystal symmetry of Sm7_{7}Pd3_{3} remains P63mcP6_{3}mc down to 5 K; however, significant changes in lattice parameters, unit cell volume, and the c/ac/a ratio occur at TCT_{\mathrm{C}}, even though this magnetoelastic transition is of second order [24]. Sm7_{7}Pd3_{3} exhibits extraordinarily large coercivity, while having nearly zero saturation magnetization—an ideal combination for efficient spin-transfer and memory applications [23, 24, 25, 26]. While the compound itself orders below room temperature, understanding the control knobs for tuning these technologically relevant properties presents an important fundamental and applied scientific problem. Our earlier study also uncovered the possible presence of antiferromagnetic interactions in the predominantly FM compound and DFT calculations showed that antiparallel moment alignment of Sm atoms from different atomic planes is energetically favourable [24]. We hypothesize that they give rise to a large exchange bias effect and a crossover from positive to negative magnetization as a function of temperature in Sm7_{7}Pd3_{3} [23, 24]. This crossover indicates strong blocking of the magnetic moments in the sample at low temperatures due to highly frustrated spin dynamics. However, the exact role of these competing interactions in shaping its observed magnetic behavior is unknown.

In addition, the Sm-based intermetallic compounds typically show a very small energy difference between the ground state (J=5/2J = 5/2), and the first (J=7/2J = 7/2, 0.1293 eV) and second (J=9/2J = 9/2, 0.2779 eV) excited multiplets of Sm3+^{3+}. This energy separation can be effectively altered by changing the crystal field through external perturbations, such as applied magnetic fields, pressure, or chemical substitution [27, 28, 29, 30, 31, 32]. As a result, the magnetic and thermodynamic properties of these compounds are governed by the degree of mixing between these multiplets, with the Van Vleck contribution playing a crucial role in their magnetic susceptibility [28, 33]. Therefore, an accurate description of such mixing is necessary for a fundamental understanding of magnetic behavior in Sm compounds.

In this work, we present a comprehensive investigation into the dynamic magnetic response of Sm7_{7}Pd3_{3} and establish that it is a novel magnetic system at the crossroads of unconventional spin-glass, domain-glass, and ferromagnetic behavior. We identify two distinct types of glassiness: one exhibiting typical spin-glass behaviour born from the interplay between dominant FM and residual AFM interactions and another showing pronounced field-induced spin relaxation reminiscent of phase-separated states observed in Nd7_{7}Rh3_{3}, originating from domain locking. The underlying mechanisms have been elucidated through detailed ac and dc susceptibility measurements under various protocols. Furthermore, we analyze the mixing of the ground-state J=5/2J = 5/2 and the excited J=7/2J = 7/2 multiplets of Sm3+^{3+} and their splitting under the influence of crystalline electric field using both magnetization and specific heat measurements.

Experimental

The polycrystalline Sm7_{7}Pd3_{3} sample was prepared by arc melting stoichiometric amounts of pure Sm (99.95%), provided by the Materials Preparation Center of Ames National Laboratory, and Pd (99.99%), supplied by Vale Acton Refinery, London, UK. A Zr getter was melted prior to the sample melting to absorb the residual oxygen in the chamber. The sample was flipped and remelted several times to ensure homogeneity. Two samples: one melted four times, and another melted six times, were prepared to ensure reproducibility. No difference in either XRD pattern or magnetization data was observed between the two samples.

The room-temperature x-ray diffraction (XRD) pattern was recorded using a Rigaku TTRAX powder diffractometer equipped with a rotating anode Mo Kα\alpha source [34]. The Rietveld refinement of the XRD data was performed using the FullProf software, which confirm an almost phase-pure (with a <1% unknown phase) Th7_{7}Fe3_{3}-type hexagonal crystal structure of the compound Figure 1(a, Figure 1b) of [35], consistent with the reported literature [23, 24]. Temperature and magnetic field (both dc and ac) dependent magnetic measurements were conducted on a randomly oriented polycrystalline sample using a superconducting quantum interference device (SQUID) from Quantum Design, USA (model MPMS XL-7). The same SQUID magnetometer was used to perform relaxation measurements (isothermal remanent and aging effects) under different protocols. The details of the various measurement protocols are provided along with their respective discussions. Temperature and magnetic field dependent specific heat measurements were performed using the conventional 2π2\pi relaxation technique in a physical property measurement system (PPMS) from Quantum Design, Inc.

Figure 1. (a–e) The temperature dependent magnetization in zero-field-cooled (ZFC) and field-cooled warming (FCW) modes at different applied magnetic fields. (f) The field dependence of the irreversibility temperature plotted as TirrT_{\mathrm{irr}} vs. H2/3H^{2/3}. The solid black line represents the linear fit of the data.

Results and Discussion

The temperature-dependent magnetization measurements have been performed under zero-field-cooled (ZFC) and field-cooled warming (FCW) conditions at different magnetic fields, as shown in Figure 1(a-Figure 1e). A clear PM to FM-like transition has been observed at TC=173±1T_{\mathrm{C}} = 173 \pm1 K, which is slightly higher than those reported in [23] (167 K) and [24] (169 K). More importantly, a large bifurcation between the ZFC and FCW curves is observed below TCT_{\mathrm{C}}, and the irreversibility temperature (TirrT_{\mathrm{irr}}) shifts to lower values with an increase in the magnetic field [see Figs. 1(a-e)]. This is commonly associated with either blocking of the spins due to magnetic frustration or random locking of domain orientations during zero-field cooling, where the higher applied magnetic fields align the moments even well within the blocked regime [37, 38, 39]. Interestingly, a decrease in magnetization has been also observed in the FCW curves [where the sample was cooled after aligning the moments in the PM (unblocked) state] below ∼130\sim130 K for all studied magnetic fields [see Figs. 1(a-e)], indicating the presence of weak but finite antiferromagnetic interactions in the sample. The presence of mixed FM and AFM interactions can give rise to magnetic frustration, which may contribute to the observed large bifurcation in the ZFC-FCW curves at low temperatures.

Another factor to consider is a strong magnetic anisotropy of the sample, predicted by earlier DFT calculations [24]. Therefore, we analyze the field dependence of TirrT_{\mathrm{irr}} using the following power-law behavior [40, 41, 42]:

Tirr(H)=Tirr(0)[1−AHp](1)T_{\mathrm{irr}}(H) = T_{\mathrm{irr}}(0)[1 - AH^p] \tag*{(1)}

where Tirr(0)T_{\mathrm{irr}}(0) is the irreversibility temperature in the zero-field regime, AA is the constant, and the power exponent pp represents the measure of the strength of anisotropy in the system with respect to the external magnetic field. In the strongly anisotropic regime, the value of pp is 2/3 in accordance with the de Almeida-Thouless (A-T) line for Ising spins [43]. On the other hand, weakly anisotropic Heisenberg spin systems are defined by the Gabay-Toulouse (G-T) line with p≈2p \approx2 [44]. We observed a significant deviation in TirrT_{\mathrm{irr}} from the G-T line, as shown in Figure 2(c) of [35]. However, good agreement with the A-T line is observed up to H=10H = 10 kOe, as evident from the linear relationship between TirrT_{\mathrm{irr}} and H2/3H^{2/3}, shown in Figure 1(f). Fitting of the Tirr(H)T_{\mathrm{irr}}(H) data using the A-T line, represented by the solid black line in Figure 1(f), gives Tirr(0)=180(6)T_{\mathrm{irr}}(0) = 180(6) K and A=0.0017(1)A = 0.0017(1) Oe−2/3^{-2/3}. The good agreement of the Tirr(H)T_{\mathrm{irr}}(H) data with the A-T line indicates the presence of strong anisotropy in the system.

Figure 2. The temperature dependent inverse magnetic susceptibility of Sm₇Pd₃ at 10 and 20 kOe fields in ZFC and FCW modes. Solid black curve represents the best fit of the 20 kOe ZFC data using the modified Curie-Weiss law.

In Figure 2, we show the temperature-dependent inverse magnetic susceptibility (χ−1=H/M\chi^{-1} = H/M) at H=10H = 10 and 20 kOe [Fig. 2(a) of [35]] in both ZFC and FCW modes. A field-independent strong non-linearity in the curves can be clearly observed in the paramagnetic state, indicating the mixing of the ground state J=5/2J = 5/2 multiplet of Sm3+^{3+} with the excited J=7/2J = 7/2 level, resulting in the non-zero second-order Zeeman splitting [29, 30, 31, 32, 33]. We fit the 20 kOe ZFC data using the modified Curie-Weiss law given as [33, 45]:

χ=χ0+CT−θP=(NAkB)(αJμB2+μeff23(T−θP))(2)\chi= \chi_0 + \frac{C}{T-\theta_P} = \left(\frac{N_A}{k_B}\right)\left(\alpha_J\mu_B^2+\frac{\mu_{\mathrm{eff}}^2}{3(T-\theta_P)}\right) \tag*{(2)}

where χ0\chi_0 is the temperature-independent Van Vleck susceptibility, CC and θP\theta_P are the Curie-Weiss constant and Curie temperature, respectively. NAN_A, kBk_B, μB\mu_B, and μeff\mu_{\mathrm{eff}} are Avogadro’s number, Boltzmann constant, Bohr magneton, and effective magnetic moment, respectively. The constant αJ\alpha_J is expressed as

αJ=kB{[J2−(L−S)2][(L+S+1)2−J2]6J(2J+1)(EJ−1−EJ)+[(J+1)2−(L−S)2][(L+S+1)2−(J+1)2]6(J+1)(2J+1)(EJ+1−EJ)}=207Δ(3)\begin{aligned} \alpha_J=k_B\left\{\frac{\left[J^2-(L-S)^2\right]\left[(L+S+1)^2-J^2\right]}{6J(2J+1)(E_{J-1}-E_J)}+\frac{\left[(J+1)^2-(L-S)^2\right]\left[(L+S+1)^2-(J+1)^2\right]}{6(J+1)(2J+1)(E_{J+1}-E_J)}\right\} \\ &=\frac{20}{7\Delta} \tag*{(3)} \end{aligned}

where Δ\Delta is the energy separation between J=5/2J=5/2 and 7/27/2 states. The best fit, shown by the solid black curve in Figure 2, yields χ0=1.11(1)×10−3\chi_0 = 1.11(1)\times10^{-3} emu/Sm mol Oe, θP=171.6(3)\theta_P = 171.6(3) K, and C=0.0718(5)C = 0.0718(5) emu K/Sm mol Oe, which gives μeff=0.76 μB\mu_{\mathrm{eff}} = 0.76\ \mu_B/Sm and Δ=965\Delta= 965 K. The calculated value of θP\theta_P is in good agreement with the TCT_C estimated from the dMM/dTT curves (173 K). The separation between J=5/2J = 5/2 and 7/27/2 states is smaller than the estimated value of ∼1500\sim1500 K for the free Sm3+^{3+} ion [46], but is in good agreement with that observed for several other Sm-based compounds [33, 45, 47], indicating the strong effect of the crystal field in Sm7_7Pd3_3. This is also evident from the observed lower value of the effective magnetic moment (0.76 μB0.76\ \mu_B/Sm) as compared to the free Sm3+^{3+} ion (0.84μB0.84\mu_B).

The reduction in the ZFC magnetization or the bifurcation in the ZFC-FCW curves at low temperatures can arise either due to the onset of a spin-glass (SG)

like disordered magnetic state [48, 49] or to the inherent magnetic hardness of the FM system [50]. To distinguish this, field-dependent magnetization (M-H) measurements were conducted at various temperatures, as shown in Figure 3(a-Figure 3o). The sample was heated to room temperature after recording each successive M-H curve to eliminate remanence. Notably, at 2 K, a giant coercivity (HCH_C) of around 55 kOe, accompanied by a step-like jump in magnetization, is observed, qualitatively replicating the results obtained in our previous study [24]. Note that the virgin magnetization curve at 2 K is not fully saturated even up to 70 kOe (see Figure 3(a)), and the value of HCH_C can be further increased by applying higher magnetic fields [24]. A small value of ∼0.15 μB\sim0.15\ \mu_B/Sm has been observed at 70 kOe (at 2 K), which is significantly lower than the theoretical saturation moment (gJJg_JJ) of 0.71 μB0.71\ \mu_B for the free Sm3+^{3+} ion, indicating the strong effect of the crystal field and/or canting of the moments. The magnetic state comprising large coercivity along with nearly zero saturation magnetization is highly desirable for efficient spintronics and memory device applications [25, 26], and, therefore, demands further investigation.

Figure 3. (a–o) The ZFC M-H curves of Sm₇Pd₃ between −70 and 70 kOe magnetic fields at different temperatures.

A negative magnetization is detected at the beginning of the virgin isotherms, possibly due to the presence of a nonzero remanent field in the SQUID coils during nominally “zero-field” cooling. Below TCT_{\mathrm{C}}, the coercivity, HCH_{\mathrm{C}}, as well as the sharpness of the MM-HH loop, decreases, whereas the remanent magnetization (MRM_{\mathrm{R}}) remains almost invariant with increasing temperature [see Figure 3(a-Figure 3m)]. In fact, the MRM_{\mathrm{R}} shows an enhancement above ∼50\sim50 K, and to quantitatively understand this, the temperature dependence of MRM_{\mathrm{R}} and HCH_{\mathrm{C}} have been plotted in Figure 4(a) and 4(b), respectively. The remanent magnetization remains nearly invariant up to around 50 K and then starts increasing with temperature [marked by the dashed line in Figure 4(a)], followed by a sharp reduction at TCT_{\mathrm{C}}. This behavior is analogous to the FCW M(T)M(T) curves discussed above [see Figure 1(a-Figure 1f) and Figure 3 of [35] for more clarity], confirming the presence of AFM coupling in the sample at low temperatures, which becomes less pronounced with an increase in temperature above 50 K. Furthermore, the coercivity decreases abruptly with increasing temperature, which for a mixed FM-AFM system, can be described by the following relation [51]:

Figure 4. The temperature dependent (a) remanent magnetization (MRM_{\mathrm{R}}) and (b) coercive field (HCH_{\mathrm{C}}) (on the semi logarithmic scale) of Sm7_7Pd3_3. The solid black line in (b) represents the best fit of the data using eq. 4. Vertical dashed line shows the deviation in the MRM_{\mathrm{R}} and HCH_{\mathrm{C}} above 50 K.

HC=HC(0)exp⁡(−T/T1),(4)H_{\mathrm{C}} = H_{\mathrm{C}}(0) \exp(-T/T_{1}), \tag*{(4)}

where HC(0)H_{\mathrm{C}}(0) is the coercivity at 0 K and T1T_{1} is a constant. Interestingly, we observe an increase in the coercivity from the exponential decay behaviour, defined by

(4), above around 50 K, as depicted by solid black line in Figure 4(b) for HC(0)=58(2)H_{\mathrm{C}}(0) = 58(2) kOe and T1=21(1)T_{1} = 21(1) K. This further confirms the enhancement in the fraction of FM interactions in the sample above this temperature. The temperature-induced change in the MRM_{\mathrm{R}} is further probed by the thermoremanent magnetization (TRM) measurements performed at different magnetic fields, as shown in Figure 5. To record the TRM data, we first cool the sample from 300 K (PM state) to 2 K in the presence of an applied magnetic field, then remove the field at 2 K and record the magnetization as a function of temperature in zero external field. A clear enhancement in magnetization is observed above ∼50\sim50 K in Figure 5 for all cooling fields. An increase in the sample’s magnetization, even in the absence of any external magnetic field, confirms the presence of AFM coupling in the sample, which increases for T≲50T \lesssim50 K and weakens with an increase in temperature. It is interesting that this process is continuous, without a well-defined spin-reorientation transition, and one may reasonably assume that the rapid but continuous change in c/ac/a ratio below the TCT_{\mathrm{C}} observed in [24] affects the relative strength, and, possibly, sign of exchange interactions between Sm atoms located on different planes, which are stacked along the cc-axis.

Figure 5. The thermoremanent magnetization (TRM) measurements performed on Sm7_7Pd3_3 at different cooling fields.

The presence of mixed FM-AFM interactions often results in magnetic frustration, which leads to SG-like behavior in the samples [52, 53, 54]. To probe this, we performed ac susceptibility measurements on Sm7_7Pd3_3 in the vicinity of the magnetic transition using ac frequencies from 0.1 to 997 Hz. The temperature dependent real (χ′\chi') and imaginary (χ′′\chi'') components of the ac susceptibility, measured at Hac=5H_{\mathrm{ac}} = 5 Oe, are presented in Figure 6(a) and (b), respectively. A clear peak in both χ′\chi'- and χ′′\chi''-TT curves around 173 K confirms the presence of spin blocking. A reduction in magnitude and a shift of the peak position toward higher temperatures have been observed with an increase in the frequency for both the real and imaginary susceptibilities, indicating the spin-glass behavior [55]. In order to quantitatively characterize this SG behavior, we calculate the shift in the peak position per decade change in the frequency, defined by the Mydosh parameter as ∂Tf=ΔTf/TfΔ(log⁡10ω)\partial T_f=\Delta T_f/T_f\Delta(\log_{10}\omega), where ΔTf\Delta T_f is the shift in freezing temperature (TfT_f) and ω=2πf\omega=2\pi f is the angular frequency [56]. It has been established that the value of ∂Tf\partial T_f lies between 0.005 – 0.01 for canonical spin glass systems, between ∼\sim0.03 – 0.08 for cluster glasses, and ∂Tf>0.2\partial T_f>0.2 for superparamagnetic systems [56, 57, 58, 59]. The ∂Tf→0\partial T_f\to0 for conventional long-range magnetic ordering (AFM/FM), where the moments align almost instantly in the direction of the applied ac magnetic field. The calculated value of ∂Tf\partial T_f for Sm7_7Pd3_3 is 0.0011, which puts it in the range between canonical spin glass and conventional long-range magnetic ordering. This means that even though magnetic glassiness is present in the sample, as evident from the frequency dispersion of χ′(T)\chi^\prime(T) and χ′′(T)\chi^{\prime\prime}(T) curves, long-range magnetic ordering coexists with the magnetic frustration.

Figure 6. The temperature dependent (a) real and (b) imaginary components of the ac susceptibility of Sm7_7Pd3_3 at different frequencies of 5 Oe ac field. (c) The frequency dependence of the freezing temperature, plotted as τ=1/f\tau=1/f vs. TfT_f. The solid black curve represents the best using the critical slowing down model.

To further understand magnetic frustration in Sm7_7Pd3_3, the frequency dependence (τ=1/f\tau=1/f) of the peak position of χ′′(T)\chi^{\prime\prime}(T) curves is fitted using the critical slowing down model, given as [56, 60]

τ=τ0(TfTSG−1)zν(5)\tau=\tau_0\left(\frac{T_f}{T_{\mathrm{SG}}}-1\right)^{z\nu} \tag*{(5)}

where τ0\tau_0 denotes the characteristic relaxation time associated with an individual spin flip, TSGT_{\mathrm{SG}} is the freezing temperature at f=0f=0 Hz, and zνz\nu is the dynamical critical exponent [56, 60]. The black solid curve in Figure 6(c) represents the best fit of the data using the critical slowing down model, which gives TSG=171.9(3)T_{\mathrm{SG}=171.9(3)} K, zν=7.3(2)z\nu=7.3(2), and τ0=1.42×10−17\tau_0=1.42\times10^{-17} sec. The value of τ0\tau_0 lies between 10−1210^{-12} – 10−1410^{-14} sec for non-interacting canonical spin glass systems and between 10−710^{-7} – 10−1010^{-10} sec for cluster-glass systems with a finite interaction between the spins [58, 59]. In the present case, the characteristic relaxation time, τ0\tau_0, is even smaller than that of canonical spin-glass systems, which confirms the presence of coexisting long-ranged FM (where τ0→0\tau_0\to0) and short-range SG-like interactions in the sample. Similar unconventional ferromagnetism produced by coexisting FM-AFM interactions has been observed in several other compounds [52, 53, 59, 61, 62, 63], where the strength of AFM is weaker compared to the FM coupling, yet sufficient to introduce frustration in the magnetic moments.

The frequency dispersion in the χ′\chi^\prime and χ′′\chi^{\prime\prime} is observed predominantly near the magnetic transition (see Figure 4 of [35]). Therefore, to further understand the spin dynamics and the distribution of the relaxation time of the glassy/frozen spins in the sample, the ac susceptibility is measured at selected temperatures with varying frequencies, as depicted in Figure 7(a–Figure 7e), in the form of a Cole-Cole plot (χ′′\chi^{\prime\prime} vs. χ′\chi^\prime). The Cole-Cole plot exhibits a perfect semicircle if all the spins have the same relaxation time, whereas a distortion in the semicircle indicates a distribution of the relaxation times of the spin freezing process [56]. We fit the χ′′(χ′)\chi^{\prime\prime}(\chi^\prime) data at different temperatures using the following relation [56, 64]:

Figure 7. (a–e) Cole-Cole plots (χ′′\chi^{\prime\prime} vs. χ′\chi^{\prime}) obtained from ac susceptibility data at the selected temperatures. The solid black curves represent the best fit using eq. 6 and the dashed curves are the extrapolation of the fitted curves. (f–h) The temperature dependence of the isothermal susceptibility (χ0\chi_{0}), adiabatic susceptibility (χs\chi_{s}), shape parameter α\alpha, respectively.

χ′′=χs−χ02tan⁡[(1−α)π2]+(χ′−χs)(χ0−χ′)+(χ0−χs)24tan⁡2[(1−α)π2](6)\chi^{\prime\prime}=\frac{\chi_s-\chi_0}{2\tan\left[(1-\alpha)\frac{\pi}{2}\right]}+ \sqrt{(\chi^\prime-\chi_s)(\chi_0-\chi^\prime)+\frac{(\chi_0-\chi_s)^2}{4\tan^2\left[(1-\alpha)\frac{\pi}{2}\right]}} \tag*{(6)}

where χ0\chi_0 and χs\chi_s are the isothermal and adiabatic susceptibilities, which represent the value of χ′\chi^\prime in the low and high-frequency limits, respectively. The parameter α\alpha is a measure of the distribution of the relaxation time, where α=0\alpha=0 represents a single relaxation time (perfect semicircle) and α=1\alpha=1 indicates a wide distribution of relaxation times (flattened semicircle). The best fit of the χ′′(χ′)\chi^{\prime\prime}(\chi^\prime) data using the above equation for selected representative temperatures is shown by the solid black curves in Figure 7(a-Figure 7e), where the dashed curves represent the extrapolation of the fit. The temperature dependences of χ0\chi_{0}, χs\chi_{s}, and (1−α)π/2(1-\alpha)\pi/2 are shown in Figure 7(f-Figure 7h), respectively. The value of χ0\chi_{0}, representing the static susceptibility, decreases with temperature [see Figure 7(f)] across the transition, analogous to the dc magnetic susceptibility, discussed above. In contrast, χs\chi_{s} increases with temperature due to the release of frozen moments, enabling them to align in the direction of the external field, even at higher frequencies. Further, the value of (1−α)π/2(1-\alpha)\pi/2 lies around 4040–45∘45^{\circ} below ∼173\sim173 K and then approaches 90∘90^{\circ} (α→0\alpha\to0) with further increase in temperature. A significant nonzero value of α\alpha in the frozen state shows the wide distribution of the relaxation time for a variety of frozen magnetic moments in Sm7_{7}Pd3_{3}. Such wide distribution reflects a potentially complex frozen magnetic structure that comes from three inequivalent Sm sites arranged in the non-centrosymmetric crystal structure. A rapid change in α\alpha around 173 K [see Figure 7(h)] indicates the onset of the PM state.

To further understand the nature of the frustration in the sample, we performed isothermal remanent magnetization (IRM) measurements at different temperatures.

To record the IRM data, we first cool the sample from 300 K (PM state) to a desired temperature in the glassy region in the presence of an applied magnetic field, wait for a certain time (twt_{w}), and then remove the magnetic field and record the magnetization as a function of time. Figure 8(a-Figure 8e) show the IRM data recorded at some representative temperatures across the magnetic transition with a cooling field of 100 Oe and tw=100t_{w}=100 sec. A clear magnetization decay is observed, in many cases even up to one hour, indicating the presence of the arrested/frustrated spin dynamics of the sample [9, 58, 59]. Note that the relaxation of the magnetic moments is observed prominently near the magnetic transition, and this effect weakens towards both the lower and higher temperatures, i.e., in the deep FM and PM states, respectively [see Figure 8(a-Figure 8e)]. This is expected for a FM state with a finite remanence (no relaxation in the ordered domain state) and a PM state having free moments (quick relaxation) [see Figure 5(a, Figure 5b) of [35] for IRM data at 5 and 300 K]. The time evolution of the magnetization for the coexisting spin glass and FM interactions can be expressed well using the stretched exponent model, given as [59, 65]:

Figure 8. (a–e) The field-cooled isothermal remanent magnetization (IRM) data at different temperatures for 100 Oe field and 100 sec waiting time. The solid black curves represent the best fit using stretched exponent model. The temperature dependence of the (f) FM moment, (g) spin-glass moment, (h) relaxation time, and (i) power exponent, η\eta.

M(t)=M0±MSG×exp⁡[(−ttr)1−η],(7)M(t) = M_0 \pm M_{\mathrm{SG}} \times\exp\left[\left(\frac{-t}{t_r}\right)^{1-\eta}\right], \tag*{(7)}

where M0M_0 and MSGM_{\mathrm{SG}} represent the ferromagnetic and spin-glass components of the magnetic moment, respectively, and trt_r and η\eta denote the mean relaxation time and relaxation rate. The minus and plus signs correspond to the model for the IRM and the aging effect (discussed below), respectively. The best fit to the IRM data using the stretched exponent model is shown by the solid black curves in Figure 8(a–Figure 8e), with the corresponding fitting parameters presented in Figure 8(f–Figure 8i) as a function of temperature.

The temperature dependence of M0M_0 exhibits a typical M–T behavior for a FM system near the FM–PM transition [see Fig. 8(f)], confirming the presence of the finite FM component in the sample. However, the transition temperature derived from the dM0dT\frac{dM_0}{dT} curve is around 165 K, which is lower than the value obtained from the dc and ac magnetic susceptibility measurements (∼173\sim173 K) presented above. Note that the time evolution of IRM in this temperature range combines the effects of both the frustrated behavior of the magnetic lattice and the relaxation of field-induced metastable states, commonly referred to as a “magnetic glass” state [13, 19, 21, 66], which will be discussed in more detail below. Therefore, we propose that this shift is associated with the field-induced decrease in the magnetic glass transition temperature of the sample. Interestingly, the glassy magnetization, MSGM_{\mathrm{SG}}, initially increases with temperature, exhibiting a pronounced maximum at ∼165\sim165 K, then decreases upon further heating, as shown in Figure 8(g) [see Fig. 6(a, b) of [35] for a direct comparison of $M_{\mathrm{IRM}}(T)$ data]. The presence of both M0M_0 and MSGM_{\mathrm{SG}} components further confirms coexistence of FM and glassy states in the sample. The relaxation time tr(T)t_r(T), plotted in Figure 8(h), decreases exponentially (note the logarithmic scale on the y-axis) with increasing temperature. A notable hump in the tr(T)t_r(T) curve is observed around 173 K [see Fig. 8(h)], where the magnetic glass state is already diminished, indicating the dominance of spin-glass behavior in this temperature regime. The exponent η\eta remains close to ∼0.5\sim0.5 [Fig. 8(i)], suggesting a similar relaxation mechanism in the sample across this temperature range.

In Figure 9(a, Figure 9b), we present the IRM and aging effect data recorded at 168 K (just below TCT_{\mathrm{C}}) for different waiting times under low magnetic fields. To perform the aging effect measurements, the sample was cooled from 300 K to 168 K in zero magnetic field, then kept at 168 K in zero field for a certain time (twt_w), and then the magnetic field was applied, and the time evolution of the sample’s magnetization was recorded. Each time, the sample was heated to room temperature to erase the prior magnetic history. In IRM, increasing waiting time in the set field enhances magnetization [see Fig. 9(a)], underscoring relatively slow alignment of the spins with time (twt_w) and confirming the spin-glass behavior of the

Figure 9. (a) The FC IRM and (b) ZFC aging effect data at 168 K using 50 Oe and 10 Oe magnetic fields, respectively, for the different waiting times.

sample [56, 58]. On the other hand, the magnetization decreases with an increase in twt_w (in zero field) during the aging effect measurements [see Fig. 9(b)], indicating enhanced spin stiffness, i.e., relaxation into deeper energy states, consistent with thermally fluctuating glassy spin dynamics [56, 67, 68]. This behavior is further supported by the shift in the relaxation rate, S=1HdM(t)dlog⁡(t)S = \frac{1}{H}\frac{dM(t)}{d\log(t)} with the waiting time, presented in Figure 7 of [35]. The solid curves in Figure 9(a, Figure 9b) represent the best fits obtained using the stretched exponent model, and the fitting parameters are summarized in Table 1 of [35].

We now discuss the ZFC relaxation (aging effect) measurements performed under higher magnetic fields to elucidate the field-induced changes in the spin dynamics of the sample. In Figure 10(a–Figure 10f), we present the ZFC relaxation data measured at a 10 kOe field for tw=600t_w = 600 sec at selected temperatures between 5 and 30 K. A clear logarithmic relaxation of magnetization is observed for up to one hour, even in the presence of a significantly higher field. Note that the sample exhibits negative magnetization at low temperatures [see Figs. 10(a, b)], likely due to a nonzero negative remanent field in the SQUID coils, which aligns the magnetic moments upon cooling through TCT_{\mathrm{C}} and subsequently freezes at T≪TirrT \ll T_{\mathrm{irr}}. However, this negative moment is not expected to affect the behavior of MSGM_{\mathrm{SG}} in the sample, which is governed by the time evolution of the magnetization. Analogous to the IRM measurements performed at 100 Oe, we observe that the relaxation behavior weakens as we move away from a characteristic temperature (say TmT_m) [see Figure 8(a, Figure 8b) of [35]]. The temperature dependence of the best-fit parameters obtained by fitting the M(t)M(t) data using Eq. (7) is shown in Figure 10(g–Figure 10j). The M0(T)M_0(T) curve in Figure 10(g) exhibits behavior similar to the ZFC curve presented in Figure 1(e). A prominent peak in the MSG(T)M_{\mathrm{SG}}(T) curve is observed at Tm∼17T_m \sim17 K [see Figure 10(h)]. More strikingly, we observe a non-logarithmic change in the characteristic relaxation time, trt_r [see Figure 10(i)], in contrast to the behavior seen in the low-field IRM data [see Figure 8(h)]. This difference indicates that the origin of the field-induced relaxation process is distinct from that of the conventional spin-glass state discussed above.

Figure 10. (a–f) The ZFC aging effect measurements performed at different temperatures using 10 kOe magnetic field and 600 sec waiting time. The solid black curves represent the best fit using stretched exponent model. The temperature dependence of the (g) ferromagnetic moment, (h) spin-glass moment, (i) relaxation time, and (j) power exponent, η\eta. The black solid curve in (h) represents the Gaussian fit of the data.

The presence of finite M0M_0 and MSGM_{\mathrm{SG}} in the low-temperature regime for H=10H = 10 kOe shows that the coexisting region of the glassy and FM components in the sample can be tuned by both the temperature and the external magnetic field. To further clarify this, in Figure 11, we show the spin-glass moment, MSGM_{\mathrm{SG}}, as a function of temperature for different magnetic fields, extracted from the ZFC relaxation data [see Figure 9(a–Figure 9i) and Figure 10(a–Figure 10i) of [35] for details]. It can be observed that the temperature regime spanning the most prominent relaxation behavior (TmT_m) shifts to lower values with an increase in the magnetic field. The peak position and full width at half maximum (FWHM) of all the curves, extracted by Gaussian fitting, are shown in the inset of Figure 11. It is interesting to note that the glassy region not only shifts to lower temperatures but also becomes more localized with an increase in the external magnetic field. A phenomenologically similar temperature and magnetic field dependence of the spin relaxation behavior has been recently reported in a magnetic glass compound Pr0.5_{0.5}Ca0.5_{0.5}Mn0.975_{0.975}Al0.025_{0.025}O3_3, where the relaxation rate first increases up to Tm=25T_m = 25 K and then decreases with a further increase in temperature for H=10H = 10 kOe [19]. The peak in the relaxation rate also shifts to lower temperatures with an increase in the magnetic field, and this behavior is attributed to the devitrification of the kinetically arrested high-temperature AFM state into the FM configuration [19]. However, the observation of a similar relaxation behavior without any first-order AFM ↔\leftrightarrow FM transition in the present case indicates a different origin for this glassiness in Sm7_7Pd3_3. In the same line, Gd5_5Ge4_4 exhibits a pronounced relaxation effect between 10–27.5 kOe, i.e., only in the rising region of the virgin magnetization curve, at 5 K due to the field-induced AFM–FM transformation [9]. However, no spin relaxation is observed in the compound either below (in the AFM state) or well above (deep in the FM state) a critical field, HcriH_{\mathrm{cri}} [9]. A partial kinetic arrest of the high-field FM state gives rise to strong spin relaxation in Nd7_7Rh3_3 below 10 K [13], further suggesting a similar relaxation behavior, albeit with a different origin, in Sm7_7Pd3_3. It is important to note that the field-induced giant relaxation observed in all these compounds originates from the metastable phase-separated AFM ↔\leftrightarrow FM regions formed due to the partially arrested first-order magneto-structural transition, and is distinct from the conventional spin-glass behavior that arises from magnetic frustration due to competing FM-AFM interactions, and termed as “magnetic glass” state [9, 13, 19, 20, 21, 66, 69].

Figure 11. The temperature dependence of the spin glass moment (MSGM_{\mathrm{SG}}) at different fields. The solid black curves represent the Gaussian fit. The inset shows the field dependence of the peak position (on the left axis) and the FWHM (on the right axis) of the MSG(T)M_{\mathrm{SG}}(T) curves.

Figure 12. The temperature dependent (a) real and (b) imaginary component of the ac susceptibility at different dc fields using a 6 Oe ac field of 10 Hz frequency. Inset of (a) shows the field dependence of the peak position of χ′(T)\chi^{\prime}(T) curves.

In Sm7_7Pd3_3, a change in the behavior of the characteristic relaxation time, trt_r, in the case of the low- and high-field relaxation measurements discussed above suggests a distinct origin for these two glassy effects. One can further distinguish them by performing the ac susceptibility measurements in the presence of dc magnetic fields, as shown in Figs. 12(a, b). If both the low-field (high-temperature) and high-field (low-temperature) glassy behaviors had the same origin, one would expect a shift in the ac susceptibility peak to lower temperatures with the applied field, analogous to the MSG(T,H)M_{\mathrm{SG}}(T, H) curves. However, contrary to this, peaks in both χ′\chi^{\prime}- and χ′′\chi^{\prime\prime}-TT curves are suppressed and shift to higher temperatures with an increase in the dc field [plotted in the inset of Fig. 12(a)]. This confirms that the origin of the glassy effects attributed to the peak of χ′/χ′′\chi^{\prime}/\chi^{\prime\prime}-TT curves in Figure 6(a, Figure 6b) and MSG(T)M_{\mathrm{SG}}(T) curves in Figure 11 is different. The former arises from the unperturbed competing FM-AFM states in the sample, whereas the latter originates from the relaxation of field-induced metastable states in the sample and more closely resembles the magnetic glass state [9, 13, 19, 20, 21, 66, 69].

Figure 12 (PDF p. 10).

However, it may be more precise to refer to it as a “domain glass” state in the present case to distinguish it from the phase-separated state reported earlier.

These experiments provide clear evidence of the unconventional field-induced spin relaxation in Sm7_7Pd3_3, even in the absence of a first-order AFM ↔\leftrightarrow FM transition. The observed relaxation in the present case fundamentally originates from the intrinsic magnetic hardness (strong magnetocrystalline anisotropy and hence large HCH_{\mathrm{C}}) of the sample. A sufficiently strong magnetic field can align the blocked magnetic moments even within a deeply frozen state. Above a critical magnetic field, HcriH_{\mathrm{cri}}, the magnetic moments or domain walls begin to orient along the field direction. However, within the polycrystalline material with intrinsically high magnetocrystalline anisotropy, there exists a distribution of domain wall energies required to overcome the pinning effect, causing a nonergodic metastable state to exist in the vicinity of (Tirr,Hcri)(T_{\mathrm{irr}}, H_{\mathrm{cri}}), thereby leading to slow magnetization relaxation. As the magnetic field is further increased (decreased), the relaxation weakens due to an increase in the number of aligned (frozen) moments. Consequently, only a specific combination of temperature and magnetic field induces these metastable states, which leads to the observed “domain glass”-like behavior in Sm7_7Pd3_3, and can be attributed to the magnetic viscosity of the sample [70, 71].

In Figure 13(a), we show the temperature-dependent zero-field specific heat (CpC_{\mathrm{p}}) data of Sm7_7Pd3_3. A sharp λ\lambda-like peak at TCT_{\mathrm{C}} is clearly observed, manifesting the presence of long-range magnetic ordering in the sample. Note that the value of specific heat at 250 K is around 280 J/mol-K, which is higher than the classical Dulong-Petit limit of 3nR=249.423nR = 249.42 J/mol-K [as shown by the horizontal dashed line in Figure 13(a)], where nn is the number of atoms per formula unit (ten in the present case), and RR is the molar gas constant. This indicates the presence of an additional contribution other than lattice in this temperature range. To estimate the electronic contribution, the low-temperature (2 – 8 K) data were fitted (not shown) using the equation Cp=γT+βT3C_{\mathrm{p}} = \gamma T + \beta T^3, where the first and second terms represent the electronic and lattice contributions, respectively, and give γ=91(2)\gamma= 91(2) mJ/mol-K2^2 and β=6.47(5)\beta= 6.47(5) mJ/mol-K4^4. The Debye temperature, θD\theta_{\mathrm{D}}, can be estimated from the extracted value of β\beta using the following relation [72]:

Figure 13. (a) The zero-field temperature dependent specific heat of Sm₇Pd₃. The dashed line shows the classical Dulong-Petit limit of CPC_{\mathrm{P}}. Black solid curve represents the sum of electronic and the lattice contribution (using Debye model), given by eq. (9). (b) The temperature dependent magnetic specific heat, CmagC_{\mathrm{mag}}, where the arrow represents the Schottky contribution and dashed red curve is the best fit using eq. (10). (c) Cmag/TC_{\mathrm{mag}}/T vs. TT (on left axis) and magnetic entropy (SmagS_{\mathrm{mag}}) vs. TT plot (on right axis). The horizontal dashed line represent the theoretical value of SmagS_{\mathrm{mag}} for J=5/2J = 5/2.

θD=(12π4nR5β)1/3(8)\theta_{\mathrm{D}} = \left(\frac{12\pi^{4}nR}{5\beta}\right)^{1/3} \tag*{(8)}

We obtained a θD=145(5)\theta_{\mathrm{D}} = 145(5) K and used these values of γ\gamma and θD\theta_{\mathrm{D}} to calculate the electronic and lattice specific heat in the entire temperature range using the Debye model [72]:

CP=γT+9nR(TθD)3∫0θD/Tx4ex(ex−1)2dx,(9)C_{\mathrm{P}} = \gamma T + 9nR\left(\frac{T}{\theta_{\mathrm{D}}}\right)^3 \int_{0}^{\theta_{\mathrm{D}}/T} \frac{x^4 e^x}{(e^x - 1)^2}dx, \tag*{(9)}

shown by the solid black curve in Figure 13(a). To estimate the magnetic contribution to the specific heat (CmagC_{\mathrm{mag}}), we subtracted the electronic and lattice specific heat data from the CPC_{\mathrm{P}} curve, and the resulting Cmag(T)C_{\mathrm{mag}}(T) curve is shown in Figure 13(b). Interestingly, apart from the λ\lambda peak at TCT_{\mathrm{C}}, we observe a broad Schottky-like anomaly in the 50 – 120 K temperature range. The crystal field splitting of the six-fold degenerate J=5/2J = 5/2 ground state of Sm3+^{3+} is expected to be responsible for this anomaly [32, 73]. A similar Schottky anomaly is also observed in the specific heat data of Sm-based compounds, Sm2_2Ni2_2Sn [73] and SmPd2_2Al3_3 [32], above their magnetic transition temperatures. We fit this broad peak in the CmagC_{\mathrm{mag}} using the following expression [74]

CS=RT2[∑igiEi2e−Ei/T∑igie−Ei/T−(∑igiEie−Ei/T∑igie−Ei/T)2],(10)C_{\mathrm{S}} = \frac{R}{T^2}\left[\frac{\sum_i g_iE_i^2e^{-E_i/T}}{\sum_i g_ie^{-E_i/T}}-\left(\frac{\sum_i g_iE_ie^{-E_i/T}}{\sum_i g_ie^{-E_i/T}}\right)^2\right], \tag*{(10)}

where EiE_i and gig_i are the energy and degeneracy of the iith level. We introduce a three-level splitting scheme with g1=g2=g3=2g_1 = g_2 = g_3 = 2 to fit the Schottky contribution; however, the fitting of the CmagC_{\mathrm{mag}} data, shown by the dashed red curve in Figure 13(b), reveals the presence of a four-fold degenerate state (E2=E3E_2 = E_3) lying 255 K above the ground state (E1E_1) doublet, where we fix E1E_1 at 0 K. These low-lying CF-split excited multiplets of Sm3+^{3+} in Sm7_7Pd3_3 are expected to give rise to the temperature-dependent changes in the density of states at the Fermi level, affecting the observed complex magnetic interactions in the sample. Further, the magnetic entropy, SmagS_{\mathrm{mag}}, is calculated using the relation Smag=∫CmagTdTS_{\mathrm{mag}} = \int\frac{C_{\mathrm{mag}}}{T}dT. The Cmag/TC_{\mathrm{mag}}/T and SmagS_{\mathrm{mag}} are shown on the left and right axes of Figure 13(c), respectively. The Schottky anomaly is more clearly visible in the Cmag/TC_{\mathrm{mag}}/T, as shown by the downward arrow. The value of SmagS_{\mathrm{mag}} at 250 K is 11.7 J/Sm-mol-K, which is nearly 78% of the theoretical value of Rln⁡(2J+1)R\ln(2J + 1) for J=5/2J = 5/2, as shown by the dashed horizontal line in Figure 13(c). This can be attributed to the underlying crystal field splitting of the J=5/2J = 5/2 levels and has been observed in several other Sm-based compounds [33]. However, a more reliable estimation of the lattice specific heat using a non-magnetic reference sample is necessary to precisely calculate the CmagC_{\mathrm{mag}} and hence SmagS_{\mathrm{mag}}. No notable change has been observed in the CPC_{\mathrm{P}} curve recorded at H=10H = 10 kOe magnetic field (see Fig. 12 of [35]) underlying the robustness of this second-order magnetoelastic transition.

Conclusion

The detailed ac and dc magnetization measurements reveal the coexistence of spin glass, domain glass, and ferromagnetism in the non-centrosymmetric compound Sm7_7Pd3_3 below its TCT_{\mathrm{C}} of 173 K. The dc susceptibility measurements performed in both zero-field-cooled and field-cooled modes exhibit strong irreversibility, which follows the de Almeida-Thouless line, confirming strong magnetic anisotropy in the sample. The temperature-dependent magnetization measurements suggest the presence of finite antiferromagnetic (AFM) coupling in the sample, particularly below ∼\sim50 K. However, the field-dependent magnetization data reveal a large coercivity, indicative of long-range ferromagnetic (FM) interactions, along with nearly zero magnetization. This mixed FM-AFM behavior gives rise to spin glass-like dynamics, which is confirmed through detailed ac susceptibility, isothermal remanent magnetization, and aging effect measurements. The analysis of the Cole-Cole plots reveals the distribution of the relaxation time of the frozen spins in the sample. At the same time, a strong field-induced spin relaxation is also observed, which shifts to lower temperatures with an increase in the magnetic field. This phenomenon suggests the presence of domain glass-like behavior in the sample, which is distinct from conventional spin glass systems, and we demonstrate that the intrinsic blocking effect in a hard FM system due to strong magnetocrystalline anisotropy is sufficient to induce this. The analysis of the paramagnetic susceptibility of the sample demonstrates the mixing of ground J=5/2J = 5/2 state of Sm3+^{3+} with the first excited J=7/2J = 7/2 multiplets, lying ∼\sim 965 K above. Moreover, the specific heat measurements reveal a λ\lambda-like anomaly at TCT_C associated with long-range magnetic ordering and a pronounced Schottky anomaly between 50 and 120 K due to the splitting of the sixfold degenerate J=5/2J = 5/2 ground state into a ground state doublet, and an excited fourfold-degenerate state situated ∼\sim255 K above the doublet, indicating the strong crystal field effect in the sample. Overall, Sm7_7Pd3_3 presents an example of the complex magnetic system, where the fundamental description of its magnetism cannot be reduced to classic textbook curriculum. There are still a number of unresolved problems, such as a potential effect of hydrostatic pressure on the devitrification of the magnetic system and its domain alignment; ideally, a single crystal would be highly desired to study magnetic anisotropy. Thus, we encourage the reader to keep an eye out for future publications on the Sm7_7Pd3_3 compound.

Acknowledgments

This work was performed at Ames National Laboratory and was supported by the Division of Materials Science and Engineering of the Office of Basic Energy Sciences, Office of Science of the U.S. Department of Energy (DOE). Ames National Laboratory is operated for the U.S. DOE by Iowa State University of Science and Technology under Contract No. DE-AC02-07CH11358.

References

  1. [1]S.-W. Cheong and F.-T. Huang, Altermagnetism with non-collinear spins, Npj Quantum Mater. 9, 1 (2024).
  2. [2]S. Khmelevskyi and L. V. Pourovskii, Non-collinear magnetism driven by a hidden multipolar order in PrO₂, Commun Phys 7, 1 (2024).DOI
  3. [3]G. Bastien, M. Roslova, M. H. Haghighi, K. Mehlawat, J. Hunger, A. Isaeva, T. Doert, M. Vojta, B. Büchner, and A. U. B. Wolter, Spin-glass state and reversed magnetic anisotropy induced by Cr doping in the Kitaev magnet α–RuCl₃, Phys. Rev. B 99, 214410 (2019).DOI
  4. [4]I. Mazin and The PRX Editors, Editorial: Altermagnetism—A New Punch Line of Fundamental Magnetism, Phys. Rev. X 12, 040002 (2022).DOI
  5. [5]J. Krempaský, L. Šmejkal, S. W. D’Souza, M. Hajlaoui, G. Springholz, K. Uhlířová, F. Alarab, P. C. Constantínou, V. Strocov, D. Ušanov, W. R. Pudelko, R. González-Hernández, A. B. Heller, Z. Jansa, H. Reichlová, Z. Šobáň, R. D. G. Betancourt, P. Wadley, J. Sinova, D. Kriegner, J. Minář, J. H. Dil, and T. Jungwirth, Altermagnetic lifting of Kramers spin degeneracy, Nature 626, 517 (2024).arxiv.org/abs/2308.10681
  6. [6]L. Šmejkal, J. Sinova, and T. Jungwirth, Emerging Research Landscape of Altermagnetism, Phys. Rev. X 12, 040501 (2022).arxiv.org/abs/2204.10844
  7. [7]F. Ma, W. Ji, J. Hu, Z.-Y. Lu, and T. Xiang, First-Principles Calculations of the Electronic Structure of Tetragonal α-FeTe and α-FeSe Crystals: Evidence for a Bicolliear Antiferromagnetic Order, Phys. Rev. Lett. 102, 177003 (2009).DOI
  8. [8]D. L. Quintero-Castro, B. Lake, M. Reehuīs, A. Niazi, H. Ryll, A. T. M. N. Islam, T. Fennell, S. A. J. Kimber, B. Klemke, J. Ollivier, V. Garcia Sakai, P. P. Deen, and H. Mutka, Coexistence of long- and short-range magnetic order in the frustrated magnet SrYb₂O₄, Phys. Rev. B 86, 064203 (2012).
  9. [9]M. K. Chattopadhyay, M. A. Manekar, A. O. Pecharsky, V. K. Pecharsky, K. A. Gschneidner, Jr., J. Moore, G. K. Perkins, Y. V. Bugoslavsky, S. B. Roy, P. Chaddah, and L. F. Cohen, Metastable magnetic response across the antiferromagnetic to ferromagnetic transition in Gd₅Ge₄, Phys. Rev. B 70, 214421 (2004).DOI
  10. [10]J. Meseguer-Sánchez, C. Popescu, J. L. García-Muñoz, H. Luetkens, G. Taniashvili, E. Navarro-Moratalla, Z. Guguchia, and E. J. G. Santos, Coexistence of structural and magnetic phases in van der Waals magnet CrI₃, Nat Commun 12, 6265 (2021).DOI
  11. [11]J. A. T. Barker, D. Singh, A. Thamizhavel, A. D. Hillier, M. R. Lees, G. Balakrishnan, D. McK. Paul, and R. P. Singh, Unconventional Superconductivity in La₇Ir₃ Revealed by Muon Spin Relaxation: Introducing a New Family of Noncentrosymmetric Superconductor That Breaks Time-Reversal Symmetry, Phys. Rev. Lett. 115, 267001 (2015).DOI
  12. [12]D. Singh, M. S. Scheurer, A. D. Hillier, D. T. Adroja, and R. P. Singh, Time-reversal-symmetry breaking and unconventional pairing in the noncentrosymmetric superconductor Nd₇Rh₃, Phys. Rev. B 102, 134511 (2020).DOI
  13. [13]K. Sengupta and E. V. Sampathkumaran, Field-induced first-order magnetic phase transition in an intermetallic compound Nd₇Rh₃: Evidence for kinetic hindrance, phase coexistence, and percolative electrical conduction, Phys. Rev. B 73, 020406 (2006).DOI
  14. [14]Arushi, D. Singh, A. D. Hillier, M. S. Scheurer, and R. P. Singh, Time-reversal symmetry breaking and multigap superconductivity in the noncentrosymmetric superconductor La₇Ni₃, Phys. Rev. B 103, 174502 (2021).DOI
  15. [15]J. M. Moreau and E. Parthé, Ferromagnetic Gd₇Pd₃ and other rare-earth-palladium compounds with non-centrosymmetric Th₇Fe₃ structure, Journal of the Less Common Metals 32, 91 (1973).DOI
  16. [16]A. F. Gubkin, A. A. Vaulin, T. Tsutaoka, A. F. Prekul, K. P. Skokov, and N. V. Baranov, Magnetic incommensurability, short-range correlations, and properties of Ho₇Rh₃, Phys. Rev. B 106, 134419 (2022).DOI
  17. [17]Y. Mudryk, C. Ritter, D. Paudyal, A. Provino, S. K. Dhar, P. Manfrinetti, F. Fauth, and V. K. Pecharsky, Magnetostructural behavior in the non-centrosymmetric compound Nd₇Pd₃, J. Phys.: Condens. Matter 31, 265801 (2019).DOI
  18. [18]K. Umeo, Y. Echizen, M. H. Jung, T. Takabatake, T. Sakakibara, T. Terashima, C. Terakura, C. Pfleiderer, M. Uhlarz, and H. V. Löhneysen, Field-induced magnetic transition in the heavy-fermion antiferromagnet Ce₇Ni₃, Phys. Rev. B 67, 144408 (2003).
  19. [19]S. Pal, K. Kumar, and A. Banerjee, Memorylike response of the magnetic glass, Phys. Rev. B 103, 144434 (2021).DOI
  20. [20]J. Dho and N. H. Hur, Thermal relaxation of field-induced irreversible ferromagnetic phase in Pr-doped manganites, Phys. Rev. B 67, 214414 (2003).DOI
  21. [21]S. Pal, K. Kumar, and A. Banerjee, Phase coexistence and nonequilibrium dynamics under simultaneously applied magnetic field and pressure: Possible role of the interface, Phys. Rev. B 108, 214409 (2023).
  22. [22]N. K. Singh, P. Kumar, Z. Mao, D. Paudyal, V. Neu, K. G. Suresh, V. K. Pecharsky, and K. A. Gschneidner Jr, Magnetic, magnetocaloric and magnetoresistance properties of Nd₇Pd₃, J. Phys.: Condens. Matter 21, 456004 (2009).
  23. [23]H. Kadomatsu, K. Kuwano, K. Umeo, Y. Itoh, and T. Tokunaga, Magnetic and electrical properties of R₇Pd₃ (R=La, Ce, Pr, Nd and Sm), Journal of Magnetism and Magnetic Materials 189, 335 (1998).
  24. [24]A. Biswas, R. K. Chouhan, O. Dolotko, P. Manfrinetti, S. Lapidus, D. L. Schlager, and Y. Mudryk, Exceptional magnetic and magnetoelastic behavior of rare-earth non-centrosymmetric Sm₇Pd₃, Acta Materialia 265, 119630 (2024).
  25. [25]T. Jungwirth, X. Marti, P. Wadley, and J. Wunderlich, Antiferromagnetic spintronics, Nature Nanotech 11, 231 (2016).DOI
  26. [26]Y. W. Windsor, S.-E. Lee, D. Zahn, V. Borisov, D. Thonig, K. Kliemt, A. Ernst, C. Schüßler-Langeheine, N. Pontius, U. Staub, C. Krellner, D. V. Vyalikh, O. Eriksson, and L. Rettig, Exchange scaling of ultrafast angular momentum transfer in 4f antiferromagnets, Nat. Mater. 21, 514 (2022).
  27. [27]H. W. De Wijn, A. M. Van Diepen, and K. H. J. Buschow, Samarium in Crystal Fields, a Case of Strong J -Mixing, Physica Status Solidi (b) 76, 11 (1976).
  28. [28]Z. Liu, Susceptibilities of SmPd₂Al₃ and SmPd₂Ga₃ studied with a crystal-field model and an ab initio approach, Phys. Rev. B 64, 144407 (2001).
  29. [29]S. K. Malik, W. E. Wallace, and R. Vijayaraghavan, Crystal-field effects on the crossover temperature of Sm³⁺ magnetization in magnetically ordered samarium compounds, Phys. Rev. B 19, 1671 (1979).
  30. [30]A. Barla, J. P. Sanchez, F. Givord, J. X. Bouchèrle, B. P. Doyle, and R. Rüffer, Crystal fields, exchange, and conduction electron polarization in SmAl₂, Phys. Rev. B 71, 012407 (2005).
  31. [31]H. W. De Wijn, A. M. Van Diepen, and K. H. J. Buschow, Nuclear Magnetic Resonance and Magnetic Susceptibility of SmAl₃, Phys. Rev. 161, 253 (1967).
  32. [32]J. Pospíšil, M. Kratochvílová, J. Prokleška, M. Diviš, and V. Sechovský, Samarium magnetism studied on SmPd₂Al₃ single crystal, Phys. Rev. B 81, 024413 (2010).
  33. [33]K. Ahn, V. K. Pecharsky, and K. A. Gschneidner, Phase relationships, and the structural, magnetic, and thermodynamic properties in the Sm₅SiₓGe₄₋ₓ pseudobinary system, Phys. Rev. B 76, 014415 (2007).
  34. [34]A. P. Holm, V. K. Pecharsky, K. A. Gschneidner Jr., R. Rink, and M. N. Jirmanus, X-ray powder diffractometer for in situ structural studies in magnetic fields from 0 to 35 kOe between 2.2 and 315 K, Rev. Sci. Instrum. 75, 1081 (2004).
  35. [35]See the supplementary information for the details of the crystal structure, magnetic relaxation, ac and dc susceptibilities, and Cₚ(T, H) data, which also includes Ref. [36].
  36. [37]D. S. Williams, P. M. Shand, T. M. Pekarek, R. Skomski, V. Petkov, and D. L. Leslie-Pelecky, Magnetic transitions in disordered GdAl₂, Phys. Rev. B 68, 214404 (2003).
  37. [38]N. Marcano, J. C. Gómez Sal, J. I. Espeso, L. Fernández Barquín, and C. Paulsen, Cluster-glass percolative scenario in CeNi₁₋ₓCuₓ studied by very low-temperature ac susceptibility and dc magnetization, Phys. Rev. B 76, 224419 (2007).
  38. [39]A. Kumar, B. Schwarz, and R. S. Dhaka, Correlation between the exchange bias effect and antisite disorder in Sr₂₋ₓLaₓCoNbO₆ (x = 0, 2), Phys. Rev. B 109, 104434 (2024).
  39. [40]G. Kotliar and H. Sompolinsky, Phase Transition in a Dzyaloshinsky-Moriya Spin-Glass, Phys. Rev. Lett. 53, 1751 (1984).
  40. [41]P. M. Shand, T. Rash, M. Streicher, T. E. Kidd, K. R. Boyle, and L. H. Strauss, Coercivity and exchange bias of Mn₀.₂₅Ti₁.₁S₂ in the cluster-glass state, Phys. Rev. B 82, 214413 (2010).
  41. [42]S. N. Kaul and S. Srinath, Irreversibility lines in the H-T phase diagram of re-entrant amorphous ferromagnet, J. Phys.: Condens. Matter 10,11067 (1998).
  42. [43]J. R. L. de Almeida and D. J. Thouless, Stability of the Sherrington-Kirkpatrick solution of a spin glass model, J. Phys. A: Math. Gen. 11, 983 (1978).
  43. [44]M. Gabay and G. Toulouse, Coexistence of Spin-Glass and Ferromagnetic Orderings, Phys. Rev. Lett. 47, 201 (1981).DOI
  44. [45]W. M. Yuhasz, N. A. Frederick, P.-C. Ho, R. A. Fisher, M. B. Maple, H. Sugawara, Y. Aoki, and H. Sato, Heavy-fermion behavior, crystalline electric field effects, and weak ferromagnetism in SmOs₄Sb₁₂, Phys. Rev. B 71, 104402 (2005).
  45. [46]J. H. Van Vleck, The Theory of Electric and Magnetic Susceptibilities (Oxford University Press, London, 1932).
  46. [47]A. M. Stewart, Paramagnetic Susceptibilities of Metallic Samarium Compounds, Phys. Rev. B 6, 1985 (1972).DOI
  47. [48]K. Jonason, J. Mattsson, and P. Nordblad, Dynamic susceptibility of a reentrant ferromagnet, Phys. Rev. B 53, 6507 (1996).pubmed.ncbi.nlm.nih.gov/9982050
  48. [49]Y. Yeshurun, M. B. Salamon, K. V. Rao, and H. S. Chen, Critical phenomena in amorphous ferromagnetic and spin-glass alloys, Phys. Rev. B 24, 1536 (1981).
  49. [50]D. A. Read, T. Moyo, and G. C. Hallam, Low temperature magnetic hardness of melt spun Fe-Zr amorphous alloys, Journal of Magnetism and Magnetic Materials 44, 279 (1984).
  50. [51]N. Moutis, C. Christides, I. Panagiotopoulos, and D. Niarchos, Exchange-coupling properties of La₁₋ₓCaₓMnO₃ ferromagnetic/antiferromagnetic multilayers, Phys. Rev. B 64, 094429 (2001).DOI
  51. [52]W. J. Feng, D. Li, W. J. Ren, Y. B. Li, W. F. Li, J. Li, Y. Q. Zhang, and Z. D. Zhang, Glassy ferromagnetism in Ni₃Sn-type Mn₃.₁Sn₀.₉, Phys. Rev. B 73, 205105 (2006).DOI
  52. [53]M. Hiroi, T. Rokkaku, K. Matsuda, T. Hisamatsu, I. Shigeta, M. Ito, T. Sakon, K. Koyama, K. Watanabe, S. Nakamura, T. Nojima, T. Nakano, L. Chen, T. Fu- jiwara, Y. Uwatoko, H. Manaka, and N. Terada, Ferromagnetism and spin-glass transitions in the Heusler compounds Ru₂₋ₓFeₓCrSi, Phys. Rev. B 79, 224423 (2009).
  53. [54]Y. D. Guan, C. H. Yan, S. H. Lee, X. Gui, W. Ning, J. L. Ning, Y. L. Zhu, J.3 M. Kotakonda, C. Q. Xu, X. L. Ke, J. W. Sun, W. W. Xie, S. L. Yang, and Z. Q. Mao, Ferromagnetic MnBi₄Te₇ obtained with low-concentration Sb doping: A promising platform for exploring topological quantum states, Phys. Rev. Materials 6, 054203 (2022).DOI
  54. [55]K. Binder and A. P. Young, Spin glasses: Experimental facts, theoretical concepts, and open questions, Rev. Mod. Phys. 58, 801 (1986).DOI
  55. [56]J. A. Mydosh, Spin Glasses: An Experimental Introduction, Vol. 125 (Taylor and Francis, London, 1993).
  56. [57]M. Giot, A. Pautrat, G. André, D. Saurel, M. Hervieu, and J. Rodriguez-Carvajal, Magnetic states and spin-glass properties in Bi₀.₆₇Ca₀.₃₃MnO₃: Macroscopic ac measurements and neutron scattering, Phys. Rev. B 77, 134445 (2008).DOI
  57. [58]A. Kumar, B. Schwarz, H. Ehrenberg, and R. S. Dhaka, Evidence of discrete energy states and cluster-glass behavior in Sr₂₋ₓLaₓCoNbO₆, Phys. Rev. B 102, 184414 (2020).DOI
  58. [59]V. K. Anand, D. T. Adroja, and A. D. Hillier, Ferromagnetic cluster spin-glass behavior in PrRhSn₃, Phys. Rev. B 85, 014418 (2012).DOI
  59. [60]P. C. Hohenberg and B. I. Halperin, Theory of dynamic critical phenomena, Rev. Mod. Phys. 49, 435 (1977).
  60. [61]G. Ceccarelli, A. Pelissetto, and E. Vicari, Ferromagnetic-glassy transitions in three-dimensional Ising spin glasses, Phys. Rev. B 84, 134202 (2011).DOI
  61. [62]J. A. Hertz, D. Sherrington, and Th. M. Nieuwenhuizen, Competition between coexistingness and order in a multispin glass, Phys. Rev. E 60, R2460 (1999).pubmed.ncbi.nlm.nih.gov/11970176
  62. [63]J. Wu and C. Leighton, Glassy ferromagnetism and magnetic phase separation in La₁₋ₓSrₓCoO₃, Phys. Rev. B 67, 174408 (2003).DOI
  63. [64]V. K. Anand, D. A. Tennant, and B. Lake, Investigations of the effect of nonmagnetic Ca substitution for magnetic Dy on spin-freezing in Dy₂Ti₂O₇, J. Phys.: Condens. Matter 27, 436001 (2015).DOI
  64. [65]D. C. Johnston, Stretched exponential relaxation arising from a continuous sum of exponential decays, Phys. Rev. B 74, 184430 (2006).DOI
  65. [66]T. Samanta, A. Us Saleheen, D. L. Lepkowski, A. Shankar, I. Dubenko, A. Quetz, M. Khan, N. Ali, and S. Stadler, Asymmetric switchinglike behavior in the magnetoresistance at low fields in bulk metamagnetic Heusler alloys, Phys. Rev. B 90, 064412 (2014).DOI
  66. [67]S. Pakhira, C. Mazumdar, R. Ranganathan, S. Giri, and M. Avdeev, A large magnetic cooling power involving frustrated antiferromagnetic spin-glass state in R₂NiSi₃ (R = Gd, Er), Phys. Rev. B 94, 104414 (2016).DOI
  67. [68]M. Suzuki, I. S. Suzuki, and M. Matsuura, Memory and aging effect in hierarchical spin orderings of the stage-2 CoCl₂ graphite intercalation compound, Phys. Rev. B 73, 184414 (2006).DOI
  68. [69]M. K. Chattopadhyay, S. B. Roy, A. K. Nigam, K. J. S. Sokhey, and P. Chaddah, Metastability and giant relaxation across the ferromagnetic to antiferromagnetic transition in Ce(Fe₀.₉₆Ru₀.₀₄)₂, Phys. Rev. B 68, 174404 (2003).DOI
  69. [70]M. Sirena, L. B. Steren, and J. Guimpel, Magnetic relaxation in bulk and film manganite compounds, Phys. Rev. B 64, 104409 (2001).DOI
  70. [71]S. J. Collocott and J. B. Dunlop, Anomalous magnetic viscosity in the bulk-amorphous ferromagnet Nd₆₀Fe₂₀Co₁₀Al₁₀, Phys. Rev. B 66, 224420 (2002).DOI
  71. [72]C. Kittel, Introduction to Solid State Physics, 8th ed. (Wiley, New York, 2005).
  72. [73]P. Kumar, N. K. Singh, K. G. Suresh, and A. K. Nigam, Magnetocaloric and magnetotransport properties of R₂Ni₂Sn compounds (R = Ce, Nd, Sm, Gd, and Tb), Phys. Rev. B 77, 184411 (2008).DOI
  73. [74]S. R. Gopal, Specific Heats at Low Temperatures (Plenum, New York, 1966).

Paper details

vlnc:W_xw4vhqnj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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