betaFeedback

WorksPhysics and Astronomy

Different critical exponents on two sides of the magnetic transition in two dimensions

Abstract

The possibility of different critical exponents on two sides of a magnetic transition has been theoretically predicted [PRB 13, 2222 (1976), PRL 115, 200601 (2015)] and experimentally realized [PRL 128, 015703 (2022)] in three-dimensional systems. However, the emergence of such a scenario in two dimensions is not yet established. Here, we report distinct two-dimensional (2D) critical exponents below and above the magnetic transition temperature (TC) through critical scaling analysis in a quasi-2D non-centrosymmetric, Dzyaloshinskii-Moriya interaction (DMI) active hybrid perovskite (C7H9NBr)2CuBr4. The system exhibits critical exponents of 2D Ising-type below TC and 2D XY-type above TC . In contrast, isostructural Cl and centrosymmetric Br analogs, in which the DMI effect is expected to be weaker, exhibit symmetric 2D critical behavior. These results highlight the crucial role of DMI in (C7H9NBr)2CuBr4 as a symmetry-breaking perturbation that alters the magnetic critical behavior differently above and below TC . Our study is the experimental realization of hitherto unreported asymmetric 2D critical behavior in structurally flexible and dimensionally tunable hybrid perovskites.

A second-order magnetic phase transition is characterized by a unique set of critical exponents β\beta, γ\gamma, and δ\delta, which govern the scaling behavior of the order parameter, susceptibility, and critical isotherm near the transition temperature TCT_\mathrm{C} respectively [1, 2, 3]. Within the renormalization group framework, the critical exponents are associated with the symmetry of order parameter (n) and lattice dimensionality (d) of the system, through their correspondence to the established universality classes (UCs) including mean-field (β=0.5\beta=0.5, γ=1\gamma=1, and δ=3\delta=3) three-dimensional (3D) Heisenberg (β=0.365\beta=0.365, γ=1.386\gamma=1.386, and δ=4.8\delta=4.8), 3D XY (β=0.345\beta=0.345, γ=1.316\gamma=1.316, and δ=4.81\delta=4.81), 3D Ising (β=0.325\beta=0.325, γ=1.24\gamma=1.24, and δ=4.82\delta=4.82), tricritical mean field (β=0.25\beta=0.25, γ=1\gamma=1, and δ=5\delta=5), and two-dimensional (2D) Ising (β=0.125\beta=0.125, γ=1.75\gamma=1.75, and δ=15\delta=15) models [4, 5, 6, 7, 8, 9, 10, 11]. Conventionally, a single set of critical exponents corresponding to a specific UC is expected to capture the critical behavior across the phase transition region. Notably, Nelson way back in 1976 proposed that symmetry-breaking perturbations such as cubic and hexagonal anisotropy in a 3D O(2) model can induce asymmetric critical behavior, with distinct exponents γ+\gamma_+ and γ−\gamma_-, above and below TCT_\mathrm{C} respectively [12]. He attributed this asymmetry to the dangerously irrelevant nature of the anisotropy, which becomes asymptotically irrelevant at the critical fixed point, but can couple differently in the two sides of transition, thereby modifying the scaling behavior. Nearly five decades later, Leonard and Delamotte proposed a generic mechanism that relies on the possibility of explicitly breaking a continuous symmetry down to discrete one through perturbations that are irrelevant at the critical fixed point, but can generate different critical exponents on the two sides of a continuous phase transition [13]. Yet, experimental realization of the above striking theoretical proposal remains sparse. The first experimental evidence of asymmetric critical behavior was reported in the skyrmion host magnet Cu2_2OSeO3_3, with the asymmetry linked to the subtle influence of underlying Dzyaloshinskii–Moriya interaction (DMI) [14]. Recent studies have also reported similar unconventional critical behavior in Cr1_1TaS2_2 [15], MnBi4_4Te7_7 [16], and FePd2_2Te2_2 [17]. However, these experimental and theoretical realizations have thus far been limited to 3D magnetic systems, while the possibility of asymmetric scaling in 2D remains elusive fig:1a.

The 2D limit with the enhanced fluctuations, provides a sensitive regime in which weak symmetry-breaking perturbation can strongly influence the critical behavior and the resulting magnetic order. For an isotropic 2D Heisenberg system, Mermin-Wagner theorem forbids spontaneous symmetry breaking at finite temperature [18]. However, magnetic anisotropy can break the continuous symmetry, suppress fluctuations, and thereby stabilize long-range magnetic order (LRMO), making it central to the study of atomically thin magnets [19, 20, 21]. Strong uniaxial anisotropy has been found to drive the system towards the 2D Ising UC in atomically thin magnet such as Cr2_2Ge2_2Te6_6 [22], CrI3_3 [23], Fe3_3GeTe2_2 [24, 25]. By contrast, easy-plane anisotropy leads to Berezinskii–Kosterlitz–Thouless (BKT) physics, where vortex–antivortex binding produces quasi-long-range order [26, 27, 28]. Recently, Parkin et al. demonstrated robust 2D XY-type long-range magnetic order (β∼0.22\beta\sim0.22) in monolayers of CrCl3_3 [26]. Between these two extreme cases, systems having fourfold symmetry belong to 2D XY4_4 UC with critical exponents lying between 2D Ising and XY limits [29, 30]. These results reveal that symmetry-breaking perturbations such as anisotropy play a decisive role in determining criticality in two dimensions. Whether symmetry-breaking perturbations can drive asymmetric critical scaling in 2D magnetic systems still remains an open question.

Schematic of symmetric and asymmetric criticality and a quasi-2D magnet

Figure 1. (a) Upper panel represents symmetric (identical γ\gamma) and asymmetric (γ+≠γ−\gamma_{+} \ne\gamma_{-}) criticality, Table illustrates the theoretical and experimental realization of asymmetric criticality in 3D systems (Ref. No.: [12–17]), which remains elusive in 2D. (b) Schematic of OIHP based quasi-2D magnet.

Leonard and Delamotte pointed out that it would be extremely interesting to investigate the possibility for reliazing asymmetric criticality in 2D [13]. Addressing this question experimentally requires a system that remains predominantly 2D while retaining the specific symmetry-breaking interaction which could generate different critical exponents below and above TCT_C. In particular, quasi-2D organic–inorganic hybrid perovskites (OIHPs) provide a natural platform for realizing 2D magnetic regime since the interlayer distance in these systems can be tuned to a large value that suppresses the interlayer exchange J′J', with J′/J∼10−4J'/J \sim10^{-4}–10−610^{-6} , such that the intralayer exchange JJ dominates the magnetic response [31, 32, 33, 34]. Besides, the structural flexibility of OIHPs enables the emergence of symmetry-breaking interactions such as DMI through broken inversion symmetry induced by the appropriate choice of organic cations.

Here, we demonstrate asymmetric criticality in a non-centrosymmetric, DMI active quasi-2D hybrid perovskite (C7_7H9_9NBr)2_2CuBr4_4. Remarkably, the system exhibits 2D Ising-type criticality below TCT_C and 2D XY-type criticality above TCT_C, providing the first experimental realization of asymmetric 2D criticality. To further substantiate our claim, we have also investigated the critical behavior of an isostructural Cl-analog (C7_7H9_9NBr)2_2CuCl4_4 and centrosymmetric Br-analog (C7_7H10_{10}N)2_2CuBr4_4, where the DMI effect is expected to be weaker, and we observe a conventional symmetric critical behavior. We attribute the asymmetric criticality in NCS-Br to the symmetry-breaking perturbation DMI that influences the magnetic critical behavior differently on either side of TCT_C in the critical regime.

The details of synthesis of the above three single crystals and structural characterization are reported in supplementary material (SM; section 1, Figure S1, Table S1), and the reference therein. Magnetic measurements were carried out on single crystals using a superconducting quantum interference device magnetometer (Quantum Design SQUID-VSM). For the critical scaling analysis, field-dependent magnetization M(H)M(H) measurements were conducted at various temperatures in the magnetic transition region [Figure S3, S7 and S10]. Demagnetization effects are found to be negligible and hence, no correction was applied.

Among the three quasi-2D OIHPs, (C7_7H9_9NBr)2_2CuX4_4 (X = Br and Cl) crystallise in noncentrosymmetric monoclinic structure with polar space group CcCc [31], while (C7_7H10_{10}N)2_2CuBr4_4 adopt centrosymmetric structure with space group C2/cC2/c [Figure S1 and Table S1 in SM]. All the compounds show dominant ferromagnetic interaction, verified from the positive Curie-Weiss temperature (TCWT_{CW}) [Insets of Figure S2(a),S6(a),S9(a)]. Note that, while both the Br analogs (C7_7H9_9NBr)2_2CuBr4_4 (NCS-Br) and (C7_7H10_{10}N)2_2CuBr4_4 (CS-Br)) exhibit easy saturation along the out-of-plane direction [Figure S2 and S9], (C7_7H9_9NBr)2_2CuCl4_4 (NCS-Cl) show an in-plane easy saturation [Figure S6]. The critical analysis is carried out along the easy saturation directions for all samples.

Presence of second-order magnetic phase transition is confirmed from the concave-downward curvature of the Arrott plots (mean field model) [Figure S4(a), S8(a), S11(a)], following the Banerjee criterion [35]. The associated critical exponents β\beta, γ\gamma and δ\delta were extracted by fitting spontaneous magnetization (MSM_S) below TCT_C, inverse susceptibility (χ0−1\chi_0^{-1}) above TCT_C, and critical isotherm (M(H) at TCT_C), respectively [2], according to the following relations:

MS(T)=M0(−ϵ)−β;ϵ<0,T<TC(1)M_S(T) = M_0(-\epsilon)^{-\beta}; \epsilon< 0, T < T_C \tag*{(1)}
χ0−1(T)=(h0m0)ϵγ;ϵ>0,T>TC(2)\chi_0^{-1}(T) = \left(\frac{h_0}{m_0}\right)\epsilon^\gamma; \epsilon> 0, T > T_C \tag*{(2)}
M=DH1δ;ϵ=0,T=TC(3)M = D H^{\frac{1}{\delta}}; \epsilon= 0, T = T_C \tag*{(3)}

where ϵ=T−TCTC\epsilon= \frac{T-T_C}{T_C} and M0M_0, h0m0\frac{h_0}{m_0}, DD are the critical amplitudes. To study the critical behavior, modified Arrott plots (MAPs) M1/βM^{1/\beta} vs. (H/M)1/γ(H/M)^{1/\gamma} are constructed based on the Arrott Noake’s relation; (H/M)1/γ=aϵ+bM1/β(H/M)^{1/\gamma} = a\epsilon+ bM^{1/\beta} where aa and bb are constants [9, 36, 37]. The β\beta and γ\gamma are so chosen that the MAPs produce a set of parallel straight lines, and the line at TCT_C passes through the origin [14]. Further the critical exponents should self consistently satisfy the Widom scaling relation δ=1+γβ\delta= 1 + \frac{\gamma}{\beta} [4] where δ\delta is determined by fitting the critical M(H) isotherm at TCT_C with the equation 3. The parallelism of the MAPs is quantified by evaluating the normalised slope (NS = S(T)/S(TCT_C)), where S(T) and S(TCT_C) are the slope of the linear fit to the MAPs at temperature T and TCT_C respectively. The accurate critical exponents are those which produce the NS value close to 1 and generally, iteration method is employed (see section 2 of SM) to determine the critical exponents β\beta and γ\gamma.

In the critical asymptotic region, the magnetic scaling equation is expressed as

M(H,ϵ)=ϵβf±(Hϵβ+γ)(4)M(H,\epsilon) = \epsilon^\beta f \pm\left(\frac{H}{\epsilon^{\beta+\gamma}}\right) \tag*{(4)}
Six-panel critical-exponent analysis plots for NCS-Br

Figure 2. NCS-Br: (a) Normalised slope NS(T) obtained from exponents of various universality classes, iteration method and distinct set of critical exponents β\beta and γ\gamma. (b) Linear fitting to the modified Arrott plots (MAPs) using β=0.125\beta=0.125, γ=1.047\gamma=1.047 (c) MAPs using β=0.19\beta=0.19, γ=1.575\gamma=1.575 for T>TCT>T_{\mathrm{C}}. Inset: MAPs using β=0.125\beta=0.125, γ=1.047\gamma=1.047 for T<TCT<T_{\mathrm{C}} (d) Log-log plot of magnetic isotherm M(H)M(H) at TC=12.5T_{\mathrm{C}}=12.5K with a linear fitting (e) Renormalised magnetization (mm) vs. field (hh) plot below and above TCT_{\mathrm{C}}. Inset: mm vs. hh in log-log scale. (f) m2m^2 vs. h/mh/m plots (Renormalised Arrott plots).

where f+f_{+} and f−f_{-} are regular functions correspond to T>TCT>T_{\mathrm{C}} and T<TCT<T_{\mathrm{C}} respectively. This equation can also be rewritten as

m=f±(h)(5)m=f_{\pm}(h) \tag*{(5)}

where m=ϵ−βM(H,ϵ)m=\epsilon^{-\beta}M(H,\epsilon) is the renormalized magnetization and h=ϵ−(β+γ)Hh=\epsilon^{-(\beta+\gamma)}H is the renormalized field [5]. The equation 5 implies that for the correct critical exponent β\beta and γ\gamma values, the renormalized mm vs. hh curves should collapse onto two different universal branches, one below TCT_{\mathrm{C}} and another above TCT_{\mathrm{C}} [2].

First, we focus on the NCS-Br compound which is reported to possess a transverse conical spiral magnetic ground state driven by the interplay of ferromagnetic exchange, magnetocrystalline anisotropy, and DMI [31]. This makes it an interesting system to examine the underlying magnetic UC in the presence of such competing interactions by analyzing the critical behavior. In this regard, MAPs are generated for all the six UCs [Figure S4 in SM], following the procedure discussed above. The MAPs show better linearity in the high-field region and NS is determined for all UCs from linear fits to the high field region [FIG. 2(a)]. The NS plot shows a large deviation from the constant slope NS=1 line for the 3D UCs and mean-field model, suggesting that the criticality in the compound is unlikely to be governed by 3D

NCS-Br modified Arrott plots for mean field, 3D Heisenberg, 3D XY, 3D Ising, tricritical, and 2D Ising universality classes.

Figure S4. NCS-Br: Modified Arrott plots for various universality classes including mean field model

critical exponents. Further, while the 2D Ising model exhibits the smallest deviation in NS below TCT_{\mathrm{C}}, the corresponding NS curve deviates more from the constant-slope NS=1 line above TCT_{\mathrm{C}}. In contrast, the tricritical mean field model produces the least deviation above TCT_{\mathrm{C}}. This primarily indicates that the critical exponents governing the critical behavior below and above the TCT_{\mathrm{C}} could be different. Note that, though the least deviated NS is produced by the tricritical mean field model above TCT_{\mathrm{C}}, the deviation is still very high compared to the NS produced by 2D Ising model below TCT_{\mathrm{C}}.

Subsequently, the iteration method is followed to evaluate the critical exponents. In this method, the initial exponents are chosen from that UC which generates the least deviated slope. Hence 2D Ising model is initially adopted. It is found that, for 2D Ising model, the MAP close to TCT_{\mathrm{C}} does not pass through the origin. Therefore, keeping β\beta (=0.125) fixed, the γ\gamma value is varied in such a way that the MAP close to TCT_{\mathrm{C}} passes through the origin. Then iteration method is employed which yields β\beta=0.147(6) and γ\gamma=0.999(25) as shown in Figure S5 of SM. However, the resulting NS from the iteration method is very much deviated from the NS=1 line, especially below TCT_{\mathrm{C}} [FIG. 2(a)]. Hence, the exponents obtained from iteration are not accurate and the iteration method is not suitable to determine the precise values of critical exponents due to the fact that single set of critical exponents can not capture the critical behavior in the present scenario. This is evident from the MAPs constructed from one set of exponents β=0.125\beta=0.125 and γ=1.04\gamma=1.04 that show nearly parallel lines below TCT_{\mathrm{C}}, however a significant deviation from parallelism is observed above TCT_{\mathrm{C}} fig:2b.

To circumvent this discrepancy, a new iterative method, proposed by Ghosh et al. [14], is adopted. In this method, first the two UCs which produce the least deviated NS below and above TCT_{\mathrm{C}} are chosen and then the critical exponents are separately iterated below and above TCT_{\mathrm{C}}, untill getting NS values near to 1. By varying β\beta and γ\gamma around the 2D Ising exponents for T≤TCT\leq T_{\mathrm{C}} and tricritical mean field exponents for T≥TCT\geq T_{\mathrm{C}}, we obtained a distinct set of β\beta and γ\gamma, below and above TCT_{\mathrm{C}} (β−\beta_{-} and γ−\gamma_{-} for T≤TCT\leq T_{\mathrm{C}} and β+\beta_{+} and γ+\gamma_{+} for T≥TCT\geq T_{\mathrm{C}}), yielding nearly parallel fitted MAPs with NS(T)(T) values closest to 1 fig:2a and the line at TCT_{\mathrm{C}} passing through the origin fig:2c. The optimized exponents are β−=0.125\beta_{-}=0.125, γ−=1.047\gamma_{-}=1.047 and β+=0.19\beta_{+}=0.19, γ+=1.575\gamma_{+}=1.575, fig:2c that leads to δ−=9.376\delta_{-}=9.376 and δ+=9.289\delta_{+}=9.289 according to the Widom scaling relation δ=1+γβ\delta=1+\frac{\gamma}{\beta}. The value of δ\delta is further experimentally determined to be 9.17(1) from the linear fits to log M vs. log H plot at TC=12.5T_{\mathrm{C}}=12.5 K, following equation 3 fig:2d, indicating that the experimental δ\delta value is closely similar to the δ−\delta_{-} and δ+\delta_{+}. fig:2e shows the mm vs. hh plot with the inset showing the same in the log-log scale by using β−=0.125\beta_{-}=0.125, γ−=1.047\gamma_{-}=1.047 for T<TCT<T_{\mathrm{C}} and β+=0.19\beta_{+}=0.19, γ+=1.575\gamma_{+}=1.575 for T>TCT>T_{\mathrm{C}}. Here, all data collapse into two different well divided branches below and above the TCT_{\mathrm{C}}. The deviation from the universal branches in the low-field region observed in the log-log plot [Inset fig:2e] can be correlated to the observed nonlinearity of the MAPs in the low-field region fig:2b. The reliability of these critical exponents and TCT_{\mathrm{C}} have been further corroborated from the renormalized Arott plots m2m^{2} vs. h/mh/m as shown in fig:2f where all the data also divided into two well separated branches, indicating the consistency of the scaling analysis.

The obtained critical exponents (β=0.125\beta=0.125, γ=1.047\gamma=1.047 below TCT_{\mathrm{C}} and β=0.19\beta=0.19, γ=1.575\gamma=1.575 above TCT_{\mathrm{C}}) lie within the range associated with 2D XY/h4h_{4} UC (β=0.125\beta=0.125–0.23) where β=0.125\beta=0.125 corresponds to 2D Ising and β=0.23\beta=0.23 to 2D XY limit [26, 29, 30]. However, it is worth noting that β=0.19\beta=0.19 has been identified to represent 2D XY UC in previous studies [38, 39, 40]. Thus, the change in β=0.125\beta=0.125 below TCT_{\mathrm{C}} to β=0.19\beta=0.19 above TCT_{\mathrm{C}}, indicates a change in UC from 2D Ising to 2D XY-like across the magnetic transition in NCS-Br. Such unconventional critical behavior has been attributed earlier to the interplay of DMI, symmetric exchange, and magnetocrystalline anisotropy in Cu2_2OSeO3_3 [14]. For NCS-Br, our recent magnetic study indicates the possible existence of a non-collinear conical spiral magnetic ground state induced by the DMI [31]. We therefore propose that the observed unconventional criticality may be linked to the underlying DMI. To establish whether DMI underlies this anomalous criticality in NCS-Br, we have undertaken detailed critical scaling analysis for the isostructural Cl analogue and centrosymmetric Br analogue, where DMI is expected to be relatively weaker.

NCS-Cl critical scaling plots

Figure 3. NCS-Cl: (a) Normalised slope NS(T)(T) obtained from exponents of various universality classes and iteration method. (b) Temperature dependent spontaneous magnetization MS(T)M_{\mathrm{S}}(T) (left) and inverse susceptibility χ0−1\chi_{0}^{-1} (right) with fitted solid curves (blue curves). Inset: Linear fitting of M vs. H at TCT_{\mathrm{C}} plot in log-log scale. (c) Modified Arrott plots (MAPs) using β=0.144\beta=0.144, γ=1.429\gamma=1.429 (d) Renormalised magnetization (mm) vs. field (hh) plot below and above TCT_{\mathrm{C}}. Inset: mm vs. hh in log-log scale. (e) m2m^{2} vs. h/mh/m plots.

Now we examine the non-centrosymmetric NCS-Cl compound, which shows a nominally uniaxial in-plane magnetic anisotropy induced by an in-plane DMI vector, as discussed in the earlier work [Figure S6 of [31]]. However, the strength of DMI in the NCS-Cl is weaker compared to the NCS-Br compound, consistent with reduced spin-orbit coupling associated with the lighter Cl ligand [32, 41, 42]. Upon analyzing the critical behavior, it is found that the NS plot (obtained from the linear fitting of MAPs in the high field region [Figure S8 in SM]) shows the least deviation for 2D Ising UC throughout the transition regime fig:3a. Iteration method produces the least deviated NS fig:3a with β=0.145(2)\beta=0.145(2), TC=6.97(2)T_{\mathrm{C}}=6.97(2) K from the fitting of MS(T)M_{\mathrm{S}}(T) and γ=1.43(2)\gamma=1.43(2), TC=6.75(3)T_{\mathrm{C}}=6.75(3) K from the fitting of χ0−1(T)\chi_{0}^{-1}(T) fig:3b. Considering this β\beta and γ\gamma, the MAP produces a set of straight lines in the high field region with the line at 7 K (TCT_{\mathrm{C}}), passing through the origin fig:3c. The linear fitting to M-H isotherm at TC=7T_{\mathrm{C}}=7 K in log-log scale yields δ=10.73(1)\delta=10.73(1) [Inset of fig:3b], indicating that exponents obey Widom scaling relation. Further, the renormalised m vs. h (along with the plots in their log-log scale) fig:3d and m2m^{2} vs. h/m fig:3e curves collapse into two different branches below and above TCT_{\mathrm{C}}, showing reliability of the critical exponents. The NS curve obtained from iterative method using β=0.145\beta=0.145, γ=1.43\gamma=1.43, deviates slightly from the NS=1 line above TCT_{\mathrm{C}} and remains close to NS=1 line below TCT_{\mathrm{C}}. This weak asymmetry in NS curve possibly arises from the non-negligible DMI present in this compound. The obtained critical exponents (β=0.144\beta=0.144, γ=1.429\gamma=1.429 and δ=10.73\delta=10.73), are characteristics of the 2D Ising UC which can be correlated to the dominant unidirectional in-plane spin-orientation in the NCS-Cl compound [Figure S6 of [31]]. Although Ising universality is often commonly associated with an out-of-plane easy axis, it can also be realized in systems with in-plane magnetization, having uniaxial magnetic anisotropy, as reported in earlier studies for Fe/W(110) and Fe(110)/Ag(111) films [43, 44]. Hence, the 2D Ising exponents consistently explain the magnetic response observed in NCS-Cl compound.

Critical exponent analysis for CS-Br

Figure 4. CS-Br: (a) Normalised slope NS(T) obtained from exponents of various universality classes and iteration method. (b) Temperature dependent spontaneous magnetization MS(T)M_{\mathrm{S}}(T) (left) and inverse susceptibility χ0−1\chi_{0}^{-1} (right) with fitted solid curves (blue curves). Inset: Linear fitting of MM vs. HH at TCT_{\mathrm{C}} plot in log-log scale. (c) Modified Arrott plots (MAPs) using β=0.128\beta=0.128, γ=1.25\gamma=1.25 (d) Log-log plot of renormalised magnetization (mm) vs. field (hh) below and above TCT_{\mathrm{C}} in the high field region.

Finally, we delve into the critical behavior of the centrosymmetric compound CS-Br, showing an out-of-plane easy axis. Here, 2D Ising UC produces NS close to 1 both below and above TCT_{\mathrm{C}}, [FIG. 4(a) and Figure S11], indicating the presence of conventional critical behavior. Further, the iteration method produces the least deviated NS [FIG. 4(a)] with β=0.128(2)\beta=0.128(2), TC=13.004(15)T_{\mathrm{C}}=13.004(15) K from the fitting of MS(T)M_{\mathrm{S}}(T) and γ=1.25(2)\gamma=1.25(2), TC=13.002(18)T_{\mathrm{C}}=13.002(18) K from the fitting of χ0−1(T)\chi_{0}^{-1}(T)[FIG. 4(b)]. The exponents β=0.128\beta=0.128 and γ=1.25\gamma=1.25 produce a set of straight lines in MAPs with the line at TC∼13T_{\mathrm{C}}\sim13 K passing through origin [FIG. 4(c)] and the exponents obey Widom scaling relation as concluded from the δ\delta value (∼10.67(3)\sim10.67(3)) shown in the inset of FIG. 4(b). Further, the log-log plot of mm vs. hh curves produces two distinct branches below and above TCT_{\mathrm{C}} in the high field region [FIG. 4(d)], which suggests the accuracy of the critical exponents. The value of β=0.128\beta=0.128 indicates the presence 2D Ising-type spin interaction, consistent with the out-of-plane easy axis observed in CS-Br compound.

Mean field Modified Arrott plot

(a)

3D Heisenberg Modified Arrott plot

(b)

3D XY Modified Arrott plot

(c)

3D Ising Modified Arrott plot

(d)

Tricritical Modified Arrott plot

(e)

2D Ising Modified Arrott plot

(f)

Figure S11. CS-Br: Modified Arrott plots for various universality classes including mean field model

Spectrum of critical behavior in NCS-Cl, NCS-Br, and CS-Br

Figure 5. Spectrum of critical behavior

The UCs obtained from the critical scaling analysis has been summarized in FIG. 5, showing the spectrum of critical phenomena in the hybrid perovskites caused by changing the crystal symmetry and halide ligands. In all three compounds, the β\beta values deviate significantly from 3D UCs with β∼0.32\beta\sim0.32 to 0.37 and lie in the range of 0.125-0.23, corresponding to 2D XY/h4h_{4} UC [10, 29]. In 2D XY/h4h_{4} UC, β=0.125\beta=0.125 represents the extreme anisotropic 2D Ising limit while β=0.23\beta=0.23 signifies the isotropic 2D XY limit. Hence, it can be unambiguously concluded that the magnetic criticality in these compounds is two dimensional. However, the δ\delta values for all three compounds (≈10\approx10) lie in between the 3D (δ≈5\delta\approx5) and 2D UCs (δ≈15\delta\approx15) [10, 45, 46], indicating the presence of non-negligible interlayer magnetic coupling.

The asymmetric critical behavior observed in NCS-Br is particularly intriguing. As discussed in the introduction, asymmetric 3D critical exponents on either side of TCT_{\mathrm{C}} have been reported in Cu2_{2}OSeO3_{3} [14], Cr1/3_{1/3}TaS2_{2} [15], MnBi4_{4}Te7_{7} [16] and FePd2_{2}Te2_{2} [17]. However, the present study highlights the asymmetric criticality in 2D regime, where 2D Ising-type criticality below TCT_{\mathrm{C}} evolves into 2D XY-type critical behavior above TCT_{\mathrm{C}} in NCS-Br. In contrast, investigation on iso-structural Cl and centrosymmetric Br analogues reveals conventional critical behavior. This highlights the role of DMI, which is more effective in NCS-Br. We conjecture that DMI influences the critical fluctuations and correlations differently on the two sides of the transition, thereby favoring distinct effective critical regimes. Interestingly, the difference between low-T and high-T 2D exponents is found to be larger in NCS-Br than that observed in most of the 3D cases. This suggests that reduced dimensionality may provide a more favorable setting to explore the effect of dangerously irrelevant perturbation, yielding different critical behavior on the two sides of TCT_{\mathrm{C}}. Our results motivate theoretical and numerical studies of asymmetric criticality in quasi-2D magnets, where controlled 2D simulations may provide direct tests of the distinct critical behavior above and below TCT_{\mathrm{C}} and the role of DMI.

In summary, we have demonstrated the presence of asymmetric critical exponents on two sides of the magnetic transition in 2D regime for a DMI-active NCS-Br compound. Notably, it exhibits 2D Ising and 2D XY type critical exponents below and above TCT_{\mathrm{C}}, respectively. In contrast, investigation on additional two hybrid perovskites NCS-Cl and CS-Br, for which the DMI contribution is expected to be less significant, reveals conventional 2D critical behavior. The systematic contrast in the critical behavior among these compounds points to DMI as a key symmetry-breaking perturbation for the observed asymmetric criticality in NCS-Br compound.

Acknowledgments- P.B. acknowledges Prof. B. Delamotte, Prof. Stuart S. P. Parkin and Prof. S.N. Kaul for valuable insights. P.B. thanks Dr. Amitava Ghosh and Dr. Hanuma Kumar Dara for useful discussions.

References

Supplementary Material: Different critical exponents on two sides of the magnetic transition in two dimensions

P. Biswal; Ruma Khatun; G. Tripathy; Diptikanta Swain; D. Samal

Synthesis and structural characterization

Single crystal synthesis and structural characterization of NCS-Br and NCS-Cl

The details of the synthesis and structural characterization of NCS-Br and NCS-Cl compound are reported in a recent paper [31].

Single crystal synthesis and structural characterization of CS-Br

CS-Br ((C7H10N\mathrm{C}_{7}\mathrm{H}_{10}\mathrm{N})2_2CuBr4_4) single crystals were synthesized from solution evaporation method. First, benzylammonium salt was synthesized from the homogeneous solution of benzylamine and HBr. Then this salt and CuBr2_2 were added in a stoichiometric ratio and a homogeneous solution was formed by further adding HBr and distilled water and stirring at 80∘^\circ C. Then the solution was kept inside an oven at 80∘^\circ C. Single crystals were obtained after 7 days. The crystal structure and structural information are presented in Figure S1 and Table S1 respectively.

Crystal structure of CS-Br

Figure S1. Crystal structure of CS-Br ((C7H10N\mathrm{C}_{7}\mathrm{H}_{10}\mathrm{N})2_2CuBr4_4)

ParametersCS-Br
Empirical formula(C7H10N)2CuBr4(\mathrm{C}_{7}\mathrm{H}_{10}\mathrm{N})_{2}\mathrm{CuBr}_{4}
Formula weight (g/mol)599.5
Temperature (K)296
Wavelength (Å)0.71073
Crystal systemMonoclinic
Space groupC 2/cC\ 2/c
Unit cell dimensiona=10.953(3)a = 10.953(3) Å
b=11.026(2)b = 11.026(2) Å
c=31.519(7)c = 31.519(7) Å
α=90∘\alpha= 90^\circ
β=100.034(8)∘\beta= 100.034(8)^\circ
γ=90∘\gamma= 90^\circ
Volume (Å3^{3})3748.4(14)
ZZ8
Crystal density (g/cm3^{3})2.125
F(000)2296
Absorption coefficient9.688
Reflections collected23923
Goodness of fit1.223
Final R indices (I>2σ(I)I>2\sigma(I))Robs=0.1510R_{\mathrm{obs}}=0.1510
wRobs=0.3831wR_{\mathrm{obs}}=0.3831
R indices (all data)Rall=0.1916R_{\mathrm{all}}=0.1916
wRall=0.4127wR_{\mathrm{all}}=0.4127

Table S1. Structural information of CS-Br obtained from the single crystal XRD measurements at room temperature.

Iteration Method

The method involves initially selecting the critical exponents β\beta and γ\gamma of that UC for which NS plot shows the least deviation from 1. Using the corresponding β\beta and γ\gamma of that particular UC, MAPs are constructed. Subsequently, the MSM_{\mathrm{S}} and χ0−1\chi_{0}^{-1} values are obtained from the yy and xx-axis intercepts of the linear fitting to the MAPs respectively. From the fitting of MSM_{\mathrm{S}}(T) and χ0−1\chi_{0}^{-1}(T) using equation 1 and equation 2 (mentioned in the main text) respectively, a new set of β\beta and γ\gamma is obtained. Then MAPs are reconstructed using the new set of β\beta and γ\gamma. The above procedure is iterated until getting stable values of β\beta and γ\gamma. The stable β\beta and γ\gamma obtained from this iteration method are independent of the initial parameters which denotes that the critical exponents obtained from the iteration method are intrinsic to the system.

M-H and M-T measurements of NCS-Br

Figure S2. NCS-Br: (a) M-H at 3K along H∥bcH\parallel bc (In-plane orientation) and along H⊥bcH\perp bc (Out-of-plane orientation). Inset: Fitting of inverse susceptibility utilizing modified Curie-Weiss law (χ(T)=χ0+CT−TCW\chi(T)=\chi_{0}+\frac{C}{T-T_{\mathrm{CW}}}; χ0\chi_{0} is temperature independent contribution to the susceptibility, CC is the Curie constant and TCWT_{\mathrm{CW}} is the Curie-Weiss temperature), showing Curie-Weiss temperature TCW∼T_{\mathrm{CW}\sim} 41 K and 39 K along in-plane and out-of-plane direction. (b) M-T at H=20H=20 Oe with inset showing transition temperature TC∼12.5T_{\mathrm{C}}\sim12.5 K in ac susceptibility χ′(T)\chi'(T) plot along H∥bcH\parallel bc (In-plane orientation) and along H⊥bcH\perp bc (Out-of-plane orientation).

NCS-Br: M-H at various temperatures in the transition regime.

Figure S3. NCS-Br: M-H at various temperatures in the transition regime.

NCS-Br spontaneous magnetization, inverse susceptibility, and modified Arrott plots

Figure S5. NCS-Br: (a)Temperature dependent spontaneous magnetization MS(T)M_{\mathrm{S}}(T) (left) and inverse susceptibility χ0−1\chi_0^{-1} (right) with fitted solid curves (blue curves). (b) Modified Arrott plots (MAPs) using β=0.147\beta=0.147, γ=0.999\gamma=0.999

NCS-Cl magnetization curves and temperature-dependent magnetization

Figure S6. NCS-Cl: (a) M-H at 3K along H∥bc (In-plane orientation) and along H⊥bc (Out-of-plane orientation). Inset: Fitting of inverse susceptibility utilizing modified Curie-Weiss law (χ(T)=χ0+CT−TCW\chi(T)=\chi_0+\frac{C}{T-T_{\mathrm{CW}}}; χ0\chi_0 is temperaure independent contribution to the susceptibility, C is the Curie constant and TCWT_{\mathrm{CW}} is the Curie-Weiss temperature), showing Curie-Weiss temperature TCW∼T_{\mathrm{CW}\sim} 21 K and 18 K along in-plane and out-of-plane direction. (b) M-T at H=20 Oe along H∥bc (In-plane orientation) and along H⊥bc (Out-of-plane orientation) with inset showing minima of the dM/dT plot as the transition temperature TC∼T_{\mathrm{C}\sim}7 K for 10 Oe applied field along in-plane orientation. The value of TC∼T_{\mathrm{C}\sim}7 K is consistent with the ac susceptibility peak value reported in the recent paper [31].

NCS-Cl: M-H curves at various temperatures in the transition regime

Figure S7. NCS-Cl: M-H at various temperatures in the transition regime.

NCS-Cl: Modified Arrott plots for various universality classes including mean field model

Figure S8. NCS-Cl: Modified Arrott plots for various universality classes including mean field model

CS-Br magnetization and susceptibility measurements

Figure S9. CS-Br: (a) M-H at 3K along H∥abH\parallel ab (In-plane orientation) and along H⊥abH\perp ab (Out-of-plane orientation). Inset: Fitting of inverse susceptibility utilizing modified Curie-Weiss law (χ(T)=χ0+CT−TCW\chi(T)=\chi_0+\frac{C}{T-T_{CW}}; χ0\chi_0 is temperaure independent contribution to the susceptibility, C is the Curie constant and TCWT_{CW} is the Curie-Weiss temperature), showing Curie-Weiss temperature TCW∼T_{CW}\sim 39 K and 33 K along in-plane and out-of-plane direction. (b) M-T at H=20 Oe along H∥abH\parallel ab (In-plane orientation) and along H⊥abH\perp ab (Out-of-plane orientation) with inset showing transition temperature TC∼T_C\sim13 K in ac susceptibility χ′(T)\chi'(T) plot along In-plane orientation and along Out-of-plane orientation.

Figure S10. CS-Br: M-H at various temperatures in the transition regime.

References

  1. [1][1–3]
  2. [2][1–3]
  3. [3][1–3]
  4. [4][4–11]
  5. [5][4–11]
  6. [6][4–11]
  7. [7][4–11]
  8. [8][4–11]
  9. [9][4–11]
  10. [10][4–11]
  11. [11][4–11]
  12. [12][12]
  13. [13][13]
  14. [14][14]
  15. [15][15]
  16. [16][16]
  17. [17][17]
  18. [18][18]
  19. [19][19–21]
  20. [20][19–21]
  21. [21][19–21]
  22. [22][22]
  23. [23][23]
  24. [24][24, 25]
  25. [25][24, 25]
  26. [26][26]
  27. [27][26–28]
  28. [28][26–28]
  29. [29][29, 30]
  30. [30][29, 30]
  31. [31]P. Biswal, S. Sarkar, S. N. Sarangi, A. K. Yogi, S. D. Mahanti, G. Tripathy, A. K. Nandy, D. Swain, and D. Samal, Ligand tunability of emergent noncollinear magnetism in cu-based layered hybrid perovskites, Phys. Rev. B 112, 224414 (2025).DOI
  32. [32]K. Taniguchi, P.-J. Huang, H. Sagayama, R. Kiyanagi, K. Ohishi, S. Kitou, Y. Nakamura, and H. Miyasaka, Tuning of spin-orbit coupling in chiral molecule-incorporated two-dimensional organic-inorganic hybrid perovskite copper halides with ferromagnetic exchange interactions, Phys. Rev. Mater. 8, 024409 (2024).DOI
  33. [33]Y. Asensio, S. Marras, D. Spirito, M. Gobbi, M. Ipatov, F. Casanova, A. Mateo-Alonso, L. E. Hueso, and B. Martín-García, Magnetic properties of layered hybrid organic-inorganic metal-halide perovskites: Transition metal, organic cation and perovskite phase effects, Advanced Functional Materials 32, 2207988 (2022).DOI
  34. [34]L. Septiany and G. R. Blake, Magnetocaloric effect and critical behavior in arylamine-based copper chloride layered organic-inorganic perovskite, Journal of Magnetism and Magnetic Materials 542, 168598 (2022).DOI
  35. [35]B. Banerjee, On a generalised approach to first and second order magnetic transitions, Physics Letters 12, 16 (1964).DOI
  36. [36]A. Arrott and J. E. Noakes, Approximate equation of state for nickel near its critical temperature, Phys. Rev. Lett. 19, 786 (1967).DOI
  37. [37]Kumar, A., Biswas, A., & Mudryk, Y. (2025). Nontrivial critical behavior at magnetic transitions: A case study of Sm₇Pd₃. Phys. Rev. B, 112, 224416.
  38. [38]Grzelakowski, K., Lyuksyutov, I., & Bauer, E. (1990). Direct observation of scaling by high-resolution low-energy electron diffraction: O on mo(110). Phys. Rev. Lett., 64, 32.
  39. [39]Baek, D. H., Chung, J. W., & Han, W. K. (1993). Critical behavior of the p(2×1)-o/w(110) system. Phys. Rev. B, 47, 8461.
  40. [40]S. T. Bramwell and P. C. W. Holdsworth, Magnetization and universal sub-critical behaviour in two-dimensional xy magnets, Journal of Physics: Condensed Matter 5, L53 (1993).DOI
  41. [41]Y. O. Kvashnin, A. Bergman, A. I. Lichtenstein, and M. I. Katsnelson, Relativistic exchange interactions in CrX₃ ($x$ = Cl, br, i) monolayers, Phys. Rev. B 102, 115162 (2020).DOI
  42. [42]J. Chang, J. Zhao, and Y. Ding, Anisotropic superexchange through nonmagnetic anions with spin-orbit coupling, The European Physical Journal B 93, 159 (2020).DOI
  43. [43]Qiu, Z. Q., Pearson, J., & Bader, S. D. (1991). Magnetic phase transition of ultrathin fe films on ag(111). Phys. Rev. Lett., 67, 1646.
  44. [44]C. H. Back, C. Würsch, A. Vaterlaus, U. Ramsperger, U. Maier, and D. Pescia, Experimental confirmation of universality for a phase transition in two dimensions, Nature 378, 597 (1995).DOI
  45. [45]Kosterlitz, J. M. (1974). The critical properties of the two-dimensional xy model. Journal of Physics C: Solid State Physics, 7, 1046.
  46. [46]K. Hirakawa, Kosterlitz-thouless transition in two-dimensional planar ferromagnet k2cuf4, Journal of Applied Physics 53, 1893 (1982).DOI

Paper details

vlnc:W_wkrnkzcy
Version
v1 · latest
Published
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