betaFeedback

WorksPhysics and Astronomy

Rare-earth oxysulfides RE2_2O2_2S as model mixed-anion frustrated magnets

Abstract

The study of low-dimensional and anisotropic magnetism is a promising avenue for the discovery of novel quantum phenomena. Mixed-anion materials, defined by the coordination of metal cations by two or more distinct anionic species, provide an intrinsically anisotropic platform for the tuning of structural, magnetic, and electronic properties. In this work, we report the synthesis and characterization of a family of rare-earth oxysulfide (RE2_2O2_2S) antiferromagnets (AFM) possessing a triangular-bilayer slab lattice geometry. Transitions consistent with long-range AFM order are observed for the majority of compositions, with several (RE = Ce, Pr, Nd, Sm) having been previously unreported. Although assumed to form stoichiometrically, pair distribution function (PDF) analysis provides evidence for significant, synthesis-dependent interslab structural disorder as RE2_2O2+x_{2+x}S1−x_{1-x}, resulting in significant variability in the bulk magnetic response of the Nd member. This work highlights the tunability of frustrated magnetic ground states in rare-earth-based mixed-anion materials, and the importance of thorough structural characterization in the discovery of new mixed-anion magnets.

Introduction

Structurally anisotropic materials often possess enhanced magnetic and electrical properties, owing to electronic confinement effects [1]. This anisotropy can be directly engineered via the substitution of entire anionic layers within a crystal structure, creating so-called mixed-anion materials defined by the anisotropic coordination of metal cations with two or more anionic species [2]. Such materials have been largely underexplored, relative to single-anion systems, owing to the increased difficulty in obtaining structurally-ordered samples in a more chemically complex phase space [3, 4]. However, the introduction of this asymmetry affords incredible tunability of overall structural dimensionality, crystal electric field (CEF) energetics, and band gap width while maintaining overall structural rigidity [2].

In the realm of frustrated magnetism, mixed anions provide a route to the controlled design of novel structures with particular connectivity and magnetic dimensionality. For instance, the mineral-inspired quantum spin liquid candidate herbertsmithite, ZnCu3(OH)6Cl2\mathrm{ZnCu_3(OH)_6Cl_2}, achieves highly two-dimensional exchange interactions, which are mediated in-plane by hydroxide ligands, with negligible out-of-plane exchange through the interlayer chlorine [5]. Another example comes from the van der Waals antiferromagnet CrSBr, where the very different chemical characters of the sulfide and bromide anions produce highly anisotropic exchange interactions, yielding a complex magnetic ground state [6, 7]. Mixtures of anions can also be employed to modify cation valence states and thereby alter their magnetic character, as achieved in the topotactic nitridation of a molybdenum oxide pyrochlore to modify its spin from S=1S = 1 to S=1/2S = 1/2 [8].

Particularly in the case of rare-earth-based magnets, the introduction of mixed-anion character enables fine control over the often complex CEF excitation spectrum, which dictates single-ion anisotropy, pseudo-spin state, and magnetic ground state stability. Until recently, literature reports of rare-earth mixed-anion materials have primarily focused on the impact of anisotropic coordination on their structural diversity and optical and luminescent properties [9]. However, several studies have highlighted the potential for interesting magnetism in certain rare-earth oxysulfides, RE2O2SRE_2O_2S (RERE = La – Lu), whose structure consists of staggered triangular-bilayer slabs of RE3+RE^{3+} cations bridged by O2−O^{2-} and S2−S^{2-} anions within and in between each bilayer, respectively [10, 11, 12]. This structure represents a quasi-two-dimensional intermediate between the three-dimensional La2O3\mathrm{La_2O_3}-type and the fully two-dimensional triangular-lattice antiferromagnets (AFMs), as realized in delafossite-type materials, both of which have seen extensive study in recent years due to their ability to stabilize intrinsic magnetic frustration [13, 14].

In this work, we investigate the effect of mixed-anion anisotropy on magnetic ground states across the RE2O2SRE_2O_2S family of AFMs possessing the La2O3\mathrm{La_2O_3}-type structure. In contrast to previous reports, the triangular-bilayer lattice geometry is found to stabilize long-range magnetic order at low temperatures for all but RERE = Eu, Er, and Tm, suggesting a greater degree of structural order in samples grown by the method described in this work. Powder magnetization and specific heat measurements are in good agreement with the calculated CEF spectrum for each composition, revealing a diversity of single-ion anisotropies. The resulting magnetic phase diagrams are strongly influenced by these anisotropies, with many members of the RE2_2O2_2S family exhibiting metamagnetic transitions. We also report strong synthesis-dependence in the bulk magnetic response. Pair distribution function (PDF) analysis of representative Nd2_2O2_2S compositions grown by three different methods suggests that this variability is due to disorder largely localized on the sulfur site, likely through partial oxygen substitution. This work highlights the utility of mixed-anion substitution in tuning single-ion properties, as well as the synthetic limitations of such materials.

Results and Discussion

Structure

Many RE2_2O2_2S synthesis techniques in the literature proceed through the direct reaction of either the rare-earth oxide (RE2_2O3_3) and sulfur (S) or the rare-earth oxide and corresponding sulfide (RE2_2S3_3) in an equimolar ratio [11, 15]. In this work, we employed the reaction of rare-earth metal, rare-earth oxide, and sulfur in a stoichiometric ratio, which was found to result in shorter reaction times and smaller oxide impurities in as-synthesized powder samples. While the A-type La2_2O3_3 structure (P3‾m1P\overline{3}m1 space group, no. 164) is only stable for the largest rare earths (RE = La - Pm) at ambient pressure [16], the substitution of 1/31/3 of coordinating O2−^{2-} anions for S2−^{2-} extends the trigonal phase stability to the full lanthanide block.

The RE2_2O2_2S structure is comprised of hemispherically-coordinated RE3+^{3+} cations arranged in an overall staggered triangular-bilayer slab geometry, a quasi-two-dimensional intermediate between the three-dimensional La2_2O3_3-type structure and the conventional triangular lattice Figure 1a,Figure 1b. Each cation is four-fold coordinated to O2−^{2-} anions within the slab and three-fold coordinated to S2−^{2-} anions between slabs, producing a highly anisotropic coordination environment.

Figure 1. Crystal structure of the RE2_2O2_2S family of mixed-anion magnets, shown (a) within and (b) normal to the triangular-bilayer slab. Although equivalent by symmetry, RE atoms in the upper (REu3+^{3+}_{\mathrm{u}}) and lower (REl3+^{3+}_{\mathrm{l}}) halves of the bilayer slab have been shaded differently for clarity. (c) Representative Rietveld refinement of powder X-ray diffraction (pXRD) data for La2_2O2_2S, where the black, orange, and gray datasets denote the experimental and modeled patterns, and the difference between them. The black and red tick marks represent reflections attributed to La2_2O2_2S and an internal Si standard, respectively. The pXRD patterns for the full lanthanide series are shown in the inset with asterisks marking the Si standard. (d) Refined aa and cc lattice parameters for the trigonal RE2_2O2_2S unit cell as a function of 7-coordinate RE3+^{3+} ionic radius.

Powder X-ray diffraction (pXRD) patterns collected from RE2_2O2_2S powder samples display a progressive rightward shift upon decreasing RE size, indicative of a shrinking of the P3‾m1P\overline{3}m1 (no. 164) unit cell, in good agreement with previous literature reports Figure 1c [17]. Rietveld refinement of the measured pXRD patterns yields an 8.5(1)% and 6.6(1)% decrease in the aa and cc lattice parameters from RE = La - Lu, respectively Figure 1d, as well as a gradual shortening of all unique RE - RE, RE - O, and RE - S bonds Table S2, Table S3. All compositions additionally exhibit a strong, non-uniformly preferred orientation which differs from that reported previously for polycrystalline RE2_2O2_2S and could not be reduced via prolonged grinding, dilution with amorphous SiO2_2, sifting, or increasing the sample volume [18, 19, 20]. These effects are most pronounced (but not limited to) reflections with a large ll component, owing to REz_z and Oz_z being the only refinable positions in the average structure. Air and overall structural stability are observed to decrease across the lanthanide series, with the measured pXRD of the smallest members of the series (RE = Tm - Lu) indicating the presence of RE2_2O3_3 and/or RE2_2S3_3 impurity phases after air exposure beyond 5-10 minutes.

REIonic Radius (Å)aa (Å)cc (Å)RE (1/3,2/3,z)(1/3, 2/3, z)zzBeqB_{\mathrm{eq}}O (1/3,2/3,z)(1/3, 2/3, z)zzBeqB_{\mathrm{eq}}S (0,0,0)(0, 0, 0)BeqB_{\mathrm{eq}}
La1.14.05243(6)6.9466(2)0.2784(3)0.61(6)0.637(2)0.1(5)1.2(3)
Ce1.074.0046(1)6.8757(3)0.2784(5)0.2(1)0.6363(1)0.1(1)
Pr1.0583.97331(5)6.8266(2)0.2796(4)0.1(1)0.636(4)1.5(8)0.1(3)
Nd1.0463.94570(6)6.7878(2)0.2792(4)0.1(1)0.635(4)1.3(8)0.1(4)
Sm1.023.89319(5)6.7152(2)0.2802(3)0.34(5)0.633(2)0.1(5)0.6(3)
Eu1.013.87162(3)6.6858(1)0.2804(3)0.39(7)0.634(2)0.5(5)0.4(3)
Gd1.003.8517(1)6.6648(4)0.2825(8)0.2(1)0.6332.7(2)0.1(1)
Tb0.983.82410(6)6.6277(2)0.2812(3)0.96(5)0.632(2)0.1(4)1.0(3)
Dy0.973.80121(8)6.5971(2)0.2811(3)1.18(6)0.631(2)0.1(5)1.2(4)
Y0.973.78412(9)6.5885(3)0.2805(4)2.5(1)0.629(2)0.1(4)2.0(3)
Ho0.973.7811(1)6.5732(3)0.2822(6)1.1(1)0.636(4)1.9(9)0.1(1)
Er0.9453.76043(3)6.5485(1)0.2822(2)1.21(4)0.639(2)0.3(3)1.6(2)
Tm0.9373.73990(6)6.5210(2)0.2820(4)2.0(1)0.641(4)1.6(9)0.1(1)
Yb0.9253.72236(7)6.5002(2)0.2823(2)1.35(7)0.640(3)1.2(3)2.2(4)
Lu0.9193.7079(1)6.4893(4)0.2816(7)3.3(2)0.641(5)0.1(1)0.1(1)

Table S2. Refined lattice parameters, atomic positions, and isotropic displacement parameters as a function of increasing seven-coordinate, trivalent ionic radius across the RE2_2O2_2S series, derived from Rietveld refinement of pXRD data collected from representative RE2_2O2_2S powder samples. All samples were measured with an internal Si standard to ensure a reliable comparison and refined in the P3ˉm1P\bar{3}m1 (164) space group.

RERE1_1-RE1_1 (Å)RE1_1-RE2_2 (Å)RE-O1_1 (Å)RE-O2_2 (Å)RE-S (Å)RexpR_{\mathrm{exp}}RwpR_{\mathrm{wp}}RpR_{\mathrm{p}}GOF
La4.0524(1)3.867(4)2.412(4)2.49(2)3.036(1)8.8410.528.571.19
Ce4.0046(1)3.825(6)2.386(7)2.46(3)3.002(3)9.5613.2311.471.38
Pr3.9733(1)3.784(5)2.365(7)2.43(3)2.984(2)11.6412.869.701.10
Nd3.9457(1)3.765(5)2.351(7)2.42(3)2.963(2)8.619.367.841.09
Sm3.8932(1)3.710(4)2.322(4)2.37(1)2.931(1)10.4512.579.021.20
Eu3.8716(1)3.690(4)2.307(4)2.36(1)2.917(1)5.806.876.181.18
Gd3.8517(1)3.654(9)2.294(1)2.34(1)2.914(4)16.219.113.91.18
Tb3.8241(1)3.645(4)2.284(4)2.33(1)2.889(1)7.088.336.611.18
Dy3.8012(1)3.627(4)2.270(4)2.31(1)2.873(1)16.9118.7914.181.11
Y3.7841(1)3.625(5)2.265(4)2.30(1)2.862(2)7.949.597.811.21
Ho3.7811(1)3.601(7)2.248(7)2.33(3)2.865(3)11.5314.6111.161.27
Er3.7604(1)3.585(3)2.232(4)2.34(1)2.851(1)6.077.725.901.27
Tm3.7399(1)3.570(5)2.217(6)2.34(3)2.836(2)10.2415.6913.261.53
Yb3.7224(1)3.554(4)2.208(5)2.33(2)2.826(1)4.775.834.801.22
Lu3.7079(1)3.552(8)2.199(8)2.52(4)2.815(3)11.0513.2610.451.20

Table S3. Interatomic distances and goodness of fit parameters derived from Rietveld refinement of pXRD data collected from representative RE2_2O2_2S powder samples. RE-O1_1 and RE-O2_2 respectively denote the distances between RE3+^{3+} cations and O2+^{2+} in the same plane and across the slab. RE1_1-RE1_1 and RE1_1-RE2_2 respectively denote nearest neighbor RE3+^{3+}-RE3+^{3+} distances in the same plane and across the slab.

RE-RE distances in RE2_2O2_2S phases are found to be intermediate between those observed in three-dimensional RE2_2O3_3 and purely two-dimensional AREO2_2 delafossite materials, in agreement with the previous literature [21]. Taking Nd2_2O2_2S as an example, the Nd-Nd distances within each triangular layer (i.e. REu_u-REu_u and REl_l-REl_l in Figure 1a) are 3.9457(1) Å, comparable to the 3.831 Å and 3.622 Å in Nd2_2O3_3 and KNdO2_2, respectively. The distance across the bilayer slab in NdO2_2S (i.e. REu_u-REl_l) is 3.765(5) Å, compared to 3.737 Å in Nd2_2O3_3. Finally, the distance between the slabs is 4.422(5) Å, placing the RE2_2O2_2S structure as a dimensional intermediate between KNdO2_2, with interlayer spacings of 7.506 Å, and Nd2_2O3_3, with an inter-bilayer spacing 3.717 Å, which is actually the shortest RE-RE distance in that structure. The substitution of the interlayer O2−^{2-} in Nd2_2O3_3 with S2−^{2-} results in only slight shifts in Nd-O-Nd bond angles within each triangular layer (106.0° to 104.3(7)°) and across the slab (112.7° to 114.1(6)°), as well as in the non-linear Nd-O/S-Nd bond angles between the slabs (91.7° to 83.5(1)° and 88.3° to 96.5(1)°). The average structural changes observed between the Nd member and Nd2_2O2_2S are assumed to be representative across the rare-earth series.

Magnetic properties

Raman scattering spectra collected at room temperature from polycrystalline RE2_2O2_2S samples also highlight the evolution of the La2_2O3_3-type structure on decreasing RE size (Figure S2). With the exception of the Er and Yb members, all samples display only the four Raman-active optical modes (2A1g_{1g} + 2Eg_g) allowed by the D3d_{3d} point group symmetry of the P3ˉm1P\bar{3}m1 (no. 164) space group. These correspond to RE-O vibrations within the triangular sublattices that comprise the overall bilayer slab, with the Eg_g and A1g_{1g} modes representing vibrations within the abab-plane and along the cc-axis, respectively. All RE-S vibrations are expected to be active only in the far-infrared, limiting spectroscopic determination of potential interslab disorder. The lower-frequency modes, here denoted as Eg1^1_g and A1g1^1_{1g}, are observed to gradually harden from RE = La - Lu. For RE = Ce, Pr, and Tb, these modes exhibit hardening beyond the overall trend. The higher-frequency modes, here denoted as Eg2^2_g and A1g2^2_{1g} instead soften upon decreasing RE size, with only RE = Ce and RE = Eu respectively displaying more significant softening and hardening. All deviations from the trends in vibrational mode frequencies notably occur for rare-earth cations possessing other stable valence states. Anharmonic behavior in materials such as these has been commonly attributed to abnormally strong interactions between phonons and crystal electric field levels. However, this may also arise as a result of higher levels of valence instability-driven disorder [22, 23].

Figure S2. Raman scattering spectra collected at room temperature from representative RE2_2O2_2S powder samples.

To explore the low-temperature magnetism of these rare-earth oxysulfides, we performed temperature-dependent DC magnetic susceptibility, χ(T)\chi(T), and field-dependent DC magnetization, M(H)M(H), measurements on all twelve polycrystalline, magnetic RE2_2O2_2S samples. The supporting data are shown in Figure 2, Figure 3, S3, and S4. Curie-Weiss (CW) fitting in both the high- (TT = 150–300 K) and low-temperature range was used to calculate the effective magnetic moment, μeff\mu_{\mathrm{eff}}, and estimate of net interaction strength, θCW\theta_{\mathrm{CW}}, within the paramagnetic regime and close to the ordered state, respectively (Table I). The θCW\theta_{\mathrm{CW}} values derived from high-temperature CW fitting in lanthanide compounds are not highlighted here, as they often significantly overestimate the net magnetic interaction strength, owing to contributions from thermally populated CEF levels [24, 25]. The frustration index, f=∣θCW∣/TNf = |\theta_{\mathrm{CW}}|/T_N, was also calculated for each composition using the θCW\theta_{\mathrm{CW}} derived from the low-temperature Curie-Weiss fits. In cases where no transition was observed, the lowest measured temperature (1.8 K) was used in place of TNT_N to set a lower bound on ff. Low-temperature fit-derived θCW\theta_{\mathrm{CW}} values more accurately capture interactions in the magnetic ground state than those derived from higher-temperature fits, where the thermal population of CEF levels dominate. However, as low-temperature fitting can still be influenced by short-range magnetic correlations or the population of low-lying CEF levels, the calculated frustration parameters reported here should be considered only as a means of qualitative comparison between oxysulfide compositions. CEF energy level splittings were calculated from a point charge model using PyCrystalField in the JJ basis for the RE2_2O2_2S besides RE = Eu and Gd, where J=0J = 0 and L=0L = 0 were assumed respectively [26].

Figure 2. (a) Representative temperature-dependent magnetic susceptibility (black), its inverse (gold), and Curie Weiss fit lines (dashed black) of a Ce2_2O2_2S powder sample. Scaled low-temperature magnetic susceptibility of representative RE2_2O2_2S powder samples for (b) RE = Ce - Sm, (c) RE = Eu - Dy, and (d) RE = Ho - Yb, highlighting the observed magnetic ordering transitions. (e) Expected and calculated effective magnetic moments (μeff\mu_{\mathrm{eff}}) and Curie-Weiss temperatures (θCW\theta_{\mathrm{CW}}) for each RE2_2O2_2S phase, derived from high- and low-temperature Curie-Weiss fitting of the measured magnetic susceptibility, respectively.

Figure 3. (a) Field-dependent magnetization of representative RE2_2O2_2S powder samples at T=1.8T = 1.8 K. Temperature-dependence of the field-dependent magnetization of (b) Ce2_2O2_2S, (c) Gd2_2O2_2S, (d) Dy2_2O2_2S, and (e) Yb2_2O2_2S.

Compared to the more three-dimensional RE2_2O3_3, magnetic exchange between bilayer slabs in RE2_2O2_2S is expected to uniformly increase, due to the hybridized, more diffuse electron density on each S2−^{2-} anion (J2J_2, J3J_3, and J5J_5 in Figure S1). While the highly-localized nature of the 4ff orbitals on each metal center precludes the clear determination of magnetic exchange strength, this strengthened interslab superexchange and intrinsic structural anisotropy is expected to produce a more anisotropic magnetic response, relative to the fully structurally and magnetically 3D RE2_2O3_3. RE2_2O2_2S compositions are thus expected to represent effective intermediates between the purely 3D and 2D analogues (i.e. KREO2_2). As a result, intermediate-strength magnetic frustration in these materials is likely to arise from both the geometric frustration intrinsic to the 2D triangular lattice and exchange frustration which dominates in 3D RE2_2O3_3.

Magnetic superexchange pathways in the RE$_2$O$_2$S structure type

Figure S1. Magnetic superexchange pathways in the RE2_2O2_2S structure type. J1J_1 and J4J_4 represent exchange through O2−^{2-} 2p2p-orbitals within the bilayer slab, while J2J_2, J3J_3, and J5J_5 represent exchange through interslab extended, hybridized S2−^{2-} orbitals.

Ce2_2O2_2S: In contrast to earlier reports on Ce2_2O2_2S, which claimed that it remains paramagnetic down to T=2T = 2 K [10], we find that Ce2_2O2_2S exhibits a sharp cusp at TN=8.5T_N = 8.5 K in χ(T)\chi(T), indicative of an antiferromagnetic (AFM) ordering transition (Figure 2b, Figure S3). High-temperature CW fitting yields μeff=2.39(2)\mu_{\mathrm{eff}} = 2.39(2) μB\mu_B,:VEVENT in good agreement with the expected μcalc=2.54μB\mu_{\mathrm{calc}} = 2.54 \mu_{\mathrm{B}} for Ce3+^{3+}. Fitting at lower temperature (T=10T = 10–2525 K) yields a reduced moment of μeff=1.99(1)μB\mu_{\mathrm{eff}} = 1.99(1) \mu_{\mathrm{B}} and θCW=−9.6(1)\theta_{\mathrm{CW}} = -9.6(1) K. The difference between the low-temperature and high-temperature paramagnetic moments is consistent with the thermal depopulation of excited crystal-field levels and a ground state Kramers doublet with dominant ∣mJ⟩=∣±3/2⟩\lvert m_J \rangle= \lvert\pm3/2 \rangle, as supported by the calculated CEF ground state shown in Table II. Unlike most of the other RE2_2O2_2S materials studied here, our point charge calculations suggest a well-separated CEF ground state with ΔE0→1=13.3\Delta E_{0 \to1} = 13.3 meV. The measured M(H)M(H) at T=1.8T = 1.8 K for Ce2_2O2_2S displays one sharp inflection around HC=6.4(1)H_{\mathrm{C}} = 6.4(1) T, marking a metamagnetic transition, which gradually broadens upon increasing temperature and is no longer visible at or above TNT_{\mathrm{N}} (Figure 3b). Up to H=14H = 14 T, the magnetization of Ce2_2O2_2S remains unsaturated, only reaching 0.67μB0.67 \mu_{\mathrm{B}}/Ce3+^{3+}, suggestive of strong anisotropy in its ordered state.

REGround state multipletDominant characterΔE0→E1\Delta E_0 \to E_1 (meV)gzg_zg⊥g_\perpμz\mu_z (μB\mu_{\mathrm{B}})μ⊥\mu_\perp (μB\mu_{\mathrm{B}})
CeDoublet∣±32⟩\lvert\pm\frac{3}{2}\rangle13.312.5701.290
PrSinglet∣±3⟩\lvert\pm3\rangle, ∣0⟩\lvert0\rangle0.89----
NdDoublet∣±72⟩\lvert\pm\frac{7}{2}\rangle, ∣±52⟩\lvert\pm\frac{5}{2}\rangle, ∣±12⟩\lvert\pm\frac{1}{2}\rangle5.610.393.450.201.73
SmDoublet∣±52⟩\lvert\pm\frac{5}{2}\rangle, ∣±12⟩\lvert\pm\frac{1}{2}\rangle1.860.630.400.310.20
TbDoublet∣±5⟩\lvert\pm5\rangle, ∣±4⟩\lvert\pm4\rangle, ∣±2⟩\lvert\pm2\rangle, ∣±1⟩\lvert\pm1\rangle0.161.9700.990
DyDoublet∣±132⟩\lvert\pm\frac{13}{2}\rangle, ∣±112⟩\lvert\pm\frac{11}{2}\rangle, ∣±72⟩\lvert\pm\frac{7}{2}\rangle, ∣±52⟩\lvert\pm\frac{5}{2}\rangle, ∣±12⟩\lvert\pm\frac{1}{2}\rangle1.493.149.621.574.81
HoSinglet∣±6⟩\lvert\pm6\rangle, ∣±3⟩\lvert\pm3\rangle, ∣0⟩\lvert0\rangle0.22----
ErDoublet∣±132⟩\lvert\pm\frac{13}{2}\rangle, ∣±112⟩\lvert\pm\frac{11}{2}\rangle, ∣±72⟩\lvert\pm\frac{7}{2}\rangle, ∣±52⟩\lvert\pm\frac{5}{2}\rangle, ∣±12⟩\lvert\pm\frac{1}{2}\rangle2.033.087.251.543.63
TmSinglet∣±6⟩\lvert\pm6\rangle, ∣±3⟩\lvert\pm3\rangle0.76----
YbDoublet∣±72⟩\lvert\pm\frac{7}{2}\rangle, ∣±32⟩\lvert\pm\frac{3}{2}\rangle, ∣±12⟩\lvert\pm\frac{1}{2}\rangle32.633.843.231.921.61

Table II. Calculated ground state multiplet and CEF splitting to the first excited state for anisotropically-coordinated, trivalent rare-earth cations in the RE2_2O2_2S structure.

Figure S3. Magnetic susceptibility as a function of temperature measured for RE2O2S\mathrm{RE}_2\mathrm{O}_2\mathrm{S} powder samples under both zero-field-cooled (ZFC - black) and field-cooled (FC - red) conditions from T=1.8−20T = 1.8 - 20 K.

Pr2_2O2_2S: Pr2_2O2_2S also exhibits a previously unobserved kink in χ(T)\chi(T) at TN=1.5T_{\mathrm{N}} = 1.5 K, below which minor splitting between the zero-field-cooled (ZFC) and field-cooled (FC) measurements is observed (Figure 2b, Figure S3). High-temperature CW fitting yields the expected paramagnetic moment, μeff=3.60(1)μB\mu_{\mathrm{eff}} = 3.60(1) \mu_{\mathrm{B}}, for Pr3+^{3+} (μcalc=3.58μB\mu_{\mathrm{calc}} = 3.58 \mu_{\mathrm{B}}), while lower-temperature fitting from T=5T = 5–1515 K yields a reduced crystal-field ground state moment of μeff=3.07(1)μB\mu_{\mathrm{eff}} = 3.07(1) \mu_{\mathrm{B}} and θCW=−7.1(1)\theta_{\mathrm{CW}} = -7.1(1) K. Our point charge calculation suggests a non-magnetic singlet ground state for Pr3+^{3+} in Pr2_2O2_2S, which is possible for a non-Kramers species. However, the calculated CEF splitting between the singlet ground state and pseudo-doublet first excited state (ΔE0→1=0.89\Delta E_{0 \to1} = 0.89 meV) is quite small, such that the CEF ground state may function as an effective triplet (Table II). The measured M(H)M(H) for Pr2_2O2_2S shows no sign of field-induced magnetic phase transitions and only a gradual magnetization plateau upon increasing field strength down to T=0.4T = 0.4 K (Figure 3a, Figure S4).

Figure S4. Field-dependent magnetization of representative RE2_2O2_2S powder samples at various temperatures from T=1.8T = 1.8 - 30 K.

Nd2_2O2_2S: Nd2_2O2_2S displays a very weak cusp in χ(T)\chi(T) at TN=2.2T_{\mathrm{N}} = 2.2 K, accompanied by minor ZFC-FC splitting upon further cooling (Figure 2b, Figure S3). Although not previously observed down to T=2.0T = 2.0 K [10, 27], the observed ordering transition appears similar to that reported for the structurally-analogous Nd2_2O3_3 at TN=0.55T_{\mathrm{N}} = 0.55 K [13]. High-temperature CW fitting yields the expected paramagnetic moment, μeff=3.64(1)μB\mu_{\mathrm{eff}} = 3.64(1) \mu_{\mathrm{B}}, which agrees well with μcalc=3.62μB\mu_{\mathrm{calc}} = 3.62 \mu_{\mathrm{B}} for Nd3+^{3+} (Table I). Low-temperature (T=5T = 5–1515 K) fitting yields a reduced μeff=3.32(1)μB\mu_{\mathrm{eff}} = 3.32(1) \mu_{\mathrm{B}} and θCW=−17.6(1)\theta_{\mathrm{CW}} = -17.6(1) K, in agreement with a Kramers doublet ground state and weaker frustration than that observed in more three-dimensional Nd2_2O3_3 (Table II) [13]. No field-induced features are visible in the MHM_H for Nd2_2O2_2S, down to T=1.8T = 1.8 K and up to H=14H = 14 T (Figure 3a, Figure S4).

Sm2_2O2_2S: A cusp is observed in the measured χ(T)\chi(T) of Sm2_2O2_2S at TN=3.5T_{\mathrm{N}} = 3.5 K, below which significant ZFC-FC splitting develops (Figure 2b, S3d), in contrast to reports of paramagnetism down to at least T=2T = 2 K [10]. High-temperature CW fitting yields a paramagnetic moment of μeff=1.73(2)μB\mu_{\mathrm{eff}} = 1.73(2) \mu_{\mathrm{B}}, consistent with the expected μcalc=1.70μB\mu_{\mathrm{calc}} = 1.70 \mu_{\mathrm{B}} for Sm3+^{3+} (Table I). Lower-temperature (T=5T = 5–15 K) CW fitting yields μeff=0.80(1)μB\mu_{\mathrm{eff}} = 0.80(1) \mu_{\mathrm{B}}, about half of the expected paramagnetic moment and in line with a reduced CEF ground state moment, and θCW=−16.9(2)\theta_{\mathrm{CW}} = -16.9(2) K, supporting moderate AFM interactions. The measured magnetization of Sm2_2O2_2S at T=1.8T = 1.8 K is the weakest of all RE2_2O2_2S phases, a product of both its highly-reduced moment and its strong single-ion anisotropy, in good agreement with the calculated gg-tensors (Table II). It shows no field-induced features observed up to H=14H = 14 T (Figure 3a, S4d).

Eu2_2O2_2S: The Eu3+^{3+} in Eu2_2O2_2S has a non-magnetic J=0J = 0 Hund’s rules ground state, due to the perfect cancellation of its spin and orbital moments. Consequently, Eu2_2O2_2S exhibits characteristic van Vleck paramagnetic behavior across the full measured temperature range, owing to thermal population of excited JJ states (Figure 2b, S3e), which is not well captured by a conventional CW fitting. As expected for a non-magnetic ground state, only slight curvature is observed in MHM_{\mathrm{H}} at low fields (Figure 3a, S4e).

Gd2_2O2_2S: We observe a sharp cusp in the χ(T)\chi(T) of Gd2_2O2_2S at TN=5.6T_{\mathrm{N}} = 5.6 K, in agreement with the literature [10, 12], with minimal ZFC-FC splitting (Figure 2b, S3f). Since L=0L = 0 for Gd3+^{3+}, CEF effects are negligible and the CW fitting could be performed over a larger temperature range than for the other compositions. Fitting from T=10T = 10–300 K yields μeff=7.95(1)μB\mu_{\mathrm{eff}} = 7.95(1) \mu_{\mathrm{B}} and θCW=−20.1(1)\theta_{\mathrm{CW}} = -20.1(1) K, in agreement with that expected for Gd3+^{3+} (μcalc=7.94μB\mu_{\mathrm{calc}} = 7.94 \mu_{\mathrm{B}}) and supporting the presence of strong AFM exchange interactions (Table I). The MHM_{\mathrm{H}} of Gd2_2O2_2S at T=1.8T = 1.8 K exhibits no field-induced features before plateauing at 6.78μB6.78 \mu_{\mathrm{B}}/Gd3+^{3+} above H=12H = 12 T, in good agreement with the expected saturated moment of μsat=gJJμB=7μB\mu_{\mathrm{sat}} = g_J J\mu_{\mathrm{B}} = 7 \mu_{\mathrm{B}} (Figure 3c).

Tb2_2O2_2S: The χ(T)\chi(T) of Tb2_2O2_2S shows a single cusp at TN=9.1T_{\mathrm{N}} = 9.1 K, the highest observed Néel ordering transition across the RE2_2O2_2S family and in agreement with the literature (Figure 2c, S3g) [10]. High-temperature CW fitting yields μeff=9.78(1)μB\mu_{\mathrm{eff}} = 9.78(1) \mu_{\mathrm{B}}, in good agreement with the expected paramagnetic moment of Tb3+^{3+}, which is only marginally reduced at lower temperature (T=25T = 25–60 K) to μeff=9.20(1)μB\mu_{\mathrm{eff}} = 9.20(1) \mu_{\mathrm{B}} (Table I). The low-temperature fit-derived θCW=−16.9(1)\theta_{\mathrm{CW}} = -16.9(1) K supports the presence of net antiferromagnetic exchange. Point charge calculations support the presence of a doublet ground state, with a low-lying singlet excited state, ΔE0−1=0.16\Delta E_{0-1} = 0.16 meV (Table II). The field-dependent magnetization of Tb2_2O2_2S at T=1.8T = 1.8 K displays a single inflection point at around H=3.2H = 3.2 T before plateauing towards its full saturated moment at higher fields (Figure 3a, S4g).

Dy2_2O2_2S: Dy2_2O2_2S displays a single cusp at TN=6.1T_{\mathrm{N}} = 6.1 K in χ(T)\chi(T) (Figure 2c, S3h), with high-temperature CW fitting yielding a paramagnetic moment of μeff=10.79(1)μB\mu_{\mathrm{eff}} = 10.79(1) \mu_{\mathrm{B}}, in good agreement with the Hund’s rules moment for Dy3+^{3+}, and no significant reduction at lower temperatures (T=25T = 25–60 K), where θCW=−19.3(1)\theta_{\mathrm{CW}} = -19.3(1) K (Table I). This low-temperature moment can be attributed to a Kramers doublet of complex character, as determined by point charge calculations presented in (Table II), as well as the thermal population of a low-lying excited crystal-field state ΔE0−1=1.49\Delta E_{0-1} = 1.49 meV, whose presence is relevant even at the ordering temperature TN=6.1T_{\mathrm{N}} = 6.1 K. The MHM_H of Dy2_2O2_2S at T=1.8T = 1.8 K displays two clear inflection points, centered around H=1.8H = 1.8 T and H=3.6H = 3.6 T, which are no longer visible by T=5.0T = 5.0 K (Figure 3d), pointing to a rich phase diagram for this phase.

REJJFit Range (K)TNT_{\mathrm{N}} (K)μeff\mu_{\mathrm{eff}} (μB\mu_{\mathrm{B}}/RE)μcalc\mu_{\mathrm{calc}} (μB\mu_{\mathrm{B}}/RE)M14 TM_{14\ \mathrm{T}} (μB\mu_{\mathrm{B}}/RE)θCW\theta_{\mathrm{CW}} (K)f=θCWTNf = \frac{\theta_{\mathrm{CW}}}{T_{\mathrm{N}}}χ0\chi_0 (emu/molRE_{\mathrm{RE}}-Oe)
Ce5/25/2150–3008.52.39(2)2.540.67−38(2)–0
10–251.99(1)−9.6(1)1.1(1)0
Pr4150–3001.53.60(1)3.581.66−41.0(7)–−2.8(3) × 10−410^{-4}
5–153.07(1)−7.1(1)4.7(1)−7.8(1) × 10−310^{-3}
Nd9/29/2150–3002.23.64(1)3.621.46−32.5(6)–0
5–153.32(1)−17.6(1)8.0(1)0
Sm5/25/215–3003.51.73(2)1.700.08−208(4)–0
5–150.80(1)−16.9(2)4.8(1)0
Gd7/27/210–3005.67.95(1)7.946.78−20.1(1)3.6(1)−7.0(3) × 10−410^{-4}
Tb6150–3009.19.78(1)9.726.91−31.9(1)–−3.0(1) × 10−310^{-3}
25–609.20(1)−16.9(1)1.9(1)0
Dy15/215/2150–3006.110.79(1)10.657.59−33.6(1)–−3.1(3) × 10−310^{-3}
25–6010.69(1)−19.3(1)3.2(1)−2.0(1) × 10−210^{-2}
Ho8150–3002.810.54(1)10.616.91−11.1(1)–0
25–6010.53(1)−9.7(1)3.5(1)0
Er15/215/2150–300–9.55(1)9.587.07−10.5(1)–−7.0(3) × 10−410^{-4}
25–609.40(1)−7.2(1)>4>4−2.5(1) × 10−310^{-3}
Tm6150–300–7.68(1)7.564.22−24.2(2)–−1.9(3) × 10−310^{-3}
Yb7/27/2150–3003.04.44(1)4.541.82−49.3(1)–0
5–153.09(1)−1.8(1)0.6(1)0

Table 1. Fitting parameters obtained from low- and high-temperature Curie-Weiss fitting of RE2_2O2_2S powder magnetization measurements. M14 TM_{14\ \mathrm{T}} is defined as the magnetization at T=1.8T = 1.8 K in a μ0H=14\mu_0H = 14 T applied field. As described in the text, θCW\theta_{\mathrm{CW}} values in the higher temperature range are excluded due to CEF effects.

Ho2_2O2_2S: The χ(T)\chi(T) of Ho2_2O2_2S shows a single kink at TN=2.8T_{\mathrm{N}} = 2.8 K (Figure 2c, Figure S3), as observed previously [10]. The high-temperature CW fitting yields μeff=10.54(1) μB\mu_{\mathrm{eff}} = 10.54(1)\ \mu_{\mathrm{B}}, in good agreement with the expected paramagnetic moment for Ho3+^{3+}, with no notable reduction in the low-temperature (T=25T = 25–60 K) limit where θCW=−9.7(1)\theta_{\mathrm{CW}} = -9.7(1) K (Table I). The crystal-field ground state of Ho2_2O2_2S is formed by a singlet with a low-lying doublet excited state, where ΔE0−1=0.22\Delta E_{0-1} = 0.22 meV (Table II). The relatively-low θCW\theta_{\mathrm{CW}} suggests the presence of weaker antiferromagnetic interactions in Ho2_2O2_2S, compared to most of the other RE2_2O2_2S phases reported here. The field-dependent magnetization of Ho2_2O2_2S at T=1.8T = 1.8 K displays two weak inflection points, centered around H=0.6H = 0.6 T and H=2.2H = 2.2 T, before plateauing at higher fields (Figure 3a, Figure S4).

Er2_2O2_2S: Er2_2O2_2S is observed to display paramagnetic behavior in χ(T)\chi(T) down to T=1.8T = 1.8 K, just above the onset of short-range order reported at T=1.5T = 1.5 K (Figure 2c, Figure S3) [17, 28]. Er2_2O2_2S is therefore the only RE2_2O2_2S phase without observed long-range order, excluding phases with non-magnetic singlet ground states. The high-temperature CW fit yields a paramagnetic moment of μeff=9.55(1) μB\mu_{\mathrm{eff}} = 9.55(1)\ \mu_{\mathrm{B}}, which again is not significantly reduced at lower-temperatures (T=25T = 25–60 K). Er2_2O2_2S also continues the trend started with Ho2_2O2_2S, showing a significantly reduced θCW=−7.2(1)\theta_{\mathrm{CW}} = -7.2(1) K (Table I), indicating weakened antiferromagnetic exchange interactions. The Kramers doublet ground state for Er3+^{3+} has a small CEF splitting to the first excited state ΔE0−1=2.03\Delta E_{0-1} = 2.03 meV (Table II). As expected for a material in its paramagnetic phase, the MHM_H of Er2_2O2_2S at T=1.8T = 1.8 K exhibits no field-induced features up to H=14H = 14 T (Figure 3a, Figure S4).

Tm2_2O2_2S: The χ(T)\chi(T) of Tm2_2O2_2S shows a largely paramagnetic response, with the exception of a very broad hump centered around T=10T = 10 K, which was not apparent in the measured χ(T)\chi(T) for samples containing small Tm2_2O3_3 impurities. This can likely be attributed to thermal depopulation of the CEF into the singlet, non-magnetic ground state, as has been observed in other Tm-based materials (Figure 2c, Figure S3) [29], followed by a weak Curie tail at lower temperature arising from paramagnetic impurities. The calculated crystal-field ground state of Tm3+^{3+} is a non-magnetic singlet, with a splitting to the first excited state of ΔE0−1=0.76\Delta E_{0-1} = 0.76 meV (Table II). We can therefore speculate that the changing curvature in the χ(T)\chi(T) of Tm2_2O2_2S is due to the thermal depopulation of these excited crystal-field states as Tm3+^{3+} effectively loses its moment. High-temperature

CW fitting yields μeff=7.68(1)μB\mu_{\mathrm{eff}} = 7.68(1) \mu_{\mathrm{B}}, which agrees well with the expected paramagnetic value (Table I), however lower-temperature fitting could not be reliably performed. The field-dependent magnetization of Tm2_2O2_2S at T=1.8T = 1.8 K displays only a gradual increase up to H=14H = 14 T (Figure 3a, Figure S4). The increasing moment can be explained by Zeeman splitting of the CEF states, changing the population of the excited CEF levels and thereby enhancing the paramagnetic moment.

Yb2_2O2_2S: Yb2_2O2_2S is observed to display a sharp cusp in χ(T)\chi(T) at TN=3.0T_{\mathrm{N}} = 3.0 K, in agreement with Quezel et. al. (Figure 2c, Figure S3) [10]. High-temperature CW fitting yields μeff=4.44(1)μB\mu_{\mathrm{eff}} = 4.44(1) \mu_{\mathrm{B}}, in agreement with the μcalc=4.54μB\mu_{\mathrm{calc}} = 4.54 \mu_{\mathrm{B}} expected for Yb3+^{3+} (Table I). However, low-temperature (T=5T = 5–15 K) fitting yields a reduced moment of μeff=3.09(1)μB\mu_{\mathrm{eff}} = 3.09(1) \mu_{\mathrm{B}} and a very low θCW=−1.8(1)\theta_{\mathrm{CW}} = -1.8(1) K, which is lower than TNT_{\mathrm{N}} and a continuation of the trend that started with Ho2_2O2_2S of decreasing CW temperatures. Yb2_2O2_2S is notable for having a large CEF splitting, ΔE0−1=32.6\Delta E_{0-1} = 32.6 meV (Table II). The MHM_H of Yb2_2O2_2S at T=1.8T = 1.8 K displays two field-induced features at H=0.2H = 0.2 T and H=2.3H = 2.3 T, before plateauing at relatively low fields (Figure 3e).

The variability in the observed magnetic response across the RE2_2O2_2S series indicates a diversity of largely AFM-ordered ground states and field-induced metamagnetic phase transitions. This is also reflected in our CEF calculations, which suggest a wide range of ground state degeneracies, ranging from non-magnetic singlets (RE = Pr, Ho, and Tm) to well-separated doublets and effective triplets, as well as three major classes of single-ion anisotropies - Ising-like for RE = Ce and Tb, XY-like for RE = Nd, and more isotropic for RE = Sm, Dy, Er, and Yb (Table II). Ce2_2O2_2S is particularly notable, as its predominantly mJ=±32m_J = \pm\frac{3}{2} ground state doublet has dipolar-octupolar character. The transitions suggestive of long-range order at low temperatures for both Pr2_2O2_2S and Ho2_2O2_2S additionally highlights the relevance of ground state mixing with low-lying excited states. Understanding the nature of magnetic ground states across the RE2_2O2_2S family will then require more thorough consideration of single-ion physics.

The observation of previously unreported transitions consistent with long-range AFM ordering for RE = Ce, Pr, Nd, and Sm, as well as sensible experimentally derived paramagnetic moments for all compositions, suggests a higher degree of structural order in RE2_2O2_2S samples grown by the method described in this work. Both Ce and Nd compositions are observed to order higher in temperature and with smaller calculated frustration parameters (using either low- or high-temperature-derived θCW\theta_{\mathrm{CW}}) than in the isostructural, more three-dimensional Ce2_2O3_3 (TN=6.2T_{\mathrm{N}} = 6.2 K, f=9.2(1)f = 9.2(1) [30]) and Nd2_2O3_3 (TN=0.55T_{\mathrm{N}} = 0.55 K, f=18.5(1)f = 18.5(1) [13]). This trend appears to originate from the overall larger distances between RE3+^{3+} moments as well as the greater difference between RE3+^{3+} distances within and across slabs, effectively reducing the degeneracy of equivalent interactions. Interestingly, ff is smallest for Ce2_2O2_2S and Yb2_2O2_2S, both of which have the largest CEF splitting to the first excited state and would thus be expected to possess less classical, more reliably Seff=1/2S_{\mathrm{eff}} = 1/2 ground states.

Despite the broad air stability of these materials described in previous reports and negligible changes observed in the measured pXRD over time, nearly all compositions showed significant changes in the measured χ(T)\chi(T) upon prolonged exposure to air. This magnetic disorder typically manifested as a broadening of the observed magnetic ordering transition, greater ZFC-FC splitting, and larger calculated effective magnetic moments and Curie-Weiss temperatures. With the exception of Ce2_2O2_2S, where irreversible oxidation to the mixed-valent Ce2_2O2.5_{2.5}S was observed beyond 30 minutes of air exposure [31], re-annealing of each sample with excess sulfur was found to gradually recover the magnetic response of the as-synthesized material, suggesting that air exposure promotes the loss of sulfur from the RE2_2O2_2S structure.

Specific heat

To further investigate the magnetic ground states of the rare-earth oxysulfides, RE2_2O2_2S, specific heat as a function of temperature was measured for several compositions (RE = Ce, Gd, Dy, Yb). The selected compositions were chosen to represent a range of anisotropies and, by extension, likely magnetic ground states. In each case, the specific heat was measured from T=2T = 2–300300 K and observed to reach a value close to the Dulong Petit limit (3nR3nR) close to room temperature. The magnetic contribution to the specific heat was isolated through the subtraction of an appropriate lattice standard, either La2_2O2_2S or Lu2_2O2_2S for the early and latter rare earths, respectively, which was scaled by a correction for the molar mass, as shown for Ce2_2O2_2S in Figure 4a. Numerical integration of the low-temperature magnetic heat capacity, CP,mag/TC_{\mathrm{P,mag}}/T, was used to determine the entropy release, ΔSmag\Delta S_{\mathrm{mag}} associated with the magnetic ordering transition.

Figure 4. a) Zero-field specific heat as a function of temperature for Ce2_2O2_2S (brown), La2_2O2_2S (black line), and their difference (gold). (b) Zero-field magnetic specific heat divided by temperature as a function of temperature for Ce2_2O2_2S (gold), Gd2_2O2_2S (red), Dy2_2O2_2S (purple), and Yb2_2O2_2S (black) powder samples. (c) Integrated magnetic entropy rise as a function of temperature for each composition up to T=50T = 50 K.

Ce2_2O2_2S: The specific heat of Ce2_2O2_2S, shown in Figure 4a, shows a single lambda-like anomaly at TN=8.5T_{\mathrm{N}} = 8.5 K, in agreement with the ordering transition observed in χ(T)\chi(T). Scaled subtraction of the measured specific heat of the La-analogue and subsequent integration of CP,mag/TC_{\mathrm{P,mag}}/T yields an initial rise in magnetic entropy, ΔSmag\Delta S_{\mathrm{mag}}, of roughly 80% of the Rln⁡(2)R \ln(2) associated with a ground state doublet up to T=10T = 10 K, and reaches 100% just above 20 K (Figure 4b,Figure 4c). This suggests that significant short-range correlations may develop well above TNT_{\mathrm{N}}, as is often observed in frustrated antiferromagnets. A gradual increase is then observed up to ΔSmag=14.81\Delta S_{\mathrm{mag}} = 14.81 J · molCe3+−1^{-1}_{\mathrm{Ce}^{3+}} K−1^{-1}, close to the Rln⁡(6)=14.90R \ln(6) = 14.90 J · mol−1^{-1} K−1^{-1} expected for a J=5/2J = 5/2 system by T=300T = 300 K, and supporting the steady population of CEF levels up to ambient temperature (Figure S5).

Figure S5. Specific heat as a function of temperature for (a) Ce2_2O2_2S (gold), (b) Gd2_2O2_2S (maroon), (c) Dy2_2O2_2S (purple), and (d) Yb2_2O2_2S (black) powder samples, as well as their respective non-magnetic analogues (gray). The estimated magnetic contribution to the total specific heat for each composition is shown in red. (e) Integrated magnetic entropy rise as a function of temperature for each composition up to T=300T = 300 K.

Gd2_2O2_2S: The specific heat of Gd2_2O2_2S is highly distinctive, containing a sharp anomaly at TN=5.6T_{\mathrm{N}} = 5.6 K, followed by a second, broader anomaly at T=2.2T = 2.2 K [32]. Close comparison with the susceptibility data shows that this second anomaly is coincident with a minor slope change in χ(T)\chi(T) (Figure 4b). This additional feature was reproducible across Gd2_2O2_2S sample batches and measurements, suggesting the presence of an intrinsic lower-temperature moment reorientation below TNT_{\mathrm{N}}. Despite the largely linear shift observed in vibrational frequencies from RE = La – Lu, the scaled CP,totalC_{\mathrm{P,total}} for La2_2O2_2S yielded convergence with the Gd curve before surpassing it above T=50T = 50 K, whereas the scaled Lu dataset failed to converge with the Gd analogue up to T=300T = 300 K. This could be due to larger changes in O/S ordering between even the best-ordered samples of each composition. A weighted average of the two scaled CP,totalC_{\mathrm{P,total}} for La2_2O2_2S and Lu2_2O2_2S, 44 La : 56 Lu, was ultimately found to best approximate the phonon contribution to the measured CP,totalC_{\mathrm{P,total}} of Gd2_2O2_2S. By T=50T = 50 K, ΔSmag\Delta S_{\mathrm{mag}} for this system peaks at 14.59 J · molGd3+−1^{-1}_{\mathrm{Gd}^{3+}} K−1^{-1}, much lower than the Rln(8) expected for Gd3+^{3+} with S=7/2S = 7/2, before plateauing up to higher temperatures (Figure 4c, Figure S5). As the lower-temperature anomaly remains far from reaching zero down to T=1.8T = 1.8 K, the significant remaining entropy must be released below it.

Dy2_2O2_2S: The specific heat of Dy2_2O2_2S, shown in Figure 4b, shows a well defined lambda anomaly centered around T=5.1(2)T = 5.1(2) K, in agreement with the ordering transition observed in χ(T)\chi(T) (Figure 4b). Above the ordering anomaly, we can notice significant additional entropy release up to T=20T = 20 K by comparison with the other measured RE2_2O2_2S samples, which we can attribute to the presence of a low-lying first excited CEF level (ΔE0→1=1.49\Delta E_{0 \to1} = 1.49 meV. Similar to the Gd member, the scaled subtraction of the CP,totalC_{P,\mathrm{total}} for either La2_2O2_2S and Lu2_2O2_2S alone failed to yield a sensible trend in ΔSmag\Delta S_{\mathrm{mag}} for the Dy member. A weighted average of the two scaled datasets, 35 La : 65 Lu, was ultimately found to reasonably approximate the phononic contribution to CP,totalC_{P,\mathrm{total}}. Although the increase in ΔSmag\Delta S_{\mathrm{mag}} is gradual over the measured temperature range, a slope change is observed around T=7T = 7 K, just above TNT_N for the system and coinciding with R ln(2), the expected ΔSmag\Delta S_{\mathrm{mag}} for a ground state doublet. ΔSmag\Delta S_{\mathrm{mag}} gradually rises over the measured temperature range, consistent with the steady population of CEF levels, before saturating around 21.9 J molDy3+−1 K−121.9\ \mathrm{J\,mol^{-1}_{\mathrm{Dy^{3+}}}\,K^{-1}} at T=300T = 300 K, close to the 23.1 J mol−1 K−123.1\ \mathrm{J\,mol^{-1}\,K^{-1}} expected for a J=15/2J = 15/2 system (Figure S5).

Yb2_2O2_2S: The specific heat of Yb2_2O2_2S shows only one sharp lambda anomaly at T=2.5T = 2.5 K, close to the χ(T)\chi(T)-derived TN=3.0T_N = 3.0 K (Figure 4b). Scaling of the CP,totalC_{P,\mathrm{total}} for Lu2_2O2_2S provides nearly exact agreement between the datasets from T=15T = 15–5555 K, below which integration yields ΔSmag=4.06 J molYb3+−1 K−1\Delta S_{\mathrm{mag}} = 4.06\ \mathrm{J\,mol^{-1}_{\mathrm{Yb^{3+}}}\,K^{-1}} (Figure 4c). This value is significantly lower than the R ln (2) for a Seff=1/2S_{\mathrm{eff}} = 1/2 Kramers doublet ground state, corresponding to only two-thirds of the expected entropy release. This suggests that there may be additional entropy release below the measured temperature range or that there is a dynamic component to the state realized below TN=3.0T_N = 3.0 K. The negligible CmagC_{\mathrm{mag}} in the intermediate temperature regime nicely reflects the large CEF splitting to the first excited state highlighted in Table II. Above this region, ΔSmag\Delta S_{\mathrm{mag}} gradually increases to 9.40 J molYb3+−1 K−19.40\ \mathrm{J\,mol^{-1}_{\mathrm{Yb^{3+}}}\,K^{-1}}, well below the expected R ln(8) = 17.28 J mol−1 K−117.28\ \mathrm{J\,mol^{-1}\,K^{-1}} for the full S=7/2S = 7/2 Yb3+^{3+} moment (Figure S5).

Strong disorder-dependence

In the course of this study, we observed significant and previously-unreported sample dependence in the measured magnetization of each oxysulfide material. This sample dependence manifests as deviations from the expected paramagnetic moment, μeff\mu_{\mathrm{eff}}, and variations in the CW temperature, θCW\theta_{\mathrm{CW}}, as determined by Curie-Weiss fits in the high-temperature limit. To identify the origin of this sample dependence, we grew the Nd member by three variations on the standard solid-state method. The first proceeded by the annealing of stoichiometric quantities of each precursor, followed by subsequent regrindings and reannealings until Nd2_2O3_3 impurities no longer appeared in the measured pXRD pattern (hereafter referred to as Sample B). The second continued in the same manner with a 5% molar excess of sulfur added between each annealing until the calculated effective magnetic moment, derived from high-temperature CW fitting of the measured magnetic susceptibility, approached the expected 3.62 μB\mu_B/Nd3+^{3+} (Sample A). The third method proceeded with the inclusion of a 0.5 mmol chunk of aluminum metal in an alumina crucible sealed above the other reagents in the same quartz tube (Sample C). Here, the aluminum acted as both an oxygen- and sulfur-getter.

Pronounced differences are observed in the measured magnetic susceptibility of each sample, shown in Figure 5a. Both Sample A and Sample B display a slight kink at TN=2.3T_N = 2.3 K and no ZFC-FC splitting, whereas Sample C exhibits a broader kink centered at TN=2.2T_N = 2.2 K and ZFC-FC splitting up to T=150T = 150 K. Curie-Weiss fitting of the inverse susceptibility from T=150T = 150–300300 K reveals a gradual increase in the degree of deviation from Curie-Weiss behavior with presumed increasing disorder. The calculated effective magnetic moment and Curie-Weiss temperature for Sample A are largely in agreement with literature values [μeff=3.62(1) μB\mu_{\mathrm{eff}} = 3.62(1)\ \mu_B/Nd3+^{3+}, θCW=−22.8(6)\theta_{\mathrm{CW}} = -22.8(6) K], while both values increase for Sample B [μeff=4.04(1) μB\mu_{\mathrm{eff}} = 4.04(1)\ \mu_B/Nd3+^{3+}, θCW=−58.6(6)\theta_{\mathrm{CW}} = -58.6(6) K] and Sample C [μeff=4.41(1) μB\mu_{\mathrm{eff}} = 4.41(1)\ \mu_B/Nd3+^{3+}, θCW=−77.6(6)\theta_{\mathrm{CW}} = -77.6(6) K] (Table III). Fitting at low temperatures (T=5T = 5–1515 K) instead yields good agreement between Sample A (μeff=3.41(1) μB\mu_{\mathrm{eff}} = 3.41(1)\ \mu_B/Nd3+^{3+} and θCW=−17.3(1)\theta_{\mathrm{CW}} = -17.3(1) K) and the more disordered compositions, suggesting that a larger effective paramagnetic moment does not necessarily indicate significant changes to the low-temperature magnetism. As observed across all RE2_2O2_2S compositions, the further annealing of Sample B with additional sulfur results in both the expected μeff\mu_{\mathrm{eff}} for Nd3+^{3+} and a more sensible θCW\theta_{\mathrm{CW}} for a low-temperature, rare-earth antiferromagnet.

Figure 5. (a) Inverse magnetic susceptibility as a function of temperature, (b) powder diffraction patterns, and (c) Raman scattering spectra collected from Nd2_2O2_2S powder samples A (gold), B (magenta), and C (purple). The stars in (a) denote peaks attributed to an internal Si standard and the solid black lines in (c) represent the high-temperature Curie-Weiss fit for each sample. (d) Reduced pair distribution function [G(r)] collected from each Nd2_2O2_2S powder sample. Experimental datasets are shown in black, theoretical fits are colored by sample as in (a)-(c), and the difference curve between them is shown in gray.

SampleTNT_{\mathrm{N}} (K)Fit Range (K)μeff\mu_{\mathrm{eff}} (μB\mu_{\mathrm{B}}/Nd3+^{3+})θCW\theta_{\mathrm{CW}} (K)χo\chi_{\mathrm{o}} (10−410^{-4})
A2.3150–3003.62(1)−22.8(6)0
B2.3150–3004.04(1)−58.6(6)−8.2(3)
C2.2150–3004.41(1)−77.6(6)0

Table III. Fitting parameters obtained from Curie-Weiss analysis of Nd2_2O2_2S powder magnetization measurements.

In contrast, room-temperature pXRD patterns measured from each sample, self-normalized to the maximum intensity and shown in Figure 5b, appear nearly identical, with only a slight broadening of reflections observed for Sample C, relative to Sample A and Sample B. The preferred orientation observed across the RE2_2O2_2S series also presents itself across all three samples, with Rietveld refinement yielding comparable lattice parameters and isotropic displacement parameters. Representative room-temperature Raman scattering spectra collected from each sample, which should be highly-sensitive to the wider distribution of bond lengths in a more disordered material, also appear identical across the three spectra, with the exception of a small hump attributed to a minimal Nd2_2O3_3 impurity in Sample A (Figure 5c).

This minimal variation in bands arising from Raman-active Nd-O vibrations suggests that sulfur content plays a significant role in determining the physical properties of RE2_2O2_2S materials, likely through intrinsic sulfur deficiencies, sulfur-oxygen site mixing, or Nd3+^{3+} - S2−^{2-} Schottky defects [11]. Interestingly, relative to Sample A, the calculated μeff\mu_{\mathrm{eff}} for Samples B and C are larger than could be attributed to changes in the effective molar mass by any of these disordering mechanisms while maintaining overall structural stability, suggesting that some additional mechanism must be considered. To further probe the structural origin of the magnetic disorder in this system, synchrotron pair distribution function (PDF) analysis was performed on the three representative samples. The reduced PDF [G(r)] data shown in Figure 5d, depicting the probability of finding an atom some distance rr from another in the Nd2_2O2_2S structure, appears quite similar between samples, similar to the pXRD data. Refinement of the three datasets against the average Nd2_2O2_2S structure yields nearly identical lattice parameters and atomic positions for each, within uncertainty (Table IV). However, refinement of both the oxygen and sulfur site occupancies suggests negligible sulfur incorporation on the oxygen site, in agreement with the uniformity of our Nd-O vibrational spectra, and significant oxygen substitution on the sulfur site coincident with the gradual increase in magnetic disorder (Figure S6). Sample A is found to possess up to 37 ±\pm 16% oxygen substitution for sulfur on the interslab anion site, greater than previously reported for La2_2O2_2S [33], while Sample B and Sample C appear even more oxygen-rich with 58 ±\pm 17% and 78 ±\pm 15% substitution, respectively. Although likely exaggerated relative to the true level of substantial disorder, these values are in general agreement with the previous assumption of structural instability for a fully stoichiometric RE2_2O2_2S phase [11]. Rare-earth oxysulfides possessing the triangular-bilayer slab geometry can then be best described by RE2_2O2+x_{2+x}S1−x_{1-x}, where the trigonal structure is stable for all but x≤0.10x \leq0.10 under standard synthesis conditions.

Sample ASample BSample C
aa (Å)3.944(2)3.944(2)3.945(2)
cc (Å)6.788(4)6.787(5)6.787(5)
Nd1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zU11U_{11}0.2803(6)0.2800(6)0.2802(6)
Nd1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zU11U_{11}0.0035(2)0.0033(2)0.0031(2)
Nd1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zU33U_{33}0.0028(5)0.0026(5)0.0026(4)
O1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zzz0.632(6)0.633(5)0.633(5)
O1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zU11U_{11}0.013(4)0.014(4)0.018(5)
O1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zU33U_{33}0.006(7)0.002(5)0.007(4)
O1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zO occ.0.99(8)0.98(8)0.99(8)
O1 (2d2d)13,23,z\frac{1}{3},\frac{2}{3},zS occ.0.01(8)0.02(8)0.01(8)
S1 (1a1a)0, 0, 0U11U_{11}0.007(2)0.008(3)0.005(2)
S1 (1a1a)0, 0, 0U33U_{33}0.004(4)0.002(4)0.001(3)
S1 (1a1a)0, 0, 0S occ.0.63(16)0.42(17)0.22(15)
S1 (1a1a)0, 0, 0O occ.0.37(16)0.58(17)0.78(15)

Table IV. Fitting parameters obtained from PDF analysis of Nd2_2O2_2S powder samples.

Figure S6. Simulated reduced pair distribution function [G(r)] for stoichiometric Nd2_2O2_2S (black) and Nd2_2O2.78_{2.78}S0.22_{0.22} (red) and their difference (gray), highlighting the interatomic distances most impacted by significant oxygen substitution on the sulfur site. The dotted black line represents the zero point of the difference curve (gray).

Attempts to quantitatively assess the compositional differences between the three Nd2_2O2_2S samples via energydispersive X-ray spectroscopy (EDS) measurements resulted in significant beam-induced decomposition. Similar effects, including the removal of sulfur from the sample surface under vacuum conditions, have been previously observed in several studies of rare-earth oxysulfides, even at beam energies as low as 1 keV [34, 35]. As such, future studies would benefit from the use of x-ray absorption spectroscopy techniques to probe the true extent of disorder on the sulfur sublattice at the sulfur K-edge and accessible RE edges.

The stability of the triangular bilayer slab geometry across nearly the full range of Nd2_2O2+x_{2+x}S1−x_{1-x} stoichiometries naturally raises the question of compositional variability in these and other RE2_2O2_2S samples synthesized by different methods [15]. Rare-earth oxysulfides have historically been grown by a wide variety of techniques, including by more diffusion-limited solid-state methods than those reported in this work [17, 18], or by more disorder-prone solution- and gas-phase decomposition and sulfurization processes [10, 12, 21, 36]. However, despite this thermodynamic tendency toward structural disorder, studies have thus far largely focused only on its impact on luminescent properties [37]. With this in mind, we can reexamine the new magnetic transitions observed in our compositions. In the case of RE = Pr and Nd, both transitions are fairly weak and occur close to the base temperature of most commercial cryostats. In previously measured samples, which may have had more substantial disorder, these transitions could have been suppressed below the window of measurement. By contrast, the higher-temperature, more pronounced transitions and increased air-sensitivity observed for our RE = Ce and Sm samples can likely be attributed to greater deviations from the ideal stoichiometry in previous reports, due to valence instability. Ce3+^{3+} and Sm3+^{3+} are particularly prone to oxidation and reduction to non-magnetic Ce4+^{4+} and Sm2+^{2+}, respectively, both of which could dilute the magnetic sublattice enough to destabilize long-range magnetic order.

Compared to the oxysulfides, other RE2_2O2_2X phases are generally expected to be less prone to substitutional disorder on the X sublattice, due to the greater difference in ionic radius and polarizability between oxygen and the heavier chalcogenides. Where more similar, such as in the rare-earth oxyfluorides (REOF), significant synthesis-dependence has been observed in both overall structural stability and defect formation [38]. Beyond monatomic anions, even more pronounced interslab disorder has been introduced via the introduction of longer, more flexible polyanionic species, such as cyanamide and carbonate in the isostructural RE2_2O2_2CN2_2 and RE2_2O2_2CO3_3 families, respectively [39, 40]. When considered alongside the broader class of mixed-anion materials, this work serves as an additional reminder of the importance of synthesis technique and local probe structural characterization in the exploration of new heteroanionic magnetic materials.

Conclusion

In conclusion, we present the synthesis and study of a series of magnetically-frustrated, mixed-anion rare-earth oxysulfides. The unique bilayer slab lattice geometry serves as a structural intermediate between the purely 2D triangular monolayer and 3D La2_2O3_3-structure type, giving rise to quasi-2D magnetism and intermediate frustration. Magnetic susceptibility measurements reveal several previously-unobserved transitions consistent with long-range AFM order at low-temperatures, as well as a variety of field-induced metamagnetic transitions. Across the RE2_2O2_2S series, a diversity of magnetic ground states are observed, owing to strong single-ion anisotropy in this highly anisotropic structure. Zero-field specific heat measurements and point charge calculations provide additional insight into the unique CEF excitation spectrum produced by the anisotropic coordination of the rare-earth cation, with the Gd and Yb compositions being particularly worthy of further study, due to the significant magnetic entropy remaining below T=1.8T = 1.8 K. We also establish a tendency towards sample dependence across the RE2_2O2_2S series, which we attribute to substitutional anionic disorder on the sulfur site, even for samples displaying minimal signs of magnetic disorder. This work demonstrates the broad tunability of magnetic states enabled by the introduction of mixed-anion character into rare-earth AFMs, as well as the importance of often overlooked structural disorder in the synthesis of new magnetic materials.

Supporting Information

Rietveld refinement tables, Raman scattering spectra, and additional magnetic susceptibility, magnetization, specific heat, and simulated reduced pair distribution function [G(r)] data.

Acknowledgments

The authors thank J. Dadap and Z. Ye for technical assistance in the collection of the Raman scattering spectra. This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canadian Institute for Advanced Research (CIFAR), the Sloan Research Fellowships program, and the Killam Accelerator Research Fellowship. Part or all of the research described in this paper was performed at the Canadian Light Source, a national research facility of the University of Saskatchewan, which is supported by the Canada Foundation for Innovation (CFI), the Natural Sciences and Engineering Research Council (NSERC), the Canadian Institutes of Health Research (CIHR), the Government of Saskatchewan, and the University of Saskatchewan. The identification of any commercial product or trade name does not imply endorsement or recommendation by the National Institute of Standards and Technology.

Supplementary Materials for: Rare-earth oxysulfides RE2O2SRE_2O_2S as model mixed-anion frustrated magnets

Austin M. Ferrenti; Ksenia Khoroshun; Mohamed Oudah; Graham King; Jonathan Gaudet; Alannah M. Hallas

Supplementary Materials for: Rare-earth oxysulfides RE2O2SRE_2O_2S as model mixed-anion frustrated magnets

Methods

Powder synthesis. Polycrystalline RE2O2SRE_2O_2S samples were grown by the solid state synthesis method. RE metal, RE oxide, and sulfur (as S - Puratronic, 99.999%, repurified by low-temperature annealing with charcoal at 100°C) were combined in a 2:2:3 (RE = La, Nd-Gd, Ho-Lu), 1:1:1 (RE = Ce), or 6:2:7 (RE = Tb) molar ratio, for a total input mass of 1.0g, and sealed under vacuum in a 14 mm ID x 16 mm OD x 10 cm long quartz ampoule (Table S1). For RE = La - Er, the samples were heated to 250°C over 5 hours, held for 5 hours, heated to 500°C over 5 hours, held for 5 hours, heated to 900°C over 5 hours, and held for 48 hours before cooling to room temperature at 100°C per hour. Samples were then reground in an argon-filled glovebox, sealed in new quartz tubes under vacuum, and reheated by the same profile until rare-earth oxide impurity phases were no longer visible in the measured powder X-ray diffraction patterns. Growth of the RE = Tm - Lu and Y phases proceeded by the same method, albeit with a maximum dwell temperature of 1100°C and slower cooling rate of 50°C per hour to force sulfur incorporation.

RE metalRE metal sourceRE oxideRE oxide source
LaSmart Elements, 99.9%La2O3\mathrm{La}_2\mathrm{O}_3Alfa Aesar, Reacton, 99.99%
CeSmart Elements, 99.9%CeO2\mathrm{CeO}_2Alfa Aesar, Reacton, 99.99%
PrSmart Elements, 99.9%Pr6O11\mathrm{Pr}_6\mathrm{O}_{11}Alfa Aesar, Reacton, 99.9%
NdSmart Elements, 99.9%Nd2O3\mathrm{Nd}_2\mathrm{O}_3Thermo Scientific, 99.9%
SmSmart Elements, 99.99%Sm2O3\mathrm{Sm}_2\mathrm{O}_3Fisher, 99.9%
EuSmart Elements, 99.9%Eu2O3\mathrm{Eu}_2\mathrm{O}_3Alfa Aesar, Reacton, 99.9%
GdSmart Elements, 99.9%Gd2O3\mathrm{Gd}_2\mathrm{O}_3Rare Earth Products(?), 99.999%
TbSmart Elements, 99.95%Tb4O7\mathrm{Tb}_4\mathrm{O}_7Alfa Aesar, 99.99%
DyAlfa Aesar, 99.9%Dy2O3\mathrm{Dy}_2\mathrm{O}_3Alfa Aesar, 99.9%
HoSmart Elements, 99.99%Ho2O3\mathrm{Ho}_2\mathrm{O}_3Alfa Aesar, Reacton, 99.99%
YAlfa Aesar, 99.98%Y2O3\mathrm{Y}_2\mathrm{O}_3Alfa Aesar, 99.999%
ErSmart Elements, 99.95%Er2O3\mathrm{Er}_2\mathrm{O}_3Rare Earth Products(?), 99.99%
TmSmart Elements, 99.9%Tm2O3\mathrm{Tm}_2\mathrm{O}_3*
YbSmart Elements, 99.99%Yb2O3\mathrm{Yb}_2\mathrm{O}_3Alfa Aesar, Reacton, 99.9%
LuSmart Elements, 99.99%Lu2O3\mathrm{Lu}_2\mathrm{O}_3Alfa Aesar, Reacton, 99.995%

Table S1. Reagents used to synthesize representative polycrystalline RE2O2S\mathrm{RE}_2\mathrm{O}_2\mathrm{S} samples. *Tm2O3\mathrm{Tm}_2\mathrm{O}_3 was produced from Tm metal (below) annealed at 1000°C.

Laboratory XRD. Powder X-ray diffraction (pXRD) patterns were collected on a laboratory Bruker D8 Advance diffractometer with Lynxeye XE-T detector and Cu Kα radiation (λKα1=1.5046\lambda_{K\alpha1} = 1.5046 Å) with Johannson monochromator in the 2θ2\theta range from 10 - 80°. Rietveld refinements on sample pXRD data with an internal silicon standard (Si powder, Acros Chem, 99.9%) were performed using Topas 5.0 (Bruker). In all cases, the inclusion of at least 10% bleed-through Kα2 radiation (λKα2=1.5444\lambda_{K\alpha2} = 1.5444 Å) was found to be necessary to account for the observed minor splitting of major Bragg reflections at higher 2θ2\theta.

Raman spectroscopy. Raman scattering spectra were collected in the backscattering geometry from 50 - 600 cm−1^{-1} using a Horiba LabRam HR800 spectrometer with integrated 632.8 nm He-Ne laser and exchangeable 514.5 Å ion laser (CVI Melles Griot model #300-001, head model #35-LAP-321-120). The spectra shown in Figure 1d were collected with an excitation wavelength of either 632.8 nm (RE = La, Ce, Pr, Nd, Eu, Gd, Tb, Tm, Lu) or 514.5 nm (RE = Sm, Dy, Ho, Er, Yb) to minimize the fluorescence commonly observed in the optical spectra of rare-earth insulators [41, 42].

Magnetization. Magnetization data was collected from T=1.8T = 1.8 - 300 K on a Quantum Design Dynacool Physical Property Measurement System (PPMS) and from T=0.3T = 0.3 - 1.8 K on a Quantum Design Magnetic Property Measurement System (MPMS3). Magnetic susceptibility was approximated as magnetization divided by the applied magnetic field (χ≈M/H\chi\approx M/H). Single-ion crystal electric field (CEF) calculations were performed from a point charge model in the JJ basis using the PyCrystalField package and limited to the first coordination sphere of the RE3+^{3+} cation [26]. The single-ion pseudo-spin anisotropy of each compound was extracted from calculations of the gg-tensor components (g⊥=gJ∣⟨−J−∣+⟩∣2g_{\perp} = g_J \lvert\langle-J_- \lvert+ \rangle\rvert^2, and gz=2gJ∣⟨+∣Jz∣+⟩∣2g_z = 2g_J \lvert\langle+ \lvert J_z \lvert+ \rangle\rvert^2) where gJg_J is the Landé gg-factor, while ∣+⟩\lvert+ \rangle and ∣−⟩\lvert- \rangle are the CEF ground-state wavefunctions. We fixed the phase gauge of the doublet wavefunctions such that the pseudo-spin frame is aligned with the principal axes, making the gg-tensor diagonal and strictly real.

Specific heat. Specific heat data was collected on a Quantum Design Physical Property Measurement System (PPMS) using the semi-adiabatic pulse technique with a 1% temperature rise and over three time constants. Subtraction of the estimated phonon contribution to the measured specific heat data of RE = Ce, Gd, Dy, and Yb was achieved by the scaling of the temperature scale of the measured specific heat of the relevant non-magnetic analogue by:

ΘMag3ΘNon−mag3=(Molar mass of Non-magMolar mass of Mag)3/2\frac{\Theta_{\mathrm{Mag}}^3}{\Theta_{\mathrm{Non-mag}}^3} = \left(\frac{\text{Molar mass of Non-mag}}{\text{Molar mass of Mag}}\right)^{3/2}

For each non-magnetic/magnetic RE2O2SRE_2O_2S pairing and assuming ideal stoichiometry, this value is: La/Ce - 0.99, La/Gd - 0.87, La/Dy - 0.82, Lu/Gd - 1.14, Lu/Dy - 1.10, Lu/Yb - 1.01.

The change in magnetic entropy as a function of temperature, ΔSmag\Delta S_{\mathrm{mag}}, was approximated as ΔSmag=∫CP/T dT\Delta S_{\mathrm{mag}} = \int C_P/T\,dT of the measured CP,magC_{P,\mathrm{mag}} for each magnetic composition (CP,total,magC_{P,\mathrm{total,mag}} - scaled CP,total,non−magC_{P,\mathrm{total,non-mag}}), from T=0T = 0 - 300 K, with the entropy rise from T=0T = 0 - 1.8 K being estimated from linear extrapolation over this range.

Synchrotron XRD. Total scattering measurements were performed at the Brockhouse High Energy Wiggler beamline at the Canadian Light Source [43]. Polycrystalline Nd2O3\mathrm{Nd}_2\mathrm{O}_3 samples were loaded into 0.81 mm inner diameter Kapton tubes in an Argon-filled glovebox and sealed at both ends with GE varnish. Data were collected using 65.65 keV X-rays and a Varex XRD 4343CT area detector. The pair distribution functions were produced with the GSAS-II software using a Qmax⁡Q_{\max} of 27 Å−1^{-1}[44].

PhaseTNT_{\mathrm{N}} (K)μeff\mu_{\mathrm{eff}}/Nd3+^{3+} (μB\mu_{\mathrm{B}})μcalc\mu_{\mathrm{calc}}/Nd3+^{3+} (μB\mu_{\mathrm{B}})θCW\theta_{\mathrm{CW}} (K)ffχo\chi_{\mathrm{o}}
Nd2_2O3_30.55[13]3.10(1)3.62-10.2(1)18.5(1)0
Nd2_2O2_2S2.33.23(1)3.62-16.3(1)7.1(1)0

Table S4. Fitting parameters obtained from Curie-Weiss analysis of powder magnetization measurements of representative Nd-containing phases from T=5T = 5 - 10 K.

References

  1. [1]A. N. Rudenko and M. I. Katsnelson, Anisotropic effects in two-dimensional materials, 2D Materials 11, 042002 (2024).DOI
  2. [2]H. Kageyama, K. Hayashi, K. Maeda, J. P. Attfield, Z. Hiroi, J. M. Rondinelli, and K. R. Poeppelmeier, Expanding frontiers in materials chemistry and physics with multiple anions, Nature Communications 9, 772 (2018).DOI
  3. [3]A. Miyoshi and K. Maeda, Recent progress in mixed-anion materials for solar fuel production, Solar RRL 5, 2000521 (2021).DOI
  4. [4]T. Katsumata, H. Yamamoto, Y. Kimura, K. Amezawa, R. Aso, S. Kikkawa, S. Yamazoe, and T. Nakamura, Development of electrochemical anion doping technique for expansion of functional material exploration, Advanced Functional Materials 33, 2307116 (2023).DOI
  5. [5]M. P. Shores, E. A. Nytko, B. M. Bartlett, and D. G. Nocera, A structurally perfect S = 1/2 kagomé antiferromagnet, Journal of the American Chemical Society 127, 13462 (2005).
  6. [6]K. Lee, A. H. Dismukes, E. J. Telford, R. A. Wiscons, J. Wang, X. Xu, C. Nuckolls, C. R. Dean, X. Roy, and X. Zhu, Magnetic order and symmetry in the 2D semiconductor CrSBr, Nano Letters 21, 3511 (2021).DOI
  7. [7]S. A. López-Paz, Z. Guguchia, V. Y. Pomjakushin, C. Witteveen, A. Cervellino, H. Luetkens, N. Casati, A. F. Morpurgo, and F. O. von Rohr, Dynamic magnetic crossover at the origin of the hidden-order in van der Waals antiferromagnet CrSBr, Nature Communications 13, 4745 (2022).
  8. [8]L. Clark, G. J. Nilsen, E. Kermarrec, G. Ehlers, K. S. Knight, A. Harrison, J. P. Attfield, and B. D. Gaulin, From spin glass to quantum spin liquid ground states in molybdate pyrochlores, Physical Review Letters 113, 117201 (2014).
  9. [9]M. Orr, G. R. Hebberd, E. E. McCabe, and R. T. Macaluso, Structural diversity of rare-earth oxychalcogenides, ACS Omega 7, 8209 (2022).DOI
  10. [10]G. Quezel, R. Ballestracci, and J. Rossat-Mignod, Proprietes magnetiques des oxysulfures de terres rares, Journal of Physics and Chemistry of Solids 31, 669 (1970).
  11. [11]Y. L. Suponitskii, G. M. Kuz’micheva, and A. A. Eliseev, Lanthanide oxide sulphides, Russian Chemical Reviews 57, 209 (1988).DOI
  12. [12]V. Biondo, P. W. C. Sarvezuk, F. F. Ivashita, K. L. Silva, A. Paesano Jr, and O. Isnard, Geometric magnetic frustration in RE2O2S oxysulfides (RE = Sm, Eu and Gd), Materials Research Bulletin 54, 41 (2014).DOI
  13. [13]B. K. Rai, A. D. Christianson, G. Sala, M. B. Stone, Y. Liu, and A. F. May, Magnetism of Nd2O3 single crystals near the Néel temperature, Physical Review B 102, 054434 (2020).
  14. [14]J. Xing, L. D. Sanjeewa, J. Kim, G. R. Stewart, M.-H. Du, F. A. Reboredo, R. Custelcean, and A. S. Sefat, Crystal synthesis and frustrated magnetism in triangular lattice CsRESe2 (RE = La - Lu): Quantum spin liquid candidates CsCeSe2 and CsYbSe2, ACS Materials Letters 2, 71 (2019).
  15. [15]C. Larquet and S. Carenco, Metal oxysulfides: From bulk compounds to nanomaterials, Frontiers in Chemistry 8, 179 (2020).DOI
  16. [16]M. Zinkevich, Thermodynamics of rare earth sesquioxides, Progress in Materials Science 52, 597 (2007).DOI
  17. [17]R. Ballestracci, E. F. Bertaut, and G. Quezel, Etude par diffraction neutronique et mesures magnetiques des oxysulfures de terres rares T2O2S, Journal of Physics and Chemistry of Solids 29, 1001 (1968).DOI
  18. [18]W. H. Zachariasen, Crystal chemical studies of the 5f-series of elements. VII. the crystal structure of Ce2O2S, La2O2S and Pu2O2S, Acta Crystallographica 2, 60 (1949).DOI
  19. [19]P. Liu, F. Wang, X. Chen, and B. Yang, Upconversion/-Downconversion luminescence and preparation of NIR-to-UV-excited Gd2O2S : Er phosphor, Journal of Luminescence 200, 126 (2018).
  20. [20]X. Sang, J. Lian, N. Wu, X. Zhang, and J. He, Synthesis, characterization and formation mechanism of Gd2O2S : Pr3+, Ce3+ phosphors by sealed triple-crucible method, Journal of Asian Ceramic Societies 8, 733 (2020).DOI
  21. [21]H. A. Eick, The preparation, lattice parameters and some chemical properties of the rare earth mono-thio oxides, Journal of the American Chemical Society 80, 43 (1958).DOI
  22. [22]J. R. McBride, K. C. Hass, B. D. Poindexter, and W. H. Weber, Raman and x-ray studies of Ce1−xRExO2−y, where RE = La, Pr, Nd, Eu, Gd, and Tb, Journal of Applied Physics 76, 2435 (1994).DOI
  23. [23]N. Dilawar Sharma, J. Singh, A. Vijay, K. Samanta, S. Dogra, and A. K. Bandyopadhyay, Pressure-induced structural transition trends in nanocrystalline rare-earth sesquioxides: a Raman investigation, The Journal of Physical Chemistry C 120, 11679 (2016).
  24. [24]D. C. Johnston, Influence of uniaxial single-ion anisotropy on the magnetic and thermal properties of Heisenberg antiferromagnets within unified molecular field theory, Physical Review B 95, 094421 (2017).
  25. [25]S. Mugiraneza and A. M. Hallas, Tutorial: a beginner’s guide to interpreting magnetic susceptibility data with the Curie-Weiss law, Communications Physics 5, 95 (2022).DOI
  26. [26]A. Scheie, PyCrystalField: software for calculation, analysis and fitting of crystal electric field hamiltonians, Applied Crystallography 54, 356 (2021).DOI
  27. [27]L. Beaury and P. Caro, Interpretation of the optical absorption spectrum and of the paramagnetic susceptibility of neodymium oxysulfide, Journal de Physique 51, 471 (1990).
  28. [28]R. Ballestracci and J. Rossat-Mignod, Structure magnetique de l’oxysulfure d’ytterbium, Solid State Communications 7, 1011 (1969).DOI
  29. [29]J. Awaka, R. Endoh, and S. Nagata, Van Vleck paramagnetism of the thulium garnet Tm₃Al₅O₁₂, Journal of Physics and Chemistry of Solids 64, 2403 (2003).DOI
  30. [30]A. Cote, J. E. Slimak, A. Sethi, D. Reigi-Plessis, Q. Zhang, Y. Zhao, D. Adroja, G. Morris, T. Kolodiazhnyi, A. M. Hallas, et al., Direct confirmation of long-range magnetic order and evidence for multipoles in Ce₂O₃, arXiv preprint 10.48550/arXiv.2303.11878 (2023), (accessed 2026-10-01).arxiv.org/abs/2303.11878
  31. [31]R. Mauricot, J. Gareh, and M. Evain, Crystal structure of Ce₂O₂.₅S, an oxidized form of Ce₂O₂S, from neutron and x-ray powder diffraction, Zeitschrift für Kristallographie 212, 24 (1997).DOI
  32. [32]J. Rossat-Mignod, G. Quezel, A. Berton, and J. Chaussy, Crystal field determination in rare-earth oxysulfides III. Specific heat measurements on Gd, Tb, Ho, and Dy oxysulfides, Physica Status Solidi B 63, 105 (1974).DOI
  33. [33]G. P. Borodulenko, A. A. Eliseev, V. A. Efremov, G. M. Kuz’micheva, I. V. Perepelkin, and N. M. Ponomarev, Defect formation in lanthanum oxide sulphide, Russian Journal of Inorganic Chemistry 30, 1255 (1985).
  34. [34]P. H. Holloway, T. Trottier, J. Sebastian, S. Jones, X.-M. Zhang, J.-S. Bang, B. Abrams, W. Thames, and T.-J. Kim, Degradation of field emission display phosphors, Journal of Applied Physics 88, 483 (2000).DOI
  35. [35]B. Jaffar, H. Swart, H. S. Ahmed, A. Yousif, and R. Kroon, Cathodoluminescence degradation of Bi doped La₂O₃ and La₂O₂S phosphor powders, Physica B: Condensed Matter 574, 411659 (2019).DOI
  36. [36]A. Le Coz, L. Cornet, G. Durand, R. Henry-Bauer, M. Even, M. Poitou, F. Chevré, Y. Suffren, and O. Merdrignac-Conanec, Energy transfer phenomena in Er³⁺, Yb³⁺, Tm³⁺ tri-doped lanthanum oxysulfide phosphors, Journal of Luminescence 281, 121195 (2025).DOI
  37. [37]M. Mikami and A. Oshiyama, First-principles study of intrinsic defects in yttrium oxysulfide, Physical Review B 60, 1707 (1999).DOI
  38. [38]T. Wen, R. Ding, Y. Zhou, Y. Si, B. Yang, and Y. Wang, Polymorphism of erbium oxyfluoride: Selective synthesis, crystal structure, and phase-dependent upconversion luminescence, European Journal of Inorganic Chemistry 2017, 3849 (2017).
  39. [39]R. Srinivasan, S. TragI, and H.-J. Meyer, Synthesis and structure of the new compound La₂O₂(CN₂)₂ possessing an interchanged anion proportion compared to the parent La₂O₂(CN₂), Zeitschrift für anorganische und allgemeine Chemie 631, 719 (2005).
  40. [40]R. P. Turcotte, J. O. Sawyer, and L. Eyring, Rare earth dioxymonocarbonates and their decomposition, inorganic Chemistry 8, 238 (1969).DOI
  41. [41]P. Fornasiero, A. Speghini, R. Di Monte, M. Bettinelli, J. Kašpar, A. Bigotto, V. Sergo, and M. Graziani, Laser-excited luminescence of trivalent lanthanide impurities and local structure in CeO₂-ZrO₂ mixed oxides, Chemistry of Materials 16, 1938 (2004).DOI
  42. [42]J. Cui and G. A. Hope, Raman and fluorescence spectroscopy of CeO₂, Er₂O₃, Nd₂O₃, Tm₂O₃, Yb₂O₃, La₂O₃, and Tb₄O₇, Journal of Spectroscopy 2015, 940172 (2015).
  43. [43]A. Rahemutulla, G. King, A. Gomez, N. Appathurai, A. F. G. Leontowich, R. Castle, N. Burns, C.-Y. Kim, B. Moreno, and S. Kycia, The high energy diffraction beamline at the Canadian Light Source, Synchrotron Radiation 32 (2025).
  44. [44]B. H. Toby and R. B. Von Dreele, GSAS-II: the genesis of a modern open-source all purpose crystallography software package, Applied Crystallography 46, 544 (2013).DOI

Paper details

vlnc:W_naczsfdk
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