Structural and magnetic properties of spin- 1/2 dimer compound Cu 2 (IPA) 2 (DMF)(H 2 O) with a large spin gap
SStructural and magnetic properties of spin- / dimer compoundCu (IPA) (DMF)(H O) with a large spin gap
S. Thamban, U. Arjun, M. Padmanabhan, and R. Nath ∗ School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram-695016, India Department of Chemistry, Amrita School of Arts and Sciences,Amrita Vishwa Vidyapeetham, Amrita University, Amritapuri Campus, Kollam-690525, India
We present the synthesis and a detailed investigation of structural and magnetic properties ofmetal-organic compound Cu (IPA) (DMF)(H O) by means of x-ray diffraction, magnetization, andheat capacity measurements. Single crystals of the title compound were synthesized by judiciousselection of organic ligand and employing a selective hydrothermal reaction route. It crystallizes inan orthorhombic structure with space group
Cmca . The structural analysis revealed that two Cu ions are held together by the organic component (-O-C-O-) in a square paddle-wheel to form spindimers which are aligned perpendicular to each other and are further coupled through organic ligands(isophthalic acid) forming two-dimensional layers. Temperature dependent magnetic susceptibility χ ( T ) could be described well using spin-1 / χ spin ( T ) showsan exponential decrease in the low temperature region, below the broad maximum, confirming thesinglet ground state with a large spin gap of ∆ /k B (cid:39)
409 K. The heat capacity C p measured as afunction of temperature also confirms the absence of magnetic long-range-order down to 2 K. PACS numbers: 75.10.Jm, 75.50.Ee, 75.30.Et
I. INTRODUCTION
The last few decades have witnessed enormous researchin the field of low-dimensional quantum magnets. Modelcompounds especially with a gap in the excitation spec-trum are extensively pursued. Spin gap is found in manylow-dimensional quantum spin systems. Some exam-ples include spin dimers, integer spin chains, evenleg spin ladders, alternating spin chains, spin-Peierlscompounds, etc. Spin dimers are the simplest magneticclusters, comprising of unified pair of magnetic ions, cou-pled antiferromagnetically. Such systems are character-ized by a gap in their excitation spectrum and the energydifference between excited triplet states ( S = 1) and sin-glet ground state ( S = 0) constitutes the spin gap.Recently, many new magnetic materials composedof antiferromagnetic spin dimers have been discov-ered. Because of different inter-dimer exchange net-works, a variety of magnetic excitations have been ob-served under external magnetic field. Bose-Einstein-Condensation (BEC) of magnons in BaCuSi O ,TlCuCl , (Ba,Sr) Cr O , NiCl -4SC(NH ) etc ,magnetization plateaus in Ba Mn O ,[Ref. 16] andmagnetization plateaus and Wigner crystallization ofmagnons in SrCu (BO ) [Ref. 17] are the best knownexamples. These discoveries have accelerated the searchfor new spin dimer compounds and interesting quantumphenomena.Unlike the inorganic materials (as discussed above),the metal-organic compounds are not explored well de-spite being easier to synthesize. In magnetically ac-tive metal-organic systems, the organic moieties bondthe metal ions in a wide variety of forms leading tostructurally and magnetically diverse systems. Wehave recently demonstrated that the organic moietiescan modulate both the structural and magnetic features between metal ions significantly as the organic ligandNIPA leads to the formation of a hourglass nanomag-net Cu (NIPA) (OH) .11H O with two antiferromag-netic (AFM) and one ferromagnetic (FM) couplings be-tween Cu ions while another similar ligand leads tothe formation of a quasi-two-dimensional (2D) AFM withweakly anisotropic spin-1 / Cu(1,3- bdc )is another well studied compound where Cu ions areconnected via IPA forming a perfect Kagom´ e lattice withstrong magnetic frustration. It undergoes a magneticlong-range-order (LRO) at T N (cid:39) .
77 K. We have re-ported that in such Kagom´ e lattices by changing theligand it is possible to tune the magnetic ground statekeeping the crystal structure intact. Thus, the organicligands play a vital role in stabilizing different groundstates in the metal-organic complexes.Herein, we report synthesis, crystal structure, andmagnetic properties of a new spin-1 / (IPA) (DMF)(H O) [IPA = isophthalic acid andDMF = dimethyl formamide]. Cu ions in the com-pound form spin dimers which are weakly coupled vialong ligands. Our magnetic susceptibility measurementsrevealed an activated behavior at low temperatures witha large spin gap of ∆ /k B (cid:39)
409 K.
II. EXPERIMENTAL DETAILS
All the reagent quality chemicals were obtained fromSigma-Aldrich and used without further purification.Single crystals of Cu (IPA) (DMF)(H O) were pre-pared by the hydrothermal route. In this process,an aqueous solution of Cu(NO ) .3H O (0.5 mM, 0.12g) was mixed with same equivalent of isopthalic (1,3-benzenedicarboxylic) acid (0.5 mM, 0.08 g) in 10 mLof DMF. The mixture was heated at 90 C in a 25 mL a r X i v : . [ c ond - m a t . m t r l - s c i ] M a r Paddle wheel IsophthalicAcidCu(2)Cu(1)Cu(3) Cu(3)abc
FIG. 1. Projection of the two-dimensional layer onthe ab -plane showing interconnecting and orthogonalnetwork of Cu dimers in the crystal structure ofCu (IPA) (DMF)(H O). Since the DMF and H O moleculesare not involved in the interaction path, we have removedthem for clarity. teflon capped autoclave. Big plate like blue crystals werecollected from the walls of the vial after 3 days. Thecrystals after washing in water and acetone and thendrying in air were found to be the phase pure form ofCu (IPA) (DMF)(H O). The yield was 72.6% (based onCu).Single-crystal x-ray diffraction (XRD) was performedusing a Bruker APEX-II diffractometer (MoK α radi-ation with wavelength λ avg (cid:39) . (IPA) (DMF)(H O) at roomtemperature and the crystal structure was solved. Thephase purity of crystals was also confirmed through pow-der XRD performed on the crushed powder sample atroom temperature using a PANalytical (Cu K α radiation, λ avg (cid:39) . χ was measured as a function of temperature(2 K ≤ T ≤
380 K) using the vibrating sample mag-netometer (VSM) attachment to the Physical PropertyMeasurement System [PPMS, Quantum Design]. Forhigh temperature measurements ( T ≥
380 K), a high- T oven (Model CM-C-VSM) was attached to the VSM. Ourmeasurements were done upto 500 K, above which thesample was found to be decomposed. From the thermogravity analysis, the decomposition temperature was in-deed confirmed to be ∼
500 K. Heat capacity, C p ( T ) wasalso measured using heat capacity option of the PPMSon a small piece of sintered pellet, adopting relaxationtechnique.
10 20 30 40 50 I n t e n s it y ( a r b . un it s ) I obs I cal I obs - I cal Bragg positions
FIG. 2. Powder XRD pattern (open circles) at room temper-ature for Cu (IPA) (DMF)(H O). The solid line representsthe Le-Bail fit, with the vertical bars showing the expectedBragg peak positions, and the lower solid line representing thedifference between the observed and calculated intensities.
III. RESULTS AND DISCUSSIONA. Crystallography
Crystal structure of Cu (IPA) (DMF)(H O) wassolved from single crystal XRD. It crystallizes in a or-thorhombic structure with space group
Cmca . The ob-tained structural parameters are tabulated in Table I andII. Figure 1 shows projection of the crystal structure inthe ab -plane obtained from the single crystal XRD. Itconsists of paddle wheel like units, each containing twoclosely held Cu ions. In each paddle wheel, Cu ionsare coupled via O-C-O path forming a spin dimer. Itis to be noted that there are three inequivalent Cu sites[Cu(1), Cu(2), and Cu(3)] in the crystal structure whichform two different dimers (type 1 and type 2). The intra-dimer distances are Cu(1)-Cu(2) (cid:39) .
641 ˚Aand Cu(3)-Cu(3) (cid:39) .
633 ˚Afor type 1 and type 2 dimers, respec-tively. Each dimer is connected to the adjacent dimersthrough the organic ligand (isophthalic acid, IPA) lead-ing to either a coupled dimer or a quasi-2D layered struc-ture in the ab -plane. The dimers are of course far apart( ∼ c -direction.Le-Bail fit of the observed powder XRD pattern wasperformed using the FullProf package. The initialstructural parameters for this purpose were taken fromthe single crystal XRD data. Figure 2 shows the powderXRD pattern of Cu (IPA) (DMF)(H O) at room tem-
TABLE I. Crystal structure data of Cu (IPA) (DMF)(H O)obtained from single crystal XRD experiment at room tem-perature.Empirical Formula C H Cu NO Formula weight 544.40Temperature 296 KWavelength 0.71073 ˚ACrystal system OrthorhombicSpace Group
Cmca
Lattice parameters a = 28 . b = 25 . c = 15 . α = 90 β = 90 γ = 90 Volume 10724(2) ˚A Z ρ cal )Absorption Coefficient 1.631 mm − F (000) 4384Crystal size 0 . × . × .
100 mm θ ranges for data collection 2.902 to 53.988 Index ranges − ≤ h ≤ − ≤ k ≤ − ≤ l ≤ R int = 0 . F R -Indices, I ≥ σ (I) R (cid:39) . wR (cid:39) . R -Indices(all data) R (cid:39) . wR (cid:39) . − perature along with the calculated pattern. All the peakscould be fitted using the orthorhombic ( Cmca ) structure.The obtained best fit parameters are a = 28 . b = 25 . c = 15 . χ (cid:39) .
27. These lattice parameters are close to thevalues obtained from the single crystal XRD.
B. Magnetic Susceptibility
Magnetic susceptibility χ ( T ) ( ≡ M/H , where M is themagnetization) measured in an applied field of H = 1 Tis shown in the upper panel of Fig. 3. As the tem-perature decreases, χ ( T ) increases and passes througha broad maximum near T max χ (cid:39)
200 K, a short rangemagnetic order which is fingerprint of low dimensionalantiferromagnetic spin systems. At very high tempera-tures ( T (cid:29) exchange coupling, J/k B ), usually spins arerandomly oriented and χ ( T ) behaves like a paramagnet.However, in our compound χ ( T ) does not seem to havereached the paramagnetic region even at 500 K. Thus,persistence of magnetic correlations upto such a hightemperature clearly reflects a strong antiferromagneticexchange coupling between Cu ions. At low tempera- TABLE II. Atomic coordinates for Cu (IPA) (DMF)(H O).The isotropic atomic displacement parameters (ADP) U eq aredefined as one third of the trace of the orthogonal U ij tensor.The errors are from the least-square structure refinement.Atoms x y z U eq ( × ) ( × ) ( × ) ( × ˚A )Cu1 5000 1150(1) 1820(1) 33(1)Cu2 5000 1904(1) 3030(1) 35(1)Cu3 2187(1) 3997(1) 1857(1) 32(1)C1 3667(1) 2656(1) 1630(3) 39(1)C2 3937(1) 2284(1) 1171(3) 42(1)C3 3816(2) 2146(2) 321(3) 52(1)C4 3419(2) 2370(2) -68(3) 60(1)C5 3145(2) 2732(2) 391(3) 52(1)C6 3265(1) 2880(1) 1235(3) 39(1)C7 2970(1) 3284(1) 1731(3) 35(1)C8 4356(1) 2017(1) 1628(3) 39(1)C9 1758(1) 5200(1) 3590(3) 37(1)C10 1326(1) 5374(1) 3250(3) 37(1)C11 6105(1) 826(1) 3613(3) 39(1)C12 6311(2) 1088(2) 4299(3) 56(1)C13 6745(2) 910(2) 4630(3) 65(2)C14 1970(2) 5473(2) 4279(3) 53(1)C15 2015(1) 4731(1) 3186(3) 35(1)C16 1359(2) 4438(2) -839(4) 98(2)C17 1960(3) 3857(3) -1584(4) 118(2)C18 1912(2) 3920(2) 7(4) 59(1)C19 5662(1) 1049(1) 3197(3) 36(1)N1 1753(2) 4067(2) -779(3) 67(1)O1 4507(1) 1601(1) 1281(2) 49(1)O2 4504(1) 2231(1) 2309(2) 50(1)O3 5495(1) 1459(1) 3559(2) 52(1)O4 5500(1) 818(1) 2537(2) 45(1)O5 1829(1) 4528(1) 2515(2) 45(1)O6 2613(1) 4581(1) 1444(2) 44(1)O7 2617(1) 3467(1) 1338(2) 43(1)O8 3118(1) 3421(1) 2461(2) 45(1)O9 1749(1) 4054(1) 710(2) 53(1)O10 5000 551(2) 804(4) 84(2)O11 5000 2456(2) 4123(3) 86(2)H1 3754 2754 2195 47H2 4002 1903 13 62H3 3336 2277 -638 72H4 2875 2879 130 62H5 1184 5193 2786 44H6 6161 1384 4544 68H7 6885 1091 5097 78H8 2261 5360 4500 64 H = 1 T Dimer fit spin ( c m / m o l - C u + ) T (K)
1/ CW-fit / ( c m / m o l - C u + ) - T (K)
FIG. 3. Upper panel: χ ( T ) measured at H = 1 T. The solidline represents the fit using Eq. (2). Lower panel: 1 /χ vs. T and the solid line is the CW fit using Eq. (1) tures ( T (cid:46)
70 K), χ ( T ) shows a clear upturn, possiblydue to extrinsic paramagnetic impurities and/or defectspresent in the sample. No signature of any magneticlong-range-order (LRO) is observed down to 2 K.To extract the magnetic parameters, χ ( T ) at high tem-peratures was fitted by the following expression χ ( T ) = χ + CT + θ CW , (1)where χ is the temperature independent contributionconsisting of core diamagnetic susceptibility ( χ core ) ofthe core electron shells and Van-Vleck paramagnetic sus-ceptibility ( χ VV ) of the open shells of the Cu ionspresent in the sample. The second term in Eq. (1)is the Curie-Weiss (CW) law with the CW tempera-ture ( θ CW ) and Curie constant C = N A µ / k B , where N A is Avogadro number, k B is Boltzmann constant, µ eff = g (cid:112) S ( S + 1) µ B is the effective magnetic moment, g is the Land´e g -factor, µ B is the Bohr magneton, and S is the spin quantum number. As shown in the lower panelof Fig. 3, we fitted the 1 /χ ( T ) data in the temperaturerange 400 K to 500 K using Eq. (1). Since the com-pound is not completely paramagnetic in this tempera-ture range, we fixed the value of C to 0.375 cm K/mol-Cu (expected C value for spin-1 / g =2 to give µ eff = 1 . µ B /Cu ) and tried to vary otherparameters. The fit yields χ (cid:39) . × − cm /mol- Cu and θ CW (cid:39) .
84 K. The positive value of θ CW indicates that the dominant exchange couplings betweenCu ions are antiferromagnetic in nature. Such a largevalue of θ CW also suggests a strong exchange interactionbetween Cu ions.In order to get an estimation of the exchange couplingand to have a better view of the spin-lattice, we fittedthe observed χ ( T ) data to the following expression χ ( T ) = χ + C imp T + θ imp + χ dimer , (2)where, C imp represents the impurity concentration and θ imp provides the effective interaction strength betweenimpurity spins. χ dimer is the expression for exact spinsusceptibility of a spin-1 / χ dimer = N A g µ k B T [3 + exp (∆ /k B T )] . (3)The dimers have a spin gap ∆ /k B (= J/k B ) between thesinglet ground state and the triplet excited states. Here, J/k B is the intra-dimer exchange coupling.The low temperature upturn of χ ( T ) below 70 K orig-inates from the isolated Cu ions possibly due to lat-tice defect. The second term (CW term) is included inEq. (2) to account for this low temperature upturn. Over-all, Eq. (2) has six fitting parameters: χ , C imp , θ imp , g ,and J/k B . As shown in the upper panel of Fig. 3, Eq. (2)fits very well to the χ ( T ) data over the whole tempera-ture range. The obtained best fit parameters are χ (cid:39) . × − cm /mol-Cu , C imp (cid:39) .
015 cm K/mol-Cu , g (cid:39) .
91, ∆ /k B (cid:39) .
83 K, and θ imp (cid:39) .
96 K.This value of C imp corresponds to a spin concentrationof nearly 4.0 %, assuming the impurity spins as spin-1 /
2. In order to demonstrate the non-magnetic groundstate, χ + C imp T + θ imp was subtracted from the χ ( T ) dataand the obtained intrinsic spin susceptibility χ spin ( T ) ispresented in the upper panel of Fig. 3. It is evident that χ spin ( T ) decreases rapidly towards zero at low tempera-tures which unambiguously establishes a spin gap in theexcitation spectrum. It is to be noted that our attempt to fit the χ ( T )data using coupled dimer model didnot improve the fit-ting significantly and the obtained value of inter-dimercoupling ( J (cid:48) ) was not at all reliable. This implies thatthe spin-lattice behaves more like isolated dimers ratherthan coupled dimers. Typically, in isolated dimers, thestrength of intra-dimer coupling quantitatively reflectsthe magnitude of spin gap. Moreover, the value of spingap can be reduced immensely by the inter-dimer cou-pling, upto a threshold value, above which there is atransition to magnetic LRO. Thus, such a large spingap in Cu (IPA) (DMF)(H O) also implies a negligibleinter-dimer coupling in the compound. This finding is inreasonable agreement with the structural data where thedimers are well separated ( ∼ dimers are formed through acomplex path of -O-C-O- and the arrangement of Cu -O-C-O-Cu is not in one line. Usually, one expects astrong exchange interaction for the interaction path in-volving only O between metal ions and ∼ metal-oxygen-metal bond angle. Despite having a complex in-teraction path within the dimers, this compound shows alarge spin gap of ∆ /k B (cid:39)
409 K which is a surprising fea-ture of the compound. It is worth mentioning that spindimer compounds such as TlCuCl , (Sr,Ba) Cr O etcwhere the super-exchange involves only oxygen/chlorineatom between metal ions gives a rather small value of thespin gap. There are a few spin-1 / reported to have paddle-wheel like dimer units but with different groundstates. For instance, [Cu (OOCC H ) (urea) ],[ { Cu (OOCC H ) (urea) } ], [Cu (OOCC H ) ] n , andCu (O CCH=CHCH ) (DMF) show singlet groundstate with a spin gap of ∼
342 K, ∼
375 K, ∼
333 K, and ∼
220 K, respectively.
The com-pound tetrakis( µ benzoato-O,O (cid:48) )-bis(dimethyl sulfox-ide)dicopper(II) which also has paddle-wheel like dimerunits but with a different ligand show one-dimensionalcharacter with a exchange coupling J/k B (cid:39)
405 K. Itundergoes a FM ordering at T C (cid:39) O) is reported to have paddle-wheel like spin dimers coupled through IPA ligand, simi-lar to our compound. The crystal lattice forms a Kagom´ e structure with a very large spin gap of ∆ /k B (cid:39)
410 K. From the above examples it is clear that though all thesecompounds are having paddle-wheel like spin dimers butdifferent connecting ligands between dimers lead to di-verse ground states. Thus, in these compounds one cantune the ground state properties and even the magni-tude of the spin gap/exchange coupling by simply chang-ing/modifying the connecting ligand.
C. Magnetic Isotherm
As one can see in the upper panel of Fig. 3, the mag-nitude of χ spin ( T ) is nearly zero below 50 K. Thus, thelow temperature upturn observed in χ ( T ) below 50 K ispurely extrinsic in nature. Therefore, one can preciselyestimate the extrinsic contribution to χ ( T ) by analyzingthe magnetic isotherms ( M vs H ) at low temperatures.Figure 4 shows the magnetic isotherms measured at dif-ferent temperatures. At T = 200 K (inset of Fig. 4),it is a nice straight line. As the temperature is low-ered, it shows a pronounced curvature typically observedfor paramagnets. In order to quantitatively estimate theparamagnetic impurity contribution, we fitted the M vs H data at T = 2.1 K and 5 K by the following equation M = χH + f imp N A g imp µ B S imp B S imp ( x ) , (4)where χ is the intrinsic susceptibility of the sample, f imp is the molar fraction of the impurities, N A is the Avo- M ( G c m / m o l -f . u . ) H ( T ) M ( G c m / m o l -f . u . ) T = 200 K
H (T)
FIG. 4. Magnetic isotherms ( M vs. H ) at T = 2.1 K and5 K. Solid lines are the fits using Eq. (4). Inset: M vs. H at T = 200 K which shows linear behavior. gadro’s number, g imp impurity g -factor, S imp is the im-purity spin, B S imp ( x ) is the Brillouin function, andthe modified argument of the Brillouin function is x = g imp µ B S imp H/ [ k B ( T − θ imp )]. C p ( J / m o l K ) H = 0 T
T (K)
FIG. 5. Temperature dependent heat capacity C p ( T ) ofCu (IPA) (DMF)(H O) measured at zero applied field.
In order to reduce the number of fitting parameters,in the Brillouin function fit using Eq. (4), we fixed g imp and S imp to 1.9 [obtained from the χ ( T ) analysis] and1 /
2, respectively. The fitting parameters ( χ , f imp , and θ imp ) obtained at T = 2 . f imp obtained from the analysis of2.1 K data was fixed while analysing the 5 K data. Usingthe value of f imp in C imp = f imp N A g imp µ B S imp ( S imp +1), we calculated the Curie constant of the paramagneticimpurities to be C imp ≈ . K/mol. This valueof C imp corresponds to ∼ .
4% of spin-1 / χ ( T ) analysis. TABLE III. Parameters obtained from the fitting of M vs. H curves by the Brillouin function [Eq (4)].Temperature χ f imp θ imp (K) (cm K/mol) (mol%) (K)2.1 ∼ × − ∼ ∼ ∼ × − ∼ ∼ D. Heat Capacity
The heat capacity C p of Cu (IPA) (DMF)(H O) ismeasured as a function of temperature at zero appliedfield in the temperature range from 2 K to 90 K (seeFig. 5). No indication of magnetic LRO was observeddown to 2 K, which is in accordance with our χ ( T ) data.In a magnetic insulator, C p ( T ) has two major contribu-tions: one due to phonon excitations and the other due tomagnetic part. We refrained from doing any quantitativeanalysis of the C p data since it was not possible to sub-tract the phonon contribution without a non-magneticanalogue compound. IV. SUMMARY
We have synthesized a novel paddle-wheel type Cu spin-1 / (IPA) (DMF)(H O) andinvestigated its crystal structure and magnetic proper-ties. It crystallizes in an orthorhombic (
Cmca ) crystalstructure. Analysis of χ ( T ) revealed dimer model for thespin lattice with a large spin gap of ∆ /k B (cid:39)
409 K. The paramagnetic impurity concentration was estimated tobe ∼
4% which is likely due to defects present in thesample. No signature of magnetic LRO was observeddown to 2 K from the C p ( T ) data, further supportingthe gapped behaviour. In magnetic insulators, strongestsuper-exchange interaction between magnetic ions is usu-ally expected via non-magnetic ions such as oxygen. However, in our compound, despite having a long andcomplex interaction path Cu-O-C-O-Cu, the intra-dimerinteraction is surprisingly very large. For a comparison,the highest value of spin gap is found to be reported forSrCu O (∆ /k B (cid:39)
420 K) among the inorganic com-pounds and for Cu(IPA)(H O) (∆ /k B (cid:39)
410 K) amongthe metal-organic compounds. Thus, the value of spingap found for Cu (IPA) (DMF)(H O) is nearly same inmagnitude to the highest values reported so far.As pointed out earlier, the crystal structure containsquasi-2D layers of orthogonal Cu dimers. Such a crys-tal lattice is reminiscent of the magnetic sub-lattice re-ported for SrCu (BO ) where Wigner crystallization ofmagnons and magnetization plateaus were observed inan applied magnetic field. Thus, a large spin gap andorthogonal arrangement of dimers in the 2D layer makeCu (IPA) (DMF)(H O) a unique system for further in-vestigations. Furthermore, the value of spin gap is verysensitive to the inter-dimer couplings which tend to re-duce the spin gap and heavily affect the properties underhigh magnetic field. In such a context, this compoundserves as a starting point to synthesize further new com-pounds where one can tune the spin gap simply by chang-ing the organic ligand.
V. ACKNOWLEDGEMENT
We would like to acknowledge IISER Thiruvanantha-puram for providing necessary experimental facilities. ∗ [email protected] P. Lemmens, G. Gntherodt, and C. Gros, Phys. Rep. ,1 (2003). S. Sebastian, N. Harrison, C. Batista, L. Balicas, M. Jaime,P. Sharma, N. Kawashima, and I. Fisher, Nature , 617(2006). C. R¨uegg, N. Cavadini, A. Furrer, H.-U. G¨udel, K. Kr¨amer,H. Mutka, A. Wildes, K. Habicht, and P. Vorderwisch,Nature , 62 (2003). H.-J. Koo and M.-H. Whangbo, J. Solid State Chem. ,263 (2000). Y. Singh and D. Johnston, Phys. Rev. B , 012407 (2007). T. Shimizu, D. MacLaughlin, P. Hammel, J. Thompson,and S. Cheong, Phys. Rev. B , R9835 (1995). M. Azuma, Z. Hiroi, M. Takano, K. Ishida, and Y. Ki-taoka, Phys. Rev. Lett. , 3463 (1994). D. C. Johnston, J. W. Johnson, D. P. Goshorn, and A. J.Jacobson, Phys. Rev. B , 219 (1987). N. Ahmed, P. Khuntia, H. Rosner, M. Baenitz, A. Tsirlin,and R. Nath, arXiv:1512.01650 (2016). K. Ghoshray, B. Pahari, B. Bandyopadhyay, R. Sarkar,and A. Ghoshray, Phys. Rev. B , 214401 (2005). A. A. Tsirlin, R. Nath, J. Sichelschmidt, Y. Skourski,C. Geibel, and H. Rosner, Phys. Rev. B , 144412 (2011). M. Hase, I. Terasaki, and K. Uchinokura, Phys. Rev. Lett. , 3651 (1993). A. A. Aczel, Y. Kohama, C. Marcenat, F. Weickert,M. Jaime, O. E. Ayala-Valenzuela, R. D. McDonald, S. D.Selesnic, H. A. Dabkowska, and G. M. Luke, Phys. Rev.Lett. , 207203 (2009). A. Aczel, Y. Kohama, M. Jaime, K. Ninios, H. B. Chan,L. Balicas, H. Dabkowska, and G. Luke, Phys. Rev. B ,100409 (2009). V. Zapf, D. Zocco, B. Hansen, M. Jaime, N. Harrison,C. Batista, M. Kenzelmann, C. Niedermayer, A. Lacerda,and A. Paduan-Filho, Phys. Rev. Lett. , 077204 (2006). M. Uchida, H. Tanaka, H. Mitamura, F. Ishikawa, andT. Goto, Phys. Rev. B , 054429 (2002). K. Kodama, M. Takigawa, M. Horvati´c, C. Berthier,H. Kageyama, Y. Ueda, S. Miyahara, F. Becca, andF. Mila, Science , 395 (2002); H. Kageyama,K. Yoshimura, R. Stern, N. V. Mushnikov, K. Onizuka,M. Kato, K. Kosuge, C. P. Slichter, T. Goto, and Y. Ueda,Phys. Rev. Lett. , 3168 (1999). R. Jain, K. Kabir, J. B. Gilroy, K. A. Mitchell, K.-C. Wong,and R. G. Hicks, Nature , 291 (2007). R. L. Carlin,
Magnetochemistry (Springer-Verlag, 1986). R. Nath, A. Tsirlin, P. Khuntia, O. Janson, T. F¨orster,M. Padmanabhan, J. Li, Y. Skourski, M. Baenitz, H. Ros-ner, et al. , Phys. Rev. B , 214417 (2013). R. Nath, M. Padmanabhan, S. Baby, A. Thirumurugan,D. Ehlers, M. Hemmida, H.-A. K. von Nidda, and A. Tsir-lin, Phys. Rev. B , 054409 (2015). E. A. Nytko, J. S. Helton, P. M¨uller, and D. G. Nocera, J.Am. Chem. Soc. , 2922 (2008); R. Chisnell, J. Helton,D. Freedman, D. Singh, F. Demmel, C. Stock, D. Nocera,and Y. Lee, Phys. Rev. B , 214403 (2016); L. Marcipar,O. Ofer, A. Keren, E. A. Nytko, D. G. Nocera, Y. S. Lee,J. S. Helton, and C. Bains, ibid . , 132402 (2009). M. Ajeesh, A. Yogi, M. Padmanabhan, and R. Nath, SolidState Commun. , 16 (2015). J. R. Carvajal, Physica B: Condens. Matter , 55(1993). D. C. Johnston, R. K. Kremer, M. Troyer, X. Wang,A. Kl¨umper, S. L. Bud’ko, A. F. Panchula, and P. C.Canfield, Phys. Rev. B , 9558 (2000). T. Nakajima, H. Mitamura, and Y. Ueda, J. Phys. Soc.Jpn. , 054706 (2006). A. A. Tsirlin, R. Nath, C. Geibel, and H. Rosner, Phys.Rev. B , 104436 (2008). U. Arjun, V. Kumar, P. K. Anjana, A. Thirumurugan,J. Sichelschmidt, A. V. Mahajan, and R. Nath, (Submit-ted). B. Kozlevˇcar, I. Leban, M. Petriˇc, S. Petriˇcek, O. Roubeau,J. Reedijk, and P. ˇSegedin, Inorg. Chim. Acta. , 4220(2004). R. F. Schlam, M. Perec, R. Calvo, L. Lezama, M. Insausti,T. Rojo, and B. M. Foxman, Inorg. Chim. Acta , 81(2000). Y. Reyes-Ortega, J. L. Alc´antara-Flores, M. C. Hern´andez-Galindo, D. Ram´ırez-Rosales, S. Bernes, J. C. Ram´ırez-Garc´ıa, R. Zamorano-Ulloa, and R. Escudero, J. Am.Chem. Soc. , 16312 (2005). M. ˇSim˙enas, R. Matsuda, S. Kitagawa, A. P¨oppl, andJ. Banys, J. Phys. Chem. C (2016). R. Nath, Y. Singh, and D. Johnston, Phys. Rev. B ,174513 (2009). C. Kittel,
Introduction to Solid State Physics , 6th ed. (JohnWiley & Sons, Inc., New York, 1986). N. Motoyama, H. Eisaki, and S. Uchida, Phys. Rev. Lett.76