WorksPhysics and Astronomy
LiSr[MnN]: a magnetically ordered, metallic nitride
Abstract
LiSr[MnN] single crystals were successfully grown out of Li rich flux. The crystal structure was determined by single crystal X-ray diffraction and revealed almost linear NMnNMn chains as central structural motif. Tetragonal columns of this air and moisture sensitive nitridomanganate were employed for electrical transport, heat capacity, and anisotropic magnetization measurements. Both the electronic and magnetic properties are most remarkable, in particular the linear increase of the magnetic susceptibility with temperature that is reminiscent of underdoped cuprate and Fe-based superconductors. Clear indications for antiferromagnetic ordering at K were obtained. Metallic transport behavior is experimentally observed in accordance with electronic band structure calculations.
Introduction
Oxides are normally associated with ionic bonding and insulating electrical transport behavior. When this intuition fails, particularly interesting physical properties and ground states can be found. Well known examples are ferrites above the Verwey transition [1], cuprate superconductors [2], or delafossites [3]. In general, nitrides tend to behave similar to oxides, however, metallic behavior is arguably even more rare.
Binary transition metal nitrides attract considerable interest due to their valuable mechanical, electrical and magnetic properties. In contrast, chemistry and physics of multinary nitrides have been far less thoroughly explored [4]. Nitridometalates of metals represent an interesting class of solid state phases, which contain nitrogen as isolated anions or feature complex anions of different dimensionalities with coordination numbers of by N typically between two and four and oxidation states of the transition metals being comparatively low. Whereas the bonding within these complex anions and frameworks is essentially covalent, nitridometalates are stabilized by predominantly ionic bonding through counterions of electropositive metals like alkali () or alkaline-earth () cations.
Nitridomanganates show a wide variety of structures and properties: manganese may be coordinated tetrahedrally (Li[MnN]) [5] or trigonal-planar (Li[MnN]N [5], Ca[MnN]N [6], Ca[MnN] [7], Sr[MnN] [8, 9], Ba[MnN] [8, 9], Sr[MnN] [10]) in isolated units; ethane-like units are observed in LiCa[MnN] [11] and LiSr[MnN] [12]. Charge ordering is observed in chains of edge-sharing tetrahedra in Ba[MnN] [13] whereas complex 2D structures of mixed valent manganese in both tetrahedral and trigonal-planar coordination are observed in Ca[MnN] [14].
Both of those semiconducting phases order antiferromagnetically; their localized magnetism originates from only a small fraction of the Mn atoms not involved in metal-metal bonding and is characterized by a large orbital contribution. Anti-rutile-type (MnLi)N [15] features 3D networks of trigonal-planar coordinated manganese. Linear substituted chains [(LiMn)N] are observed in Li[(LiMn)N] [5, 16], Li[(LiMn)N] [5] and LiCa[(LiMn)N] [17].
The single crystal growth of nitrides is often challenging due to the large dissociation energy of the N molecule, the reactivity of the starting materials and enhanced vapor pressures, but the high temperature centrifugation aided filtration (HTCAF) technique [18, 19] using Li as a flux provides a powerful tool to produce large single crystals of a large variety of phases even in complex systems such as nitridometalates (e. g. LiSr[(LiFe)N] [20], LiSr[CoN] [21]) or nitride metalides like LiSrGaN [22]. Although magnetic properties were reported for only few nitridometalates containing alkaline-earth metals, these new phases show high magnetic anisotropy (LiSr[(LiFe)N] [20]) and ferromagnetic (LiSr[CoN] [21]) or antiferromagnetic (LiSr[FeN] [23]) ordering.
Here we present a metallic nitride based on the ionic alkali-alkaline-earth-nitrogen host LiSrN substituted with Mn. Since Mn almost fully occupies one of the two Li sites of LiSrN and is absent on the other one, we refer to the material as LiSr[MnN]. The phase is isotypic to LiSr[(LiFe)N], but shows a significantly different magnetic behavior [20]. Single crystals of several millimeter along a side can be obtained from Li-rich flux. We find clear indications for magnetic ordering close to room temperature and metallic electrical transport behavior.
!Crystal structure of Li₂Sr[MnN]₂ bbox=0.10,0.07,0.91,0.28
Figure 1. Crystal structure of LiSr[MnN], the unit cell is indicated by the grey lines (space group ). Mn is substituted in linear, two-fold coordination between N. One set of the −N−Mn−N−Mn− chains is oriented vertical, the other one roughly perpendicular to the paper plane.
Methods
Starting materials were lithium rod (Evochem, 99%), manganese powder (Alfa Aesar, 99.9998%), strontium nitride (SrN) powder [prepared from strontium metal (Alfa Aesar, distilled dendritic pieces, 99.95%) and nitrogen (Westfalengas, 99.999%, additionally purified by molecular sieves and BTS catalyst)] and sodium azide (NaN) powder (Roth, 99%) as a further nitrogen source. Tantalum ampules were produced on-site from pre-cleaned tantalum tube and tantalum sheet (Plansee) in an arc furnace located within a glove box.
Laue back-reflection pattern were taken with a digital Dual FDI NTX camera manufactured by Photonic Science (tungsten anode, ). The calculated pattern as well as the picture of the crystal structure were created using CrystalMaker and SingleCrystal. Laboratory powder X-ray diffraction data of finely ground (gray) powder samples were collected on a Huber G670 imaging plate Guinier camera using a curved germanium (111) monochromator and Cu-K radiation in the range with an increment of at 293(1) K. The powder samples were placed between Kapton foils to avoid degradation in air. All preparations and sample handling were performed under inert atmosphere (Ar). Phase identification was done using the WinXPow program package[24].
The crystal structures and the compositions of several crystals obtained in different experiments of the title compound LiSr[(LiMn)N] were refined from single-crystal X-ray diffraction data. Data for single crystals 1, 2, and 4 were collected at temperatures 400 K, 300 K, and 100 K, respectively, on a Bruker APEX2 diffractometer with a D8 goniometer and AgK ( Å) employing Helios mirror optics (Incoatec). The crystal was kept at 100 K and 400 K using a Cryostream 700 low temperature device (Oxford Cryosystems) during data collection. Using Olex2 [25] the structure was solved with the ShelXS [26] structure solution program using Direct Methods and refined with the ShelXL [27] refinement package using Least Squares minimisation.
Data for single crystal 3 was collected at room temperature on a Rigaku AFC7 four circle diffractometer equipped with a redSaturn 724+ CCD detector (Mo-K radiation, graphite monochromator). For numerical absorption correction, an optimized shape of the single crystal was employed [28, 29]. After data collection the structure was confirmed by direct methods, using SHELXS [26] and subsequently refined by using the full-matrix least-squares procedure with SHELXL [27]. Further details on the crystal structure investigations may be obtained from the Fachinformationszentrum Karlsruhe, 76344 Eggenstein-Leopoldshafen, Germany (Fax: +49-7247-808-666; e-mail: [email protected]), on quoting the depository numbers CSD-xxxxx, the names of the authors, and the journal citation.
The magnetization was measured using a 7 T Magnetic Property Measurement System (MPMS3) manufactured by Quantum Design. The electrical transport properties were measured as a four-probe measurement with spring loaded contacts using a 14 T Physical Properties Measurement Systems which was also manufactured by Quantum Design. The side of the few millimeter sized samples on which the contacts were placed was ground off under argon atmosphere before each run. This served to remove insulating layers from the surface of the crystals, which are sensitive to air and moisture. Due to the limited size of the available crystals and the measurement setup, it was not possible to investigate the anisotropy of electrical transport. The data shown in this work were obtained from AC electrical transport measurements at a frequency of 117 Hz and 9 Hz. The specific electric resistivity was calculated from the measured resistance based on the sample geometry by following the procedure described in Eder [30].
Density-functional (DFT) band-structure calculations were performed in the FPLO code [31] using local density approximation (LDA) for the exchange-correlation potential [32]. Correlation effects in the Mn shell were introduced on the mean-field DFT+ level with the double-counting correction in the atomic limit. The on-site Hund’s coupling was fixed to , while the Coulomb repulsion parameter was varied between 1 and 10 eV to explore the influence of correlation effects on the electronic structure. Magnetic moments in spin-polarized calculations are estimated by the internal routine of FPLO via projections on local atomic orbitals. Experimental structural parameters at 300 K (data set 2b) were used in all calculations.
Single Crystal Growth
Due to the air and moisture sensitivity of both the reactants (Li, NaN, SrN) and the final product LiSr[(LiMn)N] all manipulations including grinding and weighing, as well as complete sample preparations for measurements were carried out in an inert gas glove box (Ar, O and HO ppm). Large single crystals of LiSr[(LiMn)N] in the form of square columns with a silverish luster were obtained by reaction of SrN, Mn, NaN, and Li in molar ratio 10 : 34 : 1 : 340 with NaN acting as a nitrogen source and Li as flux and mineralizer.
The mixtures with total mass of about 0.6 g were placed into tantalum ampules equipped with a sieve [33], sealed by arc welding under inert atmosphere of 700 mbar argon and subsequently encapsulated in a quartz tube with an internal argon pressure of 300 mbar in order to prevent oxidation of the tantalum. The samples were heated from room temperature to K with a rate of 100 K h, annealed for 2 h, subsequently cooled to 600 K with 1 K h and further with 2 K h to 400 K. After centrifugation with 3000 min employing the HTCAF method [18, 19] to separate the single crystals from the excess flux.
Large single crystals of the title compound in form of tetragonal columns with a silverish black lustre were obtained in addition to greenish black single crystals of the new phase LiSr[MnN] and a small amount of hexagonal reddish black LiN as side phases. Using a two-probe multimeter, the differently colored phases showed remarkably different resistivities further enabling the discrimination of the various phases. Small single crystals suitable for X-ray diffraction were obtained either directly selected from the product or by crushing larger specimens. One of the largest LiSr[MnN] single crystals obtained is shown as an inset of Figure 2(a).
Figure 2. X-ray diffraction pattern measured after grinding part of the product obtained by Li flux growth. Both LiSr[MnN] and LiSr[MnN] were identified (Cu-K radiation, Å). The inset shows a single crystal of LiSr[MnN] on a millimeter grid. b) Laue back-reflection pattern recorded along the c-axis of the LiSr[MnN] single crystal shown above. The calculated pattern (red spots) is slightly rotated for clarity
The HTCAF technique using a Li-rich flux has been successfully employed for the growth of large single crystals of Mn-substituted LiSrN. However, it was not possible to obtain single phase material using this reaction route up to now. Although no binary intermetallic phases seem to be stable under these conditions, other nitrides and nitridometalates may form in this quinary (Li-Sr-Mn-N-Ta) system depending on composition, temperature and reaction time. Whereas the ratio AE/N significantly influences product formation, the amount of Li as flux seems to play only a minor role. During short-term reactions, excess N results in formation of highly oxidized transition metals or LiN, whereas after long reaction times also reactions with the crucible materials come into play as evidenced by the formation of Li[TaN]. Our attempts to synthesize strontium nitridomanganates yielded a large variety of phases such as LiSr[(LiMn)N], LiSr[MnN], or Sr[MnN]. Whether larger single crystals of other known nitridomanganates as well as phases with Mn in other oxidation states can be grown by adjusting temperature profile and/or the ratio of the starting materials is subject of ongoing research. Altogether, the HTCAF method may prove a useful tool for a wide range of applications.
Crystal Structure
Figure 2 shows the X-ray powder diffraction pattern measured on a ground mixture of single crystals of LiSr[(LiMn)N] and LiSr[MnN]. Lattice parameters of Å and Å obtained from the LiSr[(LiMn)N] powder differ somewhat from single crystal data (see below) indicating a variation of the Mn concentration within the samples. The decrease in by 0.34% and increase in by 4.8% with increasing Mn content in comparison to the parent compound LiSrN ( Å, Å [34]) is in line with the trend reported for the isotypic iron (, Å, Å [20]; , Å, Å [35]), cobalt (, Å, Å [23]), nickel (, Å, Å [34]), and copper phases (, Å, Å [36]).
Smaller crystals were obtained from several different crushed samples and selected for single crystal X-ray diffraction at several temperatures. The results are summarized in Table I-IV. LiSr[(LiMn)N] crystallizes tetragonal in space group and is a substitution variant of LiSrN [34].
| sample | 1 | 2 | 3 | 4 | |
| [K] | 400 | 300 | 295(2) | 100 | |
| Mn occupancy (x) | 0.954(5) | 0.946(3) | 0.931(5) | 0.949(4) | |
| crystal system | tetragonal | tetragonal | tetragonal | tetragonal | |
| space group | (No. 141) | (No. 141) | (No. 141) | (No. 141) | |
| [Å] | 3.8151(3) | 3.8115(4) | 3.8126(2) | 3.8101(6) | |
| [Å] | 28.445(2) | 28.347(3) | 28.372(3) | 28.245(5) | |
| [Å] | 414.02(7) | 411.82(10) | 412.41(6) | 410.03(14) | |
| 4 | 4 | 4 | 4 | ||
| [gcm] | 3.769 | 3.778 | 3.740 | 3.800 | |
| crystal color, habit | metallic, black, block | metallic, black, block | silverish, tetragonal column | metallic, black, block | |
| crystal size [mm] | 0.175×0.153×0.115 | 0.223×0.16×0.107 | 0.15×0.18×0.20 | 0.192×0.185×0.126 | |
| [mm] | 9.791 | 9.821 | 18.264 | 9.875 | |
| [1/λ] | 1.00 | 1.00 | 0.89 | 1.00 | |
| diffractometer | BRUKER APEX2 | BRUKER APEX2 | RIGAKU AFC7 | BRUKER APEX2 | |
| wavelength [Å] | 0.56087 (Ag-Kα) | 0.56087 (Ag-Kα) | 0.71069 (Mo-Kα) | 0.56087 (Ag-Kα) | |
| monochromator | mirror optics | mirror optics | graphite | mirror optics | |
| scan mode | scans | scans | scans | scans | |
| measured reflections | 13690 | 7790 | 3266 | 8813 | |
| observed reflections | 525 | 509 | 370 | 512 | |
| [] | |||||
| 0.0517 | 0.0418 | 0.0765 | 0.0508 | ||
| number of parameters | 16 | 16 | 16 | 16 | |
| goodness-of-fit on | 1.345 | 1.216 | 1.161 | 1.274 | |
| 0.0616 | 0.0554 | 0.0981 | 0.0659 | ||
| R1 [] | 0.0320 | 0.0249 | 0.0376 | 0.0300 | |
| R1 (all data) | 0.0354 | 0.0291 | 0.0507 | 0.0332 | |
| residual electron density | 1.31/-1.57 | 0.920 /-0.788 | 1.66/-2.01 | 1.520/-1.343 | |
| [ pm] |
Table I. Crystallographic data and experimental details for the single crystal structure refinements of different crystals of LiSr[(LiMn)N].
Predominant structural feature of
LiSr[(LiMn)N] are 2D fragments of the LiN structure with composition Li[LiMn)N]. These slabs are connected by tetrahedrally coordinated Sr atoms and stacked along [1] in such a way that each layer is rotated 90° to the previous one. Powder as well as single crystal X-ray diffraction revealed slightly varying Mn-concentrations of ; the slight differences between powder and single crystal data may also stem from slight inhomogeneities within the samples. As indicated by the notation of the chemical formula, LiSr[(LiMn)N], the substituted Mn atoms occupy only one of the two Li-sites: whereas there is no evidence for substitution on the trigonal planar coordinated Li position, the almost linearly (∠N1-MnLi-N1 = 176(0.2)°), see Table IV coordinated chain position [(LiMn)N] is nearly completely substituted by Mn (, see Table II). Compared to a truly linear chain, the (LiMn) atoms are deflected towards the Li[LiMn)N] slabs. A designation as LiSr[MnN] is also conceivable and used in this article, since on average only every 18th manganese atom within the chains is substituted by lithium. Laue back-reflection pattern calculated based on the obtained structure are in excellent agreement with measurements recorded on large single crystals Figure 2(b).
| Sr1 | 0.01016(9)<br>0.00838(7)<br>0.01286(19)<br>0.00420(8) | ||||
| MnLi | 0 | 0.05213(2)<br>0.05210(2)<br>0.05209(3)<br>0.05213(2) | 0.01076(13)<br>0.00844(10)<br>0.0118(3)<br>0.00470(11) | 0.954(5)<br>0.946(3)<br>0.931(5)<br>0.949(4) | |
| N1 | 0.05444(10)<br>0.05452(8)<br>0.05455(15)<br>0.05450(9) | 0.0091(4)<br>0.0076(3)<br>0.0128(7)<br>0.0051(3) | |||
| Li1 | −0.0225(3)<br>−0.0226(2)<br>−0.0224(4)<br>−0.0223(3) | 0.0156(13)<br>0.0128(9)<br>0.0182(18)<br>0.0096(11) |
Table 2. Atomic positions and equivalent displacement parameters [Ų] for Li₂Sr[(Li₁₋ₓMnₓ)N]₂ for crystals 1-4. The Li occupancy on the MnLi site is obtained by , all other parameters for this site were constrained to equal values for Mn and Li.
| Atom | |||
| Sr1 | 0.01062(12)<br>0.00836(9)<br>0.0114(2)<br>0.00404(10) | 0.00923(15)<br>0.00843(12)<br><br>0.00451(13) | 0.00923(15)<br>0.00843(12)<br>0.0157(3)<br>0.00451(13) |
| MnLi | 0.00363(18)<br>0.00300(13)<br>0.0057(3)<br>0.00195(17) | 0.0145(2)<br>0.01057(16)<br>0.0131(4)<br>0.00571(19) | 0.0142(2)<br>0.01174(18)<br>0.0165(5)<br>0.00643(19) |
| N1 | 0.0039(8)<br>0.0039(6)<br>0.0109(16)<br>0.0032(8) | 0.0100(10)<br>0.0078(7)<br>0.0111(16)<br>0.0047(9) | 0.0134(9)<br>0.0110(7)<br>0.0165(18)<br>0.0076(8) |
| Li1 | 0.021(4)<br>0.016(2)<br>0.015(4)<br>0.012(3) | 0.012(3)<br>0.009(2)<br>0.018(4)<br>0.006(3) | 0.014(3)<br>0.013(2)<br>0.022(4)<br>0.010(2) |
Table 3. Anisotropic displacement parameters [Ų] for Li₂Sr[(Li₁₋ₓMnₓ)N]₂ for crystals 1-4. (, Li and Mn on the MnLi site were treated identical).
The structural feature of linear substituted chains [(Li₁₋ₓMnₓ)N] is also observed in Li₂[(Li₁₋ₓMn)N] [5, 6], Li₅[(Li₁₋ₓMn)N]₃ [5].
| 1 | 2 | 3 | 4 | ||
| Sr1-N1 | 4x | 2.769(2) | 2.761(2) | 2.762(3) | 2.756(2) |
| N1-MnLi | 2x | 1.9087(2) | 1.9070(2) | 1.9076(2) | 1.9062(3) |
| N1-Li1 | 2x | 2.113(4) | 2.110(3) | 2.113(5) | 2.111(3) |
| N1-Li1 | 1x | 2.189(8) | 2.185(6) | 2.184(11) | 2.170(7) |
| ∠N1-MnLi-N1 | 176.0(2) | 175.9(2) | 175.8(3) | 176.0(2) |
Table 4. Relevant distances (in Å) and angles in Li₂Sr[(Li₁₋ₓMnₓ)N]₂ [Å] for crystals 1-4.
and Li₂Ca[(Li₁₋ₓMnₓ)N]₂ [17]. Whereas distances decrease significantly with increasing for Fe, Co, Ni [37] and to a certain extent Cu [36], this effect is less pronounced in nitridomangantes: Å in Li₂Sr[(Li₀.₀₇Mn₀.₉₂)N]₂, 1.911 Å in Li₅[(Li₀.₃₃Mn₀.₆₇^{\mathrm{I,II}})N]₃ [5], 1.914 Å in Li₂[(Li₀.₂₇Mn₀.₇₃^{\mathrm{I}})N] [16] and 1.916 Å in Li₂Ca[(Li₀.₀₆Mn₀.₉₄)N]₂ [17] compared to Å in Li₃N [38] and 1.913 Å in Li₄SrN₂ [34].
Considering all isotypic substitution variants, the manganese phase is a special case, as there is almost complete substitution () of the linearly coordinated lithium position (compared to obtained substitution levels in Li₂Ca[(Li₁₋ₓFeₓ)N]₂: [35], Li₂Sr[(Li₁₋ₓFeₓ)N]₂: [35], Li₂Sr[(Li₁₋ₓCoₓ)N]₂: [23], Li₂Sr[(Li₁₋ₓNiₓ)N]₂: [34], Li₂Sr[(Li₁₋ₓCuₓ)N]₂: [36]). It is also interesting to note that the angle ∠N-T-N in the chain [(Li₁₋ₓTₓ)N] increases (: 174.5°) [34] with increasing for all to values between 175.5–177.5°. In contrast, Li₂Ca[(Li₁₋ₓMnₓ)N]₂ [17] exhibits a different structure type.
Electrical Transport
Initial two-probe measurements using a standard multimeter indicated metallic behavior and motivated more accurate resistivity measurements as a function of temperature, excitation current, and magnetic field. Figure 3(a) shows the temperature-dependent electrical resistivity, for measured on two different samples. Absolute values amount to μΩcm at room temperature, a value that is larger than the one of elemental metals (for example 140 μΩcm for Mn). Resistivities of this order of magnitude are somewhat smaller or comparable to what is referred to as ’bad metal’ [39]. A linear temperature-dependence - a key property of a metallic material - is observed above K. The residual resistivity ratio of indicates an acceptable sample quality. The gaps in at K that appear for sample 1 are caused by a temporary loss of contacts caused by thermal expansion and/or a mechanical blocking of the springs. Storing the measurement setup in a desiccator in between the runs helped to prevent those contact failures.
!Electrical transport properties of Li₂Sr[MnN]₂ bbox=0.08,0.07,0.48,0.47
Figure 3. Electrical transport properties of Li₂Sr[MnN]₂ obtained by using spring loaded, four wire contacts. (a) Linear temperature-dependence and relatively low absolute values at room temperature as key properties of metallic systems. (b) Ohmic current-voltage characteristics. (c) Pronounced anisotropy of the magneto-resistance at low temperatures.
Ohmic current-voltage characteristics are found up to currents of at least 0.1 A [Fig. 3(b)]. Using larger currents increases the risk of loosing the contacts presumably due to heating at the contact tips, which causes a local disintegration of the sample material.
The magneto-resistance was measured for two different orientations of the crystal: with and with . Figure 3(c) shows pronounced anisotropy and moderate absolute values obtained at K: the resistivity increases for field applied perpendicular to the -axis but decreases for field applied parallel to (that is perpendicular to the chains).
Magnetization
The magnetization was measured for two orientations: the applied field was aligned parallel () or perpendicular () to the crystallographic -axis. The obtained magnetic susceptibility is plotted in Figure 4a. Nearly isotropic behavior is found for K
!Temperature- and field dependent magnetization of Li₂Sr[MnN]₂ bbox=0.57,0.07,0.93,0.48
Figure 4. Temperature- and field dependent magnetization of Li₂Sr[MnN]₂ ( K). (a) Magnetic susceptibility exhibits low anisotropy and temperature-dependence. The magnetization is linear in field (inset). A clear anomaly is observed at K. (b) : The step-like decrease in at indicates antiferromagnetic ordering in accordance with the anomaly shifting to lower temperature in larger applied fields. (c) : a well-defined maximum is observed for T at . For larger applied fields shows a monotonic increase upon cooling with a change of slope at .
with large absolute values and a linear increase of with . The magnetization is linear in field for temperatures K as shown in the inset. A significant increase of for cooling below K indicates the presence of local magnetic moments (Curie tail).
A clear anomaly is observed around room temperature with the strongest change of slope at K (). Antiferromagnetic (AFM) ordering is inferred from the step-like decrease upon cooling and the slight shift of the anomaly towards lower temperature for increasing applied fields [Fig. 4(b), H \parallel c]. For , a local maximum in at K is found for T. As shown in Figure 4(c), a strong change of this behavior is observable in larger applied fields: a thermal hysteresis emerges and the anomaly gets broader for T. A monotonous increase upon cooling is found for T with the strongest change of slope close to .
Specific Heat and Magnetization at K
Figure 5a shows the specific heat plotted as as a function of . A linear dependence is observed for K K with a slightly enhanced Sommerfeld coefficient of mJ mol K. Towards lower temperatures, increases and a comparatively broad peak is found at K [Fig. 5(b)]. In order to estimate the corresponding magnetic entropy, , a 3rd order polynomial was fit to the data below K and above K (dashed line). We find mJ mol K.
!Specific heat and magnetization measurements of Li₂Sr[MnN]₂ bbox=0.08,0.07,0.47,0.34
Figure 5. Specific heat and magnetization of Li₂Sr[MnN]₂ for K. (a) Electron and phonon contribution lead to the expected linear behavior of as a function of . The Sommerfeld coefficient is estimated to mJ mol K. (b) A broad maximum appears at K. The magnetic contribution was estimated by fitting a polynomial to the data points above and below the anomaly (dashed, orange line). The transition is associated with a significant amount of magnetic entropy . (c) The magnetic susceptibility shows a maximum at in zero-field-cooled (ZFC) measurements for and . The field-cooled (FC) curves split from the ZFC at for both orientations. (d) Even at the lowest accessible temperature of mK, the field-dependent, isothermal magnetization is only slightly curved.
Figure 5(c) shows low temperature for and . Similar to the whole investigated temperature range, only a small anisotropy is visible. A maximum is found in zero-field-cooled (ZFC) measurements at in very good agreement with . Field-cooled (FC) runs bifurcate from the ZFCs close to .
Isothermal magnetization measurements at temperatures as low as mK are presented in Figure 5(d). There is no substantial change for increasing the temperature above and further up to room temperature
(dashed lines). Even in the largest available fields, the magnetization amounts to less then 0.03 per Mn significantly below the saturation moment of (for ). This proves the absence of (more than 1%) localized Mn magnetic moments that are free or only weakly coupled for the whole temperature range investigated.
Electronic Structure
LDA density of states (DOS) for Li₂Sr[MnN]₂ features N band between and eV followed by the Mn band that straddles the Fermi level (Fig. 6). Only a tiny gap is observed between these bands, in contrast to low-valent Mn oxides, where Mn and O bands are usually well separated from each other. We choose MnO as the reference compound and compare its LDA DOS with that of Li₂Sr[MnN]₂. The shift of the band toward higher energies is immediately visible in Figure 6. It can be quantified by the charge-transfer energy that decreases from 4.1 eV in the oxide to 2.4 eV in the nitride. Here, is estimated as the difference between centers of gravity of the Mn and O/N states.
!LDA density of states for Li₂Sr[MnN]₂ and MnO bbox=0.55,0.06,0.92,0.37
Figure 6. LDA density of states for Li₂Sr[MnN]₂ (top panel) and MnO (bottom panel). The inset shows orbital-resolved DOS for Mn atoms in the former compound. The Fermi level is at zero energy.
Orbital-resolved DOS for Li₂Sr[MnN]₂ (Figure 6, inset) suggests the splitting of Mn states into crystal-field levels with eV, eV, and eV, where the axis is chosen along the Mn–N bonds. This level scheme is consistent with the linear coordination of Mn and also shows that the charge-transfer energy for the orbital is reduced to 1.3 eV. This orbital then features a notably broader dispersion because of a much stronger hybridization with the N states.
To probe the effect of electronic correlations on the band structure, we performed DFT+ calculations for both MnO and LiSr[MnN]. Experimental antiferromagnetic structure was chosen for the former, while in the latter case an antiferromagnetic structure with the highest symmetry was considered. It keeps the periodicity of the crystal structure and comprises ferromagnetic layers in the plane coupled antiferromagnetically along the axis ( symmetry). Alternatively, a spin configuration with antiferromagnetic order along the Mn–N chains was constructed in a four-fold supercell and showed a very similar evolution of the band structure with , so we focused on a simpler antiferromagnetic structure that allowed faster and more easily converged calculations at different values of .
MnO behaves as a typical Mott insulator (Figure 7). A band gap opens at eV and grows linearly, albeit with the gradually reduced slope at higher , whereas magnetic moment on Mn systematically increases due to increased localization of the Mn states. At high values, it even exceeds the nominal value of because of the enhanced polarization of the surrounding oxygen atoms.
!Evolution of the electronic structure for MnO and Li₂Sr[MnN]₂ bbox=0.15,0.07,0.88,0.55
Figure 7. (a-c) Evolution of the electronic structure for MnO (circles) and LiSr[MnN] (diamonds) upon increasing the Coulomb repulsion parameter of DFT+: band gap (a), total DOS at the Fermi level (b), and magnetic moment on Mn (c). (d-f) Orbital-resolved DOS for a single Mn atom in LiSr[MnN] calculated at eV (d), eV (e), and eV (f). The Fermi level is at zero energy. The insets show total DOS in the vicinity of the Fermi level. The corresponding states are mostly of N origin, but their exact position and, consequently, DOS at , strongly depend on the spin polarization of Mn orbitals.
The evolution of LiSr[MnN] reveals several similarities. Its DOS at the Fermi level, , also decreases rapidly with , whereas magnetic moment increases and changes slope around eV, where drops nearly to zero. The nominal magnetic moment of is reached at eV, but in contrast to MnO, the gap never opens because a small fraction of N states always remains at the Fermi level. This residual metallicity is due to the fact that the N states are strongly mixed with Mn , especially the lowest-lying orbital. It is occupied by the minority-spin electron of Mn and remains unpolarized (Figure 7d), so it can’t be localized and gapped by adding . Only at eV does one observe the tendency toward the gradual localization of states through their spin polarization.
The steep increase in the magnetic moment of Mn above at eV reflects this tendency toward polarizing the states, with increasing as the minority-spin states associated with are shifted toward positive energies across the Fermi level (Figure 7e). This process is completed at eV, where the states are fully polarized at the cost of placing minority-spin electron onto the orbital. The resulting configuration features a pseudogap with but no gap opened even at very high –10 eV (Figure 7f).
From the experimental mJ/mol K one expects eV/f.u., which may be compatible with the weakly correlated scenario at eV or with the intermediate regime at –6 eV (Figure 7b), especially if renormalization of the effective mass and slight doping caused by the residual site mixing between Li and Mn are considered. Spectroscopic experiments and/or direct measurements of the local magnetic moment would be interesting to distinguish between these two possible regimes.
A more general outcome of our initio study is the peculiar behavior of Mn nitride as a correlated material. Here, four -orbitals show weaker hybridization with nitrogen and thus Mott behavior reminiscent of MnO, whereas the fifth -orbital has a strong charge-transfer character that causes residual metallicity due to N states remaining at the Fermi level even in the presence of strong correlations. Another unusual effect is the shift of the minority-spin electron from the lower-lying orbital with the stronger charge-transfer character to the higher-lying orbital with the weaker charge-transfer character, because the latter orbital is easier to localize. This unconventional scenario reminds of some earlier findings for a Ni oxide in the strongly correlated limit [40] and may be a general mechanism of inducing metallicity at intermediate values of . Overall, we find that Mn nitrides have a much stronger proclivity to the metallic behavior than undoped Mn oxides, and one -orbital with the strong charge-transfer character is already sufficient to trigger it.
Discussion
Absolute values and temperature-dependence of as well as DFT calculations indicate the presence of free charge carriers. There is, however, no signature in at . If the proposed AFM ordering of Mn would be caused by a spin-density-wave transition, as observed for example in the parent compounds of Fe-based superconductors [41], a clear anomaly is expected in . However, very similar transport behavior with no signature in at the AFM ordering is observed in underdoped cuprates [42] and LaMnO [43]. This is interpreted as a sign for spatial separation of conducting and magnetically ordering entities. Nevertheless, we are going to discuss and disproof the following two hypothetical scenarios:
a) the anomaly in is caused by AFM ordering of an impurity phase. Given the large absolute values of , which exceed the nearly itinerant ferromagnet Pd more than a factor of 2 [44], only a local moment bearing material seems possible. Such a phase, however, is expected to show Curie–Weiss behavior in the paramagnetic region above , which is in contrast to the linear increase of with increasing temperature observed experimentally (Figure 4(a)). Obtaining a linear as result of a superposition of two non-linear contributions (impurity Curie-Weiss and main phase with high ) seems highly unlikely. Furthermore, the anisotropy of with an increase for and a decrease for upon heating above seems incompatible with a paramagnetic local moment scenario (Figure 4(b,Figure 4c)).
b) the metallic transport is caused by a conducting impurity phase. The low absolute values of below m cm require a significant cross section of a potential current path even when highly conductive materials are considered. Optical microscopy on pristine as well as on polished surfaces never revealed indications for the presence of such a phase. Furthermore, exposure to air leads to a rapid increase of the resistivity by several orders of magnitude. We infer that also near-surface regions of LiSr[MnN] make an important contribution to the electrical transport and do not only connect spring loaded contacts and other metallic current paths. A lower air-sensitivity is expected if already insulating LiSr[MnN] merely bridges contact leads and a hypothetical conducting impurity phase.
Next, we focus on the high-temperature region ( K) of . The large absolute values of are very unusual, in particular in combination with the linear increase with temperature. The latter is normally not observed for local-moment-bearing or ferromagnetically ordered materials. All other materials we are aware of, however, show significantly smaller absolute values of . A similar linear increase of with temperature is found
in Fe-based superconductors, though with lower absolute values at around room temperature (RT): F-doped LaFeAsO with [45], BaFeAs with [45, 46], or SrFeAs [47] and CaFeAs [48] both with . The peculiar temperature-dependence was described by assuming a coexistence of localized and itinerant moments [49], frustrated local-moment model [50], or spin-density-wave scenario with strong antiferromagnetic fluctuations above [51]. Also (underdoped) cuprates show similar temperature-dependence of with even lower absolute values (), for example LaSrCuO with [52]. Note that of elemental Cr also increases with temperature with a roughly five times smaller slope and one order of magnitude lower [53] when compared to LiSr[MnN].
Observing metallic conductivity in a nitride is not unprecedented, though with significant differences to LiSr[MnN]: Antiperovskite nitrides of type [54] (with ) do show metallic conductivity and magnetic ordering (for example MnZnN [55, 56]). Those structures, however, are based on transition-metal rich hosts in strong contrast to insulating, primarily ionic LiSrN. More nitrogen-rich compounds for which a pronounced metallic temperature-dependence of was experimentally observed are (SN) [57], CaNiN [58], SrNiN [59], SrNiN [60], and CaN [61]. No indications for magnetic ordering were found in any of those materials.
Finally, we discuss the low-temperature properties with anomalies in and at . Magnetic ordering of spins ($3d^6$ spin only, according to Mn), lead to a change in entropy by . The measured magnetic entropy of amounts to only a small fraction of that value (and this does not change considering instead of ). Ordering of of Mn or a correspondingly small amount of impurities are sufficient to cause the maximum in and to account for the measured . Furthermore, the effective magnetic moment obtained from the low-temperature Curie tail amounts to per Mn (for both orientations, , see Figure 4a). Assuming that a certain fraction of Mn, , carries local magnetic moments of , according to , leads to . [Note that .] Again, a small fraction of Mn or other local moment bearing impurities is sufficient to explain the observed behavior. The perfect agreement with the fraction of isolated Mn atoms of , which follows from , however, seems to be accidental. Those Mn have two Li as next-nearest neighbor along the chain and could be decoupled from the ordered states in longer N--N--$ chains. Observing an that is one order of magnitude smaller than expected for could be caused by the large uncertainty in determining the non-magnetic background of or neglecting crystal electric field effects and exchange interactions. It could also indicate magnetic frustration or spin-glass behavior with freezing of local moments at [62]. Further similarities to spin-glasses are given by the splitting of FC and ZFC curves of and the comparatively broad feature in . Investigating the anomaly at for varying could serve as a very sensitive method to probe the Mn concentration and stoichiometry of LiSr[MnN], provided isolated Mn do indeed cause the anomaly. In particular, the question arises of whether or not the anomaly disappears for . The assumption that the vast majority of Mn particles in itinerant, magnetic ordering at and only a small fraction of isolated magnetic moments (Mn or impurities) orders at is supported by the preponderance of the obtained experimental data.
Summary
LiSr[MnN] single crystals of several millimeter along a side were grown from Li-rich flux. Clear indications for itinerant antiferromagnetic ordering at are found. The magnetic susceptibility above increases linearly with temperature and amounts to an exceptionally large value of . The magnetic ordering is not accompanied by a structural transition as shown by low temperature single crystal X-ray diffraction. Our main results can be understood based on structurally well-ordered but finite, linear chains of N--N-- that are essential for the observed metallic electrical transport and magnetic properties. A small fraction of well below $1% of isolated, local magnetic-moment-bearing Mn or a correspondingly small amount of a magnetic impurity orders/freezes at . The experimentally observed metallic resistivity is corroborated by LDA+ calculations that identified a strong mixing of N with Mn states as the cause of metallicity even in the presence of sizable electronic correlations. LiSr[MnN] is a unique magnetically ordered, metallic nitride with a highly unusual magnetic susceptibility that is reminiscent of the normal state of (underdoped) cuprate and iron-based superconductors.
Acknowledgments
The authors thank Sascha Janster, Steffen Hückmann, Klaus Wiedenmann, and Alexander Herrnberger for experimental assistance. German Hammerl is acknowledged for help with the electrical transport measurements. This work was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - grant number JE748/1.
References
- [1]F. Walz, The verwey transition - a topical review, J. Phys. Cond. Matt. 14, R285 (2002).DOI
- [2]T. Timusk and B. Statt, The pseudogap in high-temperature superconductors: an experimental survey, Rep. Prog. Phys. 62, 61 (1999).DOI
- [3]C. Putzke, M. D. Bachmann, P. McGuinness, E. Zhakina, V. Sunko, M. Konczykowski, T. Oka, R. Moessner, A. Stern, M. König, S. Khim, A. P. Mackenzie, and P. J. W. Moll, h/e oscillations in interlayer transport of delafossites, Science 368, 1234 (2020).DOI
- [4]P. Höhn and R. Niewa, Nitrides of non-main group elements, in Handbook of Solid State Chemistry (Wiley Online Library, 2017) pp. 251–359.DOI
- [5]R. Niewa, F. Wagner, W. Schnelle, O. Hochrein, and R. Kniep, Li₂₄[MnN₃]₃N₂ and Li₅[(Li₁−ₓMnₓ)N]₃, two intermediates in the decomposition path of Li₇[MnN₄] to Li₂([Li₁−xMnx]N): An experimental and theoretical study, Inorg. Chem. 40, 5215 (2001).
- [6]D. Gregory, M. Barker, P. Edwards, and D. Siddons, Synthesis and structure of Ca₆MnN₅: A new nitridomanganate with planar [MnN₃]⁶− anions, Inorg. Chem. 34, 5195 (1995).DOI
- [7]A. Tennstedt, C. Röhr, and R. Kniep, Ca₃[MnN₃], ein neues nitridomanganat (III): Anionen [MnIIIN₃]⁶− mit C₂v-Symmetrie/Ca₃[MnN₃], a new nitridomanganate (III): Anions [MnIIIN₃]⁶− with C₂v-symmetry, Z. Naturforsch. B 48, 188 (1993).
- [8]A. Tennstedt, C. Röhr, and R. Kniep, Sr₃[MnN₃] und Ba₃[MnN₃], die ersten nitridomanganate (III): Trigonal-planare anionen [MnIIIN₃]⁶−/Sr₃[MnN₃] and Ba₃[MnN₃], the first nitridomanganates (III): Trigonal-planar anions [MnIIIN₃]⁶−, Z. Naturforsch. B 48, 794 (1993).
- [9]M. G. Barker, M. J. Begley, P. P. Edwards, D. H. Gregory, and S. E. Smith, Synthesis and crystal structures of the new ternary nitrides Sr₃CrN₃ and Ba₃CrN₃, J. Chem. Soc. Dalton 1, 1 (1996).
- [10]J. K. Bendyna, P. Höhn, and R. Kniep, Crystal structure of tristrontium trinitridoferrate(III), Sr₃[FeN₃], Z. Kristallogr. - New Cryst. Struct. 223, 109 (2008).
- [11]O. Hochrein, Yu. Grin, and R. Kniep, The hexanitridomanganate (iv) Li₆Ca₂[Mn₂N₆]: Preparation, crystal structure, and chemical bonding, Angew. Chem. Int. Edit. 37, 1582 (1998).
- [12]O. Hochrein, P. Höhn, and R. Kniep, Control of oxidation states of transition elements (Cr, Mn) in nitrometallates and the solid solution series Li₆(Ca₁−xSrₓ)₂ [Mn₂N₆], Z. Anorg. Allg. Chem. 629, 923 (2003).
- [13]A. Ovchinnikov, M. Bobnar, Yu. Prots, W. Schnelle, P. Höhn, and Yu. Grin, Ba₄[Mn₃N₆], a quasi-one-dimensional mixed-valent nitridomanganate (II, IV), Crystals 8, 235 (2018).
- [14]A. Ovchinnikov, M. Bobnar, Yu. Prots, H. Borrmann, J. Sichelschmidt, Yu. Grin, and P. Höhn, Ca₁₂[Mn₁₉N₂₃] and Ca₁₃₃[Mn₂₁₆N₂₆₀]: Structural complexity by 2D intergrowth, Angew. Chem. 130, 11753 (2018).
- [15]R. Niewa, F. DiSalvo, D.-K. Yang, D. Zax, H. Luo, and W. Yelon, Synthesis, crystal structure and properties of a lithium manganese nitride, (Li, Mn)₂N, J. Alloys Compd. 266, 32 (1998).
- [16]J. Klatyk and R. Kniep, Crystal structure of dilithium (nitrilodithiate/manganate (I)), Li₂[(Li₁−xMnx)N], x = 0.73, Z. Krist. - New Cryst. Struc. 214, 445 (1999).
- [17]J. Klatyk, R. Niewa, and R. Kniep, Ca{Li₂[MnIN₂]}: Kristallchemischer brückenschlag zwischen lithiumnitrido-verbindungen und borid-carbiden des typs M[B₂C₂]/Ca{Li₂[MnIN₂]}: Link between crystal chemistry of lithium nitrido compounds and boride carbides of the type M[B₂C₂], Z. Naturforsch. B 55, 988 (2000).
- [18]M. Boström and S. Hovmöller, Preparation and crystal structure of the novel decagonal approximant Mn₁₂₃Ga₁₃₇, J. Alloys Compd. 314, 154 (2001).
- [19]A. Jesche and P. C. Canfield, Single crystal growth from light, volatile and reactive materials using lithium and calcium flux, Philos. Mag. 94, 2372 (2014).DOI
- [20]P. Höhn, T. J. Ballé, M. Fix, Yu. Prots, and A. Jesche, Single Crystal Growth and Anisotropic Magnetic Properties of Li₂Sr[Li₁−xFex]N₂, Inorganics 4, 42 (2016).
- [21]T. J. Ballé, Z. Zangeneh, L. Hožoi, A. Jesche, and P. Höhn, Ferromagnetic ordering of linearly coordinated co ions in Li₃Sr[CoN₂], Phys. Rev. B 99, 094422 (2019).
- [22]M. Pathak, D. Stoiber, M. Bobnar, A. Ovchinnikov, A. Ormeci, R. Niewa, and P. Höhn, Synthesis, characterization, and chemical bonding analysis of the lithium alkaline-earth metal gallide nitrides Li₂(Ca₃N₂)[Ga₄] and Li₂(Sr₃N₂)₂[Ga₄], Z. Anorg. Allg. Chem. 643, 1557 (2017).
- [23]P. Höhn et al., unpublished.
- [24]WinXPow, STOE & Cie GmbH: Darmstadt, Germany (2003).
- [25]O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. Howard, and H. Puschmann, Olex2: a complete structure solution, refinement and analysis program, J. Appl. Cryst. 42, 339 (2009).
- [26]G. M. Sheldrick, A short history of SHELX, Acta Crystallogr. A 64, 112 (2008).pubmed.ncbi.nlm.nih.gov/18156677
- [27]G. M. Sheldrick, Crystal structure refinement with SHELXL, Acta Crystallogr. C 71, 3 (2015).DOI
- [28]S. X-RED, Data reduction program, STOE & Cie GmbH, Darmstadt (1999).
- [29]X. STOE, Shape, crystal optimisation for numerical absorption correction, version 1.06/windows (1999).
- [30]F. X. Eder, Moderne Messmethoden der Physik, edited by R. Romp and E. Schmutzer, Hochschulbücher für Physik (VEB Deutscher Verlag der Wissenschaften, 1972).
- [31]K. Koeppernik and H. Eschrig, Full-potential nonorthogonal local-orbital minimum-basis band-structure scheme, Phys. Rev. B 59, 1743 (1999).DOI
- [32]J. P. Perdew and Y. Wang, Accurate and simple analytic representation of the electron-gas correlation energy, Phys. Rev. B 45, 13244 (1992).
- [33]P. C. Canfield and I. R. Fisher, High-temperature solution growth of intermetallic single crystals and quasicrystals, J. Cryst. Growth 225, 155 (2001).
- [34]G. Cordier, A. Gudat, R. Kniep, and A. Rabenau, LiCaN and Li₄SrN₂, derivatives of the fluorite and lithium nitride structures, Angew. Chem. Int. Ed. 28, 1702 (1989).
- [35]J. Klatyk and R. Kniep, Crystal structure of alkaline earth dilithium bis(nitridolithiate/ferrates(I)), Ca{Li₂[(Li₁−xFex)N]₂}, x = 0.30 and Sr{Li₂[(Li₁−xFex)N]₂}, x = 0.46, Z. Kristallogr. - New Cryst. Struct. 214, 449 (1999).
- [36]J. Jäger and R. Kniep, Nitridocuprate(I): Quaternære Phasen in den Systemen Li-Ca/Sr-Cu-N., Z. Naturforsch. B 47, 1290 (1992).
- [37]R. Kniep and P. Höhn, 2.06 - Low-Valency Nitidometalates, Reference Module in Chemistry, Molecular Sciences and Chemical Engineering, Comprehensive Inorganic Chemistry II (Second Edition), Volume 2, 137 (2013).
- [38]A. Rabenau and H. Schulz, Re-evaluation of the lithium nitride structure, J. Less-Common Met. 50, 155 (1976).DOI
- [39]V. J. Emery and S. A. Kivelson, Superconductivity in bad metals, Phys. Rev. Lett. 74, 3253 (1995).
- [40]K.-W. Lee and W. E. Pickett, Infinite-layer LaNiO₂: Ni¹⁺ is not Cu²⁺, Phys. Rev. B 70, 165109 (2004).DOI
- [41]H.-H. Klauss, H. Luetkens, R. Klingeler, C. Hess, F. J. Litterst, M. Kraken, M. M. Korshunov, I. Eremin, S.-L. Drechsler, R. Khasanov, A. Amato, J. Hamann-Borrero, N. Leps, A. Kondrat, G. Behr, J. Werner, and B. Büchner, Commensurate spin density wave in lafeaso: A local probe study, Phys. Rev. Lett. 101, 077005 (2008).pubmed.ncbi.nlm.nih.gov/18764570
- [42]Y. Ando, A. N. Lavrov, S. Komiya, K. Segawa, and X. F. Sun, Mobility of the doped holes and the antiferromag- netic correlations in underdoped high-Tc cuprates, Phys. Rev. Lett. 87, 017001 (2001).
- [43]S. de Brion, F. Ciorcas, G. Chouteau, P. Lejay, P. Radaelli, and C. Chaillout, Magnetic and electric properties of La1−xMnO3, Phys. Rev. B 59, 1304 (1999).DOI
- [44]W. Gerhardt, F. Razavi, J. S. Schilling, D. Hüser, and J. A. Mydosh, Pressure dependence of the magnetic susceptibility of pure Pd to 3.7 GPa from 1.3 to 300 K, Phys. Rev. B 24, 6744 (1981).DOI
- [45]R. Klingeler, N. Leps, I. Hellmann, A. Popa, U. Stockert, C. Hess, V. Kataev, H.-J. Grafe, F. Hammerath, G. Lang, S. Wurmehl, G. Behr, L. Harnagea, S. Singh, and B. Büchner, Local antiferromagnetic correlations in the iron pnictide superconductors LaFeAsO1−xFx and Ca(Fe1−xCox)2As2 as seen via normal-state susceptibility, Phys. Rev. B 81, 024506 (2010).DOI
- [46]X. F. Wang, T. Wu, G. Wu, H. Chen, Y. L. Xie, J. J. Ying, Y. J. Yan, R. H. Liu, and X. H. Chen, Anisotropy in the Electrical Resistivity and Susceptibility of Superconducting BaFe2As2 Single Crystals, Phys. Rev. Lett. 102, 117005 (2009).pubmed.ncbi.nlm.nih.gov/19392232
- [47]J.-Q. Yan, A. Kreyssig, S. Nandi, N. Ni, S. L. Bud’ko, A. Kracher, R. J. McQueeney, R. W. McCallum, T. A. Lograsso, A. I. Goldman, and P. C. Canfield, Structural transition and anisotropic properties of single-crystalline SrFe2As2, Phys. Rev. B 78, 024516 (2008).DOI
- [48]F. Ronning, T. Klimczuk, E. D. Bauer, H. Volz, and J. D. Thompson, Synthesis and properties of CaFe2As2 single crystals, J. Phys. Condens. Matt. 20, 322201 (2008).
- [49]S.-P. Kou, T. Li, and Z.-Y. Weng, Coexistence of itinerant electrons and local moments in iron-based superconductors, Europhys. Lett. 88, 17010 (2009).DOI
- [50]B. Schmidt, M. Siahatgar, and P. Thalmeier, Frustrated local-moment models for iron pnictide magnetism, Phys. Rev. B 81, 165101 (2010).DOI
- [51]G. M. Zhang, Y. H. Su, Z. Y. Lu, Z. Y. Weng, D. H. Lee, and T. Xiang, Universal linear-temperature dependence of static magnetic susceptibility in iron pnictides, Europhys. Lett. 86, 37006 (2009).DOI
- [52]T. Nakano, M. Oda, C. Manabe, N. Momono, Y. Miura, and M. Ido, Magnetic properties and electronic conduction of superconducting La2−xSrxCuO4, Phys. Rev. B 49, 16000 (1994).pubmed.ncbi.nlm.nih.gov/10010737
- [53]T. R. McGuire and C. J. Kriessman, The magnetic susceptibility of chromium, Phys. Rev. 85, 452 (1952).DOI
- [54]P. Ettmayr, Über ein mangan-quecksilber-nitrid mit perowskitstruktur, Monatshefte für Chemie 98, 1881 (1967).DOI
- [55]W. S. Kim, E. O. Chi, J. C. Kim, N. H. Hur, K. W. Lee, and Y. N. Choi, Crystallographically magnetic ordering in the antiperovskite ZnNMn3, Phys. Rev. B 68, 174202 (2003).DOI
- [56]Y. Sun, Y. Guo, Y. Tsujimoto, J. Yang, B. Shen, W. Yi, Y. Matsushita, C. Wang, X. Wang, J. Li, C. I. Sathish, and K. Yamaura, Carbon-Induced Ferromagnetism in the Antiferromagnetic Metallic Host Material Mn3ZnN, Inorg. Chem. 52, 800 (2013).DOI
- [57]V. V. Walatka, M. M. Labes, and J. H. Perlstein, Polysulfur nitride—a one-dimensional chain with a metallic ground state, Phys. Rev. Lett. 31, 1139 (1973).DOI
- [58]M. Y. Chern and F. J. DiSalvo, Synthesis, structure, electric, and magnetic properties of Ca3NiN, J. Solid State Chem. 88, 459 (1990).
- [59]T. Yamamoto, S. Kikkawa, and F. Kanamaru, Preparation, Crystal Structure, and Properties of a New Double Metal Nitride, SrNiN, and of Ca1−xSrxNiN (0 ≤ x ≤ 0.5) Solid Solutions, J. Solid State Chem. 115, 353 (1995).DOI
- [60]G. R. Kowach, N. E. Brese, U. M. Bolle, C. J. Warren, and F. J. DiSalvo, An Unusual Metallic Nitride: Br2NiN2, J. Solid State Chem. 154, 542 (2000).DOI
- [61]K. Lee, S. W. Kim, Y. Toda, S. Matsuishi, and H. Hosono, Dicalcium nitride as a two-dimensional electr ride with an anionic electron layer, Nature 494, 336 (2013).DOI
- [62]J. A. Mydosh, Spin glasses: an experimental introduction (Taylor & Francis, London + Washington, DC, 1993).