Li and Co Ordering in the Nitridocobaltate(I) SrLi2{Li[CoN2]}

SrLi2{Li[CoN2]}, an isostructural variant of Li4SrN2, has been synthesised as black single crystals from a reaction between Li2[(Li,Co)N] and Sr2N, at 973 K using a Li flux in a sealed tantalum ampoule. Single crystal diffraction refinements gave a tetragonal unit cell, which upon closer inspection showed a monoclinic supercell. This supercell allowed, for the first time, the refinement of the occupation of metal atoms along the infinite chains in the structure, resulting in the chemical formula SrLi2{Li0.65Co0.35[Co0.65Li0.35N2]}. This revealed a clear preference for the Li and Co atoms to alternate along the chains. Magnetic measurements showed the sample to be a Curie paramagnet, with Co(I) being in a high-spin S = 1 configuration.


Introduction
Lithium-containing nitrides continue to be a topic of interest for both industry and solid state chemists, in part due to the potential application of binary and ternary lithium nitrides as an anode material for batteries [1,2].The structure of the ternary nitride Li 4 SrN 2 resembles the α-Li 3 N structure [2,3], but instead of nitrogen containing hexagonal bipyramids of lithium, there are in fact distorted pentagonal bipyramids present.These are built from three lithium and two strontium atoms on the equatorial, and two lithium atoms on the axial positions.The bonding of the axial lithium with nitrogen results in one-dimensional infinite linear 1  ∞ [LiN 2 2/2 ] chains.The strontium atoms are coordinated by four nitrogen atoms to form distorted tetrahedra, which causes the 1 ∞ [LiN 2 2/2 ] chains to run along [100] and [010] directions, within the tetragonal unit cell, resulting in blocks of parallel chains that are perpendicular to one another.
Since the discovery of the ternary nitride Li 4 SrN 2 [3], there have been several examples of substitution of the lithium position by a transition metal [3][4][5][6][7][8].Some of these substitutions result in compounds isotypic to Li 4 SrN 2 , with the composition SrLi 2 [(Li 1-x M x )N] 2 (M = Fe (x = 0.46), Ni (x = 0.05) and Cu (x = 0.22-0.39))[3][4][5].Typically, quaternary nitridometalates form in the Li 4 SrN 2 -type structure, with the transition metal substituting solely with the lithium position on the infinite linear chains, resulting in a mixed occupation of lithium and transition metal atoms along these chains.To date, there has been no reported evidence or investigation into possible ordering within these partially substituted chains, except for the fully ordered chains in Sr 2 Li[CoN 2 ] [7].
The introduction of transition metals into nitride systems gives rise to a large range of oxidation states, including ones that are not often seen within other systems.For nitrides containing metals from Group 6-11, the +1 oxidation state is common, particularly for Co, Ni, and Cu.This preference for low valency means that there is solely a 2-fold coordination environment for these metals, such as dumbbells, chains or chain fragments [9].By using this preference, synthesis of Co(I)-containing nitrides has led to very intriguing structures.These compounds range from the simple substitution of This chemistry may even be seen extended to typical interstitial phases like Fe 2 CoN, with the transition metal in a 2-fold non-linear coordination [11].
Herein, we report for the first time the synthesis of the new compound SrLi 2 {Li[CoN 2 ]}, along with evidence of a preferred ordering along the linear 1  ∞ [(Li, Co)N 2 2/2 ] chains.

Materials and Methods
Due to the reactivity of the materials, all operations were conducted under inert conditions in an argon filled glovebox.
Elemental lithium rods (0.980 g, Chemetall) had the outer surface removed, using a knife, to ensure that no oxides were present.The rod was cut into pieces approximately 5 mm in length.All pieces were placed into a tantalum crucible and then into a silica tube.The tube was heated under nitrogen (Alphagaz, 99.999%) flow, which was purified by passing over a molecular sieve (4 Å) and a BTS catalyst, at 443 K for 36 hours, before increasing the temperature to 498 K at 10 K/h for 2 h to complete the reaction.The system was then allowed to cool to room temperature.This resulted in single-phase α-Li 3 N, which was confirmed by powder X-ray diffraction (PXRD).A thoroughly grounded mixture of this product and cobalt powder (Merck) with a molar ratio of 3:1 respectively, was pressed into a tablet weighing 1.003 g, with a diameter of 8 mm.The mixture was placed in a silica tube and heated, under nitrogen flow, to 573 K for 2 h, and then the temperature was increased to 973 K, at 50 K/h, for 8 h, before being allowed to cool to room temperature [12].This gave the single-phase black product Li 2 [(Li,Co)N], which was confirmed by PXRD.
Elemental strontium pieces (Sigma Aldrich, 99.99%) were reacted under nitrogen flow at 823 K for 8 h and allowed to cool to room temperature.The sample was then ground into a powder and reacted under the same conditions to ensure a complete reaction to Sr 2 N.This was then confirmed to be single-phase by PXRD.
A tantalum ampoule (Aldrich) was filled with Li 2 [(Li,Co)N] (0.127 g) and Sr 2 N (0.128 g), along with additional elemental lithium (0.250 g, Chemetall) serving as a flux.The ampoule was then sealed using an arc welder, which was integrated into an argon-filled glove box, and heated in a silica tube, under argon flow, to 973 K and allowed to cool slowly to room temperature at 1.5 K/h.The tantalum ampoule was then opened and the lithium flux was removed using liquid ammonia, leaving behind a sample containing a mixture of black and metallic silver crystals, with the metallic silver crystals having yet to be identified.Due to strong fluorescence between the X-rays and the strontium and cobalt within the sample, with both available X-ray wavelengths (Mo-K α , Cu-K α ), a reliable powder diffraction pattern has yet to be produced.
Diffraction experiments were carried out using a STOE STADI P powder X-ray diffractometer with Mythen1K micro-strip detector in a transmission geometry (Mo-K α radiation, λ = 0.7093 Å) or a STOE STADI P powder X-ray diffractometer (STOE, Darmstadt, Germany) equipped with a germanium-monochromator (Cu-K α radiation, λ = 1.5406Å).Powder diffraction data was analysed using the STOE WinXPOW software package [13].Single crystal X-ray diffraction measurements were taken with a .This chemistry may even be seen extended to typical interstitial phases like Fe2CoN, with the transition metal in a 2-fold non-linear coordination [11].Herein, we report for the first time the synthesis of the new compound SrLi2{Li[CoN2]}, along with evidence of a preferred ordering along the linear

Materials and Methods
Due to the reactivity of the materials, all operations were conducted under inert conditions in an argon filled glovebox.
Elemental lithium rods (0.980 g, Chemetall) had the outer surface removed, using a knife, to ensure that no oxides were present.The rod was cut into pieces approximately 5 mm in length.All pieces were placed into a tantalum crucible and then into a silica tube.The tube was heated under nitrogen (Alphagaz, 99.999%) flow, which was purified by passing over a molecular sieve (4 Å) and a BTS catalyst, at 443 K for 36 hours, before increasing the temperature to 498 K at 10 K/hr for 2 hr to complete the reaction.The system was then allowed to cool to room temperature.This resulted in single-phase α-Li3N, which was confirmed by powder X-ray diffraction (PXRD).A thoroughly grounded mixture of this product and cobalt powder (Merck) with a molar ratio of 3:1 respectively, was pressed into a tablet weighing 1.003 g, with a diameter of 8 mm.The mixture was placed in a silica tube and heated, under nitrogen flow, to 573 K for 2 hr, and then the temperature was increased to 973 K, at 50 K/hr, for 8 hr, before being allowed to cool to room temperature [12].This gave the single-phase black product Li2[(Li,Co)N], which was confirmed by PXRD.
Elemental strontium pieces (Sigma Aldrich, 99.99%) were reacted under nitrogen flow at 823 K for 8 hr and allowed to cool to room temperature.The sample was then ground into a powder and reacted under the same conditions to ensure a complete reaction to Sr2N.This was then confirmed to be single-phase by PXRD.
A tantalum ampoule (Aldrich) was filled with Li2[(Li,Co)N] (0.127 g) and Sr2N (0.128 g), along with additional elemental lithium (0.250 g, Chemetall) serving as a flux.The ampoule was then sealed using an arc welder, which was integrated into an argon-filled glove box, and heated in a silica tube, under argon flow, to 973 K and allowed to cool slowly to room temperature at 1.5 K/hr.The tantalum ampoule was then opened and the lithium flux was removed using liquid ammonia, leaving behind a sample containing a mixture of black and metallic silver crystals, with the metallic silver crystals having yet to be identified.Due to strong fluorescence between the X-rays and the strontium and cobalt within the sample, with both available X-ray wavelengths (Mo-Kα, Cu-Kα), a reliable powder diffraction pattern has yet to be produced.
Measurements of the magnetic susceptibility were conducted on a Magnetic Property Measurement System (MPMS3, Quantum Design, San Diego, USA).The single crystal used for the diffraction experiments, a black shard, was placed into a gelatin capsule and attached on the sample holder of a vibrating sample magnetometer (VSM) for measuring the magnetization.The sample was examined in a temperature range of between 5-300 K, under a homogenous magnetic field of 10 kOe.After which the sample was examined in a magnetic field range of between 0 and 50 kOe, with a constant temperature of 300 K.
Single crystal X-ray diffraction measurement, of one of the black crystals, resulted in reflections that were originally assigned to a body centred tetragonal crystal lattice (a = 3.7414(2) Å, c = 27.931(2)Å) and refined in the space group I4 1 /amd (Nr.141).The unit cell parameters were in agreement with other known compounds of SrLi 2 [(Li 1-x M x )N] 2 (M = Li, Fe, Ni and Cu) [3][4][5], for example Li 4 SrN 2 (a = 3.822 Å , c = 27.042Å) [3].Structure refinements were in agreement with this structure type (Table 1).However, upon closer inspection of the diffraction data, through the use of reconstructed diffraction patterns, diffuse streaks superimposed with reflections were observed, except for hkl where l = 2n (Figure 1).By overlaying the allowed reflections for a tetragonal lattice onto reconstructed diffraction patterns all reflections were described, except for the diffuse reflections (Figure 1e).In addition, the refinement gave a high value for R int = 0.2109, suggesting that the chosen symmetry is in fact too high.This, along with splitting of reflections at higher angles, resulted in the decision of going to a monoclinic system.The use of a pseudomerohedral twinned monoclinic lattice (twinning matrix of −1 0 0 0 −1 0 1 0 1) was decided, with the twinning occurring along a twofold axis parallel to the direction [1 0 2], which correctly described the diffuse reflections and splitting (Figure 1f From this comparison, it was concluded that a tetragonal space group was insufficient to describe the crystal structure correctly.A monoclinic subgroup, P2 1 /c, of the original tetragonal space group, I4 1 /amd, was determined through a Bärnighausen symmetry tree diagram (Figure 2), allowing the atomic position of the mixed Li/Co on the 1  ∞ [(Li, Co)N 2 2/2 ] chains to be split into two independent positions giving the possibility of ordering along these chains.The refinement resulted in a composition of SrLi 2 {Li 0.753 (7) Co 0.246 (7) [Co 0.748(6) Li 0.252(6) N 2 ]}; however refinement of the anisotropic displacements of the mixed valence Li/Co and nitrogen positions proved difficult.The anisotropic displacement parameters were, however, able to be refined once the value of the mixed occupation was fixed to values close to those calculated from the refinement of the tetragonal subcell.This gave a more accurate refinement (Tables 1 and 2), resulting in a final composition of SrLi 2 {Li 0.65 Co 0.35 [Co 0.65 Li 0.35 N 2 ]}, while also describing the diffuse reflections.The variation in compositions and volume between the subcell and supercell resulted in a slight discrepancy between certain refined values (Table 1).Due to the splitting of the metal position along the infinite linear chains, the refinement showed a strong indication of ordering, resulting in chains in which Li and Co are almost alternating (Figure 3).The near ordering, according to the refinement, would imply that the metal cations alternate along the chains, which would be expected when viewed from a chemical perspective.This is in agreement with the similar ordering seen in Sr 2 Li[CoN 2 ] [7], which is structurally similar to SrLi 2 {Li 0.65 Co 0.35 [Co 0.65 Li 0.35 N 2 ]}.The influence of this large amount of Co could be seen when observing the bond lengths along the 1 ∞ [(Li, Co)N 2 2/2 ] chains (Table 3).When compared with the Li 4 SrN 2 structure, the Li(1)/Co(1)-N(1)/(2) interatomic distances (1.89(2) Å/1.92(2) Å), where the position is occupied primarily with lithium, were in clear agreement with the value found for Li 4 SrN 2 (1.913 Å) [3].However, when compared with the position occupied with mainly cobalt, Li(2)/Co(2)-N(1)/(2) interatomic distances (1.87(2) Å/1.81(2) Å), the interatomic distances were shorter and tended more towards the typical Co-N interatomic distance found in Li 2 [Li 0.57 Co 0.43 N] (1.824 Å) [15].These interatomic distances were in fact also shorter than the known distances for M-N in the compounds SrLi 2 [(Li 1-x M x )N] 2 (M = Fe (1.896 Å) [4], Ni (1.912 Å) [5] and Cu (1.885 Å) [6]).

Measurements of the Magnetic Susceptibility
The presence of Co(I) cations along the 1 ∞ [(Li, Co)N 2 2/2 ] chains gave rise to the question on the spin-state and possible magnetic ordering.Measurements of the magnetization were conducted, using the identical crystal from the single crystal X-ray diffraction experiments, with a varying magnetic field (Figure 4a) and varying temperature (Figure 4b).The shape of the hysteresis and relation of magnetic susceptibility to temperature shows a Curie-Weiss paramagnetic behaviour and no indication of long-range ordering at low temperatures.
Calculations using a modified Curie law (Equation ( 1)) gave an experimental magnetic moment for Co(I) of µ exp = 2.62 µ B .This value has a high uncertainty, due to the low weight of the sample.However, the experimental magnetic moment is in close agreement with a spin-only magnetic moment calculation for S = 1 (µ cal = 2.83 µ B ), indicating a high-spin configuration of linearly coordinated Co(I).The temperature-independent susceptibility (χ 0 ) was quite high (χ 0 = 0.109 cm 3 /mol), which could indicate the presence of a small impurity of elemental ferromagnetic Co, probably on the surface of the crystal.The Weiss constant from these measurements was negative (θ = −1.87K), this suggests antiferromagnetic exchange interactions dominating among the moments of the Co(I) cations.This kind of behaviour was previously observed in Li 2 [(Li 1−x Ni x )N] (0.06 ≤ x ≤ 0.85) and Li 2 [(Li 1−x Co x )N] (0.064 ≤ x ≤ 0.36) [17,18], both of which contain infinite chains with the transition metal.In both cases an increase in the transition metal content resulted in an increase of the antiferromagnetic interactions, paired with a decrease in the experimental magnetic moment.It would be of interest if a similar trend could be observed when the cobalt content for the titular compound is also altered.magnetic field (Figure 4a) and varying temperature (Figure 4b).The shape of the hysteresis and relation of magnetic susceptibility to temperature shows a Curie-Weiss paramagnetic behaviour and no indication of long-range ordering at low temperatures.

= -+
Calculations using a modified Curie law (Equation ( 1)) gave an experimental magnetic moment for Co(I) of µexp = 2.62 µB.This value has a high uncertainty, due to the low weight of the sample.However, the experimental magnetic moment is in close agreement with a spin-only magnetic moment calculation for S = 1 (µcal = 2.83 µB), indicating a high-spin configuration of linearly coordinated Co(I).The temperature-independent susceptibility ( ) was quite high ( = 0.109 cm 3 /mol), which could indicate the presence of a small impurity of elemental ferromagnetic Co, probably on the surface of the crystal.The Weiss constant from these measurements was negative (θ = −1.87K), this suggests antiferromagnetic exchange interactions dominating among the moments of the Co(I) cations.This kind of behaviour was previously observed in Li2[(Li1−xNix)N] (0.06 ≤ x ≤ 0.85) and Li2[(Li1−xCox)N] (0.064 ≤ x ≤ 0.36) [17,18], both of which contain infinite chains with the transition metal.In both cases an increase in the transition metal content resulted in an increase of the antiferromagnetic interactions, paired with a decrease in the experimental magnetic moment.It would be of interest if a similar trend could be observed when the cobalt content for the titular compound is also altered.

Conclusions
In summary, SrLi 2 {Li 0.65 Co 0.35 [Co 0.65 Li 0.35 N 2 ]} has been synthesised, which is the first Co-containing member of the SrLi 2 [(Li,M)N] 2 family (M = transition metal).This compound crystallises in a supercell in the P2 1 /c space group and shows atomic Li/Co ordering along the infinite chains.This is the first evidence for atomic ordering along infinite chains to have been observed in such a nitride.The material is paramagnetic, exhibiting Curie-Weiss behaviour, with no indication of low temperature magnetic ordering, and the Co + cations along the infinite linear chains are in a high-spin S = 1 state.

Figure 4 .
Figure 4. SrLi 2 {Li 0.65 Co 0.35 [Co 0.65 Li 0.35 N 2 ]}: (a) Hysteresis of the magnetic moment against varying intensity of a magnetic field at 300 K; (b) Magnetic susceptibility per mol, with varying temperature at 10 kOe (black) and a fitted Curie-Weiss law curve (red); (c) Inverse magnetic susceptibility per mol with varying temperature at 10 kOe (black) and a fitted inverse Curie-Weiss law line between 40 and 150 K (red).
1Value was treated isotropically during refinement.