Synthesis , Crystal Structure , Polymorphism , and Magnetism of Eu ( CN 3 H 4 ) 2 and First Evidence of EuC ( NH ) 3

We report the first magnetically coupled guanidinate, α-Eu(CN3H4)2 (monoclinic, P21, a = 5.8494(3) Å, b = 14.0007(8) Å, c = 8.4887(4) Å, β = 91.075(6)◦, V = 695.07(6) Å3, Z = 4). Its synthesis, polymorphism, crystal structure, and properties are complemented and supported by density-functional theory (DFT) calculations. The α-, βand γ-polymorphs of Eu(CN3H4)2 differ in powder XRD, while the γ-phase transforms into the β-form over time. In α-Eu(CN3H4)2, Eu is octahedrally coordinated and sits in one-dimensional chains; the guanidinate anions show a hydrogen-bonding network. The different guanidinate anions are theoretically predicted to adopt syn-, antiand all-trans-conformations. Magnetic measurements evidence ferromagnetic interactions, presumably along the Eu chains. Finally, EuC(NH)3 (isostructural to SrC(NH)3 and YbC(NH)3, hexagonal, P63/m, a = 5.1634(7) Å, c = 7.1993(9) Å, V = 166.23(4) Å3, Z = 2) is introduced as a possible ferromagnet.


Introduction
At the beginning of the 21st century, rare-earth metals are critical materials in high-technology applications [1].Within the recent decades, several technological innovations disrupted the rare-earth market [2], in turn stimulating the scientific quest for future materials.One vibrant field is the study of Eu 2+ compounds whose complex crystal structures are coupled with application-relevant properties including, to name only some recent examples, luminescence [3][4][5][6], field-induced reversal of the magnetoresistive effect [7], and complex magnetism [8,9].The most renowned magnetic compounds are the europium chalcogenides that are considered ideal 3D Heisenberg systems [10].While EuO is a ferromagnet with a Curie temperature of 69.3 K [11][12][13], EuS is also a ferromagnet but with a far lower Curie temperature of 18.7 K, showing weak, secondary antiferromagnetic interactions [13].
In the following, we present the first magnetically coupled guanidinate, α-Eu(CN 3 H 4 ) 2 , with probable 1D ferromagnetic order.We detail its synthesis, polymorphism, crystal structure, and properties, complemented and supported by density-functional theory (DFT) calculations.Also, we present a preliminary report on EuC(NH) 3 and a first indication of its magnetic properties.
In the following, we present the first magnetically coupled guanidinate, α-Eu(CN3H4)2, with probable 1D ferromagnetic order.We detail its synthesis, polymorphism, crystal structure, and properties, complemented and supported by density-functional theory (DFT) calculations.Also, we present a preliminary report on EuC(NH)3 and a first indication of its magnetic properties.

Polymorphism of Eu(CN3H4)2
Depending on the synthetic conditions, three polymorphs of Eu(CN3H4)2 could be prepared, which we call α, β, and γ.The polymorphs show different powder X-ray diffraction (PXRD) patterns (Figure 2a).For the α-phase, the crystal structure was solved (see below).α-Eu(CN3H4)2 was prepared at temperatures around 65 °C, the β-phase at a lower 50 °C, and the γ-phase exclusively around room temperature.Under these conditions, Eu 2+ is the stable oxidation state, and Eu 3+ would only form at temperatures starting around 300 °C [32].The oxidation state was also corroborated by the magnetic measurements (see Section 2.3).Interestingly, the γ-phase spontaneously transforms to the β-phase over several weeks (Figure 2b), so γ-Eu(CN3H4)2 must be a metastable phase.IR measurements indicate a strong similarity between the α-and β-phase (Figure 3), not surprisingly so because the spectrum is dominated by the CN3H4 − anion vibrations [31].In addition, there is no trace of an IR contribution of the C(NH)3 2− unit, which supports the proposed composition.The IR spectrum of α-Eu(CN3H4)2 was also calculated by DFT from the density of phonon states and the Born effective charges as explained in references [33] and [34].All observed signals were reproduced, while the differing intensity of the simulated signals could be caused by a thermal effect (calculated 0 K vs experimental 300 K).To identify the vibrations, the IR-active phonons at the Γpoint were visualized (

Polymorphism of Eu(CN 3 H 4 ) 2
Depending on the synthetic conditions, three polymorphs of Eu(CN 3 H 4 ) 2 could be prepared, which we call α, β, and γ.The polymorphs show different powder X-ray diffraction (PXRD) patterns (Figure 2a).For the α-phase, the crystal structure was solved (see below).α-Eu(CN 3 H 4 ) 2 was prepared at temperatures around 65 • C, the β-phase at a lower 50 • C, and the γ-phase exclusively around room temperature.Under these conditions, Eu 2+ is the stable oxidation state, and Eu 3+ would only form at temperatures starting around 300 • C [32].The oxidation state was also corroborated by the magnetic measurements (see Section 2.3).Interestingly, the γ-phase spontaneously transforms to the β-phase over several weeks (Figure 2b), so γ-Eu(CN 3 H 4 ) 2 must be a metastable phase.
In the following, we present the first magnetically coupled guanidinate, α-Eu(CN3H4)2, with probable 1D ferromagnetic order.We detail its synthesis, polymorphism, crystal structure, and properties, complemented and supported by density-functional theory (DFT) calculations.Also, we present a preliminary report on EuC(NH)3 and a first indication of its magnetic properties.

Polymorphism of Eu(CN3H4)2
Depending on the synthetic conditions, three polymorphs of Eu(CN3H4)2 could be prepared, which we call α, β, and γ.The polymorphs show different powder X-ray diffraction (PXRD) patterns (Figure 2a).For the α-phase, the crystal structure was solved (see below).α-Eu(CN3H4)2 was prepared at temperatures around 65 °C, the β-phase at a lower 50 °C, and the γ-phase exclusively around room temperature.Under these conditions, Eu 2+ is the stable oxidation state, and Eu 3+ would only form at temperatures starting around 300 °C [32].The oxidation state was also corroborated by the magnetic measurements (see Section 2.3).Interestingly, the γ-phase spontaneously transforms to the β-phase over several weeks (Figure 2b), so γ-Eu(CN3H4)2 must be a metastable phase.IR measurements indicate a strong similarity between the α-and β-phase (Figure 3), not surprisingly so because the spectrum is dominated by the CN3H4 − anion vibrations [31].In addition, there is no trace of an IR contribution of the C(NH)3 2− unit, which supports the proposed composition.The IR spectrum of α-Eu(CN3H4)2 was also calculated by DFT from the density of phonon states and the Born effective charges as explained in references [33] and [34].All observed signals were reproduced, while the differing intensity of the simulated signals could be caused by a thermal effect (calculated 0 K vs experimental 300 K).To identify the vibrations, the IR-active phonons at the Γpoint were visualized (Table 1).IR measurements indicate a strong similarity between the αand β-phase (Figure 3), not surprisingly so because the spectrum is dominated by the CN 3 H 4 − anion vibrations [31].In addition, there is no trace of an IR contribution of the C(NH) 3 2− unit, which supports the proposed composition.The IR spectrum of α-Eu(CN 3 H 4 ) 2 was also calculated by DFT from the density of phonon states and the Born effective charges as explained in references [33] and [34].All observed signals were reproduced, while the differing intensity of the simulated signals could be caused by a thermal effect (calculated 0 K vs experimental 300 K).To identify the vibrations, the IR-active phonons at the Γ-point were visualized (Table 1).Preliminary thermogravimetric analysis (TGA) measurements of α-and β-Eu(CN3H4)2 show a two-step decay.The first step around 155 °C corresponds to the loss of two equivalents of ammonia, typical for guanidinates [30,31,35], to arrive at the hydrogen cyanamide Eu(NCNH)2.This phase has not been prepared before.The second step around 250 °C does not plateau in the measurement range up to 350 °C.This step could be the transformation of europium hydrogen cyanamide to the carbodiimide by releasing H2NCN, as observed, for example, in the transition-metal hydrogen cyanamides of Fe, Co, and Ni [36,37].EuNCN could not be prepared as a single-phase material in the reported synthesis at 1300 K [17].Thus, the guanidinates appear as interesting precursor materials for new (hydrogen) cyanamides.

Refinement and Crystal Structure of α-Eu(CN3H4)2
α-Eu(CN3H4)2 crystallizes in the acentric, monoclinic space group P21 with a = 5.8494(3) Å, b = 14.0007(8)Å, c = 8.4887(4) Å, β = 91.075(6)°,V = 695.07(6)Å 3 , and Z = 4 (Table 2).The asymmetric unit consists of two Eu atoms and four independent guanidinate units.The large number of parameters, the limited number of reflections, and the domination of the X-ray scattering by the heavy Eu atoms required a number of restraints and constraints to obtain a reasonable structure: while the Eu atoms were refined anisotropically, the C and N atoms of each guanidinate unit were constrained to a single Uiso value.Also, the C-N bond lengths, N-C-N angles, and CN3 torsion angles were restrained to sensible values (as obtained from similar guanidinates in the literature).The obtained structural model fits the PXRD measurement well (Figure 4).Different C-N bond lengths allowed for a distinction between amine and imine groups, while the assignment was confirmed by the DFT  Preliminary thermogravimetric analysis (TGA) measurements of αand β-Eu(CN 3 H 4 ) 2 show a two-step decay.The first step around 155 • C corresponds to the loss of two equivalents of ammonia, typical for guanidinates [30,31,35], to arrive at the hydrogen cyanamide Eu(NCNH) 2 .This phase has not been prepared before.The second step around 250 • C does not plateau in the measurement range up to 350 • C.This step could be the transformation of europium hydrogen cyanamide to the carbodiimide by releasing H 2 NCN, as observed, for example, in the transition-metal hydrogen cyanamides of Fe, Co, and Ni [36,37].EuNCN could not be prepared as a single-phase material in the reported synthesis at 1300 K [17].Thus, the guanidinates appear as interesting precursor materials for new (hydrogen) cyanamides.

Refinement and Crystal Structure of α-Eu(CN 3 H 4 ) 2
α-Eu(CN 3 H 4 ) 2 crystallizes in the acentric, monoclinic space group P2 1 with a = 5.8494(3) Å, b = 14.0007(8)Å, c = 8.4887(4) Å, β = 91.075(6)• , V = 695.07(6)Å 3 , and Z = 4 (Table 2).The asymmetric unit consists of two Eu atoms and four independent guanidinate units.The large number of parameters, the limited number of reflections, and the domination of the X-ray scattering by the heavy Eu atoms required a number of restraints and constraints to obtain a reasonable structure: while the Eu atoms were refined anisotropically, the C and N atoms of each guanidinate unit were constrained to a single U iso value.Also, the C-N bond lengths, N-C-N angles, and CN 3 torsion angles were restrained to sensible values (as obtained from similar guanidinates in the literature).The obtained structural model fits the PXRD measurement well (Figure 4).Different C-N bond lengths allowed for a distinction between amine and imine groups, while the assignment was confirmed by the DFT calculations.In addition, we detected a minor side phase of EuO, likely formed during handling of Eu metal under argon.Such an impurity can often be seen in the literature [38].The Rietveld refinement estimates the amount of EuO as 1.3(2) wt %.
Inorganics 2017, 5, 10 4 of 14 calculations.In addition, we detected a minor side phase of EuO, likely formed during handling of Eu metal under argon.Such an impurity can often be seen in the literature [38].The Rietveld refinement estimates the amount of EuO as 1.3(2) wt %.
Neutron diffraction experiments-to improve the structural model and to localize the hydrogen atoms-were not feasible owing to the high neutron absorption of Eu.Synchrotron PXRD measurements degraded the sample visibly and led to a color change from yellow to black.The obtained diffractograms were of poor quality, something which was also observed for long measurements of β-and γ-Eu(CN3H4)2 with our in-house X-ray diffractometer.For that reason, the crystal structures of β-and γ-Eu(CN3H4)2 could not be determined, very unfortunately.
Thus, we used our structural model of α-Eu(CN3H4)2 as the starting point for DFT calculations, the ultima ratio in this difficult case.First, we began with full structural optimizations and tests for dynamic stability to evaluate the plausibility of our structural model.To that end, the hydrogen positions were varied such as to find the energetically most stable structure for further phonon calculations.We will come back to the variation of the hydrogen positions below.The resulting Neutron diffraction experiments-to improve the structural model and to localize the hydrogen atoms-were not feasible owing to the high neutron absorption of Eu.Synchrotron PXRD measurements degraded the sample visibly and led to a color change from yellow to black.The obtained diffractograms were of poor quality, something which was also observed for long measurements of βand γ-Eu(CN 3 H 4 ) 2 with our in-house X-ray diffractometer.For that reason, the crystal structures of βand γ-Eu(CN 3 H 4 ) 2 could not be determined, very unfortunately.
Thus, we used our structural model of α-Eu(CN 3 H 4 ) 2 as the starting point for DFT calculations, the ultima ratio in this difficult case.First, we began with full structural optimizations and tests for dynamic stability to evaluate the plausibility of our structural model.To that end, the hydrogen positions were varied such as to find the energetically most stable structure for further phonon calculations.We will come back to the variation of the hydrogen positions below.The resulting structure shows only minor differences from the experimental structure, and it is also dynamically stable (i.e., its density of phonon states does not contain significant imaginary modes).This is no surprise because dispersion-corrected DFT calculations have been shown to be a powerful tool to validate experimental molecular crystal structures [39].The good agreement between experimental and fully optimized structure by DFT and the dynamic stability strongly support the plausibility of our experimental structure model.
In α-Eu(CN 3 H 4 ) 2 , the Eu atoms are coordinated 6-fold in distorted octahedra (Figure 5, Table 3).The octahedra are condensed to edge-sharing chains along the b-axis with short Eu-Eu distances of 3.66 and 3.77 Å, each bridged by two imine N atoms of different guanidinate units.Another guanidinate unit connects the corners of two octahedra via its N-C-N core, tilting the octahedra towards each other.This motif of two Eu octahedra, corresponding to the asymmetric unit, is repeated in a zigzag fashion along the chain, propagated by the 2 1 screw axis.
Inorganics 2017, 5, 10 5 of 14 structure shows only minor differences from the experimental structure, and it is also dynamically stable (i.e., its density of phonon states does not contain significant imaginary modes).This is no surprise because dispersion-corrected DFT calculations have been shown to be a powerful tool to validate experimental molecular crystal structures [39].The good agreement between experimental and fully optimized structure by DFT and the dynamic stability strongly support the plausibility of our experimental structure model.In α-Eu(CN3H4)2, the Eu atoms are coordinated 6-fold in distorted octahedra (Figure 5, Table 3).The octahedra are condensed to edge-sharing chains along the b-axis with short Eu-Eu distances of 3.66 and 3.77 Å, each bridged by two imine N atoms of different guanidinate units.Another guanidinate unit connects the corners of two octahedra via its N-C-N core, tilting the octahedra towards each other.This motif of two Eu octahedra, corresponding to the asymmetric unit, is repeated in a zigzag fashion along the chain, propagated by the 21 screw axis.The guanidinate units are connected with each other in a hydrogen-bonding network.While other functional groups have relatively short N-H … N contacts, only those from amine to imine groups should be considered to be hydrogen-bonded [40].
An unusual feature of α-Eu(CN3H4)2 is the conformation of the imine hydrogen atoms of the guanidinate units (Figure 6).In the gas phase, the most stable conformation was calculated as the syn-conformation [35], adopted in the solid state in KCN3H4, RbCN3H4, and CsCN3H4 [27,28,41].The energetically less favorable anti-conformation is adopted in LiCN3H4 and NaCN3H4, owing presumably to improved packing and hydrogen-bonding [28,29].In α-Eu(CN3H4)2, both the syn-and anti-conformation are adopted by guanidinate units.Furthermore, one unit also adopts an all-transconformation that has never been observed before for a guanidinate.While this conformation is unfavorable in the gas phase by 40 kJ•mol −1 [35], it is the predicted conformation for a hypothetical The guanidinate units are connected with each other in a hydrogen-bonding network.While other functional groups have relatively short N-H . . .N contacts, only those from amine to imine groups should be considered to be hydrogen-bonded [40].
An unusual feature of α-Eu(CN 3 H 4 ) 2 is the conformation of the imine hydrogen atoms of the guanidinate units (Figure 6).In the gas phase, the most stable conformation was calculated as the syn-conformation [35], adopted in the solid state in KCN 3 H 4 , RbCN 3 H 4 , and CsCN 3 H 4 [27,28,41].The energetically less favorable anti-conformation is adopted in LiCN 3 H 4 and NaCN 3 H 4 , owing presumably to improved packing and hydrogen-bonding [28,29].In α-Eu(CN 3 H 4 ) 2 , both the synand anti-conformation are adopted by guanidinate units.Furthermore, one unit also adopts an all-trans-conformation that has never been observed before for a guanidinate.While this conformation is unfavorable in the gas phase by 40 kJ•mol −1 [35], it is the predicted conformation for a hypothetical Li + CN 3 H 4 − ion pair in the gas phase [29].Substituted guanidinates can also adopt this conformation [42,43].In the case of α-Eu(CN 3 H 4 ) 2 , this conformation is taken by the guanidinate unit connecting the motif of the two tilted Eu octahedra, and this conformational change seemingly allows for a better coordination by the imine groups.Li + CN3H4 − ion pair in the gas phase [29].Substituted guanidinates can also adopt this conformation [42,43].In the case of α-Eu(CN3H4)2, this conformation is taken by the guanidinate unit connecting the motif of the two tilted Eu octahedra, and this conformational change seemingly allows for a better coordination by the imine groups.To further computationally test the calculated hydrogen positions and conformations, we optimized four additional cells with anti-or syn-conformation, replacing the all-trans-conformation.To further computationally test the calculated hydrogen positions and conformations, we optimized four additional cells with antior syn-conformation, replacing the all-trans-conformation.All results were energetically significantly less favorable by at least 12 kJ•mol −1 per formula unit of Eu(CN 3 H 4 ) 2 .Thus, we consider the DFT prediction to be reliable, but eagerly wait for further experimental corroboration.

Magnetism of α-Eu(CN 3 H 4 ) 2
Magnetic measurements of α-Eu(CN 3 H 4 ) 2 were conducted with different applied magnetic fields (Figure 7).The field-dependence of the effective magnetic moment µ eff at room temperature reveals a small ferromagnetic impurity: tiny fragments of the steel autoclaves contaminating the sample, described in reference [31].The maximum at ca. 70 K further reveals traces of EuO within the sample, consistent with the Rietveld analysis.Hence, the data for T > 20 K were corrected as detailed in the Methods section and reference [10].
Inorganics 2017, 5, 10 7 of 14 All results were energetically significantly less favorable by at least 12 kJ•mol −1 per formula unit of Eu(CN3H4)2.Thus, we consider the DFT prediction to be reliable, but eagerly wait for further experimental corroboration.

Magnetism of α-Eu(CN3H4)2
Magnetic measurements of α-Eu(CN3H4)2 were conducted with different applied magnetic fields (Figure 7).The field-dependence of the effective magnetic moment μeff at room temperature reveals a small ferromagnetic impurity: tiny fragments of the steel autoclaves contaminating the sample, described in reference [31].The maximum at ca. 70 K further reveals traces of EuO within the sample, consistent with the Rietveld analysis.Hence, the data for T > 20 K were corrected as detailed in the Methods section and reference [10].At room temperature, the μeff value of noninteracting Eu 2+ (4f 7 , gJ = 2, J = 7/2) ions is expected to be close to 7.91 μB, lowered by noticeable spin-orbit coupling contributions from the spin-only value of 7.94 μB [44,45].A Curie-Weiss fit of the corrected data (T > 90 K) gives an effective magnetic moment of 7.74 μB and a slightly negative Curie temperature with −8(2) K.Although both values imply antiferromagnetic exchange interactions between the Eu 2+ ions, they should be looked at as artifacts of the correction method since exchange interactions between lanthanide centers are usually very small (−2J < 2 cm −1 ).At room temperature, the µ eff value of noninteracting Eu 2+ (4f 7 , g J = 2, J = 7/2) ions is expected to be close to 7.91 µ B , lowered by noticeable spin-orbit coupling contributions from the spin-only value of 7.94 µ B [44,45].A Curie-Weiss fit of the corrected data (T > 90 K) gives an effective magnetic moment of 7.74 µ B and a slightly negative Curie temperature with −8(2) K.Although both values imply antiferromagnetic exchange interactions between the Eu 2+ ions, they should be looked at as artifacts of the correction method since exchange interactions between lanthanide centers are usually very small (−2J < 2 cm −1 ).
The field-dependent molar magnetization M m curve at 2.0 K hints at a saturation value of M m,sat ≤ 6 N A µ B , significantly lower than the expected saturation value of 7.0 N A µ B for noninteracting Eu 2+ (4f 7 ) ions.The ratio of these saturation magnetizations is different from the squared ratio of the effective moments µ eff at room temperature; that is, there is not a common factor that could scale both values to reach the expectation.Thus, we can conclude the existence of exchange interactions within the compound, while their nature is ambiguous.The field-dependent maxima in the µ eff vs T data in the temperature range 2-6 K indicate ferromagnetic exchange interactions.
It should be noted that Eu 3+ has a distinctly different magnetic behavior [10]: its magnetic susceptibility as a function of temperature is almost constant and hence does not exhibit Curie-Weiss behavior.Its high-temperature µ eff value is expected to be only 3.5 µ B , while µ eff shows a strong temperature dependence.This is all in stark contrast to the magnetic measurements of α-Eu(CN 3 H 4 ) 2 , thus conclusively showing that Eu is in the oxidation state +2.
In summary, the low-temperature data indicate ferromagnetic exchange interactions, most likely due to one-dimensional, weak interactions along the Eu 2+ chains of the crystal structure.Different, minor antiferromagnetic exchange pathways may additionally characterize the compound as indicated by the negative Curie temperature; they are, however, subject to speculation due to the uncertainties arising from the necessary correction for ferromagnetic impurities at T > 20 K.
Inorganics 2017, 5, 10 8 of 14 The field-dependent molar magnetization Mm curve at 2.0 K hints at a saturation value of Mm,sat ≤ 6 NA μB, significantly lower than the expected saturation value of 7.0 NA μB for noninteracting Eu 2+ (4f 7 ) ions.The ratio of these saturation magnetizations is different from the squared ratio of the effective moments μeff at room temperature; that is, there is not a common factor that could scale both values to reach the expectation.Thus, we can conclude the existence of exchange interactions within the compound, while their nature is ambiguous.The field-dependent maxima in the μeff vs T data in the temperature range 2-6 K indicate ferromagnetic exchange interactions.
It should be noted that Eu 3+ has a distinctly different magnetic behavior [10]: its magnetic susceptibility as a function of temperature is almost constant and hence does not exhibit Curie-Weiss behavior.Its high-temperature μeff value is expected to be only 3.5 μB, while μeff shows a strong temperature dependence.This is all in stark contrast to the magnetic measurements of α-Eu(CN3H4)2, thus conclusively showing that Eu is in the oxidation state +2.
In summary, the low-temperature data indicate ferromagnetic exchange interactions, most likely due to one-dimensional, weak interactions along the Eu 2+ chains of the crystal structure.Different, minor antiferromagnetic exchange pathways may additionally characterize the compound as indicated by the negative Curie temperature; they are, however, subject to speculation due to the uncertainties arising from the necessary correction for ferromagnetic impurities at T > 20 K.

Introduction of EuC(NH)3
Finally, we want to give a preliminary account of EuC(NH)3.This compound is isostructural to SrC(NH)3 [30] and YbC(NH)3 [31] and crystallizes in the hexagonal space group P63/m with a = 5.1634(7) Å, c = 7.1993(9) Å, V = 166.23(4)Å 3 , and Z = 2 (Figure 8; Table 2).As for YbC(NH)3, DFT calculations were used to locate the hydrogen atoms, a method validated in reference [46] (Table 4).So far, EuC(NH)3 was only obtained together with an unidentified side phase.Compared to both SrC(NH)3 [30] and YbC(NH)3 [31], EuC(NH)3 shows a shorter a-and a longer c-axis, while the volume falls in-between the two.The C-N bonds are found to be somewhat short at  Compared to both SrC(NH) 3 [30] and YbC(NH) 3 [31], EuC(NH) 3 shows a shorter aand a longer c-axis, while the volume falls in-between the two.The C-N bonds are found to be somewhat short at 1.328(1) Å, close to a double bond although the bond order should be 1 1 / 3 .As for the isostructural compounds, no hydrogen-bonding is expected.
The first magnetic measurements evidence ferromagnetic exchange interactions at low temperatures indicated by the occurrence of maxima in the µ eff vs T curve (Figure 9), but without phase-pure samples, this result is only tentative.In particular, the µ eff vs T curve for T > 25 K reveals, as for Eu(CN 3 H 4 ) 2 , ferromagnetic impurities that can be assigned to EuO and autoclave material.Furthermore, the unidentified side phase could be magnetic and contribute to the measured susceptibility.
Inorganics 2017, 5, 10 9 of 14 1.328(1) Å, close to a double bond although the bond order should be 1 1 /3.As for the isostructural compounds, no hydrogen-bonding is expected.The first magnetic measurements evidence ferromagnetic exchange interactions at low temperatures indicated by the occurrence of maxima in the μeff vs T curve (Figure 9), but without phase-pure samples, this result is only tentative.In particular, the μeff vs T curve for T > 25 K reveals, as for Eu(CN3H4)2, ferromagnetic impurities that can be assigned to EuO and autoclave material.Furthermore, the unidentified side phase could be magnetic and contribute to the measured susceptibility.

Syntheses
The highly moisture-sensitive compounds were handled in an argon-filled glove box (MBRAUN, Garching, Germany) to prevent degradation.Reactants were used as obtained from the manufacturers mentioned below.
Guanidine CN3H5 was prepared in a one-pot synthesis in liquid ammonia in steel autoclaves as described in [31].The autoclaves were constructed from stainless steel 1.4571 and a copper ring as a sealing gasket with a reaction volume of about 75 cm 3 .A detailed description of the autoclaves can be found in [47].
Eu(CN3H4)2 crystallizes in three different polymorphs, depending on the reaction temperature, as differentiated by PXRD.In all cases, stoichiometric reactants (0.2-1 mmol of Eu metal; Smart Elements, Vienna, Austria, 99.99%) were weighed in steel autoclaves and 15 cm 3 of dried, solid ammonia (Linde, Pullach, Germany, 99.999%, without further purification) were added.For the α-polymorph, compounds of highest crystallinity were obtained by heating for 5-10 days to 65 °C.β-Eu(CN3H4)2 could be prepared by shorter reaction times of 2-5 days at only 50 °C, but higher quality and crystallinity could be obtained after the γ-polymorph converted to β-Eu(CN3H4)2 (see below).γ-Eu(CN3H4)2 was synthesized by reacting for 4-8 days at room temperature, sometimes yielding almost amorphous samples.During storage under argon, γ-Eu(CN3H4)2 converted into β-Eu(CN3H4)2 over the course of weeks, as evidenced from PXRD.All products showed traces of EuO (about 1 wt % from PXRD), most likely formed during handling of Eu metal in the glove box.Yields were 70%-80%.All Eu(CN3H4)2 compounds were of a bright-yellow color.
EuC(NH)3 was obtained from stoichiometric reactants (0.3-1 mmol of Eu) in steel autoclaves with 5-20 cm 3 of dried, solid ammonia.Reaction times were 4-14 days at 50-70 °C to yield 60%-80% of an orange powder.In some cases, however, only amorphous samples were obtained or EuC(NH)3 was a product when Eu(CN3H4)2 was targeted at suboptimal reaction conditions.Unfortunately, no phase-pure products could be achieved, but only mixtures with an unidentified side-phase (volume fraction estimated from PXRD 10%-20%).

Syntheses
The highly moisture-sensitive compounds were handled in an argon-filled glove box (MBRAUN, Garching, Germany) to prevent degradation.Reactants were used as obtained from the manufacturers mentioned below.
Guanidine CN 3 H 5 was prepared in a one-pot synthesis in liquid ammonia in steel autoclaves as described in [31].The autoclaves were constructed from stainless steel 1.4571 and a copper ring as a sealing gasket with a reaction volume of about 75 cm 3 .A detailed description of the autoclaves can be found in [47].
Eu(CN 3 H 4 ) 2 crystallizes in three different polymorphs, depending on the reaction temperature, as differentiated by PXRD.In all cases, stoichiometric reactants (0.2-1 mmol of Eu metal; Smart Elements, Vienna, Austria, 99.99%) were weighed in steel autoclaves and 15 cm 3 of dried, solid ammonia (Linde, Pullach, Germany, 99.999%, without further purification) were added.For the α-polymorph, compounds of highest crystallinity were obtained by heating for 5-10 days to 65 • C. β-Eu(CN 3 H 4 ) 2 could be prepared by shorter reaction times of 2-5 days at only 50 • C, but higher quality and crystallinity could be obtained after the γ-polymorph converted to β-Eu(CN 3 H 4 ) 2 (see below).γ-Eu(CN 3 H 4 ) 2 was synthesized by reacting for 4-8 days at room temperature, sometimes yielding almost amorphous samples.During storage under argon, γ-Eu(CN 3 H 4 ) 2 converted into β-Eu(CN 3 H 4 ) 2 over the course of weeks, as evidenced from PXRD.All products showed traces of EuO (about 1 wt % from PXRD), most likely formed during handling of Eu metal in the glove box.Yields were 70%-80%.All Eu(CN 3 H 4 ) 2 compounds were of a bright-yellow color.EuC(NH) 3 was obtained from stoichiometric reactants (0.3-1 mmol of Eu) in steel autoclaves with 5-20 cm 3 of dried, solid ammonia.Reaction times were 4-14 days at 50-70 • C to yield 60%-80% of an orange powder.In some cases, however, only amorphous samples were obtained or EuC(NH) 3 was a product when Eu(CN 3 H 4 ) 2 was targeted at suboptimal reaction conditions.Unfortunately, no phase-pure products could be achieved, but only mixtures with an unidentified side-phase (volume fraction estimated from PXRD 10%-20%).

Powder X-Ray Diffraction
For PXRD, the samples were sealed in 0.3 mm glass capillaries and measured with a STADI MP diffractometer (STOE, Darmstadt, Germany) with monochromatic Mo Kα 1 radiation and a Mythen detector.The measurement ranges were 3 • -75 • in 2θ with a step size of 0.015 • for both EuC(NH) 3 and α-Eu(CN 3 H 4 ) 2 and limited scans of 3 • -21 • in 2θ for γand β-Eu(CN 3 H 4 ) 2 .High-resolution synchrotron powder-diffraction data were collected using beamline 11-BM at the Advanced Photon Source (APS), Argonne National Laboratory using an average wavelength of 0.414170 Å.During the measurement, darkening of the samples was observed and the diffractograms were different from those obtained from our in-house diffractometer.Apparently, the samples decomposed under X-ray radiation, even more so when exposed to intense synchrotron radiation.For γand β-Eu(CN 3 H 4 ) 2 , similar loss in crystallinity was observed for long measurements with our in-house diffractometer, so limited scans were used for identification purposes.
The crystal structure of α-Eu(CN 3 H 4 ) 2 was solved by charge-flipping with SUPERFLIP [48] as implemented in the Jana2006 suite [49] and further refined with the suite.To obtain a sensible structural model, the U iso of the C and N atoms of each guanidinate unit were constrained.Also, the C-N distances, N-C-N angles, and CN 3 torsion angles were restrained to established values for the guanidinate unit as obtained from neutron-diffraction measurements [31,41].Finally, a secondary phase of EuO [50] was added in the refinement, reaching a weight-percentage of 1.3(2) wt %.
The DFT-optimized structure of α-Eu(CN 3 H 4 ) 2 in the experimental lattice parameters also describes the PXRD pattern well.In this Rietveld refinement, only the profile parameters-a single U iso parameter for the Eu atoms, and another for all C and N atoms-were refined.These thermal displacement parameters were deposited with the calculated atomic positions as a CIF file.

DFT Calculations
DFT calculations were computed at the PBE+D3(BJ)/PAW level [51][52][53][54][55] as implemented in VASP [56][57][58][59].The cutoff energy for the plane-wave expansion was 500 eV; the k-meshes used for the calculations were sufficiently large.Phonon calculations did not use spin-polarization, and finite displacements of 0.01 Å were applied.The supercells for the phonon calculations of α-Eu(CN 3 H 4 ) 2 and EuC(NH) 3 were 3 × 1 × 2 and 4 × 4 × 3, respectively.To arrive at hydrogen positions fitting our EuC(NH) 3 experimental results at 300 K, we started out from the latter and placed the hydrogen atoms as observed for SrC(NH) 3 [30] and YbC(NH) 3 [31].Then, we selectively optimized the hydrogen positions, leaving lattice vectors and all other positions fixed.The resulting hydrogen positions are expected to be qualitatively comparable to those from neutron-diffraction experiments [46].
Field-dependent measurements of the molar magnetic susceptibility χ g allowed to correct for a small ferromagnetic impurity by applying the formula below for each temperature (T > 20 K) [10].
For this formula, the magnetization must be a linear function of the field.Therefore, we included data of fields of up to 0.3 T to rule out errors caused by a saturation of α-Eu(CN 3 H 4 ) 2 or EuC(NH) 3 .Extrapolations to infinitely high fields yield the corrected values χ m (∞) through multiplying χ g (∞) by the molar mass of α-Eu(CN 3 H 4 ) 2 or EuC(NH) 3 .

IR Spectroscopy
An ALPHA FT-IR-spectrometer (Bruker, Billerica, MA, USA) placed in an argon-filled glove box and equipped with an ATR Platinum Diamond sample holder with a measurement range of 4000−400 cm −1 was employed to measure the IR spectrum of Eu(CN 3 H 4 ) 2 .The results were compared to a DFT-based calculation of an IR spectrum of α-Eu(CN 3 H 4 ) 2 .The frequencies and eigenvectors at the Γ-point were derived by a finite displacement approach as implemented PHONOPY [60], and the Born effective charge tensor was calculated by density-functional perturbation theory as implemented in VASP ("LEPSILON=.TRUE.").The IR intensities were derived from these values as described in references [33] and [34].Also, a Gaussian broadening was applied to the spectrum.

Conclusions
In summary, we synthesized the α-, βand γ-polymorphs of Eu(CN 3 H 4 ) 2 and identified them by PXRD.The γ-phase transforms into the β-form over time.The IR spectra are dominated by the anion and are interpretable with the help of DFT calculations.Preliminary TGA measurements show that the guanidinates could be precursors for the preparation of (hydrogen) cyanamides.The crystal structure of α-Eu(CN 3 H 4 ) 2 was solved by PXRD, and DFT was used to optimize the structural model.In α-Eu(CN 3 H 4 ) 2 , Eu is coordinated in double zigzag chains that are connected by the hydrogen-bonded guanidinate anions.The CN 3 H 4 − anions are predicted to adopt the syn-, anti-, and all-trans-conformations.The all-trans-conformation is found for the first time in a guanidinate.Magnetic measurements show paramagnetism at high temperatures and ferromagnetic exchange interactions below 6 K, presumably in one dimension along the Eu chains.Finally, EuC(NH) 3 , isostructural to SrC(NH) 3 [30] and YbC(NH) 3 [31], is introduced as a possible low-temperature ferromagnet.

Figure 1 .
Figure 1.Protonation and deprotonation of guanidine from guanidinium on the left to the doubly deprotonated guanidinate on the right.

Figure 2 .
Figure 2. Diffraction patterns of the different polymorphs of Eu(CN3H4)2 (a) and the transformation from the γ-phase to the β-phase over time (b).

Figure 1 .
Figure 1.Protonation and deprotonation of guanidine from guanidinium on the left to the doubly deprotonated guanidinate on the right.

Figure 1 .
Figure 1.Protonation and deprotonation of guanidine from guanidinium on the left to the doubly deprotonated guanidinate on the right.

Figure 2 .
Figure 2. Diffraction patterns of the different polymorphs of Eu(CN3H4)2 (a) and the transformation from the γ-phase to the β-phase over time (b).

Figure 2 .
Figure 2. Diffraction patterns of the different polymorphs of Eu(CN 3 H 4 ) 2 (a) and the transformation from the γ-phase to the β-phase over time (b).

Figure 3 .
Figure 3. IR measurement of the β-(with an offset) and α-polymorphs of Eu(CN3H4)2 and simulation of the latter.

Figure 3 .
Figure 3. IR measurement of the β-(with an offset) and α-polymorphs of Eu(CN 3 H 4 ) 2 and simulation of the latter.

Figure 5 .
Figure 5. Crystal structure of α-Eu(CN3H4)2 shown along the a-axis as optimized by DFT.

Figure 5 .
Figure 5. Crystal structure of α-Eu(CN 3 H 4 ) 2 shown along the a-axis as optimized by DFT.

Figure 7 .
Figure 7. Effective magnetic moment (a) vs temperature for α-Eu(CN3H4)2 at different applied fields; effective magnetic moment in more detail (b) revealing ferromagnetic exchange interactions (lines in (b) and (c) as guide to the eye only).Molar, corrected magnetic susceptibility (c) vs temperature for different applied fields; inset: magnetization versus applied field at 2.0 K. Corrected inverse molar susceptibility (d) vs temperature, and fit to the Curie-Weiss law.

Figure 7 .
Figure 7. Effective magnetic moment (a) vs temperature for α-Eu(CN 3 H 4 ) 2 at different applied fields; effective magnetic moment in more detail (b) revealing ferromagnetic exchange interactions (lines in (b) and (c) as guide to the eye only).Molar, corrected magnetic susceptibility (c) vs temperature for different applied fields; inset: magnetization versus applied field at 2.0 K. Corrected inverse molar (d) vs temperature, and fit to the Curie-Weiss law.

Table 4 .
Atomic positions and displacement parameters of EuC(NH)3 in space group P63/m.Hydrogen position determined from DFT.

Figure 9 .
Figure 9. Magnetic susceptibility of EuC(NH)3 vs temperature (a) and effective magnetic moment vs temperature (b), assuming molar weight of a pure EuC(NH)3 sample (lines to guide the eye only).

Figure 9 .
Figure 9. Magnetic susceptibility of EuC(NH) 3 vs temperature (a) and effective magnetic moment vs temperature (b), assuming molar weight of a pure EuC(NH) 3 sample (lines to guide the eye only).

Table 3 .
Atomic positions of α-Eu(CN3H4)2 in space group P21 (all on Wyckoff position 2a) as determined from DFT.

Table 4 .
Atomic positions and displacement parameters of EuC(NH) 3 in space group P6 3 /m.Hydrogen position determined from DFT.