SHG-active Ln^III–[Mo^I(CN)₅(NO)]³⁻ (Ln = Gd, Eu) magnetic coordination chains: a new route towards non-centrosymmetric molecule-based magnets

Abstract

The rare heteroligand pentacyanidonitrosylmolybdate(I) anion was employed in the construction of d–f bimetallic cyanido-bridged {[Ln^III(dmf)₆][Mo^I(CN)₅(NO)]} (Ln = Gd, 1; Eu, 2) chains crystallizing in the non-centrosymmetric Pna2₁ space group. 1 exhibits Gd–Mo antiferromagnetic coupling giving rise to the ferrimagnetic spin chain behaviour with the onset of magnetic ordering below 2 K, and shows high second harmonic generation (SHG) activity with SH susceptibility of 1.2 × 10⁻¹⁰ esu, opening a novel family of non-centrosymmetric molecule-based magnets.

---

Molecule-based magnets which are based on paramagnetic metal complexes and/or organic radicals arouse considerable interest in chemistry, physics and materials science.¹ They are especially attractive as efficient molecular platforms for multiple functionalities, that is the combination of intrinsic magnetic phenomena (magnetic coupling, magnetic ordering, slow magnetic relaxation or spin bistability) with additional physical properties such as luminescence,² microporosity,³ and photo-induced⁴ or dehydration-driven phase transitions.⁵

The implementation of chirality or non-centrosymmetricity into molecule-based magnets is particularly fruitful due to the numerous physical cross-effects appearing as the product of coupling between the magnetic and optical properties of the material.^6–11 For instance, the non-linear optical (NLO) effect of second harmonic generation (SHG) that occurs in the solid crystallizing in a non-centrosymmetric space group⁶ can be significantly enhanced in the spin ordered state due to the magnetic contribution to the intensity of SH light. The resulting cross-effect called magnetization induced second harmonic generation (MSHG) was presented in very special examples of molecule-based magnets based on cyanido- or oxalato-bridged networks.⁷ Lately, our group has showed an even more attractive effect, a 90 degree photoswitching of the SHG polarization plane in the Fe^II–[Nb^IV(CN)₈]⁴⁻ chiral photomagnet.⁸ The chirality introduced to the spin ordered systems also enabled the observation of magnetization enhanced magnetic circular dichroism (MCD)⁹ and magneto-chiral dichroism (MChD),¹⁰ or multiferroicity.¹¹

In this context, there is great interest in searching for efficient synthetic strategies towards non-centrosymmetric molecule-based magnets.¹² They are usually prepared by a straightforward method using chiral ligands or counterions to induce non-centrosymmetricity by coordination to metal ions,^9,13 or through non-covalent interactions.^10,11,14 Using this route, a number of non-centrosymmetric magnets, usually built of cyanido- or oxalato-bridged networks, were achieved.^10,11,13,14 The alternative approach is based on the spontaneous resolution during the crystallization process.^7,15 Such a synthetic route is difficult to control due to its usually unexplained mechanism. However, during our studies on chiral [M(CN)₈]-based networks, we found that the spontaneous resolution can be induced by the selected non-chiral but sterically expanded ligands, such as 4-bromopyridine, inducing chirality in Mn^II–[Nb^IV(CN)₈]⁴⁻ ferrimagnets.^15d Following these findings, we considered the application of metal complexes that can be good precursors for the non-centrosymmetric magnets. While broadly studied [M(CN)₆]ⁿ⁻ and [M(CN)₈]ⁿ⁻ ions are isotropic and cannot be used for symmetry breaking, we found that heteroligand pentacyanidonitrosyl metallates, [M(CN)₅(NO)]ⁿ⁻, due to their decreased symmetry and structural anisotropy, can be of potential interest. We selected the rare [Mo^I(CN)₅(NO)]³⁻ anion which is paramagnetic,¹⁶ in contrast to diamagnetic nitroprusside [Fe/Ru^II(CN)₅(NO)]²⁻ ions, and its photochromic and photoswitchable properties were explored.¹⁷ [Mo^I(CN)₅(NO)]³⁻ was used only once in the synthesis of bimetallic coordination polymers, which is a K_0.9Mn_1.05[Mo(CN)₅(NO)]·n(solv) Prussian blue analogue with ferrimagnetic ordering below 39 K but only prepared in the powder form which hampers the discussion on its structure.¹⁸ Here, we report the structure and properties of bimetallic CN-bridged {[Ln^III(dmf)₆][Mo^I(CN)₅(NO)]} (Ln = Gd, 1; Eu, 2) chains crystallizing in a non-centrosymmetric space group, which can combine ferrimagnetic spin chain behaviour with SHG activity.

The yellow platelet single crystals of 1 and 2 were prepared under an argon atmosphere by crystallization from mixed solution of [Mo^I(CN)₅(NO)]³⁻ and the respective lanthanide(3+) ions in N,N′-dimethylformamide (dmf) solvent (Experimental section and Fig. S1, ESI†). Single crystal X-ray diffraction studies revealed that 1 and 2 are isostructural and composed of cyanido-bridged chains crystallizing in the non-centrosymmetric Pna2₁ space group (Fig. 1, S2 and S3, Tables 1, S1 and S2, ESI†). The coordination polymers of 1 and 2 are built of eight-coordinated [Ln^III(μ-NC)₂(dmf)₆]⁺ (Ln = Gd, 1; Eu, 2) units of a nearly ideal square antiprism geometry, which are alternately arranged with octahedral [Mo^I(CN)₅(NO)]³⁻ moieties (Fig. 1, Table S2†). Two cyanides of Mo1 complexes are bridging the neighbouring Ln1 centers while three others are terminal. The NO ligands occupying the sixth position of the Mo1 complex are distinguishable by revealing Mo1–N bond lengths of ca. 1.93 Å, which is much shorter than the 2.13–2.18 Å Mo1–C bond of Mo1–CN linkages. Thus, the structural analysis revealed that the NO ligands of [Mo^I(CN)₅(NO)]³⁻ are also terminal. Each Ln center coordinates two bridging cyanides with a N–Ln1–N angle of ca. 138° resulting in the zig-zag topology of the chains.

Fig. 1 Crystal structure of 1: (a) the representative fragment of the coordination polymer, (b) the asymmetric unit with the atoms shown as thermal ellipsoids at the 70% probability level with the atom labelling scheme, and (c) the views of supramolecular arrangement of chains in the ac and bc planes.

Table 1 Crystal data and structure refinement for 1 and 2

Compound		1	2
Formula		Gd1Mo1C23H42N12O7	Eu1Mo1C23H42N12O7
Formula weight [g mol^−1]		851.87	846.58
T [K]		90(2)
λ [Å]		0.71075 (Mo Kα)
Crystal system		Orthorhombic
Space group		Pna21
Unit cell	a [Å]	21.4968(12)	21.5317(7)
	b [Å]	9.8959(5)	9.9017(3)
	c [Å]	16.3664(8)	16.3749(5)
V [Å^3]		3481.6(3)	3491.14(19)
Z		4	4
Calculated density [g cm^−3]		1.625	1.611
Absorption coefficient [cm^−1]		2.304	2.195
F(000)		1704	1700
Crystal size [mm × mm × mm]		0.40 × 0.30 × 0.20	0.60 × 0.50 × 0.10
Crystal type		Yellow plate	Yellow plate
Θ range [deg]		3.063–27.464	3.060–27.442
Limiting indices		−27 < h < 27	−27 < h < 27
		−12 < k < 12	−12 < k < 12
		−21 < l < 17	−18 < l < 21
Collected reflections		31 381	32 294
Unique reflections		7392	7521
R int		0.0313	0.0313
Completeness [%]		99.8	99.8
Max. and min. transmission		0.630 and 0.418	0.803 and 0.485
Data/restraint/parameters		7392/1/397	7521/2/397
GOF on F^2		1.091	1.123
Flack x-parameter		0.360(3)	0.330(3)
Final R indices		R 1 = 0.0405 [I > 2σ(I)]	R 1 = 0.0451 [I > 2σ(I)]
		wR2 = 0.0886 (all)	wR2 = 0.0932 (all)
Largest diff peak/hole		3.646/−2.784 e A^−3	3.689/−2.814 e A^−3

---

The cyanido-bridged chains of 1 and 2 are arranged in parallel within the supramolecular network which is stabilized only by the steric effects on the methyl groups of dmf ligands and the weak hydrogen bonds between the terminal cyanides and C–H groups of dmf (Fig. 1c and S2c, ESI†). The crystallization solvent molecules were not detected. Thus, the inter-chain space is controlled by the alignment of dmf ligands which results in the closest inter-metallic distances between neighbouring chains of ca. 8.28 Å between Mo1 and Ln1 along the c axis. Within the supramolecular networks of 1, the nitrosyl ligands of [Mo^I(CN)₅(NO)]³⁻ ions are aligned in the directions close to the c axis, following the 2₁ screw axis, while the coordination chains are aligned along the b axis. Moreover, the Mo1–N–O fragments of 1 are oriented towards the negative coordinates of the c direction which causes the overall lack of a centre of symmetry (Fig. 1c). In such a case, two enantiomorphic forms are expected but the crystals of 1 grow as the inversion twins with a Flack parameter of 0.360(3) indicating that the twin ratio is 0.360(3) : 0.640(3) (Table 1).¹⁹ An identical behavior was found for 2 with a similar twin ratio of 0.330(3) : 0.670(3). The structural model found from the single-crystal XRD experiment is valid for the bulk sample used in other physical measurements which was proved by the powder X-ray diffraction analysis (Fig. S3, ESI†).

Due to their non-centrosymmetric Pna2₁ space group, 1 and 2 can be considered as promising functional materials for the occurrence of non-linear optical (NLO) properties. We checked for this attractive possibility by using 1 and executing the experiment of second harmonic generation (SHG), which is a well-known NLO effect of generating double frequency photons, 2ω, after interacting the photons of frequency ω with a solid medium showing at least a local lack of a centre of symmetry. After irradiating the polycrystalline sample of 1 with the fundamental light wavelength of 1064 nm, we repeatedly detected the second harmonic green light of 532 nm (see ESI,† Fig. S1 for the experimental details). The magnitude of SH susceptibility was determined to be 1.2 × 10⁻¹⁰ esu which is ca. 0.1 times the value of potassium dihydrogen phosphate (KDP), one of the most common NLO crystals. This value is, however, impressive among coordination polymers, especially those based on SHG-active cyanido-bridged materials (Table 2).¹⁵

Table 2 Second harmonic (SH) susceptibility of 1 compared with the other reported SHG-active cyanido-bridged coordination systems

Compound	SH susceptibility	Ref.
1	1.2 × 10^−10 esu	This work
[Cu(NH3)2]2[Mo(CN)8]	1.8 × 10^−10 esu	15a
[Mn(H2O)2][Mn(pyz)(H2O)2][Mo(CN)8]·4H2O (pyz = pyrazine)	6 × 10^−11 esu	15b
Rb0.94Mn[Fe(CN)6]0.98·0.2H2O	8 × 10^−11 esu	15c

---

The direct-current (dc) magnetic properties of 1 and 2 are presented in Fig. 2. The room temperature χ_MT value for the {Gd^IIIMo^I} unit of 1 is 8.1 cm³ mol⁻¹ K which is close to the value of 8.2 cm³ mol⁻¹ K expected for the free-ion contributions from Gd^III (S_Gd = 7/2, g_Gd = 2.0) and Mo^I (S_Mo = 1/2, g_Mo = 2.0). Upon cooling, the χ_MT product decreases slowly down to 50 K, and then faster below this point. A minimum of 7.5 cm³ mol⁻¹ K is observed at 10.8 K, and after that, a fast increase of χ_MT up to 10.0 cm³ mol⁻¹ K at T = 2 K is detected (Fig. 2a). The minimum in the χ_MT vs. T plot suggests the presence of antiferromagnetic coupling between Gd^III and Mo^I spin centers. This is also supported by the field dependence of magnetization at T = 2 K (Fig. 2a, the inset). With increasing field, M increases fast up to 4.9 μ_B at 10 kOe, and later much slower reaching 5.8 μ_B at H = 50 kOe. This value is close to the 6 μ_B expected for an antiparallel arrangement of Gd^III and Mo^I spins and much lower than the 8 μ_B corresponding to ferromagnetic coupling. In addition, the experimental M–H curve deviates strongly from the sum of the Brillouin functions for Gd^III and Mo^I (Fig. 2a, the inset). The rapid increase of magnetization at low magnetic field is due to magnetic ordering. The deviation at high magnetic field is explained by the saturation of magnetization in the ferrimagnetically ordered state. These all together indicate an antiferromagnetic nature of the Gd–Mo interaction leading to the ferrimagnetic spin chains of 1. Such magnetic interactions operate within the chains. In addition, at the lowest temperatures, the onset of long-range ferrimagnetic ordering is observed, as the frequency-independent peak of the out-of-phase magnetic susceptibility starts to appear. This, however, happens at the lowest available T, around 2 K, suggesting that the related critical temperature is below 2 K.

Fig. 2 Magnetic properties of 1 and 2: (a) T dependence of χ_MT of 1 at H_dc = 2 kOe (blue circles) with the best-fit curve using Seiden's model (red line), and H dependence of magnetization, M, of 1 at T = 2 K in the inset (blue squares) with the sum of the Brillouin functions of S = 7/2 and S = 1/2 (black line), (b) T dependences of the χ_M′′ and χ_M′ (the inset) components of ac susceptibilities of 1 at the indicated frequencies (H_dc = 0 Oe, H_ac = 3 Oe), and (c) χ_MT–T plot of 2 at H_dc = 5 kOe (green circles) with the best-fit curve (red line), and M versus H plot of 2 at T = 2 K in the inset (green squares) with the Brillouin function of S = 1/2 (black line).

To confirm the type of Gd^III–Mo^I magnetic coupling, the χ_MT–T plot was analysed based on Seiden's model assuming an alternate arrangement of classical spins of Gd (S_Gd = 7/2) and quantum spins of Mo^I (S_Mo = 1/2).²⁰ In this model, the spin Hamiltonian with an exchange interaction (J_exc) is defined as . Assuming g_Mo = 2.00, the values of J and g_Gd were fitted to the experimental data in the T range of 2–300 K giving J_exc = −1.02(5) cm⁻¹ and g_Gd = 2.01(1) (Fig. 2a, see the ESI†). The negative J indicates an antiferromagnetic Gd–Mo interaction leading to the ferrimagnetic spin chain of 1.

Fig. 2c shows the χ_MT vs. T plot for 2. The χ_MT at 300 K is 1.7 cm³ mol⁻¹ K, which is slightly smaller than 1.875 cm³ mol⁻¹ K calculated from isolated Eu^III ion (1.5 cm³ mol⁻¹ K) and Mo^I ion (0.375 cm³ mol⁻¹ K).²¹ On cooling, the χ_MT decreases to reach 0.35 cm³ mol⁻¹ K, which is close to the expected value for free-ion Mo^I. The magnetization curve in H = 50 kOe at T = 2 K reaches 0.93 μ_B, which is close to the 1 μ_B corresponding to the Mo^I spins. In fact, the whole M–H curve at 2 K follows the Brillouin function for S = 1/2 with g = 2.0. Such a magnetic behaviour of 2 is explained by the spin–orbit coupling effect of Eu^III. This ion reveals the diamagnetic ground state of ⁷F₀ but excited states are closely located, and can be thermally populated giving a non-zero contribution to the χ_MT. However, they are gradually depopulated on cooling, thus at 2 K only the ⁷F₀ state is occupied, and Mo^I gives the sole contribution to magnetism of 2. Assuming these, we simulated the χ_MT–T plot with the spin–orbit coupling constant (λ) as a fitting parameter (see the ESI†). Good agreement with the experiment was obtained for λ = 365(3) cm⁻¹. This indicates the negligible magnetic coupling between Mo^I and diamagnetic Eu^III.

We report two bimetallic cyanido-bridged Ln^III(dmf)–Mo^I (Ln = Gd, 1; Eu, 2) chains based on the rare [Mo^I(CN)₅(NO)]³⁻ anion. They are the second examples of coordination systems based on this exotic metalloligand, and the first obtained as single crystals. Due to the decreased symmetry and axial anisotropy of [Mo^I(CN)₅(NO)]³⁻ combined with the steric effects on the dmf ligands, 1 and 2 crystallize in the non-centrosymmetric Pna2₁ space group. This makes them attractive objects for the non-linear optical effects as shown by the SHG activity detected in 1. While 2 is a paramagnet without magnetic coupling, 1 reveals a ferrimagnetic chain behaviour due to the antiferromagnetic Gd–Mo interaction, and the onset of ferrimagnetic ordering below 2 K. Thus, 1 combines non-centrosymmetricity with the ferrimagnetic property showing that the insertion of heteroligand [Mo^I(CN)₅(NO)]³⁻ into coordination polymers is an efficient route towards non-centrosymmetric magnets with non-linear optical and magneto-optical properties.⁷ To observe such effects, a higher T_C of magnetic ordering should be induced, and it can be realized by the increase of cyanide connectivity or replacement of 4f- with d-metal ions with strong exchange coupling. Moreover, [Mo^I(CN)₅(NO)]³⁻ like other nitroprusside ions can reveal photoinduced phase transitions due to the photosensitive NO ligand which together with non-centrosymmetricity and magnetism can lead to extra cross-effects.⁸ We are currently checking these exciting possibilities.

This work was financed by the Japan Society for the Promotion of Science within the Grant-in-Aid for Specially Promoted Research, 15H05697. We thank Cryogenic Research Center and Center for Nanolithography & Analysis, The Univ. of Tokyo, supported by MEXT. M. K. acknowledges the support of the Advanced Leading Graduate Course for Photon Science (ALPS).

Notes and references

1. (a) J. A. Real, A. B. Gaspar and M. Carmen Munoz, Dalton Trans., 2005, 2062; (b) J. Miller, Chem. Soc. Rev., 2011, 40, 3266; (c) D. N. Woodruff, R. E. P. Winpenny and R. A. Layfield, Chem. Rev., 2013, 113, 5110; (d) T. Grancha, J. Ferrando-Soria, M. Castellano, M. Julve, J. Pasan, D. Armentano and E. Pardo, Chem. Commun., 2014, 50, 7569.
2. (a) J. Ferrando-Soria, H. Khajavi, P. Serra-Crespo, J. Gascon, F. Kapteijn, M. Julve, F. Lloret, J. Pasan, C. Ruiz-Perez, Y. Journaux and E. Pardo, Adv. Mater., 2012, 24, 5625; (b) S. Chorazy, J. Wang and S. Ohkoshi, Chem. Commun., 2016, 52, 10795.
3. P. Dechambenoit and J. R. Long, Chem. Soc. Rev., 2011, 40, 3249.
4. (a) S. Ohkoshi, K. Imoto, Y. Tsunobuchi, S. Takano and H. Tokoro, Nat. Chem., 2011, 3, 564; (b) C. Mathoniere, H.-J. Lin, D. Siretanu, R. Clerac and J. M. Smith, J. Am. Chem. Soc., 2013, 135, 19083.
5. (a) D. Maspoch, D. Ruiz-Molina, K. Wurst, N. Domingo, M. Cavallini, F. Biscarini, J. Tejada, C. Rovira and J. Veciana, Nat. Mater., 2003, 2, 190; (b) B. Nowicka, M. Reczyński, M. Rams, W. Nitek, J. Żukrowski, C. Kapusta and B. Sieklucka, Chem. Commun., 2015, 51, 11485.
6. L. R. Mingabudinova, V. V. Vinogradov, V. A. Milichko, E. Hey-Hawkins and A. Vinogradov, Chem. Soc. Rev., 2016, 45, 5408.
7. (a) T. Nuida, T. Matsuda, H. Tokoro, S. Sakurai, K. Hashimoto and S. Ohkoshi, J. Am. Chem. Soc., 2005, 127, 11604; (b) C. Train, T. Nuida, R. Gheorghe, M. Gruselle and S. Ohkoshi, J. Am. Chem. Soc., 2009, 131, 16838; (c) Y. Tsunobuchi, W. Kosaka, T. Nuida and S. Ohkoshi, CrystEngComm, 2009, 11, 2051; (d) D. Pinkowicz, R. Podgajny, W. Nitek, M. Rams, A. M. Majcher, T. Nuida, S. Ohkoshi and B. Sieklucka, Chem. Mater., 2011, 23, 21.
8. S. Ohkoshi, S. Takano, K. Imoto, M. Yoshikiyo, A. Namai and H. Tokoro, Nat. Photonics, 2014, 8, 165.
9. (a) K. Inoue, K. Kikuchi, M. Ohba and H. Okawa, Angew. Chem., Int. Ed., 2003, 42, 4810; (b) S. Chorazy, R. Podgajny, W. Nitek, T. Fic, E. Gőrlich, M. Rams and B. Sieklucka, Chem. Commun., 2013, 49, 6731.
10. C. Train, R. Gheorghe, V. Krstic, L. M. Chamoreau, N. S. Ovanesyan, G. Rikken, M. Gruselle and M. Verdaguer, Nat. Mater., 2008, 7, 729.
11. E. Pardo, C. Train, H. Liu, L. M. Chamoreau, B. Dkhil, K. Boubekeur, F. Lloret, K. Nakatani, H. Tokoro, S. Ohkoshi and M. Verdaguer, Angew. Chem., Int. Ed., 2012, 51, 8356.
12. C. Train, M. Gruselle and M. Verdaguer, Chem. Soc. Rev., 2011, 40, 3297.
13. (a) H. Kumagai and K. Inoue, Angew. Chem., Int. Ed., 1999, 38, 1601; (b) E. Coronado, J. Galan-Mascaros, C. J. Gomez-Garcia and A. Murcia-Martinez, Chem. – Eur. J., 2006, 12, 3484; (c) A. Beghidja, P. Rabu, G. Rogez and R. Welter, Chem. – Eur. J., 2006, 12, 7627; (d) K. Komori-Orisaku, K. Imoto, Y. Koide and S. Ohkoshi, Cryst. Growth Des., 2013, 13, 5267.
14. (a) N. S. Ovanesyan, V. Makhaev, S. M. Aldoshin, P. Gredin, K. Boubekeur, C. Train and M. Gruselle, Dalton Trans., 2005, 3101; (b) M. Clemente-Leon, E. Coronado, J. C. Dias, A. Soriano-Portillo and R. Willett, Inorg. Chem., 2008, 47, 6458.
15. (a) T. Hozumi, T. Nuida, K. Hashimoto and S. Ohkoshi, Cryst. Growth Des., 2006, 6, 1736; (b) W. Kosaka, T. Nuida, K. Hashimoto and S. Ohkoshi, Bull. Chem. Soc. Jpn., 2007, 80, 960; (c) S. Ohkoshi, S. Saito, T. Matsuda, T. Nuida and H. Tokoro, J. Phys. Chem. C, 2008, 112, 13095; (d) T. Ohno, S. Chorazy, K. Imoto and S. Ohkoshi, Cryst. Growth Des., 2016, 16, 4119.
16. A. Müller, W. Eltzner, S. Sarkar, H. Bőgge, P. J. Aymonino, N. Mohan, U. Seyer and P. Subramanian, Z. Anorg. Allg. Chem., 1983, 503, 22.
17. (a) M. D. Carducci, M. R. Pressprich and P. Coppens, J. Am. Chem. Soc., 1997, 119, 2669; (b) Z.-Z. Gu, O. Sato, T. Iyoda, K. Hashimoto and A. Fujishima, Chem. Mater., 1997, 9, 1092.
18. N. Machida, S. Ohkoshi, Z. J. Zhong and K. Hashimoto, Chem. Lett., 1999, 907.
19. T. Weber, J. Dshemuchadse, M. Kobas, M. Conrad, B. Harbrecht and W. Steurer, Acta Crystallogr., Sect. B: Struct. Sci., 2009, 65, 308.
20. J. Seiden, J. Phys., Lett., 1983, 44, 947.
21. Y. Wan, L. Zhang, L. Jin, S. Gao and S. Lu, Inorg. Chem., 2003, 42, 4985.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, setup for SHG measurement, detailed structure parameters, results of continuous shape measure analysis, additional structural views, powder X-ray diffraction patterns, and details of magnetic calculations. CCDC 1508558 and 1508559. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6ce02214f

---