English

• Metal-ligand coordination is a well-established but still intriguing research area and has attracted much attention in the past decade, especially with the flourishing crystalline coordination polymers such as metal-organic frameworks. During the construction of coordination structures, it is commonly seen that small particles would grow bigger with time, for which Ostwald ripening is responsible [1]. Ostwald ripening is an equilibrium establishing process from a meta-stable phase to its thermodynamically stable phases, and is generally used as a post-synthesis process to modulate the morphologies of particles/crystals without changing the composition [2–4]. Thus, it should be intriguing to direct the equilibrium establishing process to develop new coordination structures.

  Cyanuric acid (CA) and its ionic derivatives frequently serve as a linker in coordination chemistry due to its N- and O-sites for coordinate bonding and/or H-bonding. Many coordination complexes/polymers of CA to metal ions like Ca(Ⅱ), Sr(Ⅱ), Cu(Ⅱ), Eu(Ⅱ), Ba(Ⅱ), and Ag(Ⅰ) have been reported [5–14]. For Cu(Ⅱ), the first reported coordination crystals with CA (1, structural formula [Cu₁(H₂CA-κN,N)₂(NH₃)₂] (-κN,N indicates that the Cu atom coordinates two CA by the N-sites), synthesized by mixing cupric chloride and cyanuric acid in aqueous ammonia, is H-bonded coordination assembly formed from discrete coordination complexes (Fig. S1 in Supporting information) [13,14]. Nevertheless, the coordination polymeric structures of Cu(Ⅱ)-CA have not been discovered yet. Our group have been interested in studying the competition of H-bonding and coordinate bonding and constructed a pair of interchangeable Ag(Ⅰ)-melamine-CA coordination complex and coordination polymer [15–17]. Recently, we found that the growth of Cu(Ⅱ)-CA coordination complexes (1) underwent a classic Ostwald ripening: small crystallites gradually formed large single crystals (Figs. 1A and B) within 3–5 days. The mass transfer from crystallites dissolving to big crystals growing inspired us to explore the possibility of controlling the mass transfer direction, i.e., using the crystallites as resources to grow other coordination structures.

  Figure 1

  

  Figure 1.  (A) Photos illustrating Ostwald ripening of 1 from crystallites to large single crystals and (B) the corresponding powder X-ray diffraction (PXRD) patterns. (C) Photos of 1 transforming to 2 and 3. PXRD patterns (D) from 1 to 2 and (E) from 1 to 3.

  DownLoad: Full-Size Img PowerPoint

  In this work, we reported on how to divert the equilibrium establishing process of meta-stable crystallites of 1 to cultivate Cu(Ⅱ)-CA coordination polymers by adjusting the molar ratio of Cu(Ⅱ) to CA (hereafter denoted as M_(Cu-CA)). We obtained two novel coordination polymers with different coordination sites of CA to Cu(Ⅱ): from κN–Cu–κN to κN–Cu–κO, which displays disparate magnetic interaction in magnitude.

  The seeds, Cu(Ⅱ)-CA complex 1, were synthesized by mixing Cu(NO₃)₂ and CA at a molar ratio M_(Cu-CA) = 0.5 in glass vials with aqueous ammonia (2 wt%). After shaking, the normal Ostwald ripening would defer and the flakes of 1 were able to preserve for about three days. Then, different proportions of Cu(Ⅱ) were added into the vials with the seed crystallites to adjust the total M_(Cu-CA). After 2–3 weeks, the seeds gradually transformed into coordination polymers of Cu(Ⅱ)-CA (Figs. 1C–E). Two coordination polymers were obtained, namely 2 and 3, by adjusting the total M_(Cu-CA) to 1.0 and 2.0, respectively. We screened more intervals of M_(Cu-CA), but only found 2 and 3. Moreover, higher ammonia concentration favored the generation of another copper cyanurate complex (1*, Structural formula [Cu₁(H₂CA-κN,N)₂(NH₃)₂]·[Cu₁(H₂CA-κO,O)₂(NH₃)₄] (-κO,O indicates that the Cu atom links two CA by the O-sites), whose PXRD and crystal structure were identical as previously reported (Fig. S2 in Supporting information) [14]. Nevertheless, the crystal transformation from coordination complexes to coordination polymers was still happenable under higher ammonia concentration (Table S1), although longer reaction time and more Cu(Ⅱ) were needed. It is also found that adjusting the M_(Cu-CA) at the commencement of the reaction did not change the crystal transformation. It should be noted that the big crystals of 1 after normal Ostwald ripening (Fig. 1A) upon the addition of extra Cu(Ⅱ) were almost impossible to transform into 2 and 3 (no new crystals formed in months). Thus, the crystal transformation process from crystallites of 1 to 2 and 3 can be comprehended as a directed Oswald ripening: The addition of extra Cu(Ⅱ) favors the nucleation of Cu(Ⅱ)-CA coordination chains in solution (which is more thermodynamically stable than crystallites of 1) and the stable nuclei feeds on the discrete coordination complex to finally give rise to the coordination polymers.

  The structure of 2 was refined in the monoclinic space group C2/c with the structural formula of {[Cu₂(HCA-κN,N)₂(NH₃)₄]·H₂O}_n (Table S2 in Supporting information and Fig. 2A). The asymmetric unit comprises one Cu atom, one di-deprotonated CA (µ₂-CA), and two NH₃ (Fig. 2B). The four Cu–N dative bonds around the central Cu atom deform out of plane, constituting a flattened tetrahedral geometry. The structure shows 2 is a one-dimensional coordination polymer consisting of alternating Cu(NH₃)₂ and µ₂-CA linked via Cu–κN coordination bonds, forming a zig-zag chain structure of [Cu–(µ₂-CA)–Cu–(µ₂-CA)] (NH₃ are omitted for clarity). Neighboring chains are H-bonded via N–H···O between CA molecules, forming oval channels along c axis (Fig. 2C) which are filled by uncoordinated water molecules. The uncoordinated water molecules could be released upon heating as indicated by the gradual weight loss in thermogravimetric results (Fig. S3 in Supporting information). The CA molecules along one chain are non-coplanar, resulting in a rippled layer structure (Fig. 2D). The interlayer interactions are composed of π-π interaction (3.895 Å) and H-bonds between NH₃ and CA.

  Figure 2

  

  Figure 2.  Crystalline structure of 2: (A) Calculated PXRD pattern; (B) Basic structural unit; (C) Extended view along [021]; (D) Extended view along [101]. Color codes: Cu, Magenta; O, red; N, blue; C, gray; H, white.

  DownLoad: Full-Size Img PowerPoint

  Coordination polymer 3 crystalized in P2₁/m space group, with an empirical formula of {[Cu(HCA-κN,N)₁(NH₃)₂]Cu(HCA-κN,O)(NH₃)₂(H₂O)]·H₂O}_n, where -κN,O indicates that the Cu atom links two CA by the N-site of one CA and the O-site of the other (Table S3 in Supporting information and Fig. 3A). The uncoordinated water molecules and coordinated water molecules in 3 could be further confirmed by the two stages of weight loss upon heating before 275 ℃ (Fig. S3). The overall structure of 3 is like 2, but with different coordinate modes around Cu atoms. In 3, there are two Cu sites with different coordination numbers (Fig. 3B). One Cu is four-coordinate (Cu01) via two Cu–N(NH₃) (2.002 Å) and two Cu–κN (1.994 and 1.981 Å), forming a slightly twisted square plane. The other one, however, is five-coordinate (Cu02) via two Cu–N(NH₃) (1.969 Å), one Cu–κN (2.027 Å), one Cu–κO (2.019 Å), and one Cu–O(H₂O) (2.559 Å), making a distorted square pyramid. The alternating Cu01 and Cu02 develop the polymeric [Cu01–(µ₂—CA)–Cu02–(µ₂—CA)–Cu01–(µ₂—CA)] (NH₃ and H₂O are omitted for clarity). The supramolecular structure is spread two-dimensionally via intermolecular H-bonds (average distance 2.018 Å) (Fig. 3C), forming sheets that are vertically stabilized by the H-bonds between NH₃ and µ₂-CA. The µ₂-CA and Cu atoms along the chain are situated on a perfect plane with a dihedral angle of 0°, remarkably different from 2 (Fig. 3D).

  Figure 3

  

  Figure 3.  Crystalline structure of 3: (A) Calculated PXRD pattern; (B) Basic structural unit of 4; (C) Extended view along [100]; (D) Extended view along [100]. Color codes: Cu, Magenta; O, red; N, blue; C, gray; H, white.

  DownLoad: Full-Size Img PowerPoint

  We found that the CA tautomerized from the keto-form in 1 to the enol-form in 2 and 3. As shown in Table S4 (Supporting information), the average C=O bond length in 1 is shorter than that in 2 and 3, indicating that the C=O bonds are weakened. Moreover, in FTIR spectra the peaks assigned to C=O (1732 cm⁻¹) weaken dramatically in the spectra of 2 and 3 (Fig. S4 in Supporting information), consistent with the elongated bond length of C=O, supporting the tautomerization of CA from keto-form to enol-form [15,17,18]. It is also observed that by gradually increasing the M_(Cu-CA), the coordination modes in the final solids change from pure Cu–N (1, 2) to mixed Cu–N and Cu–O (3), then to pure Cu–O (rouaite), which might be reasoned by that Cu²⁺ is prone to attach with more electronegative sites. In mono-deprotonated CA (keto-CA), Cu²⁺ is attracted by N-sites. In di-protonated CA, keto-CA tautomerizes to enol-CA and part of the electronegativity gradually shifts from the N-sites to the O-sites, so that the Cu–O can be energetically stable. Consistently, at extreme high M_(Cu-CA), the solids are rouaite in which Cu²⁺ is attached with the even more electronegative O-sites in OH⁻, but not the CA or NH₃ molecules.

  Near edge X-ray absorption fine structure (NEXAFS) spectroscopy at Cu L_3,2 edges was used to determine the oxidation state of the copper center. As shown in Fig. 4A, for 1 the 2p_3/2 → 3d and 2p_1/2 → 3d transition of Cu is located at 931.54 and 951.24 eV, respectively, indicating that the Cu ions were in the divalent state and there was no apparent charge transfer between ligands (µ₁—CA) and cations [19]. For 2 and 3, the 2p_3/2 → 3d transition of Cu(Ⅱ) shifted to 931.04 and 931.14 eV, about 0.5 eV lower than 1. The same trend was also found in 2p_1/2 → 3d transition. Although those values for 2 and 3 still fall in the regime of Cu(Ⅱ), the slightly lower energy level of Cu(Ⅱ) might hint at partial ligand-metal charge transfer (µ₂—CA→Cu(Ⅱ)), which is reasonable due to the larger electronegativity of µ₂—CA than µ₁—CA.

  Figure 4

  

  Figure 4.  (A) NEXAFS sepctra at the Cu L_3,2 edges. The temperature dependence of χ_MT and χ_M⁻¹ for (B) 1, (C) 2 and (D) 3, respectively. The red solid lines are theoretical simulations based on the fitting procedure described in the text; The intrachain magnetic couplings and calculated spin density distributions of dimer fragments under broken symmetry (BS) states for (E) 2 and (F) 3. (isovalue = 0.002 a.u., α spin in green and β spin in blue). Color codes: Cu, orange red; O, red; N, blue; C, gray; H, white.

  DownLoad: Full-Size Img PowerPoint

  The dc magnetic susceptibility for 1 to 3 was measured at the temperature range of 2–300 K under an applied field of 1000 Oe (Figs. 4B–D), and the χ_MT products at room temperature are 0.376, 0.380 and 0.745 cm³ K/mol, respectively, which correspond to the theoretical values for spin-only Cu(Ⅱ) ion for 1 and 2 (0.375 cm³ K/mol, S = 1/2) and two non-interacting Cu(Ⅱ) ions for 3 (0.750 cm³ K/mol, S = 1). Upon cooling, the χ_MT value of complex 1 shows a quite slow declining tendency until around 20 K, then quickly drops to the minimum value of 0.282 cm³ K/mol at 2 K, which results from weak intermolecular antiferromagnetic (AFM) interactions and zero-field splitting effects [20–22]. The χ_MT products for 2 and 3 continuously decrease upon cooling, reaching almost zero at low temperatures, implying very strong intrachain AFM couplings mediated by the CA-bridges. The 1/χ_M vs. T plots for 1 to 3 indicate that their magnetic susceptibility follows the Curie-Weiss law within a certain temperature range, and the best-fitting between 100 K and 300 K gives C = 0.378 cm³ K/mol and θ = −0.308 K for 1; C = 0.493 cm³ K/mol and θ = −85.768 K for 2 and C = 1.013 cm³ K/mol and θ = −107.642 K for 3. The negative Weiss constant θ supports the dominant AFM interactions in these compounds. The field-dependent magnetization data of three complexes were also measured at 2 K under the field of 0–5 T (Figs. S5–S7 in Supporting information). The magnetization for 1 rises smoothly with magnetic field and reaches 0.94 µ_B at 5 T without saturation, while for 2 and 3, their magnetization exhibits very gradual increase and small magnitude, suggesting the existence of low residual paramagnetism and the occurrence of progressive decoupling of AFM couplings between Cu(Ⅱ) ions through the external magnetic field [23].

  To investigate the existing magnetic couplings in three compounds, we adopted different means and models to acquire the specific exchange constants with respect to their crystal structures. The temperature dependence of χ_MT of 1 was fitted using PHI program [24] and the intermolecular magnetic interactions (zJ') were treated with the mean-field approximation. The best-fitting of the magnetic data gives zJ' = −0.46 cm⁻¹ and TIP = 8 × 10⁻⁵ cm³/mol with g = 2.0, and this intermolecular exchange constant is in agreement with the Weiss constant θ. While for 2 and 3, their magnetic behaviors can be approximated as 1D uniform AFM chain with one isotropic interaction parameter J and 1D alternating AFM chain with two interaction parameters J₁ and J₂ (J₂ = αJ₁) respectively, and the Hamiltonian can be expressed as H=−J∑S_i·S_i+1 and H=−J∑_i=1^n/2(S_2i·S_2i−1+αS_2i·S_2i+1) (Figs. 4E and F), where α is alternation parameter (0 ≤ α ≤ 1), and α = 0 and α = 1 correspond to the cases of isolated pairs and a uniform chain. Therefore the experimental magnetic susceptibility data can be adequately described by the Bonner-Fisher approximation for 2 (Eq. 1, x = |J|/k_BT) [25–28] and theoretical expressions derived by Hatfield for 3 (0 ≤ α ≤ 0.4, Eqs. 2–8), x = |J|/2k_BT) [29–31], respectively. Considering the temperature-independent paramagnetism (TIP) and interchain interaction (zJ'), the temperature dependence of magnetic susceptibility for 2 and 3 can be best fitted by utilizing the parameters listed in Table S5 (Supporting information). The results demonstrated that intrachain CA-bridges were capable of transmitting strong AFM exchanges between Cu(Ⅱ) ions, while the α value closer to zero in 3 indicates that the magnetic behavior of this 1D alternating chain approaches to a dinuclear Cu(Ⅱ) system, and the different magnitude of J₁ and J₂ could be attributed to different bridging modes. Moreover, the differences of χ_MT products at low temperature range between 2 and 3 may originate from paramagnetic short chains after chain breaking in 2 and this part of magnetic susceptibility also follows the Curie law [27,28], which is reflected in the TIP terms that TIP = 1.86 × 10⁻⁴ cm³/mol for 2 is much higher than this value of 1.33 × 10⁻⁵ cm³/mol for 3.

  (1)

  (2)

  (3)

  (4)

  (5)

  (6)

  (7)

  (8)

  Then calculations based on density functional theory combined with broken-symmetry (DFT-BS) approach were performed using Orca 4.2 program [32] to probe the main source of intrachain magnetic exchanges in 2 and 3 (see Supporting information for more details). The theoretical exchange coupling constants (J_cal) can be calculated through the energy gap between high spin (HS) and broken symmetry (BS) states as well as corresponding molecular spin expectation values (Eq. 9). Ultimately, the calculated coupling constants are J_cal = −30.13 cm⁻¹ for 2, J_{1, cal} = −39.02 cm⁻¹ and J_{2, cal} = −0.54 cm⁻¹ for 3 which are very close to the experimental values. Moreover, the spin distributions from the calculations indicate only ~60% of the unpaired electrons are distributed on every Cu(Ⅱ) site, while the rest of spin densities are mainly delocalized towards surrounding N and O atoms of the CA ligands, suggesting that magnetic exchanges are produced by the delocalization mechanism and involve the d-type magnetic orbitals of Cu(Ⅱ) ions and sp² hybrid orbitals of ligand atoms. According to Kahn's model [33], the magnetic exchange constant possesses a direct proportion to the overlap density of both magnetic orbitals along the interaction pathway, and a larger degree of spin delocalization favors a stronger interaction. Checking the spin distributions of dimer fragments in 2 and 3, we found that CA molecule as bridge ligand can transmit such interactions due to the large spin densities on it, while the smaller intrachain interaction in 3 (J₂) is caused by the extended magnetic exchange pathway and staggered arrangement of magnetic orbitals, which are derived from the coordination between Cu(Ⅱ) ion and carbonyl oxygen.

  (9)

  In conclusion, we demonstrated that coordination complexes could be further transformed into coordination polymers upon proper incentives during crystal ripening. In our case, the copper cyanurate complex 1 were transformed into two different coordination polymers of copper cyanurate 2 and 3 by adjusting the molar ratio of Cu(Ⅱ) to CA. Coordination polymer 2 and 3 displays similar structures but distinct coordination site connectivity. Intrachain CA-bridges in 2 and 3 can transmit strong AFM exchanges between Cu(Ⅱ) ions. However, great disparity was observed in Cu(Ⅱ)-κN(HCA)κN—Cu(Ⅱ) and Cu(Ⅱ)-κN(HCA)κO—Cu(Ⅱ) pairs existed in 3 and the magnetic behavior of 3 is closer to a dinuclear Cu(Ⅱ) system, demonstrating the importance of site connectivity in molecular magnetism. This work provides new insights for understanding crystal transformation in coordination chemistry.

  Declaration of competing interest

  There are no conflicts to declare.

  Acknowledgments

  The authors are grateful for the financial support from the Australian Research Council (No. DP190101607) and National Natural Science Foundation of China (No. 21971203). Central Analytical Research Facility (CARF) at QUT is greatly acknowledged for technical assistance.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108542.

  
  
• 1. [1]

     T. van Westen, R.D. Groot, Cryst. Growth Des. 18 (2018) 4952–4962. doi: 10.1021/acs.cgd.8b00267

  2. [2]

     H. Xiao, C. Xue, P. Song, J. Li, Q. Wang, Appl. Surf. Sci. 337 (2015) 65–71. doi: 10.1016/j.apsusc.2015.02.064

  3. [3]

     S. Kandula, K.R. Shrestha, G. Rajeshkhanna, N.H. Kim, J.H. Lee, ACS Appl. Mater. Interfaces 11 (2019) 11555–11567. doi: 10.1021/acsami.9b02978

  4. [4]

     C.C. Yec, H.C. Zeng, J. Mater. Chem. A 2 (2014) 4843–4851. doi: 10.1039/C3TA14203E

  5. [5]

     M. Kalmutzki, M. Ströbele, F. Wackenhut, A.J. Meixner, H.J. Meyer, Angew. Chem. Int. Ed. 53 (2014) 14260–14263. doi: 10.1002/anie.201407708

  6. [6]

     M. Kalmutzki, M. Ströbele, F. Wackenhut, A.J. Meixner, H.J. Meyer, Inorg. Chem. 53 (2014) 12540–12545. doi: 10.1021/ic502035b

  7. [7]

     M. Kalmutzki, X. Wang, A.J. Meixner, H.J. Meyer, Cryst. Res. Technol. 51 (2016) 460–465. doi: 10.1002/crat.201600029

  8. [8]

     J. Tang, F. Liang, X. Meng, et al., Cryst. Growth Des. 19 (2019) 568–572. doi: 10.1021/acs.cgd.8b01782

  9. [9]

     F. Liang, L. Kang, X. Zhang, et al., Cryst. Growth Des. 17 (2017) 4015–4020. doi: 10.1021/acs.cgd.7b00677

  10. [10]

      M.J. Kalmutzki, K. Dolabdjian, N. Wichtner, et al., Inorg. Chem. 56 (2017) 3357–3362. doi: 10.1021/acs.inorgchem.6b02893

  11. [11]

      C.N.R. Rao, A. Ranganathan, V.R. Pedireddi, A.R. Raju, Chem. Commun. (2000) 39–40.

  12. [12]

      C. Xiao, Y. Li, H. Lun, C. Cui, Y. Xu, J. Solid State Chem. 208 (2013) 127–133. doi: 10.1016/j.jssc.2013.10.004

  13. [13]

      P.G. Slade, M. Raupach, E.W. Radoslovich, Acta Crystallogr. Sect. B 29 (1973) 279–286. doi: 10.1107/S0567740873002359

  14. [14]

      L.R. Falvello, I. Pascual, M. Tomás, E.P. Urriolabeitia, J. Am. Chem. Soc. 119 (1997) 11894–11902. doi: 10.1021/ja970297+

  15. [15]

      P. Meng, A. Brock, X. Wang, et al., Inorg. Chem. 61 (2022) 2086–2092. doi: 10.1021/acs.inorgchem.1c03291

  16. [16]

      P. Meng, Y. Xu, C. Yan, J. Xu, ACS Appl. Mater. Interfaces 12 (2020) 53125–53133. doi: 10.1021/acsami.0c16009

  17. [17]

      P. Meng, A. Brock, Y. Xu, et al., J. Am. Chem. Soc. 142 (2020) 479–486. doi: 10.1021/jacs.9b11336

  18. [18]

      T.J. Prior, J.A. Armstrong, D.M. Benoit, K.L. Marshall, CrystEngComm 15 (2013) 5838–5843. doi: 10.1039/c3ce40709h

  19. [19]

      K. i. Shimizu, H. Maeshima, H. Yoshida, A. Satsuma, T. Hattori, Phys. Chem. Chem. Phys. 3 (2001) 862–866. doi: 10.1039/b007276l

  20. [20]

      O. Stetsiuk, A. El-Ghayoury, F. Lloret, M. Julve, N. Avarvari, Molecules 26 (2021) 2122. doi: 10.3390/molecules26082122

  21. [21]

      Q. Zou, J.C. Liu, X.D. Huang, S.S. Bao, L.M. Zheng, Chin. Chem. Lett. 32 (2021) 1519–1522. doi: 10.1016/j.cclet.2020.10.019

  22. [22]

      W. Chen, G. Zhou, Z. Gou, et al., Chin. Chem. Lett. 32 (2021) 838–841. doi: 10.1016/j.cclet.2020.05.018

  23. [23]

      P. Pandey, B. Kharediya, B. Elrez, et al., Dalton Trans. 46 (2017) 15908–15918. doi: 10.1039/C7DT03115G

  24. [24]

      N.F. Chilton, R.P. Anderson, L.D. Turner, A. Soncini, K.S. Murray, J. Comput. Chem. 34 (2013) 1164–1175. doi: 10.1002/jcc.23234

  25. [25]

      J.C. Bonner, M.E. Fisher, Phys. Rev. 135 (1964) A640–A658. doi: 10.1103/PhysRev.135.A640

  26. [26]

      W.E. Estes, D.P. Gavel, W.E. Hatfield, D.J. Hodgson, Inorg. Chem. 17 (1978) 1415–1421. doi: 10.1021/ic50184a005

  27. [27]

      C. Kachi-Terajima, M. Ishii, T. Saito, et al., Inorg. Chem. 51 (2012) 7502–7507. doi: 10.1021/ic202708e

  28. [28]

      M.M. Dîrtu, Y. Boland, D. Gillard, et al., Int. J. Mol. Sci. 14 (2013) 23597–23613. doi: 10.3390/ijms141223597

  29. [29]

      J.W. Hall, W.E. Marsh, R.R. Weller, W.E. Hatfield, Inorg. Chem. 20 (1981) 1033–1037. doi: 10.1021/ic50218a017

  30. [30]

      I. Castro, M.L. Calatayud, W.P. Barros, et al., Inorg. Chem. 53 (2014) 5759–5771. doi: 10.1021/ic500544n

  31. [31]

      W. Huang, Y.H. Shui, X.X. Zhang, D.Y. Wu, Inorg. Chem. Commun. 33 (2013) 78–81. doi: 10.1016/j.inoche.2013.04.011

  32. [32]

      F. Neese, F. Wennmohs, U. Becker, C. Riplinger, J. Chem. Phys. 152 (2020) 224108. doi: 10.1063/5.0004608

  33. [33]

      O. Kahn, Angew. Chem. Int. Ed. 24 (1985) 834–850. doi: 10.1002/anie.198508341

• 

  Figure 1  (A) Photos illustrating Ostwald ripening of 1 from crystallites to large single crystals and (B) the corresponding powder X-ray diffraction (PXRD) patterns. (C) Photos of 1 transforming to 2 and 3. PXRD patterns (D) from 1 to 2 and (E) from 1 to 3.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Crystalline structure of 2: (A) Calculated PXRD pattern; (B) Basic structural unit; (C) Extended view along [021]; (D) Extended view along [101]. Color codes: Cu, Magenta; O, red; N, blue; C, gray; H, white.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Crystalline structure of 3: (A) Calculated PXRD pattern; (B) Basic structural unit of 4; (C) Extended view along [100]; (D) Extended view along [100]. Color codes: Cu, Magenta; O, red; N, blue; C, gray; H, white.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (A) NEXAFS sepctra at the Cu L_3,2 edges. The temperature dependence of χ_MT and χ_M⁻¹ for (B) 1, (C) 2 and (D) 3, respectively. The red solid lines are theoretical simulations based on the fitting procedure described in the text; The intrachain magnetic couplings and calculated spin density distributions of dimer fragments under broken symmetry (BS) states for (E) 2 and (F) 3. (isovalue = 0.002 a.u., α spin in green and β spin in blue). Color codes: Cu, orange red; O, red; N, blue; C, gray; H, white.

  下载: 全尺寸图片 幻灯片
•