tions, since six spins derived from one spin cage 2 and two guests 4 are accumulated in a restricted space. Various neutral molecules were also enclathrated by the spin cage, [14] but no significant change in the magnetic behavior of 2 was observed.
In summary, we have constructed a self-assembled radical cage with four spin centers. The inclusion of radical guests induced new spin-spin host-guest interactions. The water solubility of the spin cage will allow studies on organic radicals in aqueous media to enable the development of guest-tunable spin materials, such as new spin probe reagents and MRI contrast agents, that work in water.
Experimental Section
Spin cage 2: cis-(2,2'-Bipyridine)dinitratopalladium(II) (4.61 mg, 12 mmol) and 1,3,5-tri(4-pyridyl)verdazyl ( 2.54 mg, 8 mmol) were mixed in distilled water (2 mL), stirred for 1 h at RT, and dried in vacuo to give 2 quantitatively. CSI-MS: m/z 509.53 [MΓ7 PF 6 Γ ] 7+ , 618.29 [MΓ6 PF 6 Γ ] 6+ , 771.10 [MΓ5 PF 6 Γ ] 5+ , 1000.09 [MΓ4 PF 6 Γ ] 4+ . Elemental analysis calcd for C 128 H 104 N 52 O 36 Pd 6 β’18 H 2 O: C 39.33, H 3.61, N 18.63; found: C 39.46, H 3.75, N 18.38. Crystal data for 2: The very samll crystal size (80 20 20 mm) and poor crystallinity resulted in synchrotron radiation at the Photon Factory-Advanced Ring for Pulse X-rays (PF-AR) of the High Energy Accelerator Research Organization (KEK) being used. The diffraction data were measured at 80 K (l = 0.68890 ). C 129 H 102 N 47.66 O 70.85 Pd 6 , M r = 4091.71, trigonal, space group P3 Β―c1, a = b = 27.554( 1), c = 33.881(3) , V = 22 278(2) 3 , Z = 4, T = 80(2) K, 1 calcd = 1.220 g cm Γ3 ; 146 123 reflections were measured, of which 7514 were unique (R int = 0.1185) and were used in all calculations. The structure was solved by direct methods (SHELXL-97) and refined by full-matrix least-squares methods on F 2 with 721 parameters. R 1 = 0.1334 for 3201 observed data (I > 2s(I)) and wR 2 = 0.3376, GOF = 1.230, max./min. residual density 0.889/Γ0.974 e Γ3 . The trigonal crystal system results in one of the four ligands in 2 sitting on a threefold axis. Thus, the verdazyl core of this ligand was unable to be properly modeled. The refinement was done by modeling a C 3 -symmetric triazine ring instead of the verdazyl ring; see the Supporting Information.