After the mixture was stirred at 80 ℃ for 4 h, a saturated solution of NH
4PF
6 in MeOH was added to generate red precipitates, which were washed thoroughly with MeOH and H
2O. After drying
in vacuo at 50 ℃ for 18 h, the desired octagram
C was obtained in 98% yield with PF
6− counterions. The ESI-MS spectrum of octagram
C (4 mg/mL in MeCN) showed one set of dominant peaks resulting from the ions with the charge states from 18+ to 9+ generated by losing the corresponding numbers of PF
6−. The experimental isotope patterns for each charge state agreed well with the corresponding theoretical distributions (
Fig. 1c and Fig. S35 in Supporting information). The 2D TWIM-MS plot of octagram
C showed a single distribution with narrow drift times for the charge states ranging from 19+ to 13+, indicating formation of a single component. However, a charge-dependent conformation transition at 12+ charge state was observed, suggesting that 3D octagram
C has certain structural flexibility (Fig. S40 in Supporting informaiton) [
21]. The experimental average collision cross-section (CCS) of 1868.4±154.1 Å
2 for all charge states derived from the TWIM-MS analysis was in good agreement with the theoretical value of 1863.5±98.8 Å
2 generated by trajectory method (TM) in MOBCAL based on 200 annealed structures (Table S1 in Supporting information) [
30]. The relatively large standard deviation also reflected its conformational flexibility. The structure of
C was verified by
1H NMR spectroscopy as well (
Fig. 1b), which showed four singlets at
δ 9.06, 9.03, 9.02 and 8.99 with a 1:1:1:1 integration ratio assigned to 3', 5' protons of tpy-A, B, C, D respectively. The peaks for 6, 6" protons of the uncoordinated tpys in
LA (tpy-C, D) exhibited a dramatic upfield shift from 8.72 ppm to 7.84 ppm (
Figs. 1a and
b, Δ
δ=−0.88) owing to the electron shielding effect [
31]. In the non-aromatic region, a broad singlet at
δ 4.03 and two sharp singlets at
δ 3.26 and
δ 3.25 in a 4:1:1 integration ratio attributed to the methoxy groups close to tpy-C, D, tpy-A and tpy-B, respectively. In theory, self-assembly from
LA and Zn
2+ can afford two structures: octagram
C and possible
G (Fig. S48 in Supporting information). The composition (Zn
8LA4) of the possible isomer
G is the same as octagram
C which cannot be distinguished by mass spectroscopy. However, in principle, NMR spectroscopy can rule out other isomers. Compared to highly symmetric octagram
C (the chemical environmental of four
LAs is identical), the possible isomer
G is less symmetrical which suggesting two kinds of chemical environment of
LA.
1H NMR of possible
G will exhibit eight kinds of signals for tpy units. As a matter of fact,
1H NMR result revealed four kinds of tpy signals which fully validated the generation of octagram
C. All proton assignments for octagram
C were further confirmed by 2D COSY and NOESY NMR (Figs. S22 and S23 in Supporting information). The diffusion-ordered spectroscopy (DOSY) NMR spectrum of
C in CD
3CN at 25 ℃ showed a narrow band, clearly demonstrating there was only a single component in solution (diffusion coefficient log
D = -9.81 m
2/s, Fig. S24 in Supporting information) [
32]. The above evidences have proved the formation of octagram
C.