Previous work has shown that excess PPh
3 can exchange the surface thiolate of gold nanoclusters. Because PPh
3 is a relatively weak protecting ligand for gold nanoclusters compared with thiolate, such an exchange was dubbed "reverse ligand exchange" (RLE). From a microscopic point of view, PPh
3 can break the "S-Au-S-Au-S" staple-like structures on the surface of gold nanoclusters and even change the conformation of the gold core to some extent, showing the "erosion" phenomenon. However, our earlier product was not easy to be purified with a meager yield [
50]. Two independent processes were combined in this work: (i) RLE-erosion and (ii) AGR-doping. It was surprising that the difference in the degree of erosion step can directly lead to the difference in the final products. We chose a high concentration of PPh
3 and shorter erosion time (kinetical control) to synthesize a novel nanocluster (
Fig. 1a) and the opposite for Au
26Cd
5 [
50]. For the experimental details, see Supporting information. As shown in
Fig. 1b and Fig. S1b (Supporting information), the three characteristic absorption peaks of Au
25(SR)
18 at ~400, ~450, ~680 nm began to weaken with the addition of PPh
3 in the two separate reactions. However, the different extents of erosion did not make noticeable differences in the UV-vis-NIR spectra. Proceeding to the second stage, we added the same dosages and concentrations of Cd(NO
3)
2 for the two reactions, and they went quickly in entirely different directions. As shown in
Fig. 1b, the peaks belonging to Au
25(SR)
18 after deep erosion disappeared significantly within 8 min after the addition of Cd
2+. As the reaction proceeded, the characteristic absorption peaks at 390 and 440 nm belonging to a novel nanocluster gradually appeared, and the peak at 680 nm belonging to Au
25 disappeared completely. After Cd(NO
3)
2 was added for
ca. 16 min, the reaction tended to reach equilibrium, and the crude product was completely precipitated from toluene. It is noteworthy that since the two clusters are slightly soluble in toluene but better soluble in methanol, we chose toluene as the solvent in step Ⅰ and introduced a moderate amount of methanol in step Ⅱ. The adjustment of the solvent ratio reduces the difficulty of purification. Electrospray ionization mass spectrometry (ESI-MS) was employed to identify the molecular composition of the as-obtained product, and a dominant peak was observed at
m/z = 4589.23 in the mass spectrum, which is assigned to [Au
11Cd(PET)
3(PPh
3)
7NO
3]
+ (
Fig. 1c). For the other reaction with a relatively light degree of erosion, it can be found that all absorption peaks of Au
25(SR)
18 completely disappeared 10 min after the addition of Cd(NO
3)
2 (Fig. S1b in Supporting information). Still, new peaks were generated at 355 and 394 nm, which belong to Au
26Cd
5, and the doping reaction lasted about 40 min to the end. The major peak, corresponding to [Au
13Cd
2(PET)
6(PPh
3)
6(NO
3)
2]
+, was observed at
m/z = 5307.14 (Fig. S1c in Supporting information) in the mass spectrum. Single crystal X-ray diffraction (SCXRD) confirmed their compositions [Au
11Cd(PET)
3(PPh
3)
7NO
3]NO
3 and [Au
13Cd
2(PET)
6(PPh
3)
6(NO
3)
2]
2·Cd(NO
3)
4, respectively. The main structural frameworks of Au
11Cd and Au
26Cd
5 are similar to those of Au
11 and Au
13, respectively, and the Cd atom combines with three S atoms to form a "paw-like" motif structure attached to the outside of the gold core (Figs. S2 and S3 in Supporting information).