The steady-state photoluminescence (PL) spectra and transient fluorescence lifetimes were examined to characterize the charge separation efficiency. As shown in Figure
3(a), the emission peaks of CdS are approximately around 440 nm, while TPPA and the CdS@TPPA samples exhibit dual emission peaks, which is commonly existed in COF due to the excited-state intramolecular proton transfer effect [
34,
35]. The emission peaks at 650-680 nm belonging to the keto form of TPPA show lower intensity in the CdS@TPPA samples compared with that of TPPA, suggesting their decreased recombination of photogenerated charge carriers. Moreover, CdS@TPPA2 displays the weakest intensity among them, revealing the crucial role of the TPPA nanoshell thickness in boosting charge separation. It is inferred that the core-shell structure favors facilitating the transfer of photogenerated charges by allowing them to flow along the radial direction, thereby inhibiting the undesirable recombination. However, a migration threshold that is subject to the limited charge mobility [
5] gives rise to a higher recombination rate as the thickness exceeds a certain value, as illustrated by the stronger PL intensity of CdS@TPPA3. The fluorescence lifetime decay profile (Figure
3(b)), fitted by the three-exponential fitting method (supplementary Table S1), reveals that the fluorescence lifetimes of TPPA and CdS@TPPA2 are 0.890 and 1.246 ns, respectively. The longer fluorescence lifetime of CdS@TPPA2 compared with TPPA can be attributed to the formed heterojunction structure, manifesting more opportunity for photogenerated electrons transferring to the active sites [
36]. Both the decreased PL intensity and prolonged fluorescence lifetime reveal that the charge separation efficiency is greatly improved through constructing the core-shell CdS@TPPA nanospheres. Electrochemical impedance spectroscopy (EIS) was detected to explore the internal resistance in the process of charge transfer. In comparison with those of TPPA and CdS, the Nyquist curve of CdS@TPPA2 presents a smaller semicircular diameter (Figure
3(c)), indicative of the lower internal charge-transfer resistance. Moreover, the semicircle under light is much smaller than that in dark, demonstrating that more electrons produce and transfer to the conduction band (CB) under light illumination [
37]. Similarly, CdS@TPPA2 displays a much larger transient photocurrent density (Figure
3(d)), further confirming its lower internal charge-transfer resistance. The UV-vis diffusion reflectance spectroscopy (UV-vis DRS) was conducted to analyze the light capture capacity of TPPA, CdS@TPPA samples, and CdS. As illustrated in Figure
3(e), the adsorption edge of three CdS@TPPA samples is located between those of pristine CdS and TPPA and displays a redshift with the increase of the TPPA nanoshell thickness. According to the Figure
3(e) inset, the corresponding bandgaps of TPPA and CdS are identified as 2.03 and 2.37 eV, respectively. Mott-Schottky plots disclose that the flat band positions of TPPA and CdS are determined to be -1.14 and -1.00 eV
vs. Ag/AgCl (-0.94 and -0.80 eV
vs. NHE,
), separately (Figure
3(f)), and the positive slope reveals that both TPPA and CdS are
-type semiconductors. Hence, the CB potential of TPPA and CdS is estimated to be -1.04 and -0.90 eV
vs. NHE (
), respectively, which is more negative than the potential for proton reduction (
vs. NHE,
). Combined with the bandgaps obtained from UV-vis DRS analysis, the valence band (VB) potential of TPPA and CdS is calculated to be 0.99 and 1.47 eV, severally. Such staggered band structure and intimate contact on the interface can result in the formation of the type II heterojunction [
38]. The formed heterojunction of the core-shell CdS@TPPA nanospheres drives the charge transfer from one semiconductor to another, which facilitates the spatial isolation of photogenerated electron-hole pairs.