FIGURE 4 UV-vis spectra (CHCl3/CH3CN, 1:1, v/v, 298 K) of: (A) Zn3a , (B) Zn3b , and (C)Zn3c before irradiation (black), after irradiation to PSS365 (red), after partial THI (pink dashed), and after full THI (blue dashed). Partial 1H NMR spectra (500 MHz, CDCl3/CD3CN, 1:1, v/v, 277 K) of: (D) Zn3a , (E) Zn3b , and (F) Zn3c before irradiation (black) and after irradiation to PSS365(red) showing the signals of methyl protons H-c . See Figure 1 for the proton assignment. Note that the apparent triplet observed for the methyl protons of Zn3b-st-st is the result of two overlapping doublets originating from the two non-identical molecular motor substituents.
did not provide information about the relative abundance of the metastable and stable species in solution. In order to measure these PSS ratios, we resided to 1H NMR spectroscopy (Figure 4D–F). The isomerization of the motor substituents of all isomers ofZn3 with UV light (λmax = 365 nm) resulted in similar deshieldings of the methyl protons H-c (Δδ = +0.26 ppm). After the subsequent thermal helix inversion, the original1H NMR spectra were restored. The decays of the metastable states were determined by monitoring the signals of methyl protons H-c (Supporting Information, Figures S41–S44) over time, and the corresponding apparent rate constants were in reasonable agreement with those obtained by UV-vis spectroscopy (Table 2, second and third column). NMR spectroscopy was also used to probe the structural changes, as proposed in Figure 3B. The irradiation ofZn3a (two loose motors) resulted in metastable species with one or two bound motors (Δδ = −0.91 ppm for H-a , and Δδ = −0.84 ppm for H-b ). As expected, the opposite effect was observed whenZn3c (two bound motors) was irradiated: new signals for loose motors (Δδ = +0.84 ppm for H-a , and Δδ = +1.24 ppm for H-b ) were observed. For Zn3b , the situation was more complex, because of the absence of symmetry in the stable isomer and each of its three possible metastable isomers. Nevertheless, by monitoring the chemical shift changes of the aromatic protons attached to the xylylene sidewalls of the porphyrin cages, we could detect all proposed isomers (Supporting Information, Figures S49–S51). Moreover,1H NMR analysis enabled us to determine the PSS ratios for the different diastereomers, i.e. the ratios between the total amount of metastable and stable motors (Table 2). These appeared to be nearly identical (metastable:stable ~1:3) for all investigated double-motorized cages Zn3 and also comparable to those found for the single-motorized cages Zn2 . It has to be noted though that the measured PSS ratios, which are relatively low, cannot be considered to be accurate PSS ratios. In order to avoid aggregation and precipitation of the porphyrin cages, the measurements were performed at 4 °C. At this temperature, the photochemical isomerization is still in competition with the thermal helix inversion. Lower temperatures would tackle this issue, but could not be applied as it results in precipitation of the porphyrin cages.
The differences in 3D structure between the stable isomers ofZn3a , Zn3b , and Zn3c caused significant differences in their binding affinities for viologen guests (Zn3a > Zn3b >Zn3c , see Table 1). These 3D structures could be changed by photochemical isomerization of the motor substituents (Figure 3). For instance, the stepwise isomerization of Zn3a-st-st (two loose motors) gives Zn3a-ms-st (one loose motor, one bound motor), and eventually Zn3a-ms-ms (two bound motors). The latter two isomers are pseudo-identical to Zn3b-st-st andZn3c-st-st , respectively, both of which have a lower affinity for viologen guests. Therefore, we expected that the photochemical isomerization of Zn3a-st-st would also result in a stepwise lower binding affinity for viologen guests. Similarly, the stepwise isomerization of Zn3c-st-st to metastable isomers that are pseudo-identical to Zn3b-st-st and Zn3a-st-st should lead to an increase in the binding affinity for viologen guests. For the non-symmetric isomer Zn3b-st-st , photochemical isomerization to the mixture of metastable isomers is expected to have a minor influence on the binding affinity for viologen guests, as the isomerization effects of the two motor substituents would cancel out. We studied the light-gated binding of viologen guests V4V6 in the double-motorized porphyrin cages Zn3 with the help of time-resolved fluorescence quenching spectroscopy (Figure 5). The host-guest complexes are non-fluorescent, and therefore the normalized fluorescence intensity originating from the zinc porphyrin after addition of guest is a measure for the amount of free host present in solution relative to the initial amount of free host. A normalized fluorescence intensity of 0.5 resembles a 1:1 mixture of free host and host-guest complex (50% occupancy). Based on the binding constants (Table 1), we calculated the amount of each guest that should be added to Zn3b (medium affinity for V4V6 ) in order to reach ~50% occupancy with an initial host concentration of 10−5 M (Supporting Information, Table S13). We first investigated the possible