ACS Catalysis
Research Article
a
excited-state PC4* to substrates appears to be responsible for
the cycloadditions.
Scheme 5. Synthesis of Caspase-3 Inhibitors
To test the feasibility of energy transfer from photosensitizer
to 1a, we recorded the absorption and emission spectra to
determine the singlet-state and triplet-state energies of PC4
the absorption of 1a and the emission of PC4 indicated that
the singlet−singlet energy transfer from excited-state PC4* to
1a is unlikely. In addition, the phosphorescence emission
between photosensitizers and 1a is in good agreement with
deoxygenated 2-MeTHF at 77 K (Figure 2A). Therefore, a
Dexter-type mechanism is completely feasible by triplet−triplet
energy transfer from excited-state PC4* to 1a.
Stern−Volmer quenching studies were conducted to
a
Reaction conditions: (i) Maleimides (0.20 mmol), acetaldehyde (0.3
mmol), aniline (0.24 mmol), and photosensitizer PC4 (0.5 mol %) in
CH2Cl2 (4 mL), irradiated with 405 nm LEDs under argon
atmosphere at room temperature for 12 h; (ii) concentrated solvent
was added to a mixture of DDQ (0.24 mmol) and toluene (2 mL)
and heated at 100 °C for 0.5 h.
calculate the rate of EnT from excited-state photosensitizer
̀
PC4* to the partners of the thia-Paterno−Buchi reaction.
̈
These studies indicated that xanthate 2a is a modest quencher
of PC4* because the calculated Stern−Volmer constant (KSV)
is 477 M−1 that is substantially lower than KSV for quenching of
PC4* by 1a (2880 M−1) (Figure 2B). The results are in line
with the Stern−Volmer analysis of luminescence lifetime
quenching as shown in Figure 3 (for 1a, KSV = 6391 M−1, kq =
2.9 × 109 M−1 s−1; for xanthate 2a, KSV = 1128 M−1, kq = 5.1 ×
108 M−1 s−1). Therefore, the proposed reaction mechanisms
4a (4.8 mmol, 0.869 g) to obtain products 5a and 5a′ in a high
yield (1.176 g, 83% yield).
Scheme 6. Gram-Scale Reaction of Aza-
Photocycloadditions
̀
for the thia-Paterno−Buchi reaction are outlined in Scheme 7.
̈
a
After triplet−triplet EnT from excited-state photosensitizer to
maleimide 1, we postulated that maleimide 1 reacts from the
triplet excited state with thiocarbonyl 2 to form a triplet
transition state (1,4-biradical intermediate) that subsequently
proceeds intersystem crossing (ISC) and recombination to
yield product 3.
The quenching of fluorescence intensity of PC4 by imine 4a,
however, almost achieved an adjacent Stern−Volmer constant
as that by 1a (Figure 4A). In addition, the lowest triplet energy
of imine 4a is around 262 kJ/mol from its phosphorescence at
77 K (see Figure S11 for details). These results indicate that
imines could also be sensitized by PC4* in the aza-
photocyclization. Having considered whether excited-state
imines initiated the annulation reaction, control experiments
were subsequently carried out. First, we intended to investigate
the possibility of the EnT between PC4* and imine 4a by
monitoring E → Z geometric isomerization of imine 4a in the
presence of PC4 and visible light because radiationless decay
upon rotation about the CN π-bond might lead to
geometric isomerization. However, the Z-isomer of imine is
very unstable and tends to achieve thermal equilibration back
to the E-isomer at room temperature.29 Indeed, no obvious
isomerization of imine 4a was observed under the standard
conditions.
Afterward, we carried out a control experiment where 1a was
replaced by maleic anhydride 1n, which have similar electron-
deficient cyclized structure but is difficult to be sensitized due
to its high triplet-state energy around 274 kJ/mol (see Figure
S11 for details). The measured ET data of 1n were consistent
with those reported previously (ET = 275 kJ/mol).30 In this
case, similar photocyclization should occur if the reaction is
initiated by EnT between PC4* and imine 4a. Importantly, we
did not detect any photocyclized product, suggesting that the
excited state of imine 4a should not play a crucial role in the
photocyclization (Scheme 8a). Furthermore, as the reduction
potential of maleic anhydride 1n is −0.88 V vs SCE,31 lower
than the reduction potential of the photoexcited PC4* (−0.95
V vs SCE), we were more confident to exclude the SET
a
Reaction conditions: 1a (4 mmol), 4a (4.8 mmol), and photo-
sensitizer PC4 (0.5 mol %) in CH2Cl2 (80 mL), irradiated with the
compact 15 W 405 nm LEDs under argon atmosphere at room
temperature for 30 h.
Although we developed iridium photosensitizer to catalyze a
wide variety of primary photoreactions via visible-light-
mediated EnT pathway, we also considered the possibility of
SET mechanism for these reactions. First, electrochemical
studies indicated that the redox potentials of phenylmaleimide
1a and its coupling partners (xanthate 2k27 and imine 4a28) lie
well outside the potentials of the photoexcited PC4* (Table
4). Obviously, photoinduced electron transfer between these
Table 4. Redox Potentials of Representative Compounds
ox
p/2
red
compound
E
(vs SCE) (V)
E
(vs SCE) (V)
p/2
phenylmaleimide 1a
xanthate 2k
imine 4a
>1.75
+1.75
+1.67
+1.28
−1.01
−1.55
−1.91
−0.95
PC4*
substrates and photosensitizer is not thermodynamically
feasible. Besides, the excited-state Ir(ppy)3* (−1.73 V)14 has
stronger reducibility than PC4 (−0.95 V). For the electron-
deficient maleimides, Ir(ppy)3 should be a more powerful
photosensitizer if the [2 + 2] cycloaddition could proceed well
through electron transfer. But no reaction took place when
Ir(ppy)3 was employed in these systems (Table 2, entry 12;
Table 3, entry 3). Therefore, the energy transfer from the
451
ACS Catal. 2021, 11, 446−455
He, Jian
