Mahima Sneha, Aditi Bhattacherjee, Luke Lewis-Borrell, Ian P. Clark, Andrew J. Orr-Ewing
5 min
Organic photocatalysts (PCs) are increasingly used in organocatalyzed atom transfer radical polymerization (O-ATRP), yet the fundamental mechanism of the activation step—the electron transfer (ET) from the excited PC to the radical initiator—remains poorly understood. This study investigates whether the excited-state character (locally excited vs. charge transfer) and the involvement of singlet or triplet states dictate the efficiency of this activation step.
The researchers examined nine N-aryl modified photocatalysts based on dihydrophenazine, phenoxazine, and phenothiazine cores. Using transient electronic and vibrational absorption spectroscopies (TEAS and TVAS), they tracked the kinetics of the electron transfer reaction with the initiator methyl 2-bromopropionate (MBP) across subpicosecond to microsecond timescales. By varying the concentration of MBP and the solvent, they derived bimolecular rate coefficients and compared them against the predictions of a modified Marcus-Saveant theory (the sticky model of dissociative electron transfer).
The study reveals that photocatalysts with locally excited (LE) character exhibit nearly diffusion-limited electron transfer rates, whereas those with charge transfer (CT) character react 5–10 times more slowly. Contrary to common assumptions in the field, high intersystem crossing (ISC) efficiency to a long-lived triplet state is not a prerequisite for effective polymerization control. In fact, at synthetically relevant concentrations of the initiator, electron transfer predominantly occurs from the singlet (S1) state, even for catalysts with high triplet quantum yields. The observed differences in reactivity are primarily governed by the Gibbs energy of the electron transfer, with LE catalysts benefiting from more favorable thermodynamics that lower the activation energy barrier.
These findings challenge the design paradigm that prioritizes long-lived triplet states for O-ATRP catalysts. Instead, the results suggest that thermodynamic parameters, specifically the Gibbs energy of electron transfer, are more critical for predicting catalyst performance. This provides a more rational basis for designing efficient organic photocatalysts for controlled radical polymerization.
Organic photocatalysts (PCs) are gaining popularity in applications of photoredox catalysis, but few studies have explored their modus operandi. We report a detailed mechanistic investigation of the electron transfer activation step of organocatalyzed atom transfer radical polymerization (O-ATRP) involving electronically excited organic PCs and a radical initiator, methyl 2-bromopropionate (MBP). This study compares nine N -aryl modified PCs possessing dihydrophenazine, phenoxazine, or phenothiazine core chromophores. Transient electronic and vibrational absorption spectroscopies over subpicosecond to nanosecond and microsecond time intervals, respectively, track spectroscopic signatures of both the reactants and products of photoinduced electron transfer in N, N -dimethylformamide, dichloromethane, and toluene solutions. The rate coefficients for electron transfer exhibit a range of values up to ∼10 10 M –1 s –1 influenced systematically by the PC structures. These rate coefficients are an order of magnitude smaller for catalysts with charge transfer character in their first excited singlet (S 1 ) or triplet (T 1 ) states than for photocatalysts with locally excited character. The latter species show nearly diffusion-limited rate coefficients for the electron transfer to MBP. The derived kinetic parameters are used to model the contributions to electron transfer from the S 1 state of each PC for different concentrations of MBP. Comparisons of singlet and triplet reactivity for one of the phenoxazine PCs reveal that the rate coefficient k E T (T 1 ) = (2.7 ± 0.3) × 10 7 M –1 s –1 for electron transfer from the T 1 state is 2 orders of magnitude lower than that from the S 1 state, k ET (S 1 ) = (2.6 ± 0.4) × 10 9 M –1 s –1 . The trends in bimolecular electron transfer rate coefficients are accounted for using a modified Marcus theory for dissociative electron transfer.
Sam: [probing] Does this make intersystem crossing irrelevant? If electron transfer is completing from the singlet state before significant triplet population even forms, why has the field been so focused on ISC efficiency?
Alex: [measured] Intersystem crossing isn't the problem — it's just not the solution either. The authors show that for many catalysts, electron transfer is essentially complete from the singlet manifold before ISC has time to contribute meaningfully. So optimizing for triplet yield was never wrong in principle, but it was optimizing a variable that isn't rate-limiting in the systems that actually work well.
Sam: [sitting back] So the design principle inverts. Instead of asking how do we maximize triplet population, the question becomes how do we tune charge-transfer character to get the electron transfer rate into the right kinetic window. [[RP_SECTION:computational-screening-potential|Computational Screening Potential]]
Alex: [nodding] And that reframing has real practical consequences. If you can predict rate coefficients from ground-state electronic structure calculations — which the Marcus-Savéant model in principle allows — you're no longer dependent on synthesizing and testing hundreds of dye candidates empirically. You could screen computationally, filtering for catalysts that sit in that charge-transfer regime before you ever run a polymerization.
Sam: [analytical edge] That's a meaningful shift toward in silico screening. But I want to push on the model's assumptions. The reduction potentials feeding into those activation energy calculations were measured in acetonitrile or DMF. How robust is that when the actual polymerizations run in different solvents?
Alex: [cautious] It's a real simplification. Solvent-dependent shifts in reduction potential can be non-trivial, and if those shift the calculated activation energies systematically, the model's predictions could be off in specific chemical environments. The authors don't fully resolve this — it's a limitation they acknowledge but don't quantify. [[RP_SECTION:model-limitations-and-future|Model Limitations and Future]]
Sam: [nodding] And there's a second gap. The entire mechanistic analysis focuses on the activation step — the radical generation event. But dispersity is also controlled by deactivation, the process that caps radical growth between cycles. If deactivation kinetics are comparably important, this model is only giving you half the picture.
Alex: [agreeing] That's the primary methodological constraint on the current work. The activation kinetics explain why locally excited catalysts perform poorly, but a complete predictive model for dispersity will eventually need to integrate the deactivation rate as well. The authors are clear that this is the next piece — it's a limitation, not an oversight.
Sam: [measured] Still, the shift from empirical trial-and-error to a thermodynamically grounded, predictive framework is a genuine step forward. The finding that charge-transfer character — not triplet lifetime — is the key design variable gives the field a quantitative target to optimize against.
Alex: [calm, concluding] And it's the kind of mechanistic clarity that changes how you read the existing literature. A lot of structure-activity relationships in photoredox ATRP were built on the triplet lifetime heuristic. This work suggests those correlations may have been capturing charge-transfer character indirectly, without recognizing it as the operative variable.
Sam: [reflective] Which means the predictive power was always latent in the data — it just needed the right physical model to surface it. That's a satisfying kind of result. Thanks for listening to ResearchPod.