Singlet and Triplet Contributions to the Excited-State Activities of Dihydrophenazine, Phenoxazine, and Phenothiazine Organocatalysts Used in Atom Transfer Radical Polymerization

Organocatalysts Used in Atom Transfer Radical Polymerization

Article

Singlet and Triplet Contributions to the Excited-State Activities of

Dihydrophenazine, Phenoxazine, and Phenothiazine

∥

,∥

∥

Aditi Bhattacherjee,* Mahima Sneha,* Luke Lewis-Borrell, Giordano Amoruso, Thomas A.A. Oliver,

Jasper Tyler, Ian P. Clark, and Andrew J. Orr-Ewing *

Cite This: J. Am. Chem. Soc. 2021, 143, 3613 −3627

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: The photochemical dynamics of three classes of

organic photoredox catalysts employed in organocatalyzed atom-

transfer radical polymerization (O-ATRP) are studied using time-

resolved optical transient absorption and ﬂuorescence spectrosco-

py. The nine catalysts selected for study are examples of N-aryl and

core-substituted dihydrophenazine, phenoxazine and phenothia-

zine compounds with varying propensities for control of

polymerization outcomes. Excited singlet-state lifetimes extracted

from the spectroscopic measurements are reported in N,N-

dimethylformamide (DMF), dichloromethane (DCM), and

toluene. Ultrafast (<200 fs to 3 ps) electronic relaxation of the

photocatalysts after photoexcitation at near-UV wavelengths

(

318−390 nm) populates the ﬁrst singlet excited state (S ). The

S -state lifetimes range from 130 ps to 40 ns with a considerable dependence on the photocatalyst structure and the solvent. The

competition between ground electronic state recovery and intersystem crossing controls triplet state populations and is a minor

pathway in the dihydrophenazine derivatives but is of greater importance for phenoxazine and phenothiazine catalysts. A comparison

of our results with previously reported O-ATRP performances of the various photoredox catalysts shows that high triplet-state

quantum yields are not a prerequisite for controlling polymer dispersity. For example, the photocatalyst 5,10-bis(4-cyanophenyl)-

,10-dihydrophenazine, shown previously to exert good polymerization control, possesses the shortest S -state lifetime (135 ps in

DMF and 180 ps in N,N-dimethylacetamide) among the nine examples reported here and a negligible triplet-state quantum yield.

The results call for a re-evaluation of the excited-state properties of most signiﬁcance in governing the photocatalytic behavior of

organic photoredox catalysts in O-ATRP reactions.

. INTRODUCTION

and computationally directed strategies for higher eﬃciency

photoredox catalysts are built on accessing excited electronic

states with long lifetimes that enable diﬀusive bimolecular

Controlled methods for radical polymerization are of great

interest for the synthesis of polymers with a desired molecular

weight or chain length. Photoredox-catalyzed atom transfer

radical polymerization (ATRP), in particular, has emerged as

an eﬀective approach for the controlled synthesis of polymers

with a precise composition and dispersity. Recent develop-

ments in organic photoredox catalysis have improved these

radical-based polymerization reactions by, most notably,

−11

electron transfer reactions with a radical precursor.

Molecular triplet states are suﬃciently long lived to favor

intermolecular single-electron transfer to radical initiators over

intramolecular deactivation pathways. For this purpose, many

early photocatalysts deliberately incorporated transition-metal

atoms as chromophores to enhance intersystem crossing rates

−5

avoiding contamination from metal-containing catalysts.

This progress has motivated the development of new classes of

organic photocatalyst (PC) molecules based on N-aryl

phenothiazines, phenoxazines, and dihydrophenazines that

have provided the benchmark for organocatalyzed ATRP

through high spin−orbit coupling.

Received: January 9, 2021

Published: February 25, 2021

,6−8

(

denoted O-ATRP).

The redox properties of a molecular photoredox catalyst

mainly arise from the higher oxidation or reduction potential of

the excited singlet or triplet electronic states of the molecule in

comparison to its ground electronic state. Both synthetically

2021 American Chemical Society

J. Am. Chem. Soc. 2021, 143, 3613−3627

613

Journal of the American Chemical Society

separated π * SOMOs. This spatial separation between the two

Article

Seminal research on O-ATRP optimization has placed

emphasis on the energy of the lowest excited triplet state (T₁)

of a given photoredox catalyst to predict, rationalize, and tailor

SOMO electrons has been proposed to facilitate ISC and

7,11,20

intermolecular electron transfer.

Here, we present direct

counterevidence that one such optimized PC (5,10-bis(4-

cyanophenyl)-5,10-dihydrophenazine) with an electron-with-

drawing nitrile (−CN) functionality on the N-aryl group

stabilizing its CT-character S and T states repopulates the S

,10,13

photocatalytic activities.

This emphasis derives mainly

from three premises: (a) spin-forbidden relaxation to the

ground state prolongs the lifetime of a triplet state in

comparison to a singlet state; (b) back electron transfer,

which quenches photoredox activity, is inhibited in the triplet-

spin radical pairs formed by initial electron transfer from the

ground electronic state 2 orders of magnitude more quickly

than the corresponding PC (5,10-bis(4-methoxyphenyl)-5,10-

dihydrophenazine) with electron-donating substituents

(−OMe), which is detrimental to the eﬃciencies of both

ISC and bimolecular electron transfer reactions.

photocatalyst T state (although this is less of a concern for the

dissociative electron transfer pathways exploited in O-ATRP);

(

c) organic photocatalyst design should draw on the existing

In the current work, we study the solvent-dependent

photochemical dynamics of nine PCs based on dihydrophe-

nazine, phenoxazine, and phenothiazine central cores (Figure

knowledge of the photophysical behavior of organometallic

complexes, many of which exhibit ultrafast intersystem

crossing (ISC).

excited-state lifetimes has recently been challenged,

4−16

The need for organic PCs to have long

1). Such compounds have been widely applied in the recent

7,18

3,6,21,22

and we

development of O-ATRP.

These prior works charac-

instead propose a greater focus on excited-state reduction

potentials, regardless of the excited-state spin multiplicity, as a

controlling factor in O-ATRP optimization because of their

terized the steady-state spectroscopy and excited-state redox

potentials of these (and other) organic photocatalysts, the

latter by combining cyclic voltammetry data with calculated

excited-state energies, and tested their polymerization control

inﬂuence on intermolecular electron transfer rate coeﬃcients.

,23−27

Consequently, molecules with ﬂuorescence lifetimes of a few

nanoseconds and low triplet quantum yields may participate in

bimolecular electron transfer reactions driving O-ATRP

of the as-grown polymer dispersity.

Design principles

have focused on the enhanced absorption of near-ultraviolet or

visible light for photoinitiation and radical generation to make

these polymerization reactions amenable to activation by light-

emitting diodes or natural sunlight and to avoid undesirable

through their S states instead of their triplet states.

Although organic photocatalysts do not contain metal ions,

useful analogies may be drawn between the metal-centered or

metal to ligand charge transfer (CT) states common in

photoexcited metal complexes and the locally excited (LE) or

CT character excited states of organic compounds such as

side reactions.

The PCs chosen for the current study are shown in Figure 1.

They make use of modiﬁcations to N atom and core

substituents and show varying performances in O-ATRP

11,19,20

that we seek to understand. The selected PCs are ﬁve

dihydrophenazines (denoted PC-N), three phenoxazines (PC-

O), and one phenothiazine (PC-S) derivativesnote that this

nomenclature is used speciﬁcally in this work to distinguish

among the diﬀering heteroatoms. In four of these cases, (PC-

N2, PC-N3, PC-N5, and PC-O1), prior publications from our

group have reported the analysis of transient absorption

dihydrophenazines, phenoxazines, and phenothiazines.

This LE vs CT character of the frontier orbitals has been

argued to account for large discrepancies in photocatalytic

eﬃciencies of organic compounds that have similar reducing

potentials of their triplet states. A LE state descriptor is used

when the frontier π and π* orbitals are both localized over the

core (a heterocyclic ring fused between two benzene rings; see

Figure 1); CT instead describes spatially separated π and π*

orbitals that are delocalized over the core and N-aryl groups,

respectively. In CT states, electron-withdrawing (donating)

groups on the N-aryl rings stabilize (destabilize) the spatially

17,28,29

spectroscopy data to determine S -state lifetimes.

New

solvent-dependent ﬂuorescence lifetime measurements are

presented here for all nine PCs, together with previously

unreported mid-IR and UV−visible transient absorption

spectroscopy studies for ﬁve of them, in three solvents chosen

for their diﬀerent polarities (Table S1 ). The selected solvents

are N,N-dimethylformamide (DMF), dichloromethane

(

DCM), and toluene. In addition, for the four previously

studied PCs (PC-N2, PC-N3, PC-N5, PC-O1), new mid-IR

transient absorption spectroscopy measurements are presented

which extend over much greater time intervals in comparison

to those in our prior work. The combined outcomes from the

current and our prior measurements provide a consolidated

picture of the eﬀects of structural modiﬁcations and solvent

properties on the photodynamics of dihydrophenazine,

phenoxazine, and phenothiazine PCs.

The structural modiﬁcations in the chosen organic PCs

explore the eﬀects of changing one central-ring heteroatom in

the core dihydrophenazine, phenoxazine, or phenothiazine

moiety and of altering the N-aryl substituents: for example, by

comparing the eﬀects of either electron-donating or -with-

drawing groups. In the literature, PC-N3−PC-N5 have been

reported to be promising dihydrophenazine-based photo-

Figure 1. Structures of nine organic photocatalyststs studied here.

The PC-N, PC-O, and PC-S nomenclature indicates dihydrophen-

zine, phenoxazine, and phenothiazine core structures, respectively.

The core heteroatoms are emphasized by colored labels. The PCs

N1−N5 and O1−O3 further diﬀer in their N-aryl substituents. Only

one phenothiazine catalyst is studied. PC-O3 further incorporates a

core modiﬁcation by the addition of biphenyl groups.

catalysts with high eﬃciency and polymer dispersity control.

PC-O1 and PC-O2 are both eﬀective phenoxazine-based

catalysts with a phenyl or naphthyl substituent at the N atom.

PC-O3 additionally involves substitution at the phenoxazine

614

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

Article

core with biphenyl groups, added to enhance visible-light

absorption via extended conjugation. Finally, PC-S is a

phenothiazine-based catalyst representing a class of organic

PCs reported to have reducing properties superior to those of

repetition rate was set at 1 kHz, and that of the probe was 100 kHz.

The pump−probe time delays were continuously tuned from 1 ps to

12 ns using an optical delay line, from 12 ns to 10 μs using electronic

delays and probe−pulse selection, and every 10 μs utilizing the 100

kHz repetition rate. The UV pump power was set to deliver a pulse

energy of 150−300 nJ/pulse at the sample, and the pulse energies of

both mid-IR probes were ∼0.05 μJ/pulse.

2.2. Fluorescence Lifetime Spectroscopy. Sub- and few-

nanosecond dynamics in transient absorption spectroscopy measure-

ments were supplemented by time-correlated single photon counting

3,10,13

metal-centered PCs and high ISC quantum yields.

. METHODS

The PCs indicated in Figure 1 were synthesized according to

6,7,30

published protocols in the literature,

and a detailed description of

(TCSPC) determinations of photocatalyst ﬂuorescence lifetimes.

the methods and characterization can be found in Section S1 of the

TCSPC measurements used a custom-built apparatus at the

University of Bristol. The fundamental laser output at 680 or 740

nm of a narrow-band high-power Ti:sapphire oscillator (3.7 W, 80

MHz, Chameleon Ultra II, Coherent) was frequency-doubled in a 2

mm thick BBO crystal (θ = 29.2°, Eksma), giving an excitation

wavelength of 340 or 370 nm. The residual near-IR fundamental light

was removed using two dichroic beam splitters (106160, Layertec).

To avoid re-excitation of samples, the repetition rate of the laser pulse

train was reduced to 3 MHz using an acousto-optic modulator

Supporting Information. The solvents N,N-dimethylformamide,

dichloromethane, toluene (all anhydrous, ≥99.8%), and toluene-d

(

99 atom % D) were purchased from Sigma-Aldrich and used without

further puriﬁcation. The relevant solvent properties are summarized in

Table S1 of the Supporting Information.

The time-resolved absorption spectroscopy measurements em-

ployed solutions of the PCs at various concentrations, adjusted for the

sample path length (see below), that were chosen not to exceed an

absorbance of 0.5 at the excitation wavelength. Typical concen-

trations, unless otherwise indicated, were 4.2 mM (PC-N1), 5 mM

(AOM) based pulse picker (APE cavity dumper kit). Samples were

diluted to have an absorbance of 0.2 at the excitation wavelength, and

they were not circulated in TCSPC experiments. The diﬀracted light

from the AOM was subsequently focused through a 10 μm pinhole to

remove any undiﬀracted parent beam and then collimated and

focused into the sample with a second lens. Photoluminescence was

collected from a 1 cm path length cuvette, at 90° relative to the

excitation laser, with an inﬁnity-corrected microscope objective (4×/

(

PC-N4), 3.2 mM (PC-O1), 2.9 mM (PC-O2), 1 mM (PC-O3), and

.6 mM (PC-S). To ensure a fresh part of the sample was excited and

probed with every laser measurement, the solutions were circulated

from a 10 mL sample vial using a peristaltic pump, and sample cells

were spatially rastered. For most measurements, samples were not

purged by nitrogen because the excited-state dynamics of interest

were too rapid to be aﬀected by dissolved oxygen. However, to assist

in the assignment of triplet-state contributions to the photochemistry

of the PCs, the eﬀects of quenching by oxygen over longer

0.2 NA Plan Apochromat, Nikon). The collected ﬂuorescence was

ﬁltered to acquire emissions at wavelengths longer than 375 or 395

nm (Schott Glass long-pass ﬁlter) or 450 nm (long-pass ﬁlter

FELH0450, Thorlabs) to eliminate laser scatter. A polarizer in the

detection line was set to the magic angle relative to the excitation laser

polarization to remove rotational anisotropy eﬀects from the decay

kinetics. The emitted light was focused onto an avalanche-photodiode

detector (ID100-20-REG, IDQ) with an achromatic doublet (AC508-

(

nanosecond to microsecond) time scales were tested by comparing

with measurements made on samples purged with nitrogen. Even with

the use of a sealed ﬂow system reported previously, interference by

dissolved oxygen could not be completely suppressed.

2.1. Transient Absorption Spectroscopy. Transient electronic

absorption spectroscopy (TEAS) was carried out at the University of

Bristol using a UV pump pulse (at wavelengths of 370 nm for PC-N1

and PC-N2, 318 nm for PC-O1, PC-O2, and PC-S, and 389 nm for

PC-O3) and a white-light continuum (WLC) probe (340−700 nm)

pulse, the relative polarizations of which were at magic angle

75-A, Thorlabs). The photon counts from the detector were

acquired by a time to digital converter (Time Tagger 20, Swabian

Instruments) and binned into histograms with a bin width of 10 ps.

The wavelength-dependent instrument response functions of the

TCSPC spectrometer were determined to be ∼170 and 260 ps for

31,32

(

54.7°).

The WLC was generated by focusing 800 nm pulses into

370 and 340 nm excitations, respectively. TCSPC traces were ﬁtted to

a 3 mm thick CaF window that was rastered both horizontally and

an analytical solution of a Gaussian instrument response function

convoluted with single- or multiple-exponential decays.

vertically to prevent optical damage. The WLC probe was

subsequently recollimated by an oﬀ-axis parabolic mirror before

being focused and spatially overlapped with the pump beam in a

noncollinear geometry at the sample. The sample ﬂowed continuously

between two CaF windows separated by 380 μm. The transmitted

probe was dispersed in a spectrometer (Andor, Shamrock 163)

.3. Electronic Structure Calculations. Computational charac-

terizations of the ground- and excited-state properties of the PCs were

performed using Gaussian 09 software. Geometries and harmonic

vibrational frequencies of the ground electronic states of the

photocatalysts were computed using the restricted Kohn−Sham

density functional theory. The pure hybrid-GGA functional PBE0 was

used with the polarized and augmented double-ζ basis set 6-31+G(d)

equipped with a 1024-element photodiode array (Entwicklungsburo

Stresing) to measure the transmitted white-light spectra in the

presence and absence of the pump pulse. The electronic absorption

spectrum of holmium oxide was used to provide a pixel to wavelength

calibration for the spectrometer. The time delay between the pump

and probe pulses was controlled by an optical delay line and limited to

a maximum of 1.3 ns for this experimental setup.

(

6d,7f). Solvent eﬀects for toluene, DCM, and DMF were included by

implicit simulation of solvation as a continuous polarizable medium

using the total solute density model (SMD).

37−39

The PBE0

functional long-range London dispersion interactions were simulated

with Grimme’s D3 dispersion correction, supplemented by the

Transient vibrational absorption spectroscopy (TVAS) measure-

ments were carried out at the LIFEtime facility (and in a few cases,

the ULTRA facility) at the Rutherford Appleton Laboratory.

These experiments made use of the same pump wavelengths for the

various PCs as indicated for TEAS experiments. However, the mid-

infrared probe (comprising two separately tunable IR laser pulses,

0,41

Becke−Johnson damping function (GD3BJ).

Vertical transition

31,33−35

energies, oscillator strengths, and excitation amplitudes were

computed with time-dependent density functional theory (TD-

DFT) using the Coulomb attenuated variant of the B3LYP functional

42,43

(CAM-B3LYP).

CAM-B3LYP was used with the 6-311++G-

−1

(2d,p) basis set and nonequilibrium SMD solvation to characterize

the excited electronic states of the photocatalyst in the Franck−

Condon region.

each of ∼200 cm bandwidth) was tuned to the ring C−C stretch/

−1

C−H bend region and spanned 1400−1650 cm . The linear

polarizations of the pump and probe beams were ﬁxed at the magic

angle (54.7°). After they passed through the sample, the probe pulses

were dispersed in spectrometers ﬁtted with 128-element MCT

3. RESULTS AND DISCUSSION

(

mercury−cadmium−telluride) detector arrays (IR Associates). A

Solutions of the organic PCs were ﬁrst characterized by steady-

state absorption and ﬂuorescence spectroscopy. The photo-

chemical dynamics of the PCs were then explored using

pair of CaF windows separated by 100 μm spacers conﬁned the

ﬂowing samples. The infrared spectra were wavenumber-calibrated

using the known infrared absorption bands of polystyrene. The pump

615

J. Am. Chem. Soc. 2021, 143, 3613−3627

Article

Figure 2. UV−vis absorption and FTIR spectra of six of the PCs in DCM. Gray circles identify the excitation wavelengths used in the time-resolved

spectroscopic investigations. Computed ground-state structures of the organic photocatalysts (PCs) are also shown. The corresponding data for

17,28

PC-N2, PC-N3, and PC-N5 have been reported elsewhere.

ultrafast transient absorption spectroscopy, with sub-pico-

second time resolution, and time-resolved ﬂuorescence spec-

troscopy. The time-resolved spectroscopy methods determined

excited-state lifetimes, propensities for ISC to triplet states, and

time scales for recovery of ground-state molecules by a

combination of radiative and nonradiative pathways.

strengths, LE or CT characters of the S states, and S and T

1 1 1

reduction potentials are summarized in Table 1. These

properties are estimated from computed excited-state energies

,7,20

and reported ground-state cyclic voltammetry data.

The

ﬂuorescence emission spectra of all six PCs measured in DMF,

DCM, and toluene are shown in Figure S1.

3.2. Time-Resolved Spectroscopy of UV-Excited

Photocatalysts. Figure 3 shows the TVAS spectra of six

PCs in three diﬀerent solvents (DMF, DCM, and toluene-d₈).

Similar measurements for PC-N2, PC-N3, and PC-N5 have

.1. Steady-State Spectroscopy Data. The UV−vis

absorption spectra of six of the PCs, measured in DCM, are

shown in Figure 2 for wavelengths between 275 and 425 nm,

together with the optimized geometries of the PCs in their S₀

states. Corresponding data for the other three dihydrophena-

7,28

been reported elsewhere.

The probed IR region from 1400

17,28

−1

zine PCs studied have been reported elsewhere.

In their

to 1650 cm corresponds mainly to the −CC− ring

stretching/deformation modes of the photocatalysts. For

comparison, the FTIR spectra of selected PCs in their ground

electronic states in DCM are shown in Figure 2. The presence

of strong absorption bands in DMF restricted the observable

spectral window for TVAS to 1525−1625 cm in this solvent.

Toluene-d was chosen in place of normal toluene for these

preferred structures, the N-aryl groups of all the PCs are

oriented perpendicularly to the central cores to reduce steric

repulsion between the H atoms of the aromatic rings. A similar

perpendicular orientation is preferred for the biphenyl core

substituents in PC-O3. PC-S diﬀers structurally from all other

PCs in that it has a puckered phenothiazine core to

accommodate the larger S atom. The UV−vis spectra show

that the strongest absorption bands for all PCs lie in the

ultraviolet region from 200 to 300 nm; however, we focus here

on photoexcitation via the longer wavelength spectral features

in view of their applications in visible/near-UV initiated

photoredox catalysis.

−

experiments because it has fewer interfering infrared features in

this region; however, the TEAS and TCSPC experiments made

use of normal toluene.

Figure 4 shows the time dependences of the integrated

TVAS band intensities, derived from spectra of the PC

solutions in DMF and obtained by ﬁtting the various excited-

state absorption (ESA) and ground-state bleach (GSB)

features to Gaussian functions. The corresponding analyses

were also performed for TVAS measurements made with

The frontier orbitals involved in the long-wavelength (>300

nm) photoexcitations of the PCs are predominantly π and π*

orbitals. These absorption bands (Figure 2) may consist of

multiple, unresolved, S → S (n = 1−4) electronic excitations.

solutions of the PCs in DCM and toluene-d . For each

Further details of the contributing orbital excitations, oscillator

combination of a PC and solvent, these types of time-

616

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

Article

Table 1. Computed and Experimental Photophysical Data

for the Studied PCs in DMF

following discussion considers in turn the photodynamics of

the dihydrophenazine, phenoxazine, and phenothiazine PCs

deduced from all the excited-state population growth and

decay time constants summarized in Tables 2−4.

reduction potential (V) and

electronic LE or CT character

3.2.1. Dihydrophenazine Catalysts. PC-N1 and PC-N4 are

S → S

examples of dihydrophenazine catalysts with electron-donating

(−OMe, PC-N1) and electron-withdrawing (−CN, PC-N4)

N-phenyl para substituents and have been chosen as the focus

of the discussion because of these contrasting electronic

properties. Where appropriate, further comparisons are drawn

with the other dihydrophenazine PCs studied. The triplet

wavelength/

excitation

energy (nm/ oscillator

photocatalyst

PC-N1

eV)

strength

S₁

T₁

374/3.3 (S₁)

32/3.7 (S₂)

376/3.3 (S₁)

33/3.7 (S₂)

374/3.3 (S₁)

37/3.7 (S₄)

402/3.1 (S₁)

34/3.7 (S₄)

377/3.3 (S₁)

14/3.9 (S₄)

307/4.0 (S₁)

0.0000

0.2056

0.000

0.206

0.000

−2.50 (LE)

−2.15 (LE)

PC-N2

PC-N3

PC-N4

PC-N5

PC-O1

−2.40 (LE)

−1.83 (CT)

−2.41 (CT)

−1.83 (CT)

−2.44 (CT)

−2.15 (LE)

•+ 3

excited-state reduction potentials, E°(PC / PC*), of PC-N1

and PC-N4 obtained by combining DFT calculations of

excitation energies and ground-state redox potentials (versus

saturated calomel electrode) are −2.15 and −1.89 V,

−2.08 (CT)

−1.89 (CT)

−1.92 (CT)

−2.01 (CT)

0.154

0.0000

0.1508

0.0024

0.2176

0.017

respectively. These values are compared with the correspond-

ing estimates for the S -state reduction potentials in Table 1.

Without any N-phenyl substituents (i.e., PC-N2), the

corresponding T -state reduction potential is −2.15 V. All of

these dihydrophenazine derivatives are more reducing from

their T₁states (and by implication, their S₁states) in

comparison to commonly used metal-centered PCs such as

03/4.1 (S₂)

90/4.3 (S₃)

0.056

0.088

PC-O2

PC-O3

PC-S

337/3.7 (S₁)

06/4.1 (S₂)

354/3.5 (S₁)

03/4.1 (S₂)

320/3.9 (S₁)

0.0000

0.0303

0.7692

0.3734

0.0018

0.0165

0.0261

−1.75 (CT)

−2.02 (CT)

−2.16 (CT)

−1.76 (CT)

−1.72 (CT)

−1.65 (CT)

polypyridyl iridium complexes following photoexcitation.

The electron-withdrawing substituents stabilize excited

states of CT character and lower their reduction potentials

in comparison to states with LE character. In these ππ* CT

states, the electron density in the lower-lying π SOMO is

mainly on the dihydrophenazine core, whereas that in the

higher π* SOMO is mainly on the N-phenyl ring(s). This

08/4.0 (S₂)

82/4.4 (S₃)

The S → S photoexcitation wavelengths (nm) and energies (eV)

delocalization is augmented by −CN (in PC-N4) and −CF

and the corresponding oscillator strengths were computed by TDDFT

methods at the CAM-B3LYP/6-311++g(2d,p)/SCRF=(SMD, sol-

(

in PC-N3) groups in the para position. Miyake and co-

workers hypothesized that the superior photocatalytic ATRP

performances of PC-N4 and PC-N3 in comparison to PC-N1

derive from their excited-state CT character and used this

vent, NonEq) level of theory. The reduction potentials (vs SCE) for

•

the S and T excited electronic states were calculated using E°(PC /

PC) − E(PC*), where E(PC*) denotes the energy (in eV) of the S

,20

or T state. S energies were estimated from the wavelengths of

principle as the basis for further PC design.

Our

maxima in experimental emission bands, and T energies were

quantiﬁcation of the eﬀects of the N-phenyl ring substituents,

•+

calculated using DFT methods. E°(PC /PC) values measured by

cyclic voltammetry were taken from refs 6, 7, and 20. Locally excited

and hence of the CT or LE character of the T (or S ) states,

on the rate coeﬃcients for intermolecular electron transfer

reactions that initiate an ATRP cycle is the subject of a

forthcoming publication.

(

LE) or charge-transfer (CT) characters of the excited states were

deduced from the computed molecular orbitals. The method of

determination of E°(PC / PC(S )) appears to underestimate their

values.

•+ 1

The TDDFT calculations indicate that photoexcitation is to

the S state in PC-N1 and to the S state in PC-N4 at the near-

UV wavelengths used here. Internal conversion to S proceeds

dependent ESA and GSB band intensities were globally ﬁtted

to single- or double-exponential functions to extract the time

constants summarized in Tables 2 and 3. The tables also report

ﬂuorescence lifetimes for the photoexcited PCs determined

from the TCSPC measurements shown in Section S5 of the

Supporting Information. Fluorescence lifetimes of a few tens of

nanoseconds reﬂect the low oscillator strengths for the S → S

on ultrafast time scales, faster than intermolecular electron-

transfer reactions or radiative decay, in accordance with

Kasha’s rule. This ultrafast relaxation S → S is mostly

n 1

comparable to, or faster than, the temporal resolution of our

TVAS experiments but can be observed by this spectroscopic

method in the case of PC-N1 as a combination of internal

conversion and vibrational cooling kinetics (see Tables 2 and

4).

transitions in some of these photocatalysts (Table 1), whereas

signiﬁcantly shorter ﬂuorescence lifetimes, some of which are

sub-nanosecond, indicate competing nonradiative decay path-

ways from S to S or T or perhaps an electron transfer

When photoexcited at 370 nm, PC-N1 shows strong,

moderate, and weak GSB features at 1485, 1511, and 1613

−

cm , respectively, in the TVAS spectra of Figure 3. In contrast

−

reaction with the solvent in the case of measurements made in

to these sharp depletion bands, which have 5−10 cm full

DCM.

widths at half-maximum (fwhm), the feature due to ESA is

−

Complementary TEAS measurements for DMF, DCM, and

toluene solutions of the PCs are shown in Figure 5. The

methods for spectral decomposition into S , S and T ESA

unusually broad, spanning from 1530 to 1590 cm . This width

is likely to derive from overlapping vibrational bands of the S₁

state (as discussed below and in Section 3.3, there is no clear

evidence for T₁population in our data for PC-N1). In

comparison, PC-N4 shows one strong and two weak GSB

bands, performed using the KOALA program, are reported in

Section S4 of the Supporting Information. Kinetic ﬁtting of the

time-dependent intensities of the spectral components gave the

exponential time constants summarized in Table 4. The

−

features at 1488, 1503, and 1595 cm , respectively, when it is

photoexcited at 370 nm. An observed ESA band at 1547 cm⁻

617

J. Am. Chem. Soc. 2021, 143, 3613−3627

Article

Figure 3. TVAS measurements for six PCs in solution in DMF, DCM, and toluene-d , shown at the selected time delays in nanoseconds as

indicated by the inset keys. In each case, measurements were also made at multiple time delays of up to 10 μs (not shown). A negative time trace

(

−0.01 ns) is shown as a dashed gray line in all of the panels to provide a reference for the baseline. The concentrations of the solutions were 4.2

mM (PC-N1), 10 mM (PC-N4), 3.2 mM (PC-O1), 2.9 mM (PC-O2), 1 mM (PC-O3), and 7.5 mM (PC-S). Gaps in the data (PC-N4 in DCM

and in toluene-d at 1510 cm ; PC-O3 in DCM at 1480 cm ) resulted from damaged pixels in the MCT array detector. A discontinuity in the

spectrum for PC-N1 in toluene-d at 1520 cm was an artifact of joining two separate MCT detectors. The spectral noise in the TVAS

measurements in toluene-d at 1575 and 1585 cm was caused by low counts on the detector because of solvent infrared absorption features.

−

−1

−

is much narrower than for PC-N1; for PC-N4, all peaks have a

signiﬁcant involvement of any intermediate electronic state

(e.g., T ).

Both prior and current measurements for DMF solutions of

−

width of 15−20 cm . The ground-state FTIR spectral

absorptions in this region match the wavenumbers of the

−

PC-N3, which has para, electron-withdrawing −CF sub-

GSB features and have a fwhm of 10 cm , which is ascribed to

inhomogeneous broadening by interaction with the solvent

stituents on the N-phenyl groups, show fast (∼700 ps) internal

conversion to the ground state, on a time scale approaching

(

see Section S3 of the Supporting Information for steady-state

that for PC-N4 relaxation. On the basis of this observation,

we posit that the polar solvent stabilizes the CT-character S₁

states of PC-N4 and PC-N3, thereby reducing the S −S

FTIR spectra).

N,N-Dimethylacetamide (DMAc) is the solvent medium

used in several reported O-ATRP photopolymerization studies.

Of the three solvents used here, DMF most closely resembles

DMAc in terms of polarity (Table S1), and it is a useful

analogue because the IR spectrum of DMAc masks all the

transient absorption features of interest. Inspection of Figures

energy gap and opening up a rapid internal conversion (IC)

route to S (most likely through a conical intersection), which

outcompetes ﬂuorescence decay or ISC. Such pathways are not

accessible for the LE-character S state of PC-N1 or for PC-N2

studied both here and previously by Koyama et al. The S₁

and 4 shows that all the spectral features described above

excited-state lifetime for PC-N4 in DMF was independently

veriﬁed using TEAS to be 136 ± 9 ps (cf. 135 ps by TVAS),

following a 2.4 ± 0.2 ps growth by IC from the photoexcited S₄

which fall in the narrow IR-transparent window of DMF decay

completely to zero. However, ground-state recovery is much

faster for PC-N4 than for PC-N1. The reasonable agreement

between time constants deduced from the analysis of TVAS

state. The experimentally determined S lifetime for PC-N4 in

DMAc was found to be ∼180 ps by TEAS, consistent with

expectations based on the similar polarities of the two solvents.

Thus, for PC-N4 and PC-N3, both of which have been shown

to provide good dispersity control in the polymer products of

O-ATRP, we ﬁnd that the excited-state lifetime is shorter than

^aⁿ^d^T^C^S^P^C^d^a^t^a⁽^T^a^b^l^e²⁾^c^oⁿ^ﬁ^r^m^s^a^f^a^s^t^e^r^d^e^c^a^y^o^f^t^h^e^S1

population in PC-N4. The recovery of the GSB mirrors the

decay of the ESA, which indicates a predominantly direct

relaxation pathway from the S₁to the S₀state without

618

J. Am. Chem. Soc. 2021, 143, 3613−3627

Article

S₁state, respectively. A shift of the band maximum to shorter

wavelength at times longer than 5 ps is consistent with this

latter interpretation.

For PC-N1 in toluene-d , TVAS measurements show that

the ground-state recovery is complete within 100 ns. The time

constant for the decay (11.7 ± 0.6 ns) matches that obtained

from TCSPC (10.95 ± 0.17 ns), and in DMF the ground-state

recovery is even slower (time constants of 38.5 ± 1.3 ns from

TVAS and 19.55 ± 0.17 ns from TCSPC, with unclear reasons

for the discrepancy in the values). However, in DCM the

photophysical behavior is noticeably diﬀerent, with an ESA

decay time constant of 2.2 ± 0.1 ns. Although the ESA bands

assigned to photoexcited PC-N1 in DCM decay completely to

zero, the GSB recovery remains 70% incomplete at time delays

up to 10 μs, indicating a long-lived intermediate state or

reactive loss. Four small positive diﬀerential absorbance peaks

−

are noted at 1466, 1508, 1552, and 1601 cm , and we assign

these bands to the PC-N1^•radical cation, pointing to an

electron transfer reaction of the photoexcited PC-N1 with the

DCM solvent. This interpretation is consistent with prior

reports of halogenated solvents such as DCM participating in

electron transfer reactions with photoexcited organic mole-

cules. No such indications of cation formation are seen for

photoexcited PC-N4 in DCM, most likely because the sub-

nanosecond relaxation to S (Table 3) outcompetes inter-

molecular electron transfer to the solvent.

The extended time scale TVAS measurements performed in

this work also reveal PC-N2 behavior which is similar to that of

PC-N1 in the three solvents studied. In DMF and toluene-d₈

solutions, the PC-N2 S -state population decays with time

Figure 4. Kinetic traces from TVAS for six UV-photoexcited PCs in

DMF obtained by integrating band intensities at diﬀerent time delays.

Positive amplitude traces with red or purple symbols denote excited-

state absorption (ESA), and negative amplitude traces with blue

symbols denote ground-state bleaches (GSB). Insets show to which

PCs the traces correspond. Note that the magnitudes of all peak

amplitudes are normalized to 1. For a given PC, the decays of both

ESA and GSB intensities were globally ﬁtted (solid black lines) to

single or double-exponential functions to extract a consistent set of

photophysical time constants (see Table 3).

constants of 16.8 ± 2.0 and 25.5 ± 7.6 ns, respectively, whereas

in DCM, PC-N2 has a shorter S lifetime and incomplete

ground-state recovery, indicating a reaction with the solvent.

Both PC-N1 and PC-N2 have LE character in their S states,

and the large S -state reduction potentials can drive ET

reactions with the chlorinated solvent. However, for PC-N2

solutions in DMF and toluene-d , the GSB recovery includes a

second, slower component with a time constant of ∼100−300

ns, which we assign to quenching of T -state population by

dissolved oxygen.

nanosecond time scales for diﬀusional bimolecular electron

transfer reactions at the concentrations used in O-ATRP

synthesis.

The TEAS data for PC-N1 shown in Figure 5 for solutions

in DMF conﬁrm the kinetics of excited-state relaxation derived

from the TVAS measurements. The observed bands are

In DCM, the PC-N4 ground-state recovery is also complete

assigned to S ESA, and spectral decomposition and kinetic

(

Figure 3), and the time scale matches that of the decay of the

analysis give time constants for the growth of this ESA as 2.8 ±

0.4 and 134 ± 12 ps. The two time constant values most likely

S₁ESA bands observed by TVAS, indicating little or no

branching into the triplet manifold by ISC. A comparison of

the time constants in Tables 2 and 4 shows that the relaxation

dynamics in DCM (510 ± 30 ps) are comparable to those in

DMF (135 ± 3 ps). In toluene, however, PC-N4 behaves

represent steps of internal conversion to S

photoexcited S state and subsequent vibrational cooling in S

but a component of direct excitation to the S state followed by

vibrational relaxation cannot be discounted. Although it is

formally an electronically forbidden transition (Table 1), the S

→ S excitation in PC-N1 might become weakly allowed by

from the initially

diﬀerently: while the S ESA decays with a time constant of 5.7

0.3 ns, the ground state bleach shows a double-exponential

recovery with time constants of 5.7 ± 0.3 and 398 ± 16 ns. We

vibronic coupling. In toluene, the corresponding time

assign this second time constant to the quenching of T -state

constants for growth of the S absorption are 2.7 ± 0.6 and

population by dissolved oxygen, the details of which are

54 ± 6 ps; the latter time constant is also obtained from a

TVAS band intensity analysis (Table 3). We again interpret

these processes as signatures of IC from the photoexcited S₂

state and vibrational relaxation of internally hot PC-N1(S₁)

discussed in Section 3.3. The longer S lifetime in toluene is

supported by our TCSPC data and by observation of strong

ﬂuorescent emission from PC-N4 in this solvent, but not in

DMF or DCM under the same measurement conditions. TEAS

molecules. In DCM, the PC-N1 S population growth with a

measurements show that the main S ESA band of PC-N4

time constant of 30 ± 4 ps is indicative of S → S internal

1 1

population loss is consistent with TVAS measurements, with

centered at 450 nm grows with 1.0 ± 0.1 and 17.7 ± 1.4 ps

time constants in toluene (and 1.1 ± 0.1 and 18.4 ± 1.6 ps in

DCM), which we attribute to IC and vibrational cooling in the

conversion and vibrational cooling in S . The rate of S

a time constant of 2.2 ± 0.1 ns. A band consistent with PC-

619

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

Article

Table 2. Time Constants (in ns) Obtained from the Fitting of Kinetic Data from TVAS and TCSPC Measurements for UV-

Photoexcited Dihydrophenazine PCs in the Three Solvents DMF, DCM, and Toluene-d (Normal Toluene for TCSPC)

time constant from TVAS and TCSPC data (ns)

photocatalyst

PC-N1

method

TVAS

DMF

DCM

toluene-d₈

0.15 ± 0.07

8.5 ± 1.3

0.030 ± 0.004

0.077 ± 0.007

11.7 ± 0.2

10.95 ± 0.17

2.2 ± 0.1

TCSPC

TVAS

19.55 ± 0.17

0.30 ± 0.17 (43%)

.30 ± 0.17 (57%)

PC-N2

16.8 ± 2.0

0.013 ± 0.002

25.5 ± 7.6

87 ± 21

2.77 ± 0.40

104 ± 15

TCSPC

TVAS

12.98 ± 0.17

0.24 ± 0.17 (37%)

8.83 ± 0.17

.55 ± 0.17 (63%)

PC-N3

PC-N4

0.635 ± 0.022

2.85 ± 0.1

70 ± 160

3.07 ± 0.24

36.4 ± 5.7

TCSPC

TVAS

0.76 ± 0.17 (94%)

0.135 ± 0.003

2.62 ± 0.17

0.51 ± 0.03

3.25 ± 0.17

5.7 ± 0.3

98 ± 16

TCSPC

0.20 ± 0.17 (58%)

0.72 ± 0.17 (93%)

5.31 ± 0.17

.58 ± 0.17 (42%)

PC-N5

TVAS

5.2 ± 0.4

0.060 ± 0.008

25 ± 1

7 ± 1

00 ± 100

TCSPC

5.33 ± 0.17

10.85 ± 0.17

6.13 ± 0.17

70 nm excitation was used for all measurements on dihydrophenazines. Single or double entries denote time constants derived from single- and

double-exponential ﬁts, and percentages in parentheses show relative amplitudes of the components for TCSPC data. Quoted uncertainties are

from the experimentally determined IRF, with ﬁts to data returning smaller errors. The time constant corresponds to the growth of the S

population, perhaps with a vibrational cooling component. S lifetime shortened by reaction with the solvent (dichloromethane). Assigned to T

lifetimes and likely to be aﬀected by O quenching. Suspected to be ﬂuorescence from UV excitation of ground-state PC generated in the solution

by the light-induced ET reaction of PC(S ) with DCM. Once formed, the PC is stable in solution and can absorb the 370 nm pump light. Data

obtained using the ULTRA laser system at the RAL Central Laser Facility. The ﬂuorescence spectrum shows unusual behavior in DMF, with an

·+

unexplained second decay component. Time constants from TVAS data were previously reported and are added here for comparison.

Table 3. Time Constants (in ns) Obtained from the Fitting of Kinetic Data from TVAS and TCSPC Measurements for UV-

Photoexcited Phenoxazine and Phenothiazine PCs in the Three Solvents DMF, DCM, and Toluene-d (Normal Toluene for

TCSPC)

time constant from TVAS and TCSPC data (ns)

photocatalyst

PC-O1

method

DMF

DCM

toluene-d₈

TVAS

2.1 ± 0.1

0.02 ± 0.01

0.014 ± 0.002

2.4 ± 0.2

6 ± 12

1.5 ± 0.2

3 ± 9

27 ± 2

TCSPC

TVAS

2.61 ± 0.26 (79%)

1.20 ± 0.26 (84%)

0.040 ± 0.008

2.77 ± 0.26

0.035 ± 0.008

2.4 ± 0.3

PC-O2

0.07 ± 0.04

.6 ± 1.2

6.4 ± 0.3

5 ± 13

112 ± 4

57 ± 4

TCSPC

TVAS

11.59 ± 0.26

0.58 ± 0.18

6.55 ± 0.26

0.25 ± 0.02

2.57 ± 0.05

2.40 ± 0.17

0.014 ± 0.003

2.57 ± 0.26

0.50 ± 0.12

2.4 ± 0.3

PC-O3

PC-S

.3 ± 0.4

TCSPC

TVAS

2.72 ± 0.17

2.01 ± 0.17

0.007 ± 0.001

1.77 ± 0.08

5.9 ± 0.6

8 ± 30

3.8 ± 0.2

95 ± 10

673 ± 18

TCSPC

6.27 ± 0.26

3.90 ± 0.26

1.85 ± 0.26

Excitation wavelengths were 318 nm for TVAS and 340 nm for TCSPC, with the exception of PC-O3, which was excited at 389 nm for TVAS and

70 nm for TCSPC. Single or double entries denote time constants derived from single- and double-exponential ﬁts, and percentages in

parentheses show relative amplitudes of the components for TCSPC data. Quoted uncertainties are from the experimentally determined IRF, with

ﬁts to data returning smaller errors. The time constant corresponds to the growth of the S population, perhaps with a vibrational cooling

component. Assigned to T lifetimes and likely to be aﬀected by O quenching. Extends to 840 ± 31 ns with improved N purging. Minor

contributions from a second time constant of 6.78 ± 0.26 ns in DMF and 6.10 ± 0.26 ns in DCM are also observed in the TCSPC experiments.

Time constants for PC-O1 were previously reported from our laboratory and are added here for comparison. Extends to 1.63 ± 0.03 μs with

improved N purging.

620

J. Am. Chem. Soc. 2021, 143, 3613−3627

Article

Figure 5. Transient electronic absorption spectra measured for six PCs in DMF, DCM, and toluene up to a maximum time interval of 1 ns. The

legend indicates representative pump−probe time delays in picoseconds. The solid gray line (−1 ps, where the probe pulse arrives at the sample

before the pump pulse) provides a reference for the baseline. Figure S3 of the Supporting Information further shows TEAS data measured in

DMAc.

N1^•production grows on this same time scale, providing

PC-O3) wavelength, respectively, in comparison to the

absorption maxima of the dihydrophenazines discussed

above. Consequently, PC-O3 is deemed a more promising

visible-light-mediated photoredox catalyst than PC-O1 and

PC-O2. Three close-lying excited singlet states of LE and CT

character underlie the 280−360 nm absorption band in PC-O1

further evidence for charge transfer from the PC-N1 S state to

the solvent DCM competing with relaxation to S₀.

On the 1.3 ns time scales of the TEAS measurements in

DMF and toluene, there is no evidence for the growth of

additional features that might correspond to ESA from the T₁

states of the dihydrophenazine PCs PC-N1 and PC-N4.

Similar conclusions were drawn in our prior studies of PC-N3

(Table 1). Our previous analysis of PC-O1 showed that the

ordering of these states is sensitive to solvent polarity and that,

at our excitation wavelength, this PC is predominantly excited

to an S state of LE character in DCM or DMF but to both S

7,28

and PC-N5,

although spectroscopic evidence for some

branching into T was reported for PC-N2 in both DMF and

DCM solutions and is discussed further in Section 3.3.

An interesting trend emerges when we compare the ﬁve

dihydrophenazine PCs for solutions in DMF: the PCs with CT

(LE) and S₃(CT) states in toluene. For PC-O2, the

calculations reported in Table 1 suggest our 318 nm laser

excitation populates the S state, which has mixed LE/CT

excited-state character have shorter S -state lifetimesPC-N3

character.

(

136 ps), PC-N4 (500 ps), and PC-N5 (5 ns)than the PCs

In PC-O3, which was designed to have an absorption

spectrum that extends into the visible region, the direct S →

with LE characterPC-N1 (11 ns) and PC-N2 (17 ns). These

observations may hold clues to the S -state relaxation

S₁transition at wavelengths approaching 400 nm (Table 1)

involves electronic excitation from a HOMO localized on the

phenoxazine core to a SOMO delocalized across the core and

both biphenyl substituents. The emissive singlet state was

assigned by Sartor et al. as having symmetry-breaking CT

character involving electron density redistribution from the

core to one core-attached biphenyl group (denoted CT-

Biphenyl).

TVAS measurements of all the phenoxazine PCs (Figure 3)

show a ground-state depletion feature near 1487 cm .

mechanisms, but further theoretical work is required to

identify the details of the decay pathways.

.2.2. Phenoxazine Catalysts. The photophysical studies of

PC-O1 and PC-O2 use excitation at 318 nm, whereas that of

PC-O3 is at 389 nm. These excitation wavelengths were

chosen because they lie close to the maxima of the ﬁrst strong

absorption bands of the respective molecules in their ground

states (Figure 2). Note that these excitation wavelengths are at

much shorter (∼50 nm, PC-O2, PC-O1) or longer (∼19 nm,

11,24

−

621

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

Article

Table 4. Exponential Time Constants for Growth and Decay

of Bands in the TEAS Spectra of the Dihydrophenazine,

Phenoxazine, and Phenothiazine PCs in DMF, DCM and

Toluene Solutions

initially photoexcited singlet state and vibrational relaxation in

−

the S state. A band at 1484 cm in the TVAS spectra of PC-

O2 in DCM grows at later times and is interpreted as a

signature of population of the T₁state on the basis of

computed T absorption band frequencies, and because of the

observed sensitivity of its decay rate to dissolved O₂.

−

time constant (ps)

photocatalyst

PC-N1

DMF

DCM

toluene

Figures 3 and 4 show that, in DMF solutions, the IR ESA

and GSB spectral signatures decay with double-exponential

behavior for PC-O1 and PC-O2, but the ESA decay is single-

exponential in PC-O3, with a time constant of 3.3 ± 0.4 ns.

The double-exponential decay behavior for PC-O2 persists in

DCM and toluene: the faster components (8.6 ± 1.2, 6.4 ±

0.3, and 2.4 ± 0.3 ns in DMF, DCM, and toluene-d₈,

respectively) are consistent with TCSPC measurements and

2.8 ± 0.4 (↑S₁)

34 ± 12 (↑S )

13 ± 2 (↑S₁)

30 ± 4 (↑S₁)

2.7 ± 0.6 (↑S₁)

2200 ± 100 (↓ S₁) 54 ± 6 (↑S₁)

18 ± 3 (↑S₁) 23 ± 2 (↑S₁)

2770 (↓S , ↑T )

PC-N2

PC-N3

PC-N4

∼

613 ± 54

0.400 ± 0.01(↑S₁) 1.7 ± 0.1 (↑S₁)

1.0 ± 5.0 (↑S₁)

1.1 ± 0.1 (↑S₁)

18.4 ± 1.6 (↑S₁)

∼3000 (↓S , ↑T )

2.4 ± 0.2 (↑S₁)

36 ± 9 (↓S₁)

1.0 ± 0.1 (↑S₁)

17.7 ± 1.4 (↑S₁)

correspond to the loss of S population. The larger time

∼510(↓S₁)

constant values are found to be sensitive to N purging of the

PC-N5

PC-O1

10.4 ± 2.6 (↑S₁)

11 ± 8 (↑S₁)

14.8 ± 0.3 (↑S₁)

3.4 ± 0.7 (↑S₁)

54.6 ± 3.3 (↑S₁)

14.8 ± 0.4 (↑S₁)

4.7 ± 0.3 (↑S₁)

91 ± 5 (↑S₁)

22.8 ± 1.0 (↑S₁)

solutions, which suggests that these correspond to triplet-state

lifetimes and possible overlap of the T and S ESA features in

the spectra. Similar considerations apply to PC-O1 in the three

solvents, with evidence of S and T states participating in the

100 ± 100 (↓S , 1500 ± 200 (↓S , 2400 ± 200 (↓S ,

1 1 1

↑

T )

↑T )

photodynamics. For PC-O1 in toluene-d , we recently

PC-O2

1.9 ± 0.2(↑S₁)

7 ± 6(↑S₁)

0.60 ± 0.01(↑S₁) 2.1 ± 0.1(↑S₁)

estimated a T quantum yield of Φ(T ) = 0.56 ± 0.03. We

∼

20 ± 0.5(↑S₁)

21.0 ± 1.5 (↑S₁)

∼2400(↓S , ↑T )

also observed the reaction of photoexcited PC-O1 with DCM.

The photodynamics for PC-O3 are broadly consistent for all

three solvents, with the decay time constants showing modest

shifts from 2.4 to 3.3 ns. Agreement with time constants from

TCSPC measurements conﬁrms the assignment to S₁

population decay, and the measured ﬂuorescence lifetime of

8600(↓S , ↑T )

∼6400(↓S , ↑T )

PC-O3

PC-S

2.9 ± 0.1 (↑S₁)

2500 (↓S₁)

1.4 ± 0.1 (↑S₁)

250 (↑S₁)

2400 (↓S₁)

0.9 ± 0.1 (↑S₁)

21 ± 2 (↑S₁)

500 (↑S₁)

2400 (↓S₁)

2.0 ± 0.1(↑S₁)

19.4 ± 3.2(↑S₁)

≥

1.0 ± 5.2 (↑S₁)

900 (↓S , ↑T )

∼3900 (↓S , ↑T ) ∼1770(↓S , ↑T )

.72 ns in DMF agrees well with a prior TCSPC study of PC-

The symbols ↑ and ↓ denote growth and decay, respectively, of the

intensities of ESA bands assigned to the speciﬁed states. Recovery of

the S population may be double exponential but always shows one

time constant in agreement with S₁population decay. Time

constants greater than 1000 ps are poorly determined from ﬁts to

the TEAS measurements, which extend only to 1.3 ns. Instead, time

constants from TVAS (Tables 2 and 3) were shown to be consistent

O3 in DMAc, for which a time constant of 2.87 ns was

reported. The features observed by TVAS all decay to

baseline within 10 ns, suggesting ISC to the manifold of triplet

states is inoperative or at most plays a minor role.

^L^o^s^s^o^f^p^o^p^u^l^a^tⁱ^oⁿ^f^r^o^m^t^h^e^P^C^-^O¹^S1 state and

contributions to the photodynamics from ISC are conﬁrmed

by distinct bands observed in the time-resolved electronic

spectra (Figure 5). The TEAS spectrum of PC-O1 in DMF

clearly shows an intermediate triplet state developing on a

with the TEAS data. Value obtained in DMAc.

However, in the case of PC-O3 it is barely discernible against

pronounced ESA bands, although the asymmetry in the 1475

cm peak is a consequence of the overlapping GSB on the

higher wavenumber side. For PC-O1 and PC-O2, the 1487

nanosecond time scale, as the band assigned to S

grows and then decays. The early time growth of the S ESA

initially

−

feature (with time constant 14.8 ± 0.3 ps) is attributed to

internal conversion after photoexcitation to a higher-lying

singlet state. Similar photodynamics are observed in the TEAS

spectra of PC-O1 in DCM and in toluene. For PC-O2 in

toluene and DMF, while a triplet-state feature is not as readily

discernible as for PC-O1, the later time TEAS traces can be

−

cm GSB feature is both strong and sharp. A second weak

−1

GSB band is observed near 1600 cm for both PC-O1 and

PC-O2, and a third such feature at 1508 cm⁻is only seen for

PC-O2. On time scales of a few nanoseconds, the GSB

recovery is incomplete for PC-O1 and PC-O2. Although this

ground-state recovery appears to be complete for PC-O3

within 2−3 ns, the extent is diﬃcult to quantify because of the

spectral overlap by the numerous ESA features. These

observations suggest ISC to the triplet manifold plays a role

decomposed into a broad S band and a T band centered at

1 1

415 nm (Figure S6 ). However, for PC-O2 in DCM or PC-O3

in all solvents, no new time-dependent ESA peaks are seen in

the TEAS spectra up to 1.3 ns. Instead, the changes involved

are more subtle and relate to evolution in the peak shapes

alone.

in S relaxation in PC-O1 and PC-O2, but it is less important

for PC-O3. Estimates of ISC quantum yields and rate

coeﬃcients are reported in Section 3.3.

Only two weak ESA bands at 1547 and 1575 cm are

observed for PC-O2 in DMF and DCM. In toluene-d , the

The TEAS spectra for PC-O3 are similar to those reported

previously, with the prior PC-O3 study being conducted in

−

11,29

DMAc;

the bands of photoexcited PC-O3 observed with

latter peak is obscured by a solvent band. For PC-O1, we ﬁnd a

centers at 440 and 625 nm are assigned to ESA from the S₁

CT-biphenyl state. Our observations of PC-O3 show rapid

growth of these ESA features, with the associated decay of a

shoulder to the 625 nm band located around 540 nm (Figure

S4 of the Supporting Information). These changes occur with

the time constant τ = 2.9 ± 0.1 ps in DMF. We attribute the

changing band shapes to vibrational cooling from the Franck−

broad and weak ESA feature lying between 1500 and 1575

−

cm , the assignment of which is addressed in detail

elsewhere. Only the higher wavenumber tail of this

absorption band is observed in DMF. For both PC-O1 and

PC-O2, we see evidence of a fast initial growth of S absorption

bands over a few picoseconds, as reported in Table 3, which we

attribute to a combination of internal conversion from the

Condon region of the directly photoexcited S state. The bands

622

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

Article

Table 5. Estimated Intersystem Crossing Quantum Yields, Φ_I_S_C, and Rate Coeﬃcients, k_I_S_C, Derived from Ground-State

a b

Bleach Recovery Kinetics Measured by TVAS

solvent

DCM

photocatalyst

S and T character

DMF

toluene-d₈

PC-N1

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

Φ_I_S_C

k_I_S_C/10 s

−

−1

PC-N2

PC-N3

PC-N4

PC-N5

PC-O1

PC-O2

PC-O3

PC-S

0.71 ± 0.03

0.42 ± 0.06

−

0.84 ± 0.21

4.0 ± 1.0

0.74 ± 0.09

0.9 ± 0.2

−

0.82 ± 0.16

2.7 ± 0.6

0.60 ± 0.06

1.2 ± 0.2

0.57 ± 0.06

1.0 ± 0.1

−

0.19 ± 0.01

0.68 ± 0.04

−

0.14 ± 0.02

0.08 ± 0.01

0.43 ± 0.05

2.9 ± 0.5

0.56 ± 0.03

0.9 ± 0.1

−

0.56 ± 0.03

2.3 ± 0.2

0.70 ± 0.03

2.9 ± 0.4

−

Φ_I_S_C

k_I_S_C/10 s

0.37 ± 0.07

0.6 ± 0.1

0.63 ± 0.03

1.7 ± 0.1

0.57 ± 0.05

3.2 ± 0.3

A blank entry indicates no evidence of T -state population, suggesting Φ ≈ 0. Uncertainties are from the ﬁtted errors in the amplitudes of the

components of multiple-exponential ﬁts to single kinetic data sets.

ISC

then decay with a ≥2.5 ns time constant that is consistent with

TVAS and TCSPC data (Table 3). In DCM and toluene

solutions, the TEAS measurements indicate similar changes in

Time constants extracted from kinetic ﬁts to TVAS data

(Table 3) show S relaxation with time constants of 5.9 ± 0.6,

3.8 ± 0.2, and 1.77 ± 0.08 ns in DMF, DCM, and toluene-d ,

the shapes of the S ESA bands over a few picoseconds

respectively. This interpretation is conﬁrmed by TCSPC

measurements of ﬂuorescence lifetimes (Table 3) and by a

prior report of a singlet-state lifetime for PC-S of 7.6 ns in 3-

(

although the evolution of the shoulder on the 625 nm band is

less pronounced in toluene) before decay with ∼2.4 ns time

constants, suggesting photodynamics similar to that in the

DMF solution. Prior to this decay, there may be a small

component of growth (or spectral band shift) with 250−500 ps

time constant, also observed by TVAS (Tables 3 and 4). Sartor

et al. reported an intersystem crossing quantum yield for PC-

methylpentane at 298 K.

The second, slower time

components observed in GSB recovery kinetics in all three

solvents indicate quenching of T -state population, most likely

accelerated by dissolved O in the samples.

The TEAS data for PC-S in DMF (Figure 5) show three

broad and overlapping peaks, with continued growth of some

components up to the longest measured delay time of 1.3 ns.

In DCM, the spectra are best described by a decaying exponent

with time constant 0.9 ± 0.1 ps and an intermediate with two

time constants (0.9 ± 0.1 and 21 ± 2 ps) for growth, which

then decays as a third Gaussian band centered at 460 nm

grows in with an ∼3.9 ns time constant. The kinetics of the

intermediate component are in reasonable agreement with the

O3 in DMAc of Φ = 0.11, involving singlet and triplet states

ISC

both with CT-Biphenyl character. Our spectroscopic data

show no clear evidence for a triplet intermediate state in the

three solvents studied.

.2.3. Phenothiazine Catalyst. PC-S is a phenothiazine

catalyst that possesses a puckered central ring. There are

conﬂicting views about the excited state(s) responsible for the

photoredox reactivity of phenothiazine PCs, with prior studies

pointing to either the short-lived singlet state or the much

4 ± 3 ps growth and 3.8 ± 0.2 ns decay time constant for PC-

S (S ) ESA bands observed by TVAS (Table 3), and this

3,48

longer lived triplet state.

At our chosen excitation

feature is therefore assigned to ESA from the S state. The

wavelength of 318 nm, it is likely that we excite the S state

initial population of S is by internal conversion from the

of PC-S (Table 1) with a weaker contribution from the S →

optically excited S states, and the band emerging at later times

S transition and perhaps also from the S → S transition

and centered near 460 nm is likely to correspond to ESA from

the T state of PC-S, populated by ISC from S . Similar

computed to lie at higher energy.

The TVAS spectra of PC-S (Figure 3), photoexcited at 318

behavior is observed for PC-S in toluene, with a T absorption

nm in DCM or toluene-d , show a strong GSB feature at 1463

band centered at 460 nm (Figure 5). This band grows as the

−

corresponding to depletion of the ground-state

broad intermediate S ESA band decays, albeit more quickly

−

population and two weak GSBs at 1573 and 1591 cm .

Only the last two depletions are observed in DMF, with the

ﬁrst being masked by strong solvent absorption. ESA bands are

than in DCM, with a time constant of 1.77 ns. Further analysis

of the kinetics of ISC in PC-S is presented in Section 3.3.

3.3. Triplet-State Branching. Where we have spectro-

evident for PC-S solutions in DCM and toluene-d near 1500

scopic evidence for triplet-state population in the various PCs

studied, either from distinct absorption bands in the TEAS

measurements or from biexponential kinetics for GSB

recovery, we deduce rate coeﬃcients for ISC and the triplet-

−

and 1530 cm , which decay completely to zero within a few

tens of nanoseconds. In DMF, only the higher energy ESA

feature can be resolved and decays to zero within 40 ns after

photoexcitation.

state branching from the S state in competition with relaxation

623

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

Article

to S . The T -state lifetime can be several hundred micro-

naphthyl, and N-ﬂuoronaphthyl substituents, with the

seconds. In such a case, we expect the PC ground-state

depletion to not recover completely on a nanosecond to

microsecond time scale. However, T₁populations are

quenched more rapidly by small amounts of dissolved O₂,

which happened to be persistent in our sample ﬂow system

even in solutions purged by N . Under these conditions, the T

corresponding k values in DMAc in the range (1.0−2.0)

ISC

−1

× 10 s . These values are comparable to those we report for

PC-S in Table 5, although the mechanism for ISC explored by

Sartor et al. requires CT states localized on core substituents

such as biphenyl groups and does not apply to PC-S, which has

no such core substituents. Our measurements across

dihydrophenazine, phenoxazine, and phenothiazine PCs

suggest that CT character is not a prerequisite for ISC in

these classes of compounds because we see triplet-state

population growth from S₁states of both LE and CT

character. The ISC rate coeﬃcients are comparable, and

lifetimes shortened to tens of nanoseconds or longer. Another

contributing factor toward partial or incomplete recovery of

the ground state could be a reaction from the photoexcited

state, for example by electron transfer to a solvent molecule,

and we see evidence for this in some of the experiments

conducted in DCM, as discussed earlier. Here, the observed

amplitudes of components of double-exponential or triple-

exponential ground-state recovery in the TVAS experiments

are used to guide determinations of an upper limit for the

branching into triplet states. For example, for a double-

exponential GSB recovery with amplitudes of the faster (S₁

decay) and slower (T decay) components denoted by A and

both these and Φ values are sensitive to the properties of the

ISC

solvent, indicating that tuning of excited-state energies by the

environment plays an important role.

3.4. Structural and Solvent Eﬀects on Excited-State

Lifetimes and Electronic Properties. In all of the

dihydrophenazine PCs studied, the S -state lifetimes are highly

sensitive to the electron-withdrawing or -donating nature of

the N-aryl substituents, as well as to the properties of the

solvent. The excited states of the phenoxazine PCs also show

sensitivity to these types of structural and environmental

changes. In DCM solutions, the highly reducing excited states

of these PCs facilitate electron transfer reactions with the

A , respectively, we estimate the T quantum yield from Φ

ISC

A /(A + A ). Uncertainties in Φ

are obtained by

ISC

appropriate propagation of the ﬁt uncertainties in A and A

values. Our analysis assumes that any dissolved O does not

inﬂuence the rates of ISC or S quenching to S . Table 5

summarizes the resulting quantum yields (Φ_I_S_C) and rate

solvent, but no such reactions shorten the S -state lifetimes in

coeﬃcients (k ) for ISC in the PC solutions studied. The

DMF and toluene. Instead, the diﬀerent dielectric constants of

the two solvents modify to greater or lesser extents the energies

of the CT and LE character excited states, as do the N-aryl

substituents. In the cases of PC-N3 and PC-N4, which have

electron-withdrawing N-aryl substituents and hence CT-

character S₁states, the eﬀect of the solvent is most

ISC

latter values are evaluated using k_I_S_C= Φ /τ(S ), where τ(S )

ISC

is the measured S -state lifetime (Tables 2 and 3).

Intersystem crossing is a minor pathway in the relaxation of

most of the photoexcited dihydrophenazine PCs studied,

particularly in DCM and DMF. For a PC-N5 solution in

DCM, we recently reported Φ = 0.13 ± 0.02, consistent

pronounced. In these two PCs, which are known to exert

ISC

with this general conclusion. The exception is toluene, where

three out of the ﬁve dihydrophenazine PCs (PC-N2, PC-N3,

and PC-N4) show signiﬁcant intersystem crossing eﬃciencies.

PC-N4 in toluene is an interesting case, as in this solvent the

good control over polymer dispersity in O-ATRP, the S

lifetimes in DMF (and DMAc) are less than 1 ns. The

likelihood is that the tuning in energy of the CT states by these

intramolecular and environmental inﬂuences makes eﬃcient

pathways for nonradiative S → S decay accessible via conical

greater S -state lifetime (5.31 ± 0.17 ns from TCSPC)

apparently allows ISC to compete with radiative decay to S . In

intersections between the electronic states. Further inves-

tigation of this supposition requires electronic structure

calculations which map the structures and energy dependences

of seams of intersection between the excited- and ground-state

potential energy surfaces.

contrast, despite PC-N1 having S₁lifetimes of several

nanoseconds in the three solvents (only shortened to ∼2 ns

in DCM by electron transfer reactions with the solvent), we

see no evidence for ISC, indicating weak spin−orbit coupling

and/or no accessible regions of near-degeneracy between the

singlet and triplet manifolds. Higher triplet-state quantum

yields are observed for the phenoxazine and phenothiazine PCs

Changes to S -state lifetimes have consequences for the

quantum yields for ISC, because short-lived S states favor

relaxation within the singlet spin manifold over spin-changing

dynamics. From a combination of the experimental evidence

reported here for dihydrophenazine PC-N1, PC-N3, and PC-

N4 photocatalysts and experimental polymerization studies

using these PCs, we suggest that population via ISC of a long-

lived triplet state of the optically excited PC is not an essential

prerequisite for eﬀective ATRP control of polymer dispersity.

These and the other dihydrophenazine compounds show little

or no propensity for ISC (Table 5), with the exception of PC-

N2, which was reported to be an inferior PC for photoredox-

studied, which may reﬂect the longer S lifetimes of these

compounds, as well as the inﬂuence of a larger spin−orbit

interaction in the sulfur-containing phenothiazine.

Sartor et al. recently reported a time-resolved spectroscopy

study of four N-aryl phenothiazines with biphenyl core

substituents. They showed the importance of singlet states

with CT character involving the nearly orthogonal N-aryl

group in promoting ISC to a triplet state characterized by

partial charge transfer to one biphenyl group. The large change

in orbital angular momentum associated with the change in

electron density between these nearly orthogonal molecular

orbitals of CT character enhances the spin−orbit interac-

catalyzed control of polymer dispersity in O-ATRP. PC-N3

shows solvent-dependent ISC quantum yields and rate

coeﬃcients, but its S lifetime in DMF is short enough to

9,50

tion.

ISC quantum yields therefore depended sensitively

outcompete ISC. Hence, PC-N3(T ) is unlikely to play a

on the electron-withdrawing character of the N-aryl substituent

and, by implication, the choice of solvent because of the

inﬂuence of solvent polarity on the energies of the CT states.

For measurements made in DMAc, their deduced quantum

yields Φ_I_S_Cwere in the range 0.4−1.0 for N-phenyl, N-

signiﬁcant role in radical generation for O-ATRP synthesis in

solvents such as DMF and DMAc.

Several photocatalysts based on a phenoxazine core have

been reported to show strong solvent dependence in their

ﬂuorescence spectra, indicating CT character of the emissive S₁

624

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

https://pubs.acs.org/doi/10.1021/jacs.1c00279.

Article

state. Our TVAS and TCSPC data (Table 3) reveal not only

nonradiative relaxation pathways via conical intersections. We

posit that detailed quantum chemical calculations can help

ascertain the state-speciﬁc relaxation mechanisms in the future.

singlet lifetimes of a few nanoseconds in the three solvents

studied but also a slower component to the ground-state

recovery in two of the chosen phenoxazine PCs, indicating ISC

to the triplet manifold of states. Reduction by electron transfer

from either the S or T state may therefore contribute to

photoredox-catalyzed O-ATRP activity in these two cases.

Instead of focusing on ISC propensities, we propose that the

Gibbs energy change for the electron transfer reaction is an

important metric for assessing the eﬃcacy of these PCs in O-

ATRP control, as outlined in Marcus theory and suggested in

The S -state lifetimes range from sub-nanoseconds to a few

tens of nanoseconds for all the OPCs studied. ISC is found to

be a minor relaxation pathway for PCs with short-lived S₁

states. In these cases, bimolecular, excited-state electron

transfer reactions must compete with the (ultra)fast intra-

molecular energy relaxation time scale if the PCs are to play an

eﬀective role in O-ATRP. We propose that rapid loss of the

excited-state population may be partially compensated for by

the high reduction potentials of the excited singlet states,

which minimize activation barriers for bimolecular electron-

7,47

earlier work.

This argument is developed further in a

companion paper in which supporting evidence is presented

from direct measurements of intermolecular electron transfer

for the nine PCs studied here. However, we note here that

whether electron transfer occurs from the S or T state will

transfer reactions. Crucially, the T state need not be the sole

electronic state responsible for electron transfer, as is often

assumed in O-ATRP studies.

inﬂuence the Gibbs energy changes and hence ET rate

coeﬃcients, and the results presented in this article unravel the

PC structure and solvent-dependent propensities for S₁

Evidence for competition between electron transfer from the

S and T states to a radical initiator is presented in a

companion paper, in which bimolecular electron transfer rate

coeﬃcients are reported for the nine dihydrophenazine,

phenoxazine, and phenothiazine PCs studied here. The

population to undergo ISC. The S lifetime also appears to

be an important factor for consideration, with consequences

for both the propensity for diﬀusive intermolecular electron

transfer and ISC quantum yields. In several of the PCs studied

here, and more so for the dihydrophenazine derivatives, S₁-

state lifetimes are controlled by nonradiative and radiative

relaxation pathways to S , not by ISC dynamics to populate T .

observed trends are discussed in terms of Marcus−Saveant

theory, and they indicate that both static and dynamic

(diﬀusional) rate coeﬃcients for the photoinduced electron

transfer will be lower from the T than from the S state

because of the smaller excited-state reduction potentials. In

general, a combination of (i) short, ﬁrst-excited-singlet-state

lifetimes, (ii) low propensities for intersystem crossing, and

(iii) slower photoinduced electron transfer rates are expected

to maintain a low steady-state concentration of radical

intermediates, thereby exercising a better control on polymer

dispersity in O-ATRP reactions. Our spectroscopic data with

sub-picosecond time resolution underline the importance of

considering both S - and T -state reduction potentials in

The sensitivity of the excited-state dynamics of the various

PCs to the choice of solvent raises interesting questions about

the eﬀects of the environment on PC performance in O-ATRP.

Under the types of conditions used for polymer growth, the

solvent is typically DMAc or DMF, but the solution also

includes a signiﬁcant fraction of a lower polarity monomer. As

our solvent-dependent measurements reported in Tables 2−5

show, the local solvation environment of a photoexcited PC in

these mixed solutions will aﬀect its S -state lifetime and ISC

yield.

assessing the usefulness of organic photocatalysts for ATRP.

These results will further guide the design principles and

applications of these general and widespread classes of PC

35,51,52

. CONCLUSIONS

compounds.

Study of the femtosecond to nanosecond time scale photo-

physical dynamics of organic molecular photocatalysts is

essential to identify the electronic states that underlie their

useful photocatalytic properties. We report the lifetimes and

relaxation pathways of the excited electronic states of nine

representative molecular photocatalysts for O-ATRP. Time-

resolved spectroscopic measurements performed with sub-

picosecond time resolution in three solvents (DMF, DCM, and

toluene) directly explore the states involved in the photoredox

behavior of these compounds. In conjunction with quantum-

chemical calculations and global kinetic ﬁtting, the observed

spectroscopic features reveal the electronic states populated

upon photoexcitation and internal conversion, the eﬃcacies for

triplet-state branching and ground-state recovery, and their

dependence on solvent. The results are consolidated with our

previous publications on this topic to outline general

ASSOCIATED CONTENT

Supporting Information

The Supporting Information is available free of charge at

■

sı

Synthetic methods for preparation of the PCs, selected

physical properties of the solvents, steady-state ﬂuo-

rescence and FTIR spectra of the organic PCs, an

analysis of transient electronic absorption spectra of the

organic PCs, and time-correlated single photon counting

(TCSPC) data (PDF)

AUTHOR INFORMATION

■

Corresponding Authors

17,28,29

photophysical principles.

Aditi Bhattacherjee − School of Chemistry, University of

Bristol, Bristol BS8 1TS, U.K.; orcid.org/0000-0001-

7146-1128 ; Email: aditi-bhattacherjee@uiowa.edu

Mahima Sneha − School of Chemistry, University of Bristol,

Bristol BS8 1TS, U.K.; orcid.org/0000-0003-4896-6915;

Email: mahima.sneha@bristol.ac.uk

For solutions of dihydrophenazine PCs in DMF, catalysts

possessing CT excited-state character have relatively short S₁-

state lifetimes (e.g., 136 ps for PC-N3 and 500 ps for PC-N4),

whereas those with LE character have longer S lifetimes (e.g.,

1 ns for PC-N1 and 17 ns for PC-N2). The experimental

results point toward polar solvents shortening the excited-state

lifetimes, most likely by stabilizing the CT-state energies

relative to the corresponding ground states and opening

Andrew J. Orr-Ewing − School of Chemistry, University of

Bristol, Bristol BS8 1TS, U.K.; orcid.org/0000-0001-

5551-9609 ; Email: a.orr-ewing@bristol.ac.uk

625

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

J.; Clark, P. G.; Read de Alaniz, J.; Hawker, C. J. Simple Benchtop

Article

Authors

precious-metal-based photoredox catalysts. Aldrichimica Acta 2019,

2, 7−21.

9) Discekici, E. H.; Pester, C. W.; Treat, N. J.; Lawrence, J.;

(

Luke Lewis-Borrell − School of Chemistry, University of

Bristol, Bristol BS8 1TS, U.K.; orcid.org/0000-0002-

Mattson, K. M.; Narupai, B.; Toumayan, E. P.; Luo, Y.; McGrath, A.

Approach to Polymer Brush Nanostructures Using Visible-Light-

Mediated Metal-Free Atom Transfer Radical Polymerization. ACS

Macro Lett. 2016, 5, 258−262.

198-722X

Giordano Amoruso − School of Chemistry, University of

Bristol, Bristol BS8 1TS, U.K.; orcid.org/0000-0003-

244-4804

Thomas A.A. Oliver − School of Chemistry, University of

Bristol, Bristol BS8 1TS, U.K.; orcid.org/0000-0003-

(10) Sartor, S. M.; Chrisman, C. H.; Pearson, R. M.; Miyake, G. M.;

Damrauer, N. H. Designing High-Triplet-Yield Phenothiazine

Donor −Acceptor Complexes for Photoredox Catalysis. J. Phys.

Chem. A 2020, 124, 817−823.

979-7857

Jasper Tyler − School of Chemistry, University of Bristol,

Bristol BS8 1TS, U.K.; orcid.org/0000-0002-3155-1847

Ian P. Clark − Central Laser Facility, Research Complex at

Harwell, Science and Technology Facilities Council,

(11) Sartor, S. M.; McCarthy, B. G.; Pearson, R. M.; Miyake, G. M.;

Damrauer, N. H. Exploiting Charge-Transfer States for Maximizing

Intersystem Crossing Yields in Organic Photoredox Catalysts. J. Am.

Chem. Soc. 2018, 140, 4778−4781.

Rutherford Appleton Laboratory, Didcot, Oxfordshire OX11

(12) Prier, C. K.; Rankic, D. A.; MacMillan, D. W. C. Visible Light

QX, U.K.

Photoredox Catalysis with Transition Metal Complexes: Applications

in Organic Synthesis. Chem. Rev. 2013 , 113 , 5322 −5363.

(13) Jockusch, S.; Yagci, Y. The Active Role of Excited States of

Phenothiazines in Photoinduced Metal Free Atom Transfer Radical

Polymerization: Singlet or Triplet Excited States? Polym. Chem. 2016,

7, 6039−6043.

Complete contact information is available at:

https://pubs.acs.org/10.1021/jacs.1c00279

Author Contributions

∥

A.B., M.S., and L.L.-B. contributed equally.

(14) Bressler, C.; Milne, C.; Pham, V.-T.; ElNahhas, A.; van der

Notes

Veen, R. M.; Gawelda, W.; Johnson, S.; Beaud, P.; Grolimund, D.;

Kaiser, M.; Borca, C. N.; Ingold, G.; Abela, R.; Chergui, M.

Femtosecond XANES Study of the Light-Induced Spin Crossover

Dynamics in an Iron(II) Complex. Science 2009 , 323 , 489 −492.

(15) Auböck, G.; Chergui, M. Sub-50-fs Photoinduced Spin

Crossover in [Fe(bpy)3]2+. Nat. Chem. 2015, 7, 629−633.

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

This work was supported by EPSRC grant EP/R012695/1.

The ultrafast laser laboratory at the University of Bristol was

established with funding from ERC Advanced Grant CAPRI

(16) Monni, R.; Capano, G.; Aubock, G.; Gray, H. B.; Vlcek, A.;

Tavernelli, I.; Chergui, M. Vibrational Coherence Transfer in the

Ultrafast Intersystem Crossing of a Diplatinum Complex in Solution.

Proc. Natl. Acad. Sci. U. S. A. 2018, 115, E6396−E6403.

90966. M.S. is supported by Marie Skłodowska-Curie

fellowship MARCUS 793799. L.L.-B and J.T. thank the Bristol

Chemical Synthesis Centre for Doctoral Training, funded by

the EPSRC (EP/L015366/1), and the University of Bristol, for

Ph.D. studentships. G.A. thanks the EPSRC for Ph.D.

studentship funding through EP/N509619/1. T.A.A.O.

acknowledges the Royal Society University Research Fellow-

ships UF1402310 and URF\R\201007. Data are available at

the University of Bristol data repository, data.bris at 10.5523/

bris.jud6zf0aehpn2pjwv1klmba58.

(17) Koyama, D.; Dale, H. J. A.; Orr-Ewing, A. J. Ultrafast

Observation of a Photoredox Reaction Mechanism: Photoinitiation in

Organocatalyzed Atom-Transfer Radical Polymerization. J. Am. Chem.

Soc. 2018, 140, 1285−1293.

(18) Noto, N.; Takahashi, K.; Goryo, S.; Takakado, A.; Iwata, K.;

Koike, T.; Akita, M. Laser Flash Photolysis Studies on Radical

Monofluoromethylation by (Diarylamino)naphthalene Photoredox

Catalysis: Long Lifetime of the Excited State is Not Always a

Requisite. J. Org. Chem. 2020, 85, 13220−13227.

(19) Ryan, M. D.; Theriot, J. C.; Lim, C.-H.; Yang, H.; Lockwood,

A. G.; Garrison, N. G.; Lincoln, S. R.; Musgrave, C. B.; Miyake, G. M.

Solvent Effects on the Intramolecular Charge Transfer Character of

N,N-Diaryl Dihydrophenazine Catalysts for Organocatalyzed Atom

Transfer Radical Polymerization. J. Polym. Sci., Part A: Polym. Chem.

REFERENCES

■

(

1) Matyjaszewski, K.; Xia, J. Atom Transfer Radical Polymerization.

Chem. Rev. 2001, 101, 2921−2990.

(

017, 55, 3017−3027.

(

2) Romero, N. A.; Nicewicz, D. A. Organic Photoredox Catalysis.

Chem. Rev. 2016, 116, 10075−10166.

3) Treat, N. J.; Sprafke, H.; Kramer, J. W.; Clark, P. G.; Barton, B.

20) Lim, C.-H.; Ryan, M. D.; McCarthy, B. G.; Theriot, J. C.;

Sartor, S. M.; Damrauer, N. H.; Musgrave, C. B.; Miyake, G. M.

Intramolecular Charge Transfer and Ion Pairing in N,N-Diaryl

Dihydrophenazine Photoredox Catalysts for Efficient Organocata-

lyzed Atom Transfer Radical Polymerization. J. Am. Chem. Soc. 2017,

139, 348−355.

(21) Pearson, R. M.; Lim, C.-H.; McCarthy, B. G.; Musgrave, C. B.;

Miyake, G. M. Organocatalyzed Atom Transfer Radical Polymer-

ization Using N-Aryl Phenoxazines as Photoredox Catalysts. J. Am.

Chem. Soc. 2016, 138, 11399−11407.

(22) Dadashi-Silab, S.; Pan, X.; Matyjaszewski, K. Phenyl Benzo-

[b]phenothiazine as a Visible Light Photoredox Catalyst for Metal-

Free Atom Transfer Radical Polymerization. Chem. - Eur. J. 2017, 23,

5972−5977.

(23) Miyake, G. M.; Theriot, J. C. Perylene as an Organic

Photocatalyst for the Radical Polymerization of Functionalized

Vinyl Monomers through Oxidative Quenching with Alkyl Bromides

and Visible Light. Macromolecules 2014, 47, 8255−8261.

(24) Sartor, S. M.; Lattke, Y. M.; McCarthy, B. G.; Miyake, G. M.;

Damrauer, N. H. Effects of Naphthyl Connectivity on the Photo-

(

E.; de Alaniz, J. R.; Fors, B. P.; Hawker, C. J. Metal-Free Atom

Transfer Radical Polymerization. J. Am. Chem. Soc. 2014, 136,

4) Chen, M.; Zhong, M.; Johnson, J. A. Light-Controlled Radical

6096−16101.

Polymerization: Mechanisms, Methods, and Applications. Chem. Rev.

016, 116, 10167−10211.

5) Corrigan, N.; Shanmugam, S.; Xu, J.; Boyer, C. Photocatalysis in

Organic and Polymer Synthesis. Chem. Soc. Rev. 2016, 45, 6165−

212.

6) Theriot, J. C.; Lim, C.-H.; Yang, H.; Ryan, M. D.; Musgrave, C.

B.; Miyake, G. M. Organocatalyzed Atom Transfer Radical Polymer-

ization Driven by Visible Light. Science 2016, 352, 1082−1086.

7) McCarthy, B. G.; Pearson, R. M.; Lim, C.-H.; Sartor, S. M.;

(

Damrauer, N. H.; Miyake, G. M. Structure −Property Relationships

for Tailoring Phenoxazines as Reducing Photoredox Catalysts. J. Am.

Chem. Soc. 2018, 140, 5088−5101.

(

8) Corbin, D. A.; Lim, C.-H.; Miyake, G. M. Phenothiazines,

dihydrophenazines, and phenoxazines: sustainable alternatives to

626

J. Am. Chem. Soc. 2021, 143, 3613−3627

Journal of the American Chemical Society

physics of Compact Organic Charge-Transfer Photoredox Catalysts. J.

Article

Phys. Chem. A 2019, 123, 4727−4736.

(38) Miertus, S.; Scrocco, E.; Tomasi, J. Electrostatic Interaction of a

Solute with a Continuum. A Direct Utilizaion of Ab Initio Molecular

Potentials for the Prevision of Solvent Effects. Chem. Phys. 1981, 55,

117−129.

(25) Wang, J.-S.; Matyjaszewski, K. Controlled/″Living ″ Radical

Polymerization. Atom Transfer Radical Polymerization in the

Presence of Transition-Metal Complexes. J. Am. Chem. Soc. 1995,

(39) Tomasi, J.; Mennucci, B.; Cammi, R. Quantum Mechanical

Continuum Solvation Models. Chem. Rev. 2005 , 105 , 2999 −3094.

(40) Goerigk, L.; Grimme, S. Efficient and Accurate Double-Hybrid-

Meta-GGA Density Functionals Evaluation with the Extended

GMTKN30 Database for General Main Group Thermochemistry,

Kinetics, and Noncovalent Interactions. J. Chem. Theory Comput.

(

17, 5614−5615.

26) Rolland, M.; Whitfield, R.; Messmer, D.; Parkatzidis, K.;

Truong, N. P.; Anastasaki, A. Effect of Polymerization Components

on Oxygen-Tolerant Photo-ATRP. ACS Macro Lett. 2019, 8, 1546−

(

551.

41) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and

011, 7, 291−309.

27) Martinez, M. R.; Sobieski, J.; Lorandi, F.; Fantin, M.; Dadashi-

Silab, S.; Xie, G.; Olszewski, M.; Pan, X.; Ribelli, T. G.; Matyjaszewski,

K. Understanding the Relationship between Catalytic Activity and

Termination in PhotoATRP: Synthesis of Linear and Bottlebrush

Polyacrylates. Macromolecules 2020, 53, 59−67.

Accurate Ab Initio Parametrization of Density Functional Dispersion

Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010,

32, 154104.

42) Yanai, T.; Tew, D. P.; Handy, N. C. A New Hybrid Exchange −

Correlation Functional using the Coulomb-Attenuating Method

CAM-B3LYP). Chem. Phys. Lett. 2004, 393, 51−57.

43) Peach, M. J. G.; Benfield, P.; Helgaker, T.; Tozer, D. J.

(28) Lewis-Borrell, L.; Sneha, M.; Bhattacherjee, A.; Clark, I. P.; Orr-

Ewing, A. J. Mapping the Multi-Step Mechanism of a Photoredox

Catalyzed Atom-Transfer Radical Polymerization Reaction by Direct

Observation of the Reactive Intermediates. Chem. Sci. 2020, 11,

Excitation Energies in Density Functional Theory: An Evaluation and

a Diagnostic Test. J. Chem. Phys. 2008, 128, 044118.

(

475−4481.

29) Sneha, M.; Lewis-Borrell, L. J.; Shchepanovska, D.;

(44) Chaudhuri, S.; Rudshteyn, B.; Prémont-Schwarz, M.; Pines, D.;

Bhattacherjee, A.; Tyler, J. L.; Orr-Ewing, A. J. Solvent-Dependent

Photochemical Dynamics of a Phenoxazine-Based Photoredox

Catalyst. Z. Phys. Chem. 2020, 234, 1475−1494.

Pines, E.; Huppert, D.; Nibbering, E. T. J.; Batista, V. S. Ultrafast

Photo-Induced Charge Transfer of 1-Naphthol and 2-Naphthol to

Halocarbon Solvents. Chem. Phys. Lett. 2017, 683, 49−56.

(30) Pan, X.; Lamson, M.; Yan, J.; Matyjaszewski, K. Photoinduced

(45) Grubb, M. P.; Orr-Ewing, A. J.; Ashfold, M. N. R. KOALA: a

Metal-Free Atom Transfer Radical Polymerization of Acrylonitrile.

Program for the Processing and Decomposition of Transient Spectra.

Rev. Sci. Instrum. 2014, 85, 064104.

ACS Macro Lett. 2015, 4, 192−196.

(31) Bhattacherjee, A.; Sneha, M.; Lewis-Borrell, L.; Tau, O.; Clark,

(46) Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. Modern Molecular

I. P.; Orr-Ewing, A. J. Picosecond to Millisecond Tracking of a

Photocatalytic Decarboxylation Reaction Provides Direct Mechanistic

insights. Nat. Commun. 2019, 10, 5152.

Photochemistry of Organic Molecules; University Science Books: Mill

Valley, CA, 2010.

(47) Pan, X.; Fang, C.; Fantin, M.; Malhotra, N.; So, W. Y.; Peteanu,

(32) Roberts, G. M.; Marroux, H. J. B.; Grubb, M. P.; Ashfold, M. N.

L. A.; Isse, A. A.; Gennaro, A.; Liu, P.; Matyjaszewski, K. Mechanism

of Photoinduced Metal-Free Atom Transfer Radical Polymerization:

Experimental and Computational Studies. J. Am. Chem. Soc. 2016,

R.; Orr-Ewing, A. J. On the Participation of Photoinduced N-H Bond

Fission in Aqueous Adenine at 266 and 220 nm: a Combined

Ultrafast Transient Electronic and Vibrational Absorption Spectros-

copy Study. J. Phys. Chem. A 2014, 118, 11211−11225.

48) Theriot, J. C.; McCarthy, B. G.; Lim, C.-H.; Miyake, G. M.

38, 2411−2425.

(33) Greetham, G. M.; Donaldson, P. M.; Nation, C.; Sazanovich, I.

Organocatalyzed Atom Transfer Radical Polymerization: Perspectives

on Catalyst Design and Performance. Macromol. Rapid Commun.

2017, 38, 1700040.

V.; Clark, I. P.; Shaw, D. J.; Parker, A. W.; Towrie, M. A 100 kHz

Time-Resolved Multiple-Probe Femtosecond to Second Infrared

Absorption Spectrometer. Appl. Spectrosc. 2016, 70, 645 −653.

(49) Benniston, A. C.; Harriman, A.; Li, P.; Rostron, J. P.; van

Ramesdonk, H. J.; Groeneveld, M. M.; Zhang, H.; Verhoeven, J. W.

Charge Shift and Triplet State Formation in the 9-Mesityl-10-

methylacridinium Cation. J. Am. Chem. Soc. 2005, 127, 16054−16064.

(50) Dance, Z. E. X.; Mickley, S. M.; Wilson, T. M.; Ricks, A. B.;

Scott, A. M.; Ratner, M. A.; Wasielewski, M. R. Intersystem Crossing

Mediated by Photoinduced Intramolecular Charge Transfer:

Julolidine −Anthracene Molecules with Perpendicular π Systems. J.

Phys. Chem. A 2008, 112, 4194−4201.

(34) Koyama, D.; Donaldson, P. M.; Orr-Ewing, A. J. Femtosecond

to Microsecond Observation of the Photochemical Reaction of 1,2-

Di(quinolin-2-yl)disulphide with Methyl Methacrylate. Phys. Chem.

Chem. Phys. 2017, 19, 12981 −12991.

(35) Orr-Ewing, A. J. How can Ultrafast Laser Spectroscopy Inform

the Design of New Organic Photoredox Catalysts for Chemical and

Materials Synthesis? Struct. Dyn. 2019, 6, 010901.

(36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;

Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,

B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.

P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;

Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.;

Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.;

Montgomery, J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.;

Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi,

R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar,

S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox,

J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;

Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;

Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.;

Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A.

D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.

Gaussian 09; Gaussian Inc.: Wallingford, CT, 2009.

(51) Enciso, A. E.; Fu, L.; Russell, A. J.; Matyjaszewski, K. A

Breathing Atom-Transfer Radical Polymerization: Fully Oxygen-

Tolerant Polymerization Inspired by Aerobic Respiration of Cells.

Angew. Chem., Int. Ed. 2018, 57, 933−936.

(37) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal

52) Buzzetti, L.; Crisenza, G. E. M.; Melchiorre, P. Mechanistic

Studies in Photocatalysis. Angew. Chem., Int. Ed. 2019, 58, 3730−

747.

Solvation Model Based on Solute Electron Density and on a

Continuum Model of the Solvent Defined by the Bulk Dielectric

Constant and Atomic Surface Tensions. J. Phys. Chem. B 2009, 113,

6378−6396.

627