Photoredox Catalysis Using Heterogenized Iridium Complexes**

Article abstract of DOI:10.1002/chem.202101651

Heterogenized photoredox catalysts provide a path for sustainable chemical synthesis using highly tunable, reusable constructs. Here, heterogenized iridium complexes as photoredox catalysts were assembled via covalent attachment to metal oxide surfaces (I

Full text of DOI:10.1002/chem.202101651

Accepted Article
Accepted Article
Title: Photoredox Catalysis Using Heterogenized Iridium Complexes
Authors: Kelly L. Materna and Leif Hammarström
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202101651
Link to VoR: https://doi.org/10.1002/chem.202101651
0
1/2020

Chemistry - A European Journal
10.1002/chem.202101651
Photoredox Catalysis Using Heterogenized Iridium Complexes
Dr. Kelly L. Materna*, Prof. Leif Hammarström
Department of Chemistry-Ångström Laboratories, Uppsala University, Box 523, SE75120
Uppsala, Sweden.
*
Corresponding author: kelly.materna@kemi.uu.se
Abstract: Heterogenized photoredox catalysts provide a path for sustainable chemical
synthesis using highly tunable, reusable constructs. Here, heterogenized iridium complexes as
photoredox catalysts were assembled via covalent attachment to metal oxide surfaces (ITO,
ZrO
2
, Al
2
3
O ) in thin film or nanopowder constructs. The goal was to understand which materials
provided the most promising constructs for catalysis. To do this, reductive dehalogenation of
bromoacetophenone to acetophenone was studied as a test reaction for system optimization. All
catalyst constructs produced acetophenone with high conversions and yields with the fastest
reactions complete in fifteen minutes using Al
2
3
O supports. The nanopowder catalysts resulted
in faster and more efficient catalysis, while the thin film catalysts were more robust and easily
reused. Importantly, the thin film constructs show promise for future photoelectrochemical and
electrochemical photoredox setups. Finally, all catalysts were reusable 2-3 times, performing
at least 1000 turnovers (Al
2
3
O ), demonstrating that heterogenized catalysts are a sustainable
catalyst alternative.
Keywords: heterogenized; iridium; photocatalysis; photoredox catalysis; sustainable
chemistry
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Introduction:
Due to the active occurrence of climate change and its impending consequences,
the world today requires renewable energy sources to drive components of daily life.[1] Much
progress has been made in the solar fuels and photovoltaics communities, where sunlight is
used to generate electricity using solar cells, and produce renewable fuels, such as H , using
2
[2]
photoelectrochemical cells. For example, water-splitting dye sensitized photoelectrochemical
cells (WS-DSPEC) incorporate molecular photosensitizers to harvest sunlight and perform
charge separation and use molecular catalysts to drive fuel production; these molecular
components are covalently attached to semiconducting metal oxide surfaces. The molecular
nature of the components in WS-DSPECs makes them highly tunable, while the metal oxide
surface provides a robust platform for photoelectrocatalysis. However, because the devices are
complex, with multiple molecular components and a semiconducting surface that participates
during catalysis, fuel production can be challenging as multiple charges must be accumulated
for the catalyst to turnover. [3]
Complementary to this, the photoredox catalysis field has demonstrated that a
variety of chemicals (natural products, pharmaceuticals, feedstock chemicals, etc.) can be
produced in an environmentally friendly fashion under mild conditions using common
[4]
household light bulbs. Upon illumination, molecular photoredox catalysts generate long-lived
excited states (ns-µs), which are able to react with organic substrates in a reaction mixture.
Typically in photoredox catalysis, catalysts are homogenous, molecular species such as
ruthenium, iridium, or copper coordination complexes; these inorganic coordination complexes
undergo a metal to ligand charge transfer (MLCT) and intersystem crossing (ISC) upon
excitation, producing long-lived excited states, which are simultaneously highly reducing and
oxidizing in nature (Scheme 1).[4e, 4f, 5] Given this duality of the excited state, the complexes can
act as effective photoredox catalysts in a library of organic transformations.
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
III
]+ with light, where ppy=2-
Scheme 1. Example reaction of photoredox catalyst [Ir (bpy)(ppy)
2
phenylpyridine, and bpy=2,2′-bipyridine.[6] The catalyst undergoes a metal to ligand charge
transfer (MLCT) and intersystem crossing (ISC) to form the charge separated excited state
species. Excited state potentials displayed in the scheme are based on literature data collected
in acetonitrile under argon.[6]
Furthermore, homogeneous molecular photoredox catalysts have the advantage
of being highly tunable due to their molecular nature, allowing a variety of organic
transformations to be studied. However, as they are usually in the same phase as the organic
substrate and product, a purification step is needed to separate them from the reaction mixture.
To make the catalyst easier to separate from the reaction mixture, heterogeneous solid-state
materials have also been used as photoredox catalysts.[4a] These catalysts are both robust and
easily reusable catalysts; although, they are less tunable than molecular, homogeneous
catalysts.
To assemble a catalyst that has the advantages of both a homogeneous and
heterogeneous one, inspiration can be taken from the solar fuels field, by designing a
photoredox catalyst that is molecular in nature, but covalently bound to a solid-state support,
called a heterogenized catalyst (Figure 1).[2e, 2g, 7] To covalently attach the molecular catalyst to
the solid-state support, a surface anchor is used, which is incorporated into the molecular
catalyst structure. Upon immobilization, this reacts with a metal oxide (MOx) solid-state
support to form a covalent bond; this method is well-established in the solar fuels field and dye
[8]
sensitized solar cell (DSSC) literature. By assembling a catalyst in this fashion, we gain the
tunability of a molecular catalyst and also the reusability of heterogeneous catalyst to design
the ultimate environmentally-friendly photoredox catalyst.
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Figure 1. Differences between homogeneous, heterogeneous, and heterogenized catalysts.
In the photoredox catalysis field, this idea has not been extensively explored. A
few examples exist using derivatives of tris(2,2´-bipyridine)-ruthenium coordination complexes
[9]
[10]
[11]
immobilized onto silica , glass wool , or polymer supports. Other unique examples have
also been tested such as Rose Bengal[12] dye or perylene diimide (PDI)[13] molecules
immobilized on silica, and porphyrins immobilized onto cotton threads[14]. Heterogenized
iridium photoredox catalysts are even less explored, with only a few examples existing where
polypyridyl-based iridium complexes have been incorporated into polymer supports[15]; note
that in polymer supports, iridium leaching can be an issue leading to catalyst instability.[15b] All
of these studies demonstrate a promising future for heterogenized photoredox catalysts since
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
they provide tunable and reusable catalysts; however, there lacks an organized set of guidelines
for how to design a heterogenized photoredox catalyst.
To address this issue, we will perform a systematic study to analyze the effect of
the catalyst support in both composition and architecture. To do this, polypyridyl iridium
coordination complexes will be used and immobilized onto three different metal oxide supports
in the form of nanopowders or thin films. The results here will provide a strong foundation to
the heterogenized photoredox catalysis field by identifying what materials are optimal when
using these catalysts. The overall goal is to develop a heterogenized catalyst that is highly
functional, easily reusable, tunable, and useful in a variety of photoredox applications. In this
study, photoredox catalysis for the reductive dehalogenation of 2-bromoacetophenone (BrAPN)
to acetophenone (APN) was chosen as the model reaction to analyze the effect of the photoredox
catalyst´s support, in both content (composition of the metal oxide) and in architecture (thin
film or nanopowder). This reaction is advantageous as it is simple, well-known, and is easy to
follow using H NMR spectroscopy, making it an ideal reaction for initial catalytic studies.[16]
Furthermore, the product APN is non-toxic and highly applicable in the perfume industry.
Ultimately, we aim to generate an initial understanding of what materials and assembly methods
provide the most functional and robust heterogenized photoredox catalysts by performing a
baseline, systematic study. Once this foundation is established, additional reaction and substrate
scopes can be analyzed in future work to test for increased applicability.
1
Results and Discussion
Catalyst design and rationale. To design the iridium catalyst, we were inspired
by the varieties of tris(2-phenylpyridine)iridium and bis(2-phenylpyridine)(2,2’-
bipyridine)iridium coordination complexes that are widely used in homogeneous photoredox
catalysis.[4c-f, 6, 17] To be able to bind the complex to a metal oxide surface, carboxylic acid
surface anchors were incorporated in the catalyst structure (Ir, Figure 2) by replacing a 2,2′-
bipyridine ligand of the latter complex for a 2,2′-bipyridine-4,4′-dicarboxylic acid ligand.
Carboxylic acids are commonly used surface anchors in the solar fuels and dye sensitized solar
cell (DSSC) communities, community, as they can form covalent surface bonds to a metal oxide
surface.[8] In addition, the photophysical properties of Ir has been previously studied.[18]
Absorption features in the visible region and a long excited state lifetime (69.5 ns in acetonitrile)
make it a good candidate for photoredox catalysis.
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Figure 2. Structure of Ir (left complex) and proposed structure of heterogenized Ir (MO -Ir).
x
Preparation method for catalysts is depicted. The metal oxide nanopowders or thin films are
soaked in an Ir/methanol solution overnight in the dark to provide the heterogenized catalysts
(
MOx-Ir). Pictures of the nanopowder catalysts are shown in the top right and thin films in the
bottom right.
Ir synthesis and characterization. To prepare the heterogenized catalysts, Ir
was synthesized via a previously reported procedure.[18] Briefly, Ir was synthesized via reaction
of [Ir(ppy) (µ-Cl)] with 2,2′-bipyridine-4,4′-dicarboxylic acid in dichloromethane followed by
workup with ammonium hexafluorophosphate, NH
2
2
4
PF
6
(more details in the Supporting
1
Information). Ir was characterized by proton nuclear magnetic resonance ( H NMR), UV-Vis,
photoluminescence, and attenuated total reflectance Fourier transform infrared (ATR-FTIR)
spectroscopic measurements supporting that Ir was successfully synthesized (see Supporting
Information). The UV-Vis spectrum of Ir shows a MLCT band at 370 nm and a broad band at
4
97 nm (Figure S10). ATR-FTIR confirms the presence of the carboxylic acid anchoring group,
-1
having a signature peak at 1709 cm , suggesting the presence of a C=O group (Figure S19).
The photoluminescence spectrum after excitation at 370 nm shows a band with maximum at
6
35 nm, typical for this class of complexes (Figure S36).[18-19]
Metal oxide rationale. Three different metal oxide (MOx) supports were chosen
to be examined: Al
2
O
3
, ZrO
2
, and ITO. Al
2
O
3
and ZrO
2
are wide band gap semiconductors,
while ITO is a conductive metal oxide. Since Al
2
O
3
(8.45-9.9 eV) and ZrO
2
(5 eV) have wide
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
band gaps and high energy conduction bands,[20] and Ir´s redox potentials lie far from the
valence and conduction band potentials, Ir should not be able to inject charges into the metal
oxide upon excitation. As a result, no redox events should occur between the catalyst and metal
oxide. Therefore, the metal oxide should only act as a support for the molecular catalyst and
should not actively participate in catalysis (unlike semiconductors with lower energy
conduction bands, such as TiO and ZnO where charge injection could occur). This allows us
2
to probe Ir´s activity in photoredox catalysis with less interference from the surface. To increase
catalyst applicability in future heterogenized setups, ITO (indium doped tin oxide) was also
chosen as it is a conducting metal oxide, having a continuous electronic band structure- (i.e. no
band gap); ITO itself should not undergo light induced reactions, and thus, should mainly act
as a catalyst support (although electron injection from the catalyst might occur[2h, 21]). Further,
ITO is advantageous as it can be applied as an electrode material in (photo)electrochemical
setups in future studies,[7b] which would allow sacrificial reagents to be eliminated from
reactions as charges could be provided by the counter electrode in an external circuit; this would
create a more environmentally-friendly way to do photoredox catalysis.
Metal oxide architectures. The effect of the metal oxide architecture was also
examined, either as nanopowders or thin films. Some differences in reactivity, reusability, and
applicability may be found with the different architectures (Table 1). For example,
nanopowders can be stirred in a reaction, minimizing concentration gradients by enabling fast
catalyst transport to the BrAPN substrate, which should result in quick reactions. In contrast,
thin films reactions have no catalyst diffusion or transport via stirring as the catalyst is
immobilized onto a metal oxide thin film annealed to a glass slide. The substrate must be
transported to the catalyst within the film via stirring and diffusion; thus, slower reactions with
thin films are expected. The quantity (equivalents) of Ir in the reactions with nanopowder
supports can be more easily modified than thin films, as more nanopowder MO -Ir can simply
x
be added to the reaction. Whereas, to increase the quantity (equivalents) of catalyst with the
thin films (if the catalyst loading is already maximized), a larger geometric surface area and/or
thicker films would be needed to add more MOx-Ir. Since the films are placed in the reaction
vessel, the vessel must be large enough to fit a larger film, which could require a custom
designed vessel. Regarding reusability, thin films are advantageous since they can be easily
removed from the reaction without losing any catalyst. In contrast, nanopowder-based catalysts
require centrifugation or filtration to remove the catalyst and several rinsing steps, which could
result in some catalyst loss over time, thus limiting the long-term reusability. Thin films can be
7
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
easier to characterize by spectroscopic methods (UV-Vis, ATR-FTIR, TCSPC, XPS) as they
are transparent in nature, possibly having a more evenly distributed catalyst loading, and can
be studied as prepared unlike nanopowders. Finally, thin films have the more applicable
architecture when it comes to future electrochemical or photoelectrochemical setups as an FTO-
coated glass electrode is used to support the thin film in these devices and provide a conductive
contact in the electrical circuit.
Table 1. Advantages and disadvantages of thin films versus nanopowders.
Thin Films
Nanopowders
Advantages:
Disadvantages:
Advantages:
Disadvantages:
•
Easily removed with
tweezers from reaction
No sample loss
Easy to characterize
catalyst
• Slower
• Faster reactions- • Sample loss due
reactions- no
diffusion or
convective
transport (via
stirring) of
catalyst
diffusion and
convective
transport via
stirring
to centrifugation
and rinsing
• Harder to
•
•
characterize
catalyst
•
Applicable in
catalyst
• Potential for
more catalyst
loading
• One step
preparation
(
photo)electrochemical
setups
• Less catalyst
loading (due to
size restriction
of film)
•
Extra
preparation step
Preparation of heterogenized nanopowder catalysts. To prepare the
heterogenized catalysts, approximately 25 mg of nanopowder (Al , ZrO , ITO) was stirred
2
O
3
2
in the presence of 2.5 mL of a 0.25 mM solution of Ir in methanol overnight in the dark. Post
sensitization, the catalyst was centrifuged off from the sensitization solution and rinsed three
times with methanol via subsequent rinsing and centrifugation steps. Finally, the catalyst was
dried under vacuum for several hours, affording the MOx-Ir nanopowders (Figure 2).
Preparation of heterogenized thin film catalysts. To prepare the thin film based
catalysts, metal oxide thin films were first prepared. One layer of metal oxide paste was doctor
bladed onto FTO-coated glass slides, followed by annealing at high temperatures for several
hours (see Supporting Information for further details). The thin films were then soaked in a 0.1
mM Ir sensitization solution in methanol overnight, rinsed with methanol three times, and air-
dried (Figure 2).
8
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Catalyst characterization. Several spectroscopic methods were used to
characterize the catalysts. First, UV-Vis spectroscopy was used to analyze the Ir surface
loadings on the nanopowders and thin films. For the nanopowders, the depletion method was
used to quantify the Ir surface loadings, which is a well-established procedure in the solar fuels
field.[7b, 22] To do this, the UV-Vis spectrum of the sensitization solution was collected before
and after exposure to the nanopowders to get the initial and final iridium concentrations in the
-1
-1
solutions using the molar extinction coefficient, ε(370 nm) = 8730 cm M (Figure S10). After
sensitization of the nanopowders and centrifugation, a decrease in absorbance at 370 nm was
observed in the supernatant, suggesting that the iridium complex in solution had bound to the
nanopowders (Figure S12). The difference in iridium concentration before and after
sensitization was calculated and approximated to be the loading on the surface. For the samples,
an average loading of 6.0 ± 0.9 nmol/mg for Al
2
O
3
-Ir, 7.1 ± 0.8 nmol/mg for ZrO
2
-Ir, and 6.7
±
0.9 nmol/mg for ITO-Ir was obtained (Table S1). For the thin films, UV-Vis spectra was
collected directly on the thin films, showing similar absorption features to the Ir in solution,
suggesting the molecular structure had been retained upon surface binding (Figure 3). The
−
2
loadings (Γ) were obtained using the formula, Γ(mol cm ) = A(λ)/ (1000ε), where A is the
absorbance at wavelength λ, and ε is the molar extinction coefficient at wavelength λ.[23]
2
Average loadings were calculated and found to be 28.4 ± 6.9 nmol/cm for Al
2
3
O -Ir, 31.1 ±
2
2
8
.7 nmol/cm for ZrO -Ir, and 35.8 ± 7.0 nmol/cm for ITO-Ir (Table S2). Finally, if larger
2
catalyst loadings were desired, they could be potentially increased by increasing the thin film
thickness, increasing the concentration of the catalyst sensitization solution, or by changing the
sensitization solution solvent.
Figure 3. UV-Vis spectra of thin films of ZrO
2
-Ir (purple), Al
2
3
O -Ir (pink), ITO-Ir (blue).
9
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
To further characterize the heterogenized catalysts, ATR-FTIR, and XPS were
collected on the nanopowders and thin films. In the ATR-FTIR spectra, C=C and C=N aromatic
-1
-1
stretches at 1607 cm and 1583 cm , respectively, are observed in both Ir powder and MO
x
-Ir
samples, suggesting that the molecular structure is retained upon binding (Figure S20-21). In
-1
addition, the C=O stretch in Ir at 1709 cm disappears upon surface binding, suggesting that
the carboxylic acid anchor binds in a bridging bidentate binding mode to covalently attach the
catalyst the metal oxide supports.[8] X-ray photoelectron spectroscopic (XPS) measurements
also show characteristic peaks for Ir in the iridium 4f region at 61.8 eV and 64.7 eV, which
match well with the Ir powder XPS spectrum, supporting the presence of iridium on the metal
oxides (Figure S26-27). Furthermore, to demonstrate that the carboxylic acid group in Ir was
necessary for covalent attachment to the metal oxide support, we tested to see if an iridium
complex without a carboxylic acid surface anchor would also attach to the metal oxide support.
To do this, Ir(ppy)
3
(Tris[2-phenylpyridinato-C2,N]iridium(III), which has no surface anchor,
thin film in methanol. Post soaking, no iridium
was soaked in the presence of an Al
2
O
3
absorption feature were detected in the UV-Vis spectrum of the thin film (Figure S11B),
suggesting that the carboxylic acid anchor was necessary to covalently attach Ir to the supports.
Based on the UV-Vis, ATR-FTIR, and XPS spectroscopic data, Ir has successfully bound to
the metal oxides and is likely a molecular complex on the surface.
Electrochemistry. To ensure that Ir is capable of reducing the BrAPN substrate,
cyclic voltammetric (CV) measurements were performed on ITO-Ir to get the reduction
potentials of the heterogenized catalyst. The CVs in acetonitrile showed quasi-reversible
+
+
reduction and oxidation waves, at -1.93 V vs. Fc/Fc and 0.98 V vs. Fc/Fc , respectively (Figure
+
S32), while the CV of BrAPN showed an irreversible reduction at -1.73 V vs. Fc/Fc (Figure
−
S33). These potentials confirm that the reduced Ir should be thermodynamically able to reduce
BrAPN.
Excited state potentials. Excited state potentials for Ir were estimated from the
UV-Vis absorption spectrum, photoluminescence spectrum, and ground state redox potentials
using Weller approximations.[24] Intersection of the normalized absorption and emission spectra
give an estimated transition energy from the ground state to the lowest excited state (E0-0) of
2
.21 eV (Figure S36). Using the Weller approximations, excited state potentials were estimated
+
+
0
+
0
+
+
to be -1.23 V vs. Fc/Fc for Ir /Ir* (E (Ir /Ir*) = E (Ir /Ir) – E0-0) and 0.28 V vs. Fc/Fc for
1
0
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
−
0
−
0
-
Ir*/Ir (E (Ir*/Ir ) = E (Ir/Ir ) + E0-0). Thus, the Ir excited state is thermodynamically unable
to reduce the substrate during photoredox catalysis. Furthermore, one should note that
triethanolamine (TEOA) is also present during the reductive dehalogenation reactions as a
sacrificial electron donor and proton source. The TEOA reduction potential is estimated to be
.3 V vs. Fc/Fc [0.7 V vs. SCE, ref. [25]] but as its´ oxidation is irreversible, this value should
+
0
be taken as an approximation and not a fixed value; hence, the Ir excited state is most likely
quenched by TEOA during photoredox catalysis.
Time correlated single photon counting. Prior to catalytic testing, the catalyst
excited state lifetimes were quantified using time correlated single photon counting (TCSPC)
measurements to ensure they did not change significantly upon surface binding and had long
enough lifetimes for photoredox reactions (>1 ns needed). To verify this, thin film catalysts
were used for these measurements and were placed in the sample holder as a dry thin film in
air. Nanopowders were not tested, as the particles settle to the bottom of a cuvette without
stirring making catalyst excitation challenging. For the measurements, Ir was excited with a
4
70 nm laser pulse, and the photoluminescence from the excited state was monitored over time
via collection of single photon events (Figure S34).[26] Decays were fit to mono or biexponential
functions and the results can be found in Table S5. Using monoexponential fits, τ =220 ns was
found for ZrO -Ir. This lifetime is similar to that of the homogeneous version of Ir in
acetonitrile, having τ = 210 ns. Biexponential fits were used for Al -Ir and ITO-Ir as these
1
2
1
2
O
3
did not fit well to single exponential decays; this could be due to surface inhomogeneity or non-
innocent surface behavior. For these surfaces, an average excited state lifetime (τaverage) of 1.5
ns was found for ITO-Ir, and 3.3 ns for Al
lifetime was then ZrO -Ir > Al -Ir > ITO-Ir. It is possible that the conductivity of ITO
interferes some with the Ir excited state (via electron injection into ITO), causing a faster
excited state decay,[2h, 21] while potential surface trap states in Al
, may also shorten the
2
3
O -Ir. The overall trend in catalyst excited state
2
2
O
3
2
O
3
lifetimes.[27] We note that the lifetimes reported here should be taken as estimates as the
measurements were collected and fitted under shorter time scales than the catalyst lifetime;
longer timescales will be measured in future work to obtain more precise values during
mechanistic studies. Recall, the main goal was to demonstrate that the catalyst excited state
persisted long enough to participate in catalytic reactions. Furthermore, all samples show
excited state lifetimes longer than 1 ns, which the minimum lifetime needed to react with a
substrate diffusing in a reaction mixture. Thus, all catalysts are good candidates for photoredox
tests.
1
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Catalyst stability. As catalyst desorption can occur in heterogenized systems,[8]
a quick check was performed prior to catalytic tests to ensure that Ir was stable on the surface
in the reaction solvent, acetonitrile. No catalyst loss from the metal oxide was observed after
soaking MOx-Ir in acetonitrile overnight (Figure S17). In addition, no light induced desorption
was observed when the metal oxide was illuminated with a white light lamp in acetonitrile
(
Figure S17).
Initial catalytic tests. Since Ir had been successfully characterized on the three
surfaces, and had properties promising for photoredox catalysis, initial reductive
dehalogenation tests were conducted. A reaction scheme is depicted in Scheme 2 for the
reductive dehalogenation of BrAPN to APN. We were inspired by results from a
Tris(bipyridine)Ru(II) metal organic framework (MOF) photoredox catalyst reported to
catalyze reductive dehalogenation reactions.[16a] In the reaction mixture, triethanolamine
(
TEOA) was also present to act as a sacrificial electron donor and proton source. During each
reaction, BrAPN, TEOA, and MO -Ir (0.2 mol%) were added to a vial with deuterated
acetonitrile (CD CN) and a stir bar. The reaction was sealed and degassed with Ar prior to
x
3
measurement and kept as dark as possible (more reaction details in SI). Reactions were stirred
rapidly at 1000 RPM, and illuminated with a blue Kessil LED with a 435 long pass filter, placed
2
precisely 3 cm away from the vial (~125 mW/cm ) ; the reaction was also kept cool with a fan
blowing on it at all times. These conditions were kept constant to minimize error between
measurements (extra details in SI).
Scheme 2. Reaction conditions for reductive dehalogenation of bromoacetophenone (BrAPN)
to acetophenone (APN).
The photoredox catalytic reactions were initially examined using the nanopowder
catalysts over two hours of reaction time (Table 2, entry 1-3). Reactions were performed in
1
CD
3
CN in order to analyze the reaction mixture by H NMR spectroscopy (details in SI).
1
Importantly, a main advantage of this reaction is that it is easily followed by H NMR as BrAPN
and APN have singlets in non-overlapping regions; BrAPN has a singlet at 4.68 ppm for its
1
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
CH
2
while APN has a singlet at 2.56 ppm for its CH
3
(Figure S7). Amazingly, all nanopowder
catalysts showed full conversion to APN after two hours of reaction.
Table 2. Reaction yields for reductive dehalogenation of bromoacetophenone to acetophenone
with MO -Ir nanopowders and control reactions after two hours.
x
Entry
Sample
Light / Dark
Light
Light
Light
Light
Light
Light
Dark
% Yield APN
1
2
3
4
5
6
7
8
ITO-Ir
100
Al
ZrO
ITO
Al
ZrO
ITO-Ir
Al -Ir
ZrO -Ir
ITO-Ir
Al -Ir
ZrO -Ir
ITO-Ir
Al -Ir
ZrO -Ir
2
O
3
-Ir
100
2
-Ir
100
0
2
O
3
4
2
2
0
2
O
3
Dark
0
9
2
Dark
0
1 [
1 [
1 [
1 [
1 [
1 [
0
a]
Light
Light
Light
Light
Light
Light
Light
Light
23
a]
a]
b]
b]
b]
1
2
O
3
25
2
2
20
3
0
4
2
O
3
0
5
2
0
1
6
No catalyst
No catalyst
0
1 [
7
c]
16
[
a]reactions used a ratio of BrAPN:TEOA:MO
x
-Ir of 1:0.5:0.002; [b] 0 equivalents of
triethanolamine in reaction; [c] reaction performed without 435 nm long pass filter.
Reaction controls. Since the reactions showed all catalysts were functional,
several controls were then performed to make sure catalysis was occurring as expected (Table
2
). First, we tested to ensure the catalyst was necessary for the reaction to occur (entry 16). No
APN was formed when the catalysts were removed from the reaction mixture, suggesting they
were necessary for catalysis. To ensure that the molecular catalyst was necessary, controls were
performed with just the metal oxide nanopowders (without catalyst bound) in the reaction (entry
4
-6). Insignificant APN was produced, suggesting that the molecular component is necessary
for photoredox catalysis. Furthermore, the reactions do not work in the dark, suggesting that
1
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
light is necessary to excite the catalyst and promote turn over (entry 7-9). In addition, the
necessity for triethanolamine was tested. No APN was formed without TEOA present (entry
1
3-15), as this is likely needed to i.) reduce the iridium excited state, Ir*, and ii.) donate a
proton to the reduced BrAPN species, which forms the APN product. When the concentration
of TEOA is half that of BrAPN (entry 10-12), reaction yields are only 20-25%, suggesting there
is not enough reagent around to finish the catalytic reaction. Finally, when the concentration of
TEOA is three times that of BrAPN (entry 1-3), reactions yields go to completion. This trend
suggests that we do need TEOA in the reaction mixture and in excess to the BrAPN substrate.
Finally, as there is always a possibility that catalyst desorption can occur in
[8]
heterogenized systems, we wanted to check that the active catalyst was indeed heterogenized
and not a desorbed homogeneous catalyst. To do this, the reaction yield was analyzed after a
few minutes (to ensure the reaction was not complete), catalyst removed from the mixture, and
APN yield quantified. The reaction was then continued without the catalyst present to see if the
yields changed when the catalyst was removed. If the yield increased, this would suggest that
the active catalyst was actually a desorbed Ir species and not the heterogenized catalyst. If the
yield did not change, then the active catalyst is likely heterogenized. Indeed, for all catalysts
the yield did not increase significantly, suggesting that the active catalyst was heterogenized
(
Figure 4). Moreover, if the catalyst was added back into the reaction, yields increased again,
further confirming that the catalyst was heterogenized (Figure 4).
Figure 4. Heterogenized tests for (A) ITO-Ir (blue) (B) Al
2
O
3
-Ir (pink), (C) ZrO -Ir (purple).
2
Reactions were run for a few minutes, catalyst removed from the reaction mixture and yields
1
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
quantified. The reaction was continued without the catalyst for a few more minutes, and the
yield checked again (light colored circle). This was compared to when the catalyst was added
in the reaction for the same amount of time.
Mechanistic insight. Since the controls suggested that the catalyst was
functioning as a heterogenized catalyst, we postulated what mechanistically could be occurring
in the reaction mixture. Based on the above controls and electrochemical data, we proposed that
upon illumination i.) Ir becomes photoexcited forming the excited state species ii). the excited
state species is reduced by TEOA forming the reduced iridium catalyst, and the reduced catalyst
reduces the BrAPN substrate, which goes on to form the product through subsequent reduction
and protonation steps (Figure 5). To further support this hypothesis, TCSPC measurements
were performed on Ir in the presence of TEOA, which showed evidence for excited state
quenching; this is seen in the decrease in the excited state lifetimes from 210 ns to 80 ns in the
presence of TEOA (Figure S35). The excited state decay of Ir shows a much smaller change in
the presence of BrAPN, with a lifetime of 180 ns (Figure S35, Table S5). Concentrations of the
Ir, BrAPN, and TEOA were kept the same as the reaction conditions during the TCSPC
measurements to keep conditions as realistic as possible. These additional experiments suggest
that in the catalytic cycle, excited Ir is first reductively quenched, which is then followed by
BrAPN reduction as shown in Figure 5. Recall from the electrochemical and photophysical data
that the reduced state of Ir is thermodynamically capable of reducing BrAPN, while the excited
Ir* is not, further supporting the cycle in Figure 5. We note that these TCSPC experiment were
done in the homogeneous phase for simplicity, and mechanistic results could change in the
heterogenized system. This will be evaluated more extensively in a later study.
1
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Figure 5. Proposed catalytic cycle based on TCSPC data and electrochemical data. First, Ir
III
]+, where ppy=2-phenylpyridine, and bpy=2,2′-
(
depicted more precisely as [Ir (bpy)(ppy)
2
bipyridine-4,4′-dicarboxylic acid) becomes photoexcited upon illumination, forming the
[6]
excited iridium species , which is then reduced by TEOA to form the reduced iridium state.
Finally, the reduced iridium species reduces BrAPN, which returns Ir to its ground state. The
reduced BrAPN then goes on to form the final product after subsequent reduction and
protonation via reaction with the triethanolamine radical cation.[25]
The need for speed. With a grasp of what could be happening during catalysis,
and as the reactions reached completion during two hours, we wondered if the reactions were
1
complete prior to the two-hour mark. To do this, we tracked the reaction over time by H NMR
and found that indeed, the nanopowder-based catalysts were remarkably faster (Figure 6A). For
example, Al
2
3
O -Ir samples showed the fastest APN formation with reactions complete by 15
minutes doing 385 turnovers (limited by substrate consumption) with a turnover frequency
-1
(
TOF) of ~0.4 s . ZrO
2
-Ir samples were also fast, reaching completion after 30 minutes, being
-1
slightly slower with a TOF of ~0.2 s . ITO-Ir was the slowest, taking the full two hours to
-1
reach completion giving it a TOF of 0.05 s . Clearly, the fastest catalysts are those using wide
bandgap semiconductor supports, which suggests these are the optimal surfaces for the
nanopowder reactions. ITO supports may have resulted in slower reactions because the Ir
1
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
excited state lifetime is much shorter on these surfaces, or if the surface is non-innocent during
catalysis (e.g. from electron injection into ITO by Ir* or Ir ).[2h, 21] Table 3 summarizes these
findings.
−
Figure 6. Reaction yield over time during photoredox catalysis using (A) nanopowder and (B)
thin film catalysts (ITO-Ir (blue), Al
2
O
3
-Ir (pink), and ZrO -Ir (purple)).
2
Table 3. Reaction times, TOFs, TONs for the nanopowder and film catalysts.
Nanopowder
Catalysts
Reaction Time
-
1
TOF (s )
TON (post 3x uses)
(min) [a]
Al
2
O
3
-Ir
15
0.4
0.2
942[b]
663
ZrO
2
-Ir
30
120
ITO-Ir
0.05
777[b]
Reaction Time
-
1
Film Catalysts
TOF (s )
TON (post 3x uses)
(
min) [a]
Al
ZrO
ITO-Ir
2
O
3
-Ir
240
0.025
0.025
0.05
984[b]
798
2
-Ir
240
120
558
[
a]Time for the reaction to reach completion; [b]More turnovers possible.
Thin film reactions. Since the nanopowder reactions were highly functional
using heterogenized iridium photoredox catalysts, we wanted to test if thin films could also be
used for these reactions as they are easier to characterize, easier to remove from the reaction,
1
7
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
and have possible applications in future (photo)electrochemical setups. Reactions were
performed with the thin films placed in the reaction vial, face up, to prevent the stir bar from
removing the thin film (see Figure S3). The equivalents of the reagents (BrAPN, TEOA, and
Ir) in the thin film and nanopowder reactions were kept constant in order to compare the two
1
systems, and the time needed to reach reaction completion was monitored by H NMR (Table
3
). Since there is no catalyst diffusion or transport via stirring with thin films, we expected the
reactions to take longer than the nanopowders. Indeed, reactions for both Al -Ir and ZrO -Ir
films were much slower (Figure 6B), taking four hours to reach completion and perform 385
2
O
3
2
-1
turnovers, giving them a TOF of 0.025 s . Interestingly, ITO-Ir remained the same, taking 2
hours still to complete. Since ITO-Ir nanopowders and films took similar reaction times, this
suggests that ITO may not be acting solely as a catalyst platform; perhaps electron transfer
events between the catalyst and ITO (such as electron injection) are also occurring during these
reactions, which can alter the reaction times.
We note that the catalyst concentration in the reaction mixture was lower with the
thin films, which may contribute to the slower reaction times. However, catalyst architecture
effect (nanopowder versus film) on reaction completion time, is likely more significant as the
nanopowders reactions for Al
2
3
O -Ir are complete sixteen times faster than the films, even
though the catalyst concentrations are only approximately four times lower. This suggests that
the film architecture is likely contributing more to the slower catalysis than the catalyst
concentration.
Comparison to homogeneous Ir. Finally, we were curious how our photoredox
catalysts compared to Ir in a homogenous system. When this reaction was followed, we found
it was quite similar to the ZrO
Figure S5). Interestingly, the fastest nanoparticle-based heterogenized catalyst, Al
operates more efficiently than the homogeneous catalyst. This is noteworthy as Al -Ir also
2
-Ir nanopowder system taking 30 minutes to reach completion
(
2
O
3
-Ir,
2
O
3
has the advantage of being easier to separate from the reaction mixture, making it the more
reusable option.
Catalyst integrity. Since both nanopowder and film catalysts showed excellent
photoredox activity, we wanted to check the catalyst integrity after reactions to see if they could
be good candidates for reusability measurements. UV-Vis of the thin films showed that Ir was
retained on the surface (Figure 7A, S13). ATR-FTIR of the catalysts also indicated that the C=C
and C=N stretches were retained (Figure 7B, S22-23), and XPS showed that iridium remained
1
8
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
on the metal oxide supports (Figure 7C, S28-29). These experiments suggest retention of the
Ir structure on the surface and that they have the potential to be reusable.
Figure 7. (A) UV-Vis spectra of Al
photoredox catalytic test. (B) ATR-FTIR spectra of ZrO
purple) one photoredox catalytic test; Ir powder is shown in grey. (C) XPS spectra of Al
dark pink) and Al blank (grey). The XPS spectra of Al -Ir after one photoredox test is
shown in red and after three tests is shown in black.
2
O
3
-Ir before (dark pink) and after (light pink) one
-Ir before (dark purple) and after (light
-Ir
2
2
O
3
(
2
O
3
2
O
3
Catalyst reusability. Since Ir remained mostly surface-bound after catalysis,
reusability experiments were performed to see if the catalysts could be used more than once
(
Figure 8, Table 3). All catalysts showed good yields over two uses, with yields only dropping
slightly to ~90% for Al -Ir and ZrO -Ir catalysts. ITO-Ir saw a greater drop after two uses
to around ~60% yield (Figure 8). After three uses, ZrO -Ir nanopowder and film catalysts as
well as ITO-Ir nanopowders did not produce APN, suggesting that they had reached maximum
turnover; ZrO -Ir nanopowder performed 663 turnovers, ZrO -Ir films 798 turnovers, and ITO-
2
O
3
2
2
2
2
Ir films 588 turnovers prior to losing activity. ITO-Ir nanopowders continued to function after
three uses, dropping steadily in yield over the uses to 60% yield by the end of the third reaction
performing 777 turnovers as tested (more possible). The best catalysts for reusability in both
films and nanopowders were those of Al
2
3
O -Ir films, with yields staying above ~80% yield
over three uses, and just at ~80% for the nanopowders. Clearly, these are the most robust metal
oxides for this reaction, and conveniently the fastest catalysts in the nanopowder form. After
three uses, Al
2
3
O -Ir films performed 984 turnovers and nanopowders 942 turnovers, with more
1
9
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
turnovers possible as they may be able to be reused further. Table 3 summarizes the TONs from
the reusability tests.
Figure 8. Reaction yields for three sequential photoredox tests for (A) nanopowder catalysts
and (B) thin film catalysts; ITO-Ir (blue), Al
2
O
3
-Ir (pink), ZrO -Ir (purple).
2
Furthermore, Al
oxide film loss was observed after the reactions, unlike ZrO
the reaction conditions, likely limiting the reusability. Nanopowder reactions, although faster,
do lose some of the initial MO -Ir starting mass after each use due to losses during
centrifugation and rinsing steps, even up to 50% of the initial catalyst mass by the start of
reaction three (Figure S6). Clearly, as the Al and ITO films are retained after each use, these
2
O
3
and ITO films were the most robust supports, as no metal
2
films, which peeled away under
x
2
O
3
are the most robust and easily reusable platforms. Regardless, as all catalysts were reusable at
least twice, this highlights that heterogenized catalysts have the potential to be used as an
environmentally friendly alternative way to do photoredox catalysis when fully optimized.
Catalyst integrity after reusability. To understand why some of the catalysts
were not as reusable, they were characterized again after three uses. Characterization of the film
samples by UV-Vis showed catalyst loss or degradation after each use with Ir signature bands
decreasing after each use (Figure S14), and similarly when ATR-FTIR was used (Figure S24-
2
5). XPS still showed iridium content on most of the metal oxides, demonstrating that the
catalyst had not fully desorbed (Figure S30-31). Based on these results, some of the iridium
complex has been removed from the metal oxide surface after three uses, which could be a
reason for the loss in activity with some catalysts.
2
0
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Reaction mixtures. To further understand what was happening during catalysis,
we examined the UV-Vis spectra of the reaction mixtures post catalysis during the reusability
tests. The UV-Vis spectra of the reaction mixtures after the first two reactions showed a species
present in the UV-Vis spectra (Figure S15-16). The species observed differed depending on
what catalyst was used, and could be either a desorbed, degraded, or reduced Ir species. Note
that the spectra did not match that of Ir for nearly all of the catalysts with the exception being
from the ITO-Ir nanopowder reactions where the UV-Vis spectra from the first reaction
appeared to mimic Ir, suggesting some Ir desorption in this specific case.
To understand the origin of these species, we tested to see if they were formed
before or during catalysis. First, controls were performed to examine if the species was formed
prior to catalysis by soaking the MO -Ir films in the dark in the presence of acetonitrile
x
solutions of BrAPN, TEOA, a mixture of both (Figure S18). Recall, that solvent and light alone,
does not desorb Ir (Figure S17). We found that a species desorbs when films are exposed to
both BrAPN and TEOA prior to catalytic measurements (prior to LED exposure), which has
absorption bands around 400 nm and 500 nm (Figure S18F); this spectrum is similar to that of
Ir, but has shifted slightly, which could suggest some changes in the ligand environment of the
iridium complex due to exposure to both BrAPN and TEOA. This suggests that prior to
illumination, some desorbed/altered iridium species is already present in the reaction mixture.
However, these spectra do not match that of the reaction mixtures post catalysis, suggesting
that i) the species that desorbs prior to catalysis may change some during photoredox catalysis
or ii) that another degraded/changed species desorbs from the metal oxide surface during
catalysis. We note that soaking the catalyst in an APN acetonitrile mixture does not match these
spectra either, suggesting that APN does not aid in forming the species. For now, it is clear that
an iridium species is appearing in the reaction mixtures, but it is different from Ir. Importantly,
recall that when the catalyst is removed from the reaction mixture during catalysis (Figure 4),
the reaction mixture alone does not show catalytic activity. Thus, we conclude that these
desorbed/degraded species are likely innocent during catalysis as they are possibly i) an inactive
complex slightly different than Ir or ii) too low in concentration to contribute to catalysis.
Nonetheless, loss of the surface species or degradation is likely a reason for some loss in
catalytic activity for some of the samples, and will be optimized and investigated further in
future work.
2
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Conclusions
In this study, we have performed a systematic study to analyze the effect of the
catalyst support in heterogenized iridium photoredox catalysts. Three metal oxide supports
were examined in both nanopowder and thin film form, and all catalysts were found to be
functional for reductive dehalogenation reactions to form acetophenone. Significantly, the
fastest catalyst, nanopowder Al
2
3
O -Ir, was able to reach reaction completion in nearly 15
minutes, which is slightly faster than the homogenous system. Thin film-based catalysts also
gave high yields, but operated slower (2-4 hours) than their nanopowder counterparts due to no
catalyst diffusion or transport via stirring. However, thin films provide more applicable
architectures in future (photo)electrochemical photoredox catalytic studies and additionally,
were more easily reused. Post catalysis, all catalysts showed good surface stability and were
able to be reused at least 2-3 times. The most reusable catalysts were Al
2
3
O -Ir (both film and
nanopowder) showing a limited drop in yield after recycling, which approached 1000 turnovers
after three uses, further highlighting that these catalysts can be used to do efficient photoredox
catalysis sustainably. Thus, of the three surfaces, Al
2
3
O is the optimal metal oxide, and thin film
supports are preferred over the nanopowders as they are the most easily reused and greener
options; in future work, catalysts will be further explored with these architectures. Since some
surface instability was observed during the reactions, catalyst stability on the metal oxide will
be improved in future investigations by using stronger surface anchors or through the addition
of surface protecting layers. We are currently working to expand the substrate and reaction
scope with the optimized catalysts, to show broader applicability in future studies. Finally, we
hope to have shown here that heterogenized catalysts have a great potential to be used in the
photoredox community, bringing together both the tunability of homogeneous systems and
reusability of heterogeneous systems in one complete, environmentally friendly catalyst.
Experimental Section: Further experimental details on catalyst preparation and
characterizations, reaction conditions, and NMR quantifications can be found in the Supporting
information. Additional UV-Vis, ATR-FTIR, photoluminescence, XPS, and NMR spectra and
TCSPC results can also be found in the SI.
Acknowledgements: We thank the Foundation Olle Engkvist Byggmästare (SOEB, grant
2
00-0565) and Knut and Alice Wallenberg Foundation (grant 2019.0071) for funding this
research. We also thank Dr. Anna Arkhypchuk for aiding in experimental setups and Nidhi
2
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Kaul for helping with TCSPC measurements. We also thank Dr. Anna M. Beiler, Dr. Anna
Arkhypchuk, and Dr. Haining Tian for providing feedback on the manuscript.
References
[
[
1]
2]
a) U. S. E. I. Administration, (Ed.: U. D. o. Energy), DC, 2017; b) IEA, IEA, Paris,
020.
a) E. A. Gibson, Chem Soc Rev 2017, 46, 6194-6209; b) J. R. Swierk, T. E. Mallouk,
Chem Soc Rev 2013, 42, 2357-2387; c) D. L. Ashford, M. K. Gish, A. K. Vannucci,
M. K. Brennaman, J. L. Templeton, J. M. Papanikolas, T. J. Meyer, Chem. Rev. 2015,
2
1
15, 13006-13049; d) Z. Yu, F. Li, L. Sun, Energy Environ. Sci. 2015, 8, 760-775; e)
M. K. Brennaman, R. J. Dillon, L. Alibabaei, M. K. Gish, C. J. Dares, D. L. Ashford,
R. L. House, G. J. Meyer, J. M. Papanikolas, T. J. Meyer, J. Am. Chem. Soc. 2016,
1
1
3085–13102; f) M. K. Nazeeruddin, E. Baranoff, M. Grätzel, Solar Energy 2011, 85,
172-1178; g) S. Zhang, H. Ye, J. Hua, H. Tian, EnergyChem 2019, 1; h) D. Wang, B.
H. Farnum, C. J. Dares, T. J. Meyer, J. Chem. Phys. 2020, 152, 244706.
[
3]
a) L. J. Antila, P. Ghamgosar, S. Maji, H. Tian, S. Ott, L. Hammarström, ACS Energy
Letters 2016, 1, 1106-1111; b) L. D'Amario, L. J. Antila, B. Pettersson Rimgard, G.
Boschloo, L. Hammarstrom, J. Phys. Chem. Lett. 2015, 6, 779-783; c) L. D'Amario, J.
Föhlinger, G. Boschloo, L. Hammarström, Chemical Science 2018, 9, 223-230; d) L.
D'Amario, R. Jiang, U. B. Cappel, E. A. Gibson, G. Boschloo, H. Rensmo, L. Sun, L.
Hammarstrom, H. Tian, ACS Appl. Mater. Interfaces 2017, 9, 33470-33477; e) T.
Daeneke, Z. Yu, G. P. Lee, D. C. Fu, N. W. Duffy, S. Makuta, Y. Tachibana, L.
Spiccia, A. Mishra, P. Baeuerle, U. Bach, Advanced Energy Materials 2015, 5; f) S.
Mori, S. Fukuda, S. Sumikura, Y. Takeda, Y. Tamaki, E. Suzuki, T. Abe, J. Phys.
Chem. C. 2008, 112, 16134-16139; g) B. Shan, A. Nayak, M. K. Brennaman, M. Liu,
S. L. Marquard, M. S. Eberhart, T. J. Meyer, J. Am. Chem. Soc. 2018, 140, 6493-6500;
(
h) B. Shan, B. D. Sherman, C. M. Klug, A. Nayak, S. L. Marquard, Q. Liu, R. M.
Bullock, T. J. Meyer, J. Phys. Chem. Lett. 2017, 8, 4374-4379.
[
4]
a) J. Z. Bloh, R. Marschall, Eur. J. of Org. Chem. 2017, 2017, 2085-2094; b) R. C.
McAtee, E. J. McClain, C. R. J. Stephenson, Trends in Chemistry 2019, 1, 111-125; c)
M. Reckenthäler, A. G. Griesbeck, Advanced Synthesis & Catalysis 2013, 355, 2727-
2
744; d) N. A. Romero, D. A. Nicewicz, Chem. Rev. 2016, 116, 10075-10166; e) M.
H. Shaw, J. Twilton, D. W. MacMillan, J. Org. Chem. 2016, 81, 6898-6926; f) C. Y.
Stephenson, T.; MacMillan, D.W.C., Wiley‐VCH Verlag GmbH & Co. KGaA 2018.
B. M. Hockin, C. Li, N. Robertson, E. Zysman-Colman, Catal. Sci. Technol. 2019, 9,
[
[
[
5]
6]
7]
8
89-915.
C. Lentz, L. Marcélis, L. Troian-Gautier, K. Robeyns, E. Cauët, B. Elias, Journal of
Photochemistry and Photobiology A: Chemistry 2021, 405.
a) K. L. Materna, J. Jiang, K. P. Regan, C. A. Schmuttenmaer, R. H. Crabtree, G. W.
Brudvig, ChemSusChem 2017, 10, 4526-4534; b) K. L. Materna, B. Rudshteyn, B. J.
Brennan, M. H. Kane, A. J. Bloomfield, D. L. Huang, D. Y. Shopov, V. S. Batista, R.
H. Crabtree, G. W. Brudvig, ACS Catal. 2016, 6, 5371-5377; c) H. Tian,
ChemSusChem 2015, 8, 3746-3759.
[
[
8]
9]
K. L. Materna, R. H. Crabtree, G. W. Brudvig, Chem. Soc. Rev. 2017, 46, 6099-6110.
a) G. J. Barbante, T. D. Ashton, E. H. Doeven, F. M. Pfeffer, D. J. D. Wilson, L. C.
Henderson, P. S. Francis, ChemCatChem 2015, 7, 1655-1658; b) B. Tambosco, K.
2
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
Segura, C. Seyrig, D. Cabrera, M. Port, C. Ferroud, Z. Amara, ACS Catal. 2018, 8,
383-4389.
4
[
10] R. I. Teixeira, N. C. de Lucas, S. J. Garden, A. E. Lanterna, J. C. Scaiano, Catal. Sci.
Technol. 2020, 10, 1273-1280.
[
[
11] Y. Liang, D. E. Bergbreiter, Catal. Sci. Technol. 2016, 6, 215-221.
12] S. M. Soria-Castro, B. Lebeau, M. Cormier, S. Neunlist, T. J. Daou, J.-P. Goddard,
Eur. J. Org.Chem. 2020, 2020, 1572-1578.
[
[
[
13] A. I. Carrillo, A. Elhage, M. L. Marin, A. E. Lanterna, Chemistry 2019, 25, 14928-
1
4934.
14] Y. Chu, N. Corrigan, C. Wu, C. Boyer, J. Xu, ACS Sustainable Chem. Eng. 2018, 6,
5245-15253.
1
15] a) P. Zhi, Z.-W. Xi, D.-Y. Wang, W. Wang, X.-Z. Liang, F.-F. Tao, R.-P. Shen, Y.-M.
Shen, New J. Chem. 2019, 43, 709-717; b) W.-J. Yoo, S. Kobayashi, Green Chem.
2
014, 16, 2438-2442.
[
16] a) A. Santiago-Portillo, H. G. Baldoví, E. Carbonell, S. Navalón, M. Álvaro, H.
García, B. Ferrer, J. Phys. Chem. C. 2018, 122, 29190-29199; b) Z. J. Wang, S.
Ghasimi, K. Landfester, K. A. I. Zhang, J. Mater. Chem. A 2014, 2, 18720-18724; c)
A. Santiago-Portillo, S. Remiro-Buenamanana, S. Navalon, H. Garcia, Dalton Trans.
2
019, 48, 17735-17740; d) R. Martinez-Haya, M. A. Miranda, M. L. Marin, Eur. J.
Org. Chem.2017, 2017, 2164-2169.
[
[
17] a) C. K. Prier, D. A. Rankic, D. W. MacMillan, Chem. Rev. 2013, 113, 5322-5363; b)
T. M. S. Monos, C., in Photoredox Catalysis of Iridium(III)‐Based Photosensitizers,
Ch11 (Ed.: E. Zysman‐Colman), John Wiley & Sons Ltd., 2017.
18] J. B. Waern, C. Desmarets, L. M. Chamoreau, H. Amouri, A. Barbieri, C. Sabatini, B.
Ventura, F. Barigelletti, Inorg. Chem. 2008, 47, 3340-3348.
[
[
19] D. Rehm, A. Weller, Ber. Bunsenges 1969, 73, 834-830.
20] E. Palomares, J. N. Clifford, S. A. Haque, T. Lutz, J. R. Durrant, J. Am. Chem. Soc.
2
003, 125, 475-482.
[
[
21] L. Alibabaei, R. J. Dillon, C. E. Reilly, M. K. Brennaman, K. R. Wee, S. L. Marquard,
J. M. Papanikolas, T. J. Meyer, ACS Appl. Mater. Interfaces 2017, 9, 39018-39026.
22] K. L. Materna, N. Lalaoui, J. A. Laureanti, A. P. Walsh, B. P. Rimgard, R. Lomoth, A.
Thapper, S. Ott, W. J. Shaw, H. Tian, L. Hammarstrom, ACS Appl. Mater. Interfaces
2
020, 12, 4501-4509.
[
[
[
[
23] S. A. Trammell, T. J. Meyer, J. Phys. Chem. B 1999, 103, 104-107.
24] A. Weller, Zeitschrift für Physikalische Chemie 1982, 133, 93-98.
25] Y. Pellegrin, F. Odobel, Comptes Rendus Chimie 2017, 20, 283-295.
26] a) K. S. Kjaer, N. Kaul, O. Prakash, P. Chabera, N. W. Rosemann, A. Honarfar, O.
Gordivska, L. A. Fredin, K. E. Bergquist, L. Haggstrom, T. Ericsson, L. Lindh, A.
Yartsev, S. Styring, P. Huang, J. Uhlig, J. Bendix, D. Strand, V. Sundstrom, P.
Persson, R. Lomoth, K. Warnmark, Science 2019, 363, 249-253; b) A. El-Zohry, A.
Orthaber, B. Zietz, J. Phys. Chem. C Nanomater. Interfaces 2012, 116, 26144-26153.
27] a) V. V. Afanas’ev, A. Stesmans, J. Appl. Phys. 2004, 95, 2518-2526; b) Y. N.
Novikov, V. A. Gritsenko, K. A. Nasyrov, Appl. Phys. Lett. 2009, 94.
[
2
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101651
2
5
This article is protected by copyright. All rights reserved.