A Zirconium Photosensitizer with a Long-Lived Excited State: Mechanistic Insight into Photoinduced Single-Electron Transfer

Article abstract of DOI:10.1021/jacs.8b00742

Time-resolved emission spectroscopy for the luminescent zirconium complex Zr(^MePDP)₂ (^MePDP = 2,6-bis(5-methyl-3-phenyl-1H-pyrrol-2-yl)pyridine) revealed a long-lived excited state with a lifetime τ = 325 ± 10 μs. Computational studies using time-dependent density functional theory were conducted to identify the nature of the luminescent excited state as a mixed triplet intraligand/ligand-to-metal charge-transfer state. Stern-Volmer experiments showed a strong dependence of the quenching rate on the redox potential of the quencher indicating photoinduced single-electron transfer (SET) as the quenching pathway. Mechanistic investigations of the photocatalytic homocoupling of benzyl bromide allowed the detection of organic radical intermediates during turnover and provided further evidence for SET mediated by Zr(^MePDP)₂. Isolation of the one-electron-reduced form of the photosensitizer, [Zr(^MePDP)₂]^-, enabled studies of its electronic structure by a combination of experimental and computational techniques and confirmed its role as a strong reductant. Additionally, the role of the benzimidazolium hydride derivatives as two-electron sacrificial reductants during photoredox catalysis was investigated. In combination, the results presented in this report establish a detailed mechanistic picture of a photoredox catalytic reaction promoted by an earth-abundant early transition metal photosensitizer.

Full text of DOI:10.1021/jacs.8b00742

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
A Zirconium Photosensitizer with a Long-Lived Excited State:
Mechanistic Insight into Photo-Induced Single Electron Transfer
Yu Zhang, Tia S. Lee, Jeffrey L. Petersen, and Carsten Milsmann
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 19 Apr 2018
Downloaded from http://pubs.acs.org on April 19, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
A Zirconium Photosensitizer with a Long-Lived Excited State: Mechanistic
Insight into Photo-Induced Single Electron Transfer.
†
‡
†
,†
Yu Zhang, Tia S. Lee, Jeffrey L. Petersen, and Carsten Milsmann*
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
C. Eugene Bennett Department of Chemistry, West Virginia University, Morgantown, West
Virginia 26506, United States
‡
Department of Chemistry, Princeton University, Princeton, New Jersey 08544, United States
camilsmann@mail.wvu.edu
Abstract
Me
Timeꢀresolved emission spectroscopy for the luminescent zirconium complex Zr( PDP)2
Me
( PDP = 2,6ꢀbis(5ꢀmethylꢀ3ꢀphenylꢀ1Hꢀpyrrolꢀ2ꢀyl)pyridine) revealed a longꢀlived excited state
with a lifetime τ = 325 ± 10 ꢁs. Computational studies using timeꢀdependent density functional
theory (TDꢀDFT) were conducted to identify the nature of the luminescent excited state as a
3
3
mixed triplet intraligand/ligandꢀtoꢀmetal charge transfer ( IL/ LMCT) state. SternꢀVolmer
experiments showed a strong dependence of the quenching rate on the redox potential of the
quencher indicating photoꢀinduced single electron transfer (SET) as the quenching pathway.
Mechanistic investigations of the photocatalytic homocoupling of benzyl bromide allowed the
detection of organic radical intermediates during turnover and provided further evidence for SET
Me
mediated by Zr( PDP) . Isolation of the oneꢀelectron reduced form of the photosensitizer,
2
Me
1ꢀ
[
Zr( PDP) ] , enabled studies of its electronic structure by a combination of experimental and
2
computational techniques and confirmed its role as a strong reductant. Additionally, the role of
R
the benzimidazolium hydride derivatives ( BIH) as twoꢀelectron sacrificial reductants during
photoredox catalysis was investigated. In combination, the results presented in this report
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
establish a detailed mechanistic picture of a photoredoxcatalytic reaction promoted by an earthꢀ
abundant early transition metal photosensitizer.
Introduction
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Replacing precious metal catalysts with complexes based on earthꢀabundant metals is an
attractive goal considering the economic and environmental benefits resulting from the lower
costs and reduced toxicity of these elements, respectively. For traditional organometallic
1
,2
3–7
8–12
applications such as polymerization, olefin functionalization,
crossꢀcoupling,
or CꢀH
13–16
functionalization,
significant advancements in base metal catalysis have been reported over
the last decades. In contrast, photochemical and photovoltaic applications relying on the use of
molecular transition metal photosensitizers are still dominated by precious metal complexes with
2+
17–20
21,22
derivatives of [Ru(bpy)3] (bpy = 2,2’ꢀbipyridine),
[Ir(ppy) ] (ppy = 2ꢀphenylpyridine),
3
2
3,24
and [ReCl(CO) (bpy)]
as the most prominent examples. While the design of photoactive
3
2
5,26
coordination compounds from earthꢀabundant metals has been a longꢀstanding target,
further
27–29
urgency has been added by the increasing demand for efficient solar energy conversion
and
The
3
0–35
the rise of photoredox catalysis as a rapidly expanding field in organic synthesis.
production of solar fuels and photovoltaic applications on a global scale in particular are limited
3
6,37
by the low abundance of traditional molecular photosensitizers materials.
The popularity of precious metal photosensitizers can be attributed to their favorable and wellꢀ
understood photochemical and electron transfer properties: Efficient light absorption in the
visible region of the spectrum via metalꢀtoꢀligand charge transfer (MLCT) from an electron rich
6
8
10
metal center (d , d , d electron configurations) to a π accepting ligand; facile intersystem
crossing (ISC) to a longꢀlived excited state due to strong spinꢀorbit coupling in secondꢀ or thirdꢀ
3
8
row metals; reversible outerꢀsphere electron transfer in both excited state and ground state. The
2
ACS Paragon Plus Environment

Page 3 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
most critical obstacle in transferring these design principles to analogous complexes of the more
2
+
abundant firstꢀrow transition metals, such as [Fe(bpy)3] and related polypyridyl complexes, is
the lower ligand field splitting in firstꢀrow metals compared to their precious metal
3
9,40
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
congeners.
The resulting lowꢀlying dꢀd excited states provide readily available nonꢀradiative
deactivation pathways and result in drastically decreased excited state lifetimes in the picosecond
4
1–43
I
range.
One exception to this phenomenon are Cu diimine complexes that have gained
4
4–50
10
increased attention over the last years.
These complexes possess d configurations that
preclude ligand field excited states and allow excited state lifetimes in the photocatalytically
useful nanoꢀ to microsecond range. A second class of MLCT photosensitizers avoiding the use of
0
0
0 51–55
precious metals are isocyanide complexes of Cr , Mo , and W .
Despite significant advances
in recent years, several challenges remain for this class of photosensitizers. While the high
reducing power of the excited states provides large driving forces for excited state electron
transfer, the molecules are easily oxidized in their ground state resulting in undesired side
reactions. Additionally, the highest quantum yields and longest excited state lifetimes are
observed for W, the element with the lowest abundance among the group VI elements.
In pronounced contrast to MLCT photosensitizers, luminescent metal complexes relying on
alternative mechanisms for light absorption such as ligandꢀtoꢀmetal charge transfer (LMCT),
ligandꢀtoꢀligand charge transfer (LLCT), or metalꢀcentered excitations (i.e. dꢀd, fꢀf, or dꢀf) are
not as wellꢀstudied, but offer the potential to expand the palette of metals suitable for
5
6
photosensitizer development.
Recent reports of organic photoredox reactions using
photoluminescent lanthanide complexes highlight the potential of “nonꢀtraditional” metals for
57–64
the design of photosensitizers.
Early transition metals from groups IIIꢀV are particularly
attractive targets for the development of nonꢀprecious metal photocatalysts due to their increased
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
abundance in the Earth’s crust compared to most late transition metals and lanthanides. Among
these relatively earthꢀabundant elements, titanium and zirconium occupy a prominent position as
3
6,37
the second and fourth most abundant among all transition metals, respectively.
The electronꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
deficient nature of these species requires a reversal of the chargeꢀseparation step upon light
0
absorption from MLCT to ligandꢀtoꢀmetal charge transfer (LMCT). The readily accessible d
electron configuration commonly found for early transition metals prevents detrimental ligand
field deactivation pathways and allows for extended luminescence lifetimes. Prompted by the
III
65,66
initial discovery of longꢀlived luminescent LMCT states in cyclopentadienyl Sc complexes,
0
several examples of luminescent group IIIꢀV metal complexes with d configuration have been
6
7–77
reported.
However, most of these complexes do not exhibit reversible redox properties and
often require excitation by UV light. In contrast, we recently reported the synthesis of the
Me
Me
zirconium
complex,
Zr( PDP)2
(H2 PDP
=
2,6‐ bis(5‐ methyl‐ 3‐ phenyl‐ 1H‐ pyrrol‐ 2‐ yl)ꢀpyridine), that shows photoluminescence
7
8
upon visible light excitation and can undergo reversible electron transfer in the ground state.
Me
This report also demonstrated the utility of Zr( PDP) as a photosensitizer in visible light
2
promoted, reductive photoredox reactions previously performed using precious metal catalysts.
While these studies provided a proof of concept for photocatalysis by group IV metal complexes,
the mechanistic details of the photocatalytic reactions as well as several key properties of
Me
Zr( PDP) such as the exact nature of the emission (phosphorescence vs. fluorescence) and the
2
lifetime of the luminescent excited state remained unclear.
Herein, we report a detailed mechanistic study into photoꢀinduced singleꢀelectron transfer
Me
(
SET) by Zr( PDP) and provide further insight into the excited state properties of the complex.
2
The previously reported reductive homocoupling reaction of benzyl bromide to bibenzyl was
4
ACS Paragon Plus Environment

Page 5 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
chosen as a model system to firmly establish SET reactivity rather than energy transfer pathways
via inꢀsitu trapping and characterization of radical intermediates. Isolation and characterization
of the oneꢀelectron reduced form of the zirconium photosensitizer provide further insight into the
Me
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
electron transfer characteristics of Zr( PDP) .
2
Results and Discussion
Re-Evaluation of the Photocatalytic Model System. To establish a reliable protocol for the
Me
mechanistic investigation of photoꢀinduced SET by Zr( PDP) , our study commenced with a reꢀ
2
evaluation of the reaction conditions for the photoredox catalytic reductive homocoupling of
benzyl bromide (Table 1). As previously demonstrated, this reaction can be conducted under
Me
irradiation with commercially available green LEDs (λmax = 520 nm) using Zr( PDP) as the
2
Me
photosensitizer, 1,3ꢀdimethylꢀ2ꢀphenylꢀ2,3ꢀdihydroꢀ1Hꢀ7ꢀmethylbenzoꢀ[d]imidazole, BIH, as
the sacrificial reductant, and 2,6ꢀlutidine as the base. Control experiments conducted in the
Me
absence of light and/or Zr( PDP) resulted in quantitative recovery of starting materials, while
2
78
reactions without addition of base resulted in low turnover and rapid catalyst decomposition.
One potential complication for a thorough mechanistic study under these initial conditions is the
Me
presence of benzylic CꢀH bonds in BIH which can undergo unwanted side reactions with
benzyl radical intermediates in the proposed SET mechanism. This hypothesis is supported by
the low yield of bibenzyl (30% at >99% conversion of benzyl bromide) and concomitant
Me
formation of BIH derived side products during photocatalytic turnover. The benzimidazolium
H
Cl
hydride derivatives BIH and BIH studied previously provide more selective product
formation, but result in slow photocatalytic reactions which can be attributed to their more
positive redox potentials (Table 1, entries 2ꢀ4).
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MeO
Building on these observations,
BIH – a more reducing BIH derivative without benzylic
protons ꢀ was introduced for the present study to improve the selectivity and rate of the reaction.
Despite an improved reaction time (Table 1, entry 5), the yield of bibenzyl was poor (19%) and a
significant amount of toluene (15%) was observed as a byꢀproduct. This competing
hydrodehalogenation reaction was proposed to result from facile hydrogen atom abstraction from
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MeO
●+
the oxidized form of the sacrificial reductant,
BIH , which exhibits a very weak CꢀH bond
BDFE = 29.3 kcal mol ) that can easily be attacked by a potential benzyl radical intermediate.
This reactivity highlights the complex protonꢀcoupled electronꢀtransfer (PCET) properties of
ꢀ1 79
(
R
79
BIH derivatives and prompted further studies of the influence of the base on the photocatalytic
performance.
A significant improvement was achieved by changing the base from 2,6ꢀlutidine to the
stronger base triethylamine, NEt , increasing the yield of bibenzyl (64%) while simultaneously
3
decreasing the amount of toluene (Table 1, entry 6). This improvement is consistent with more
MeO
●+
facile deprotonation of
BIH by the stronger base NEt resulting in the formation of the
3
MeO
●
strong single electron reductant
BI and preventing hydrogen atom transfer (HAT). While the
formation of toluene was suppressed, small amounts of the ammonium salt [BnNEt ]Br resulting
3
from nucleophilic attack of NEt on benzyl bromide were detected as a new byꢀproduct. The use
3
of the bulkier trialkylamine base N,Nꢀdiisopropylethylamine did not alleviate this issue but
resulted in slightly lower yields of the desired homocoupling product. Unfortunately, the use of
the even stronger base DBU resulted in rapid formation of ammonium salts via nucleophilic
substitution and completely shut down the desired photocatalytic pathway.
Table 1. Optimization of the reaction conditions for the photoredox catalytic coupling of benzyl bromide with
Me
Zr( PDP) as the photosensitizer.
2
6
ACS Paragon Plus Environment

Page 7 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
%
Yield a/b
R
a
Entry
BIH
Base (equiv.)
pK of the base
Time (h)
a
b
(% Conv.)
6:/ (47)
c
Me
1
BIH
ꢀ
ꢀ
16
108
48
24
12
12
12
ꢀ
c
Cl
2
BIH
BIH
2,6ꢀlutidine (2.0)
2,6ꢀlutidine (2.0)
2,6ꢀlutidine (2.0)
2,6ꢀlutidine (2.0)
TEA (2.0)
14.13
14.13
14.13
14.13
18.82
ꢀ
29:/ (31)
20:/ (21)
30:/ (99)
19:15 (100)
64:10 (100)
53:13 (100)
ꢀ
c
3
c
Me
4
BIH
MeO
5
BIH
BIH
BIH
BIH
BIH
d
MeO
MeO
MeO
MeO
6
d
7
DIPEA (2.0)
DBU (2.0)
d
8
24.23
18.82
d
9
TEA (10.0)
12
27:5 (100)
a
80 b 1
pK value in MeCN from ref.
.
H NMR yield and conversion of benzyl bromide using 1,3,5ꢀtrimethoxybenzene
a
c
78 d
as the internal standard. Data from ref. . Ammonium salts were formed.
Intermittent Illumination and Quantum Yield of Photocatalysis. With improved reaction
conditions in hand, the photocatalytic nature of the reaction was initially investigated via an
intermittent illumination experiment. A correlation of irradiation time and bibenzyl yield is
shown in Figure 1. No product formation was observed in the absence of light which clearly
establishes the lightꢀdependent nature of the reaction. However, this type of experiment is not
8
1
suitable to rigorously rule out a potential photoꢀinitiated radical chain mechanism. Even though
it is difficult to envision such a pathway for the homocoupling reaction studied herein, the
quantum yield of the reaction was determined via actinometry to provide further evidence for a
photoredox catalytic mechanism. To establish the photonꢀflux of the illumination setup, the
photodissociation of Reinecke’s salt, K[Cr(NH ) (NCS) ], was used as the chemical actinometer
3
2
4
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
due to its wellꢀestablished quantum yield at 520 nm. The overall quantum yield of the photo
reaction, Φ , was then determined by dividing the number of product molecules formed by the
cat
number of photons absorbed over the illumination period (see Supporting Information for
details). The low value of Φcat = 0.004 obtained in this way for bibenzyl formation clearly favors
a photoredox catalytic mechanism over a radical chain process. The significantly lower reaction
quantum yield compared to the luminescence quantum yield of 0.08 is likely due to back electron
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MeO
●+
transfer from the reduced photocatalyst to
BIH . This nonꢀproductive pathway should be
enhanced under the reaction conditions, because the two ionic products of photoꢀinduced SET
MeO
●+
Me
1ꢀ
(
BIH and [Zr( PDP) ] ) are likely to remain in close enough proximity for facile outerꢀ
2
sphere back electron transfer as ionꢀpair formation is favorable in the nonꢀpolar reaction medium
benzene.
Figure 1. Intermittent illumination experiment for the photoredox catalytic homocoupling of benzyl bromide
1
monitored by H NMR in C D .
6
6
8
ACS Paragon Plus Environment

Page 9 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Excited State Lifetime and TD-DFT Studies. Our initial studies of the photoluminescence
Me
and electron transfer properties of Zr( PDP) did not investigate the emission lifetime or the
2
multiplicity (singlet vs. triplet) of the excited state. To address these important questions, timeꢀ
Me
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
resolved emission data were collected for Zr( PDP) . Excitation of the chromophore was
2
performed at a wavelength of 516 nm, close to the absorption maximum in the visible region of
the spectrum, and emission was detected at the maximum of the emission band at 595 nm for
Me
maximum signal intensity. A typical timeꢀresolved emission trace for Zr( PDP) in rigorously
2
deoxygenated and anhydrous THF solution at room temperature is shown in Figure 2and can
readily be fit by a single exponential decay (see also Figure S9). Data collection and fits for
multiple independently prepared samples yielded an excited state lifetime, τ , of 325 ± 10 µs for
0
Me
Zr( PDP) in THF. The strikingly long lifetime is indicative of phosphorescent emission due to
2
a spinꢀforbidden T ꢀ S transition. This result is also consistent with strong quenching of the
1
0
emission in the presence of triplet O (Figure S14). Phosphorescence lifetimes in the hundreds of
2
µs range at room temperature are rare for transition metal complexes, but have been reported for
2
III
3+
82,83
dꢀd excited states ( E ) in [Cr (phen) ] derivatives (phen = 1,10ꢀphenanthroline)
and intraꢀ
g
3
3
84
ligand chargeꢀtransfer states ( ILCT) in ZnTPP (TPP = tetraphenylporphyrin).
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
Figure 2. Timeꢀresolved emission data for Zr( PDP) in THF at room temperature (black line) detected at 595 nm
2
upon excitation at 516 nm (see inset). The red line shows the fit to a single exponential decay with τ = 325 µs and
the green trace shows the excitation pulse (516 nm).
0
The majority of phosphorescent transition metal complexes for photochemical applications
3
relies on MLCT excited states, which exhibit lifetimes in the ns to low µs range. In contrast, our
Me
previous computational analysis of the lowest energy absorption band of Zr( PDP) centered at
2
5
28 nm indicated that this feature is due to a transition with significant ligandꢀtoꢀmetal chargeꢀ
7
8
transfer (LMCT) character. With the results from timeꢀresolved emission spectroscopy in hand,
a more detailed analysis of the computational data obtained in these preliminary theoretical
studies was conducted to further characterize the full UV/vis absorption profile of the complex
and complement our initial report. Additional calculations of the electronic structure of the
lowest energy triplet state were conducted to further investigate phosphorescent emission. The
Me
ground state (S ) geometry of Zr( PDP) was obtained previously via geometry optimization
0
2
and is in excellent agreement with the geometric parameters determined by single crystal Xꢀray
7
8
diffraction (Table 3). As is often seen in transition metal complexes, the metal ligand bond
10
ACS Paragon Plus Environment

Page 11 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
85
distances in the computed structure are slightly overestimated by ca. 0.04 Å. Even though no
symmetry constraints were applied during optimization, the calculated structure exhibits D2d
Me
local symmetry with two equivalent PDP ligands if the orientation of the phenyl substituents is
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
disregarded. The maximum deviation for corresponding metalꢀligand and intraligand bond
distances between the two PDP moieties is 0.001 Å and the angle between the two ligand planes
is 89.83°. The calculated absorption spectrum obtained via timeꢀdependent density functional
theory (TDꢀDFT) at the computed S geometry is shown in Figure 3. For all TDꢀDFT
0
calculations solvent effects were taken into account using the conductorꢀlike screening model
8
6
with parameters for THF. The predicted spectrum is in good agreement with the experimental
data showing four major absorption bands in the range of 285ꢀ1000 nm, which are labeled with
Roman numerals in Figure 3. The most intense TDꢀDFT states (predicted oscillator strength fosc
>
0.03) contributing to these bands are labeled according to their state number. The transition
energies and contributing single electron excitations (including their weight) for each TDꢀDFT
state are given in Table 2.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
Figure 3. Electronic absorption spectrum of Zr( PDP) obtained via TDꢀDFT calculations (red line, FWHM of
2
2
−1
000 cm ). The stick plot indicates the positions and relative intensities of individual transitions. Transitions with
calculated oscillator strengths larger than 0.03 are labeled according to their TDꢀDFT state number. The major
contributions to each numbered state are listed in Table 2. The experimental spectrum is shown as a dotted black
line for comparison.
Table 2. Vertical Electronic Excitation Energies and Main Excitations Contributing to the Absorption Bands
of Zr( PDP) obtained via TD-DFT Calculations.
Me
2
ꢀ1
TDꢀDFT Energy / cm
Excitations
(weight)
Character
(%LMCT)
Band
fosc
a,b
State
(λ / nm)
1
1
1
18,196
0.031 223 ꢀ 224 (0.80)
0.032 223 ꢀ 225 (0.79)
IL/ LMCT
(
549.6)
(31%)
1
1
1
2
3
18,230
(548.5)
IL/ LMCT
(31%)
1
19,975
0.140 222 ꢀ 224 (0.71)
222 ꢀ 225 (0.15)
IL/ LMCT
I
(
500.6)
(31%)
2
23 ꢀ 225 (0.10)
1
1
1
4
19,982
(500.4)
0.142 222 ꢀ 225 (0.71)
222 ꢀ 224 (0.15)
IL/ LMCT
(31%)
223 ꢀ 224 (0.10)
1
1
3
26,739
0.044 219 ꢀ 224 (0.83)
219 ꢀ 225 (0.11)
IL/ LMCT
II
(
374.0)
(31%)
12
ACS Paragon Plus Environment

Page 13 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
4
26,745
373.9)
0.044 219 ꢀ 225 (0.83)
219 ꢀ 224 (0.11)
IL/ LMCT
(
(31%)
1
15
27,956
357.7)
0.136 218 ꢀ 225 (0.50)
217 ꢀ 224 (0.36)
IL/ LMCT
(
(31%)
1
2
3
1
0
29,753
0.371 219 ꢀ 226 (0.92)
IL/ LMCT
III
IV
(
336.1)
(45%)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
31,901
(313.5)
0.081 222 ꢀ 229 (0.44)
223 ꢀ 230 (0.20)
IL/ LMCT
(41%)
221 ꢀ 227 (0.16)
1
1
1
1
3
3
1
2
31,981
0.082 222 ꢀ 230 (0.41)
223 ꢀ 229 (0.18)
IL/ LMCT
(
312.7)
(41%)
2
20 ꢀ 227 (0.18)
1
32,322
0.522 222 ꢀ 229 (0.38)
221 ꢀ 227 (0.20)
IL/ LMCT
(
309.4)
(41%)
2
18 ꢀ 226 (0.16)
1
33
32,325
(309.4)
0.517 222 ꢀ 230 (0.37)
220 ꢀ 227 (0.20)
IL/ LMCT
(41%)
217 ꢀ 226 (0.16)
a
b
only excitations with a weight larger than 0.1 are shown. HOMO 223,
LUMO 224, 225 (degenerate).
A qualitative MO diagram for the S state is shown in Figure 4 and allows to characterize the
0
transitions according to the metal and ligand contributions of the donor and acceptor orbitals for
each excitation. Notably, the doublyꢀoccupied molecular orbitals 223ꢀ217 (HOMO – HOMOꢀ4)
that act as donor orbitals in the main TDꢀDFT excitations observed between 285ꢀ1000 nm are
exclusively ligand centered. In contrast, the unoccupied molecular orbitals 224ꢀ230 (LUMO –
LUMO+4) that act as acceptor orbitals in the major TDꢀDFT excitations in the same wavelength
range exhibit significant Zr character with the exception of orbital 227 (LUMO+2). The lowest
Me
energy absorption band (I) observed in the electronic absorption spectrum of Zr( PDP) is
2
comprised of four TDꢀDFT states that can formally be described as the result of electronic
transitions from HOMO and HOMOꢀ1 (orbitals 223 and 222, respectively) to a set of degenerate
LUMOs (orbitals 224 and 225). Due to the D symmetry of the molecule the four transitions can
2
d
ꢀ1
be grouped into two pairs of degenerate excitations with energies of approximately 18,200 cm
ꢀ1
(549 nm) and 19,980 cm (500 nm). Minor deviations from perfect degeneracy are due to the
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
absence of D symmetry constraints during geometry optimization. Because the metal character
2
d
1
1
of the LUMOs was calculated to be 31%, these transitions can be classified as mixed IL/ LMCT
transitions. While absorption band I (λmax(exp) = 528 nm, λmax(calc) = 500 nm) is the most
relevant feature for photocatalysis using visible light excitation, similar assignments were
performed to elucidate the nature of the remaining absorption bands IIꢀIV in the higher energy
region of the spectrum. Careful analyses of the molecular orbitals involved in the major
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
excitations for these spectral features showed mixed IL/ LMCT character for all major
1
excitations observed in the calculated spectrum between 300ꢀ400 nm. The amount of LMCT
contribution for these TDꢀDFT states is summarized in Table 2.
14
ACS Paragon Plus Environment

Page 15 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
Figure 4. Qualitative MO diagram for Zr( PDP) showing the donor and acceptor orbitals for TDꢀDFT excitations
2
computed between 285ꢀ1000 nm. Symmetry labels shown to the left of each energy level were assigned according to
the D2d point group of the core structure. The gray numbers to the right correspond to the MO number with 223
being the HOMO and 224 and 225 forming a degenerate set of LUMOs. The dotted arrows represent the single
electron excitations with the highest weight for the main TDꢀDFT states of the absorption bands IꢀIV shown in
Figure 3 and Table 2.
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Because the long emission lifetime and quenching by triplet oxygen indicated phosphorescent
emission from a triplet excited state, the energies of low lying triplet states were investigated
computationally. Consistent with the long luminescence lifetime, the lowest excited state of the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
complex in the S geometry is a triplet state composed of two degenerate TDꢀDFT excitations
0
ꢀ1
with a predicted energy of 16,080 cm (621 nm). This value is in agreement with the
experimental emission maximum at 595 nm. Three additional triplet excited states were
predicted within the energy range of the photocatalytically relevant absorption band I. A
Jablonski diagram of the main transitions at calculated wavelengths above 500 nm is shown in
Figure 5.
Me
Figure 5. Jablonski diagram of vertical electronic excitations (FranckꢀCondon states) of Zr( PDP) in the visible
2
region obtained via TDꢀDFT calculations. The isosurface plots represent difference density maps for each transition
(
red: increase in electron density, yellow: decrease in electron density). The superscript e indicates a degenerate
state. The second difference density for these states can be obtained via a C ’ operation under D symmetry.
2
2d
The structure of the lowest energy triplet state, T , was optimized starting from the S0
1
geometry using spinꢀunrestricted DFT methods (UKS3, S = 1). Important bond distances of the
16
ACS Paragon Plus Environment

Page 17 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
T geometry are shown in Table 3 alongside the parameters for S . Despite the D symmetric
1
0
2d
starting geometry, the triplet state optimization converges to a lower symmetry structure
belonging to the C point group. Most significantly, the ZrꢀN bond distance of one of the
2
v
pyridine
Me
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
PDP units is contracted by more than 0.08 Å compared to the second ligand resulting in an
axial compression along the C axis of the molecule. This lowering of the molecular symmetry
2
can be understood as a JahnꢀTeller distortion. In a simple oneꢀelectron picture, excitation of an
electron from the HOMO or HOMOꢀ1 to one of the degenerate LUMOs (corresponding to S1
and S ) followed by intersystem crossing results in a T state with a degenerate electron
2
1
configuration that undergoes an axial compression lowering the overall energy of the system and
reducing the symmetry of the molecule. At the same time, the axial distortion results in an
Me
electronic structure in which the spin density is localized on zirconium and one of the PDP
ligands (Figure 6). The spin density of 0.32 on the zirconium center for T is consistent with a
1
3
3
1
mixed IL/ LMCT state and mirrors the 31% LMCT character computed for the S and S states.
0
1
Me
Table 3. Selected Bond Distances for the S and T Geometries of Zr( PDP) obtained via DFT Calculations.
0
1
2
a
a
Singlet (RKS)
Triplet (UKS3)
Ligand 1
Ligand 2
Ligand 1
Ligand 2
Zr(1)ꢀN(1)
Zr(1)ꢀN(2)
Zr(1)ꢀN(3)
2.195
2.335
2.194
2.195
2.335
2.194
2.221
2.320
2.225
2.234
2.237
2.237
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
N(1)ꢀC(1)
C(1)ꢀC(2)
C(2)ꢀC(3)
C(3)ꢀC(4)
N(1)ꢀC(4)
C(4)ꢀC(5)
1.375
1.390
1.423
1.406
1.393
1.445
1.375
1.390
1.423
1.406
1.394
1.445
1.371
1.392
1.422
1.409
1.396
1.449
1.367
1.404
1.408
1.430
1.388
1.432
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N(2)ꢀC(5)
C(5)ꢀC(6)
C(6)ꢀC(7)
C(7)ꢀC(8)
C(8)ꢀC(9)
N(2)ꢀC(9)
1.365
1.405
1.394
1.394
1.405
1.365
1.365
1.405
1.394
1.394
1.405
1.365
1.366
1.405
1.394
1.394
1.405
1.366
1.385
1.404
1.405
1.395
1.408
1.394
C(9)ꢀC(10)
N(3)ꢀC(10)
C(10)ꢀC(11)
C(11)ꢀC(12)
C(12)ꢀC(13)
1.445
1.394
1.406
1.423
1.390
1.375
1.445
1.393
1.406
1.423
1.390
1.376
1.448
1.396
1.410
1.421
1.392
1.370
1.422
1.383
1.448
1.399
1.410
1.373
N(3)ꢀC(13)
a
No symmetry constraints were applied during geometry
optimization.
Me
Figure 6. Spin density distribution for the lowest energy triplet state of Zr( PDP) obtained via Mulliken
2
population analysis.
18
ACS Paragon Plus Environment

Page 19 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Based on our (TDꢀ)DFT studies, we propose the following hypothesis for the generation of a
Me
longꢀlived triplet excited state in Zr( PDP)2 during photoredox catalytic turnover.
Photoexcitation of the S state with visible light (green LED, λ = 521 nm, FWHM = 36 nm)
0
max
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
generates IL/ LMCT excited states, S and S , with 31% LMCT character. Under Franckꢀ
1
2
Condon conditions the initial excited state retains D2d symmetry and the change in charge
Me
distribution is delocalized over both PDP ligands via the zirconium d and d orbitals. The
xz
yz
degenerate nature of the S and S states allows for strong spinꢀorbit coupling mediated by the Zr
1
2
d and d contributions to the degenerate LUMOs and facilitates rapid intersystem crossing to
xz
yz
the triplet manifold. Vibrational cooling results in a localized T state with reduced C
1
2v
symmetry. The lowered symmetry lifts the degeneracy of the d and d orbitals and reduces
xz
yz
spinꢀorbit coupling resulting in the remarkably long lifetime observed experimentally.
Stern-Volmer Quenching Experiments. Having established the emission lifetime and a
computational model for the electronic structure of the excited state, we extended our previous
SternꢀVolmer quenching studies to obtain further insight into the mechanistic steps following
photon absorption during bibenzyl formation. A comparison of the SternꢀVolmer plots for the
R
MeO
three previously reported BIH derivatives and the new sacrificial reductant
BIH are shown
in Figure 7. The strictly linear dependence of the luminescence intensities on the concentration
of the quencher is indicative of diffusionꢀcontrolled dynamic quenching processes without preꢀ
Me
R
association of Zr( PDP) and BIH. The obtained SternꢀVolmer constants, K , and the
2
SV
MeO
corresponding quenching rate constants, k , are summarized in Table 4. The inclusion of
BIH
q
R
further establishes the strong correlation between oxidation potential of the BIH derivative and
the quenching rate supporting a photoꢀinduced SET step. Individual quenching experiments for
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the remaining components of the reaction mixture established that neither benzyl bromide nor
Me
NEt are suitable quenchers for the excited state of Zr( PDP) . The absence of quenching by
3
2
NEt is an important distinction from classic precious metal photoredox catalysts such as
3
2+
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
[
Ru(bpy) ] and Ir(ppy) and highlights the substantially more negative ground and excited state
3
3
Me
redox potentials of Zr( PDP) , which necessitate more powerful sacrificial reductants. The
2
absence of quenching by benzyl bromide rules out a potential oxidative quenching pathway. As a
R
whole, the SternꢀVolmer experiments strongly support reductive quenching by BIH as the first
step following photoexcitation.
Figure 7. SternꢀVolmer plots obtained via steadyꢀstate emission spectroscopy.
Table 4. Quenching parameters from SternꢀVolmer analysis and redox potentials.
ꢀ1
7
ꢀ1 ꢀ1
+/0
KSV / L mol
k / 10 L mol s
Eox / V vs. Fc
q
MeO
b
BIH
150,000 ± 2000
46.15
ꢀ0.25
20
ACS Paragon Plus Environment

Page 21 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Me
H
a
b
b
BIH
47,900 ± 600
14.74
1.08
0.10
ꢀ0.16
ꢀ0.10
a
BIH
3,500 ± 100
340 ± 50
Cl
b
BIH
NEt3
BnBr
0.00
0.47
n.a.
c
no quenching
no quenching
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
c
from reference 78. from reference 79. from reference 87.
Synthesis, Characterization, and Reactivity of the Reduced Photosensitizer. The
reductive quenching pathway indicated by the SternꢀVolmer quenching studies prompted
attempts to isolate the oneꢀelectron reduced form of the photosensitizer. The cyclic
voltammogram (Figure S8) guided the choice of a sufficiently strong single electron reductant
Me
+
78
for the first reversible reduction event in Zr( PDP) at ꢀ2.16 V vs. Fc /Fc.
2
Me
1ꢀ
Me
Scheme 1. Synthesis of [Zr( PDP) ] via alkali metal reduction of Zr( PDP) .
2
2
Addition of one equivalent of a freshly prepared sodium naphthalenide in THF solution to
Me
Zr( PDP) in the presence of 1,4,7,10,13,16ꢀhexaoxacyclooctadecane, 18ꢀcrownꢀ6, resulted in
2
an immediate color change from pink to dark purple. Removal of solvent and recrystallization
from THF/toluene at ꢀ35 °C yielded a dark purple crystalline solid identified as [Na(18ꢀcrownꢀ
Me
1
6)(thf) ][Zr( PDP) ] in 84% yield (Scheme 1). The H NMR spectrum recorded in THFꢀd
8
2
2
exhibits several paramagnetically broadened signals consistent with the formation of D2d
Me
1ꢀ
symmetric [Zr( PDP) ] in solution. A comparison of the electronic absorption spectra of
2
2
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Me
Me
[
Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] in THF and the starting material Zr( PDP) is shown
2 2 2
Figure 8. The most notable difference is the appearance of a broad absorption band with a
ꢀ1
ꢀ1
maximum at 649 nm (ε = 2,200 M cm ). Two additional features are observed as a band with a
ꢀ1
ꢀ1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
maximum at 516 nm (ε = 19,070 M cm ) and a shoulder around 550 nm. These two features
Me
appear at similar wavelengths as the main visible absorption band in Zr( PDP) albeit with
2
slightly reduced intensity.
Me
Me
Figure 8. Comparison of the electronic absorption spectra of [Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] and Zr( PDP) in
2
2
2
THF at room temperature.
Me
The solid state structure of [Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] was determined by Xꢀray
2
2
diffraction and representations of the molecular structures of the cationic and anionic portions
are shown in Figure 9. Comparison of the structural parameters with those obtained previously
for the neutral form of the photosensitizer allowed further evaluation of the effects of oneꢀ
electron reduction on the geometry of the complex and provided insight into the electronic
22
ACS Paragon Plus Environment

Page 23 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Me
1ꢀ
structure of [Zr( PDP) ] (Table 5). We have recently shown that the PDP scaffold can act as a
2
redoxꢀactive ligand under highly reducing conditions, which raised the question whether
Me
Me
1ꢀ
reduction in the Zr( PDP) /[Zr( PDP) ] redox couple is metalꢀ or ligandꢀcentered. The
2
2
Me
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
overall changes to the molecular structure upon reduction of Zr( PDP) are remarkably small.
2
For the metalꢀligand bond distances a slight decrease by 0.039 Å is observed in the average Zrꢀ
N bond length while an even smaller increase by 0.030 Å can be noted for the average Zrꢀ
py
Npyrrole bond length. The changes in the intraꢀligand bond distances are even subtler. Previous
Me
z
studies on the related group VI complexs [M( PDP) ] (M = Cr, Mo; z = 1ꢀ, 2ꢀ) established that
2
changes in the pyridine CꢀN bond lengths are the most reliable reporter of the redox state of the
py
8
8
PDP ligand. While the dianionic form of the ligand typically exhibits CꢀN bond lengths of
py
IV Me
2ꢀ
1
.355 – 1.365 Å, as observed for the neutral complex Zr ( PDP ) (average CꢀN = 1.361 Å),
2 py
Me
●3ꢀ
formation of a metalꢀbound radical trianion, PDP , is associated with CꢀN bond elongation
py
Me
1ꢀ
to 1.390 – 1.410 Å. The average CꢀNpy bond lengths of 1.380 Å in [Zr( PDP) ] falls inꢀ
2
between these two ranges, while changes in the remaining intraꢀligand bond distances are < 0.02
Å and cannot be considered significant within the resolution of the experiment. The experimental
data are consistent with two electronic structure descriptions: a) ligandꢀcentered reduction, in
which the radical is delocalized over both ligands or b) strong metalꢀligand covalency, where the
Me
unpaired electron occupies an orbital with significant contributions from both Zr and the PDP
ligand. This ambiguity emphasizes the difficulty in unequivocally establishing an electronic
structure based on geometric parameters alone and prompted further spectroscopic and
computational studies.
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1+
Me
1ꢀ
Figure 9. Representation of the molecular structures of [Na(18ꢀcrownꢀ6)(thf)2] (left) and [Zr( PDP) ] (right) at
2
3
0% probability ellipsoids. Hydrogen atoms were omitted for clarity.
Me
z
Table 5. Selected Bond Lengths (Å) and Angles (°) for [Zr( PDP) ] (z = 0, 1ꢀ).
2
Me
a
Me
1ꢀ b
Zr( PDP)2
[Zr( PDP) ]
2
Zr(1)ꢀN(1)
Zr(1)ꢀN(2)
Zr(1)ꢀN(3)
Zr(1)ꢀN(4)
Zr(1)ꢀN(5)
Zr(1)ꢀN(6)
2.151(3)
2.288(3)
2.183(3)
2.143(3)
2.300(3)
2.171(3)
2.181(2)
2.266(2)
2.192(2)
2.199(2)
2.243(2)
2.197(2)
N(2)ꢀC(5)
C(5)ꢀC(6)
C(6)ꢀC(7)
C(7)ꢀC(8)
C(8)ꢀC(9)
N(2)ꢀC(9)
1.364(5)
1.387(5)
1.376(6)
1.383(6)
1.383(5)
1.362(5)
1.375(2)
1.391(3)
1.394(3)
1.393(3)
1.392(3)
1.382(2)
N(5)ꢀC(32)
C(32)ꢀC(33)
C(33)ꢀC(34)
C(34)ꢀC(35)
C(35)ꢀC(36)
N(5)ꢀC(36)
1.361(5)
1.384(5)
1.381(5)
1.385(5)
1.387(5)
1.358(5)
1.389(2)
1.384(3)
1.389(3)
1.405(3)
1.379(3)
1.372(2)
24
ACS Paragon Plus Environment

Page 25 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Dihedral
88.30
88.81
N(2)ꢀZr(1)ꢀN(5)
171.76(8)
174.84(6)
a
78 b
from reference . from [Na(18ꢀcrownꢀ6)(thf) ]
Zr( PDP) ]·3THF.
2
Me
[
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
The Xꢀband EPR spectrum of [Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] recorded in frozen
2
2
THF/toluene solution at 80 K is shown in Figure 10. The observed axial signal clearly establishes
Me
1ꢀ
the S = 1/2 ground state of [Zr( PDP) ] . The data were simulated using a spin Hamiltonian
2
9
1
with g values of g = 1.998 and g = 1.971. Coupling to Zr (I = 5/2, natural abundance =
┴
ǁ
91
1
1.22%) was taken into account by a hyperfine coupling constant of A ( Zr) = 30 MHz and
┴
nicely reproduces the additional feature in the low field region of the spectrum. No hyperfine
9
1
splitting was resolved in the high field region of the spectrum. The magnitude of the Zr
hyperfine coupling and the significant deviation from the g value of the free electron (g =
e
2
.0023) towards lower g values are consistent with significant Zr contributions to the SOMO of
Me
1ꢀ
[
Zr( PDP) ] .
2
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
1ꢀ
Figure 10. Xꢀband EPR spectrum of [Zr( PDP) ] in frozen THF/toluene solution recorded at 80 K (experimental
2
parameters: frequency = 9.463 GHz, power = 0.4 mW, modulation: 2 G). See text for simulation parameters.
To support the spectroscopic studies, full molecule density functional theoretical (DFT)
calculations were conducted at the B3LYP level. The geometric parameters from geometry
optimization are in excellent agreement with the experimental structure from Xꢀray
crystallography. The spin density distribution obtained via Mulliken population analysis is
Me
1ꢀ
shown in Figure 11 and reveals the highly delocalized nature of the SOMO in [Zr( PDP) ]
2
Me
with near equal contributions of 48% from the metal center and 52% from the two PDP ligands
consistent with the structural data. Further inspection of the molecular orbital manifold shows
that the metal character in the SOMO results from close to equal contributions of the d and d
xz
yz
orbitals that form a degenerate e set under the idealized D symmetric ligand field. This result is
2d
Me
surprising given that the T state calculated for neutral Zr( PDP) , which can also be described
1
2
26
ACS Paragon Plus Environment

Page 27 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
as containing a single electron in the e set, undergoes a JahnꢀTeller distortion resulting in a
symmetry lowering to C . While this effect might be weaker for the doublet ground state of
2
v
Me
1ꢀ
[Zr( PDP) ] compared to a triplet system, calculations using the conductorꢀlike screening
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
model (COSMO) were conducted to simulate the effects of a dielectric medium. This approach
has previously been shown to favor localized electronic structures in systems containing redoxꢀ
8
9
active ligands. Consistent with these previous studies, the optimized geometry using COSMO
exhibits C symmetry most clearly established by the two different ZrꢀN bond distances
2
v
pyridine
(2.333 Å and 2.235 Å) for the two ligands. This structure is the result of delocalization of the
Me
unpaired electron between the Zr center (42%) and only one of the two PDP ligands (58%) and
removes the degeneracy of the d and d orbitals. It should be noted however, that the electronic
xz
yz
structures with D2d and C2v symmetry are likely to interconvert via vibronic coupling under
experimental conditions.
Both electronic structures with degenerate or nearꢀdegenerate dxz and dyz orbitals are
consistent with the observed axial EPR spectrum. Strong spinꢀorbit coupling in the d , d
xz
yz
manifold results in significant orbital angular momentum contributions and deviation of the gǁ
value from 2.0023. In contrast, spinꢀorbit coupling involving the dxy orbital that result in
deviations for g are reduced due to the strong destabilization of this orbital by the four pyrrolate
┴
Me
1ꢀ
π donors. Reassuringly, DFT (CPꢀSCF) calculations of the g values for [Zr( PDP) ] (g =
2
┴
2
.000, g = 1.957) are in excellent agreement with the experimental parameters. Based on the
ǁ
combined experimental and computational results the electronic structure of the reduced from of
Me
1ꢀ
the photosensitizer, [Zr( PDP) ] , is best described as a hybrid of the limiting resonance
2
III Me
2ꢀ
1ꢀ
IV Me
●3ꢀ Me
2ꢀ 1ꢀ
structures [Zr ( PDP ) ] and [Zr ( PDP )( PDP )] .
2
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
1ꢀ
Figure 11. Spin density distribution obtained via Mulliken population analysis for [Zr( PDP) ] with (bottom) and
without (top) using the COSMO approximation.
2
Me
With the isolated complex [Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] in hand, the reactivity of the
2
2
reduced photosensitizer towards benzyl bromide was explored to establish whether
Me
1ꢀ
[
Zr( PDP) ] can act as the active singleꢀelectron reductant in the photocatalytic homocoupling
2
Me
reaction. Slow addition of a dilute solution of [Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] in THF to a
2
2
solution containing 10 equivalents of benzyl bromide at room temperature resulted in an
Me
1
immediate reaction yielding bibenzyl and Zr( PDP) as the major products identified by H
2
Me
1ꢀ
NMR spectroscopy (Scheme 2). Slow addition and low concentrations of [Zr( PDP) ] ,
2
simulating the slow but continuous generation of this species under photoredox catalytic
conditions, proved to be critical. High concentrations of the reduced zirconium species resulted
28
ACS Paragon Plus Environment

Page 29 of 57
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
in poor yield of bibenzyl and formation of a new diamagnetic zirconium species in addition to
Me
Zr( PDP) . While this new complex has so far eluded isolation and further characterization,
2
1
Me
preliminary H NMR data indicate a compound with a modified PDP ligand likely due to
Me
1ꢀ
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
reaction of benzyl radical with a second equivalent of [Zr( PDP) ] . These reactivity studies
2
Me
1ꢀ
clearly show that [Zr( PDP) ] is a potent oneꢀelectron reductant that can facilitate the
2
formation of bibenzyl from benzyl bromide.
Me
Scheme 2. Reactivity of [Na(18ꢀcrownꢀ6)(thf) ][Zr( PDP) ] with benzyl bromide.
2
2
R
Synthesis, Characterization, and Reactivity of ( BI) Derivatives. The versatile PCET
2
79
pathways reported for a wide variety of benzimidazolium hydrides motivated further studies of
the fate of the sacrificial reductant after photoꢀinduced singleꢀelectron transfer (SET). The
critical role of the base observed in the optimization of the catalytic conditions indicated that
R
●
R
●+
benzimidazolyl radicals, BI , formed via deprotonation of the oxidized reductant, BIH , could
play a keyꢀrole during catalytic turnover. Electrochemical studies of 2ꢀarylꢀbenzimidazolium
halide salts suggest that the oneꢀelectron reduced species could serve as potent singleꢀelectron
+
reductants with redox potentials ranging from ꢀ1.80 V to ꢀ2.10 V vs. Fc /Fc depending on the
79
substitution patterns. While synthetic routes to 2ꢀalkyl substituted benzimidazolyl radicals via
9
0,91
reduction of benzimidazolium halide salts have been reported,
isolation and characterization
of 2ꢀaryl substituted derivatives proposed and targeted in this study have remained elusive.
29
ACS Paragon Plus Environment