Propagation chains in photocatalyzed radical nucleophilic substitutions of thiolates with perfluoroalkyl groups

Article abstract of DOI:10.1016/j.cplett.2020.137790

Quantum yields and radical chain lengths have been investigated in photocatalyzed radical nucleophilic perfluorobutyl group substitution reactions of 2-mercaptothiolate anion under three different photocatalytic systems operating in reductive and quenching manners. Important propagation chain lengths have been uncovered with photocatalysts under both scenarios, reductive and oxidative. The different overall quantum yields and chain lengths observed demonstrate the balance between the efficiency of the photoinitiation events, propagation and termination steps. Theoretical calculations provide a better understanding and throw light into whether a closed or an open catalytic cycle is operating in photoinitiated perfluoroalkyl group- thiolate coupling reaction.

Full text of DOI:10.1016/j.cplett.2020.137790

Journal Pre-proofs
Research paper
Propagation chains in photocatalyzed radical nucleophilic substitutions of thi-
olates with perfluoroalkyl groups
Al Postigo
PII:
S0009-2614(20)30705-3
DOI:
https://doi.org/10.1016/j.cplett.2020.137790
CPLETT 137790
Reference:
To appear in:
Chemical Physics Letters
Received Date:
Revised Date:
Accepted Date:
25 May 2020
10 July 2020
13 July 2020
Please cite this article as: A. Postigo, Propagation chains in photocatalyzed radical nucleophilic substitutions of
thiolates with perfluoroalkyl groups, Chemical Physics Letters (2020), doi: https://doi.org/10.1016/j.cplett.
2020.137790
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.

Propagation chains in photocatalyzed radical nucleophilic substitutions of thiolates with
perfluoroalkyl groups
Al Postigo
Departamento de Ciencias Químicas, Facultad de Farmacia y Bioquímica, Universidad
de Buenos Aires, Junín 954 CP 1113, Buenos Aires, Argentina
apostigo@ffyb.uba.ar; tel.: +54 11 4587-4312
Abstract
Quantum yields and radical chain lengths have been investigated in photocatalyzed
radical nucleophilic perfluorobutyl group substitution reactions of 2-mercaptothiolate
anion under three different photocatalytic systems operating in reductive and quenching
manners. Important propagation chain lengths have been uncovered with photocatalysts
under both scenarios, reductive and oxidative. The different overall quantum yields and
chain lengths observed demonstrate the balance between the efficiency of the
photoinitiation events, propagation and termination steps. Theoretical calculations
provide a better understanding and throw light into whether a closed or an open
catalytic cycle is operating in photoinitiated perfluoroalkyl group- thiolate coupling
reaction.
Keywords: oxidative quenching cycle; reductive quenching cycle; photoinitiated
perfluorobutylation; radical nucleophilic substitution; radical chain lengths

Introduction
Visible light photocatalysis can enable a growing number of organic transformations
through the employment of (organo)metallic photocatalysts [1] that garner light in the
way of energy in order to generate reactive species responsible for numerous
substitution or addition reactions.
The incorporation of fluoroalkyl groups by means of visible-light photocatalytic
methods has experienced a resurgence due to the facile formation of electrophilic
fluoroalkyl radicals from sources such as fluoroalkyl iodides.[2] Fluoroalkyl-substituted
compounds bear significant contrasting properties confronted with their alkylated
analogs, such as solubility, metabolic stability, and lipophilicity, which are considered
sought-after properties in pharmaceutical, agrochemical, and materials science for
designing targets of biological importance and building blocks.
In order to distinguish light-triggered processes, the term ‘photocatalytic
process’ is used when an intervening photocatalyst PC is operating in both initiation and
propagation cycles at least, whereas a ´photoinitiated process´ refers to a PC
responsible for the initiation event that produces the first radical to enter the chain loop,
and is not participating in the propagation events of the catalytic cycle.
In order to distinguish between an open or a closed catalytic cycle in a
photoredox process, we shall consider a generic mechanism for an oxidatively-induced
photoredox perfluoroalkylation of a thiolate anion (SRN reaction), as illustrated in
Figure 1A, and a reductively-induced photoredox process as shown in Figure 1B below.
Figure 1. A: The oxidative photocatalytic quenching cycle for the perfluorobutylation of
a thiolate anion, where closed (in red) and open (in blue) catalytic loops are shown. B:
Idem for a reductive photocatalytic quenching cycle
Once the photocatalyst PC is excited by visible light produces a long-lived
redox-active triplet state PC* (in the case of organometallic photocatalysts, Metal-to-
Ligand-Charge Transfer Complexes). This PC* can be oxidized by most perfluoroalkyl
iodides RF-I (Figure 1A, path I), or reductively suppressed (Figure 1B, path III) by a
wide range of organic substrates or radical ion intermediates.
In Figures 1A and 1B, closed (red) and open (blue) photoinitiated oxidative and
reductive quenching cycles are depict, respectively. The difference between the open

and closed cycles rely that on the open catalytic loop, the generation of the reacting RF
radical is also taking place in the propagation steps (blue lines, steps C or D, depending
on the photoredox process) other than in the initiation step (paths I and II, in red,
Figures 1A and 1B, respectively). In this manner, in the open catalytic loop (blue), a
radical chain reaction is taking place, whereas the closed catalytic cycle, results in no
radical propagating chain. For both quenching cycles (Figures 1A and 1B), propagation
events are equally described (steps C and D).
In the literature, either for oxidative [3-5] and reductive [6,7] photoredox
processes, depictions of closed and open catalytic cycles can be found. Arguments for
and against open catalytic processes can be encountered in photoredox reactions.[8-12]
The necessity for discriminating between an open (with chain propagation) or
closed (no radical chains) catalytic cycle is determining in optimizing reaction
conditions. In this sense, if product formation arises from closed catalytic cycles, the
structure and electrochemical properties of the photocatalysts play crucial roles. On the
other hand, if propagating chains are present, electron transfer ET rates such as those
from intermediate E to RF-I (steps C or D, Figure 1), play important roles (formation of
radical anions from the substitution/addition products can warrant chain propagation as
the transfer of the odd electron to the perfluoroalkyl halide RF-X generates more RF
radicals (vide infra)).
As for the termination steps, the encounter of two reactive intermediates (such as
E and PC.+ or PC*, Figure 1A and 1B, respectively) will derive in a non-chain catalytic
cycles. Considering that most photoinitiated / photocatalyzed reactions are considerable
fast processes, and that low catalyst loadings are required (even sub-catalytic
quantities), the presence of fast chains in photoredox reactions should be considered.
Consequently, the exploration of chain lengths in photocatalytic reactions should be
considered a relevant mechanistic tool to uncover the actual photocatalytic /
photoinitiated process.
Some authors resource to the so-called light on/off experiments to reveal the
presence of chains in photoredox processes.[13,14] However, ordinary lifetimes for
radical chain events may commonly be on the second or sub-second time scale;[15] and
if product conversion is discontinued during dark intervals is only in agreement with
chain processes that cease faster than the timescale of the analytical measurement
employed.[16,17]
Quantum yields are also invoked as the presence or lack of chain reactions in
photoredox processes. In a closed catalytic cycle (Figures 1A (red) and 1B (red)), a
maximum quantum yield of 1 is expected for the reaction, whereas in the presence of an
open cycle, a quantum yield higher than unity is observed. However, low quantum
yields can also be consequence of non-productive events responsible for significant loss
of energy.[9]
In this paper we disclose that the lower than-unity quantum yields observed can
disguise radical propagation events operating in a photoinitiated mechanism. This
finding is supported by the employment of different photocatalysts operating under
distinctive reductive and oxidative quenching manners (different PCs), where the
contrasting efficiencies of the different photocatalysts is brought to light while
comparing the initiation, propagation and termination steps.[18a]
These studies will make use of quantum yield, luminescence measurements and
computational results to study chain processes in photoinitiated radical nucleophilic
substitutions of thiolate anions with perfluoroalkyl radicals. The investigation will make
use of three different types of photocatalysts absorbing in the violet, green and red
regions of the electromagnetic spectrum, respectively.

Results and Discussion
2.3-Perfluorobutylation of thiols
The reaction type investigated is the radical nucleophilic substitution of thiolate anion
(derived form 2-mercaptoethanol 1) with perfluorobutyl radical, generated from ET-
reduction of perfluorobutyl iodide. This perfluoroalkylation reaction has been informed
by other authors to proceed under flow system very efficiently.[17b] Table 1 depicts the
yields of the perfluorobutylation product obtained and the mass balance under the three
photoinitiated systems. Optimization of reaction conditions is illustrated in ESI.
The photoinitiated perfluorobutylation reactions of 2-mercaptoethanol 1 have
been carried out employing commercially available perfluorobuyl iodide n-C4F9-I as
source of C4F9 radicals under three different photocatalysts, i.e.:
Ir[(dF(CF3)ppy]2(dtbbpy)+ (PC-1), Rose Bengal (4,5,6,7-Tetrachloro-3',6'-dihydroxy-
2',4',5',7'-tetraiodo-3H- spiro[isobenzofuran-1,9'-xanthen]-3-one (PC-2), and Zinc
phthalocyanine ZnPh (PC-3). Table 1 depicts the reaction conditions employed under
the three types of photocatalysts for substrate 1, showing the photocatalysts, maximum
irradiation wavelengths, the solvents employed, and the electron mediator or additive
used.
Table 1. Reaction conditions for the perfluorobutylation of substrate 1 (0.2 mmol) with
n-C4F9-I (3 equiv) employing different PCs under irradiation in indicated Ar-
deoxygenated solvent (3 mL), in the presence of additives at room temperature, under
constant stirring.
entry system
Photocatalyst
solvent
CH3CN
CH3CN
Additive (equiv)
Cs2CO3 (1.5)

max,a nm
1
1
Ir[(dF(CF3)ppy]2(dtbbpy)+
(PC-1)
Rose Bengal (PC-2)
Zinc phthalocyanine (PC-3)
392
2
3
2
3
525
657
Cs2CO3 (1.5)
CH3CN:DM Ascorbic acid (1.5)
collidine (1.5)
F
a.- max of irradiation (LED) employed
Inorganic base such as Cs2CO3 or collidinium ascorbate have been used as
electron mediators. Organic bases such as N,N,N`,N`-tetramethylethylene diamine

(TMEDA) and others have not been employed in order to minimize EDA complexes
formation. This latter base forms electron donor acceptor complexes with perfluoroalkyl
iodides RF-I which can be activated by visible light in the absence of photocatalysts.[17]
Table 2 illustrates the studied reaction. A good mass balance and similar product
distributions along the three photoinitiated systems is observed.
Table 2. Reaction product form the perfluorobutylation of substrate 1 (0.2 mmol) with
n-C4F9-I (3 equiv) employing different photocatalysts PCs (PC-1-3) in indicated Ar-
deoxygenated solvent (3 mL), with additives at room temperature for 20 hrs under
constant irradiation and stirring
Entry System
PC
(mol%)
Additive (equiv) Solvent
Mass
max
nm
,
Balancea %
1
2
3
1
2
3
PC-1
(0.5)
PC-2
(5)
PC-3
(5)
392b
525c
657d
Cs2CO3 (1.5
equiv)
Cs2CO3 (1.5
equiv)
Ascorbic acid
(1.5 equiv)
2,4,6-collidine
(1.5 equiv)
CH3CN
CH3CN
>99
>99
>99
CH3CN:
DMF
a.- ca. 100% mass balance was obtained, indicating that substrate converts only to product shown.
b.-violet LEDs, 3 Watt, ET = 8 mWatts
c.-green LEDs, 3 Watt, ET = 10 mWatts
d.-red LEDs, 3 Watt, ET = 16 mWatts
The presence of radical scavenger 2,2,6,6-tetramethylpiperydine oxide
(TEMPO) in the photoinitiated reactions (systems 1-3) (Scheme 1A) and 1-methyl-1-
phenyl-ethylene (Scheme 1B), suggests the presence of radicals, as these two species
are excellent radical acceptors. The photoinitiated reactions of 1 with
Ir[(dF(CF3)ppy]2(dtbbpy)+ (PC-1) and Zn-phthalocyanine (PC-3) as photocatalysts are
suppressed in the presence of 1,4-dinitrobenzene (Scheme 1C), which means that an ET
process involves radical anions (as in Figure 1B, paths II and III). This quenching is a
result of being 1,4-dinitrobenzene a more suitable electron scavenger, accepting an
electron either from Ir(II) or ZnPh radical anion (which prevents the transfer of the odd
electron to n-C4F9-I). On the other hand, the photoinitiated reaction of 1 with RB (PC-2)
as PC is not suppressed with 1,4-dinitrobenzene (Scheme 1D), showing the absence of a
better reducing species than RB* (the excited state of the photocatalyst is acting as a
reductant to C4F9-I, Figure 1A, path I). The fluorescence spectra of
Ir[(dF(CF3)ppy]2(dtbbpy)+ are quenched upon addition of a DMF solution of Cs2CO3,
which implies that an ET from CO32- to PC-1* is occurring, as in Figure S2A. On the
contrary, the fluorescence spectra of RB are suppressed upon addition of n-C4F9-I,
which is evidence of an ET step between excited RB* and n-C4F9-I (Figure S2B). On

the other hand, the fluorescence spectra of Zn-phthalocyanine PC are suppressed when
adding 2,4,6-collidinium ascorbate (Figure S2C). Neither fluorescence of
Ir[(dF(CF3)ppy]2(dtbbpy)+ (PC-1) nor Zn-phthalocyanine (PC-3) are suppressed upon
addition of n-C4F9-I (Figure S3 for PC-1), which means that no ET takes place between
PC* and C4F9-I. Therefore, PC-1 (i.e.: Ir(III)) would not oxidize to Ir(IV) transferring
an electron to C4F9-I in an oxidative quenching manner, neither PC-3 would oxidize to
ZnPhCyanine radical cation and reduce C4F9-I to C4F9 radicals in an ET process.
Scheme 1. Probe experiments for the perfluorobutylation of 1A under three
photoinitiated systems. A.-Use of TEMPO under PCs-1-3. B.-Use of 1-methyl-1-phenyl
ethelyne as probe under PCs-1-3. C.-Use of p-dinitrobenzene under PC-1 and PC-3. D.-
Use of p-dinitrobenzene under PC-2
Figures 2A-C depict the cycles for the production of C4F9 radicals with PCs-1-3.
Gibbs energy differences for the production of C4F9 radicals with the three PCs are
calculated according to the Rehm Weller equation 1.
G = Eored (D/D+) - Eored (A/A-) - E*(0,0) - w
(1)
Where Eored (D/D+) is the redox potential (reduction) of the species that is being
oxidized, Eored(A/A-) the redox potential (reduction) of the species which is reduced,
Eo,o the excited state triplet energy, and w the coulombic term.
In Figure 2A, the use of PC-1 (i.e.: Ir[(dF(CF3)ppy]2(dtbbpy)+) in a reductive
quenching cycle affords C4F9 radicals from the reduction of n-C4F9-I by ground state
Ir(II) (reductive ET). The GET is calculated to be exergonic by -0.1 V.

Figure 2. Cycles for the reductive (A and B) and oxidative (C) quenching of the
photocatalysts. In cycle A , excited (Ir[(dF(CF3)ppy]2(dtbbpy)+)* is reduced to Ir(II),
responsible for the production of C4F9 radicals. In cycle B, the radical anion of ZnPC
produces the initiating radical, whereas in cycle C excited PC* is the reductive species
that forms C4F9 radicals.
PC-1, a metal-organo-photoredox catalysts,[18b] could act in both oxidative and
reductive quenching manners. However, the luminescence of PC-1* is not quenched by
addition to n-C4F9-I (Supporting Information, Figure S3), which means that electron
transfer oxidation of Ir(III)* to n-C4F9-I does not take place, ruling out an oxidative
quenching process. A quenching of the fluorescence of PC-1* is observed by adding a
DMF-solution of Cs2CO3 (see Figure S2A, Supporting Information), which means that
an electron transfer reduction occurs between PC-1* and Cs2CO3, yielding the
.–
photocatalyst into its reduced state, i.e.: Ir (II) , and carbonate radical anion (ECO3
/CO3 2– = +1.23 +/–0.15 V). The Gibbs energy difference for this process is exergonic
by -0.08 V (see Supporting Information, section IX.2).[19]
In Figure 2B, the employment of PC-3 (zinc-phthalocyanine, red light) is also
postulated to operate in a reductive quenching cycle, where the excited state of PC-3* is
suppressed by 2,4,6-collidinium ascorbate, confirmed by Stern Volmer quenching
experiments (Supporting Information, Figure S2C). C4F9 radicals are afforded from the
reductive electron transfer of the radical anion of PC-3.- to n-C4F9-I. The calculated
Gibbs energy difference for this process is quite endergonic by +0.38 V, as depicted in
Table 3 (see also Supporting Information, Table S10)
Despite that the reduction of n-C4F9I by PC-3 (Figure 2B) is a
thermodynamically unfavorable ET [20] by about +0.38V, the base 2,4,6-collidinium
ascorbate could act as Lewis acid in order to activate the C-I bond of n-C4F9I. In fact,
Shibata and co-workers [21] reported that the photocatalyzed trifluoromethylation of
olefins with sub-phthalocyanine as photocatalyst under red light, the Gibbs energy
difference, ΔGET of the process was in excess of +0.5 V.[21] We have carried out
theoretical calculations [22] revealing that G from PC-3 to n-C4F9I is indeed an
exergonic event with a value of -12.11 Kcal/mol, suggesting the role of PC-3 in the
production of C4F9 radicals, according to Figure 2C.
Figure 2C depicts the production of C4F9 radicals by ET reduction from excited
state Rose Bengal to n-C4F9-I. This is supported by Stern Volmer quenching
experiments, where the luminescence of RB* is effectively suppressed by addition of n-
C4F9-I (Supporting Information, Figure S2B). The Gibbs energy difference GET
calculated for this process is +0.18 V, which is somewhat endergonic.

In Figure 2C, PC-2 is shown to be regenerated from its oxidized radical cation
form PC+. by electron transfer reduction from carbonate ion, yielding the carbonate
radical anion. This process is calculated to be very exergonic by G ET = +0.49 V, see
ESI, Section IX.2. We will next show (see Figure 1A), that this is not a likely electron
transfer mechanism for the re-generation of the active species of the photocatalyst PC-2.
Table 3. Difference in Gibbs energies for the ET step producing C4F9 radicals,
according to Figure 2
.+
Entry
System
PC
EPC/PC.-
(V)
EPC
ET (eV)
GET
(eV)
/PC
(V)
1
2
3
1
2
3
PC-1
PC-2
PC-3
-1.37
-0.99
-0.89
+1.69
+0.74
-
2.62
1.77
1.12
-0.1
+0.18
+0.38
We next focused our attention to investigating the presence of chain propagation
mechanisms in radical nucleophilic substitution of thiolate anion by perfluorobutyl
radicals, under different photoinitiated systems (systems I-III, as in Table 2), knowing
that quantum yields are characteristic of photoreactions under given conditions and vary
as a function of wavelength and nature of the photocatalyst.
We therefore determined the quantum yields measurements the three
photoinitiated systems, utilizing calibrated violet, green, and red LEDs (with emission
maxima at 392 nm, 525 nm, and 657 nm, respectively).
Absolute measurement of incident photon flux was achieved by means of
calibrated photodiodes which are widely used for the detection of electromagnetic
radiation in the ultraviolet, visible, and infrared range.[23] Silicon photodiodes for light
power measurement with resolution of 1 nW were employed. For LEDs with emission
maxima at 392 nm a calibrated UV-enhanced silicon sensor for 250-400 nm range
(Coherent OP-2 UV) was used; and for LEDs with emission maxima at 525 nm and 657
nm a calibrated silicon sensor for 400-1064 nm (Coherent OP-2 VIS) was employed. In
all cases, measurements of the light that passed through the bottom of the vial were
performed at the mouth of the vial (vial walls were covered with a bright white optical
film to ensure total light reflection, Figures S4-S6).The incident photon flux measured
by photodiodes was corroborated by chemical actinometry (potassium ferrioxalate for

max = 392 nm for PC-1, and max = 525 nm for PC-2, and Reinecke salt, for max
=
657 nm, for PC-3, ESI).[23,24]
For the 392 nm LED (3 Watts, which radiant power ET is 8 mWatt), the photon
flux is calculated to be 1.5787 x 1016 photons/sec. For the 525 nm LED (radiant power
ET = 10 mWatt), the value obtained is 2.6429 x 1016 photons/sec., and for the 657 nm
LED (ET = 16 mWatt), the photon flux is 5.2916 x 1016 photons/sec. The calculations
of photon flux obtained with calibrated diodes and that obtained from chemical
actinometry is illustrated in ESI, Section VII a-b. The limiting assumption that the
incident photon flux is solely absorbed by the photocatalyst can be made as substrate 1
and additives are transparent at the maxima absorption wavelength of photocatalysts
(absorption fraction  0.99).
After irradiation, the number of moles of product 1A formed is quantified by 1H
and 19F NMR spectroscopy using an internal standard (benzotrifluoride, see Table S7,
ESI, for number of moles calculation for each PC). An average of three determinations

were taken. The overall quantum yield for product formation from the perfluo
roalkylation of 1 can be calculated by dividing the moles of 1A formed by the einsteins
of photons consumed (eq 2).
The einsteins of light absorbed (mol x s-1) equal the mole of photons per seconds
times the irradiation time (s). From these data, we calculate a quantum yield value of Φ
= 40.7  0.5 for PC-1, 228  5 for PC-2, and 0.34  0.03, for PC-3 (Table 4A),
respectively.
moles of product formed
(2)
 =
einsteins of light absorbed
The observation of a quantum yield greater than unity gives conformation that
there exists a propagating chain reaction in the actual mechanism. However, the
observation of a quantum yield lower than unity does not always translates into lack of
chains, as much loss of energy can be caused by non-productive processes.[9]
According to the procedure of Yoon and colleagues,[14] the quenching fraction
Q is calculated according to eq 3:
Where KSV is the Stern-Volmer coefficient which is determined by Stern Volmer rate
quenching experiments (Supporting Information, Figure S2-S3). Applying eq. 3, the Q
values are calculated and illustrated in Table 4A. The chain length (L) is approximated
by dividing the measured quantum yield  by the quenching fraction Q. This ratio (i.e.:
/Q ) is governed by both  and Q values, and could result in large L numbers should
the quenching fraction be small. A more accurate measurement of L should be taken as
the ratio between the propagation rate and the sum of the rates of all termination steps.
If the photoinitiator is involved in a termination step (as in steps IV and II, Figures 1A
and 1B, respectively, green lines), then the photoinitiator (i.e.: photocatalyst) should be
playing a role in the actual chain length.
Quantum yield measurements (), calculation of quenching fractions (Q) and
chain lengths (L) for the perfluorobutylation reaction of 2- mercaptoethanol 1 are
summarized in Table 4A.
Table 4A. Measurements of quantum yields (), quenching fraction (Q) and chain
lengths (L) for the photoinitiated production of product 3A from the perfluorobutylation
reaction of 1 under three photoinitiated systems
a
entry
System
PC
Qb
Lc
40.70.5d
2285e
1
2
3
1
2
3
PC-1
PC-2
PC-3
0.979
0.340
0.951
41.6
670
0.36
0.34 0.03f
a.- electronic actinometry employed. Calculation using equation 1. Moles of product formed calculated
employing an internal standard by NMR integration (19F and 1H NMR spectroscopy experiments) after
b.-using equation 2
c.- L =  / Q

d.- measured at  = 392 nm. Irradiation time 150 sec (see ESI)
e.- measured at  = 525 nm. Irradiation time 150 sec. (see ESI)
f.- measured at  = 668 nm. Irradiation time 1200 sec. (see ESI)
Quantum yields much higher than unity are observed for the perfluorobutylation
reaction of 1 with PCs-1-2. A large radical chain length (i.e.: 670) is calculated for PC-
2, and a significant chain propagation length is encountered for PC-1 (i.e.: 41.6) as well.
No chain length is apparent for PC-3.
In Figure1A (oxidative quenching cycle), radicals C4F9 (RF) are postulated to be
produced either through path I (ET- reduction from excited PC-2* to n-C4F9-I, in a
closed catalytic loop (red), or ET reduction from the radical anion of the substitution
product (intermediate E) to n-C4F9-I (step C, Figure 1A, open catalytic cycle, blue). In
reductive quenching cycles (Figure 1B, for PC-1 and PC-3), radicals C4F9 (RF) are
shown to be generated either from the ET reduction from the reduced form of the
photocatalyst (either Ir(II) for PC-1, or ZnPhthalocyanine radical anion for PC-3.-) to n-
C4F9-I in a closed catalytic cycle loop (red) that re-establishes thermoneutral PC to its
active redox state or through ET reduction from the radical anion of the substitution
product (intermediate E) to n-C4F9-I (step D, Figure 1B, open catalytic cycle, blue). The
calculated exergonic nature of step C (Figure 1A) or step D (Figure 1B) (i.e.: -15.2
Kcal/mol, Table 5, first row) accounts for the open catalytic cycle and the large chain
lengths observed for PC-1 and PC-2. On the other hand, for PC-3, the endergonic nature
of step III, Figure 2B, or path II, Figure 1B, which amounts to GET +0.38 V in the
initiation event, drives the inefficiency of the photoreaction, rendering a low quantum
yield for the reaction with PC-3.
The excited state quenching fractions Q for photocatalysts PC-1 and PC-3 do not
represent the actual percentage of the excited state fraction of the photocatalysts that
give rise to the formation of the primary reacting radical C4F9 (RF), as C4F9 radicals are
produced from ET reduction from the reduced forms of these photocatalysts (i.e.: Ir(II)
for PC-1, and ZnPC.- for PC-3, respectively) as is shown in path II, Figure 1B (vide
supra). In fact, the actual fraction of the excited state of PC-1 and PC-3, Q, that give
rise to C4F9 radicals should be smaller than the calculated by equation 3, as the excited
state of PC-1 and PC-3 are reduced by electron donors (ET reduction of PC-1* and PC-
3* by the additives Cs2CO3 for PC-1, Figure S2A, and collidinium ascorbate for PC-3,
Figure S2C) to their radical anions forms (or Ir(II) for PC-1 and ZnPC.- for PC-3,
respectively), which in turn effect the ET reduction of n-C4F9-I to afford C4F9 (RF)
radicals. The scenario given by Q is overestimating the actual quenching fraction of
PC* that is responsible for C4F9 radical production. Given that process for PC-3
depicted in path II, Figure 1B very endergonic (GET +0.38 V) the actual Q´ (which
will not consider path III, Figure 1B, is expected to be smaller than the one calculated
(i.e: 0.951, entry 3, column 5, Table 4A) resulting in a larger L. Consequently, the true
value of L, i.e.: L´, is disguised by the inefficiency of the initiation event, in the case of
PC-3.
For the three photocatalysts, ET reduction from the radical anion of the
substitution product (intermediate E, Figure 1) to n-C4F9-I (steps C and D, Figure 1, in
blue) is likely to play an important role in the actual mechanism, generating more C4F9
radicals and driving the mechanism into an open catalytic cycle.[25] This is supported
by calculations (row 1, Table 5).
Electron transfer from the radical anion of the substitution product, intermediate
E (Figure 1) to the excited state of PC-3* (PhC*) has been calculated by theory to be

highly exergonic (-43.2 Kcal/mol) and exothermic event (-44.5 Kcal/mol) (Table 5, row
2). Even more exergonic is the ET reduction of the radical cation of PC-2 .+ (i.e.: RB.+)
by the radical anion of the substitution product (as in path IV, Figure 1B), intermediate
E (Figure 1), which suggests that this step is likely to be the one responsible for the
regeneration of the active state of the photocatalysts PC-2 (GET = -68.67 Kcal/mol,
Table 5) rather than the electron transfer reduction of RB.+ by carbonate anion (step IV,
Figure 2B), which has a G ET = +0.49 V (see Supporting Information, Section IX.2)
This latter has important implications in terms of mechanistic details, as the
postulated and described manners by which photocatalysts are re-generated into heir
active redox states can derive into postulating open or closed catalytic cycles.
Table 5. Simulated G, H, and S at the B3LYP/LANL2DZ theory level at 298 K
Ga
H a
Sa
Reaction
-15.2
-7.0
+27.7
-43.2
-44.5
-4.5
-68.67
-70.41
-5.51
a.- in KCal/mol
Conclusions
Being postulated propagation events similar (steps C and D, Figures 1A and 1B,
respectively) the calculations of chain lengths are biased by the efficiency of the
initiation (path I and II, Figure 1A and 1B, respectively) and termination (steps IV and
III, Figure 1A and 1B, respectively). The higher efficiency of the photoreaction with
PC-2, as compared with the other two PCs, might be attributed to a better efficient loop
(thermo-neutralization of Rose Bengal radical cation by E, step IV, Figure 1A as
opposed to path II, Figure 1B), this latter can be suggested by the redox potentials of
RB.+/RB and PC*/PC.- (+0.74 and +1.02, respectively). As a matter of fact, from theory,
it is calculated that step IV, Figure 1A, is a highly exergonic event by -69 Kcal/mol
(Table 5, row 3), compared to step II (Figure 1B).
One could assume that radical chains in the SRN reactions of substrate 1 with
C4F9 radicals are significant. The fact that quantum yield under PC-2 is the largest could
signify that for the other two PCs (i.e.: PC-1 and PC-3) other inefficient termination
events are playing significant roles.
We show in a concise manner how one could tell the presence of a radical chain
length operating in a mechanism of a certain reaction being the quantum yield of the
process lower than unity. One way around, which our manuscript shows, would be to
study the mechanism with different photocatalysts, that operate under oxidative and
reductive quenching manners. This could be useful, specially so for chemists that lack
of quantitative methods that can be easily implemented in synthetic laboratories. We
suggest that this approach to studying the mechanism of photoinitiated reactions should
be generally considered to the growing body of literature involving photoredox
catalysis.
Acknowledgements

I wish to thank Dr. Sebastian Barata-Vallejo and Dr. Damian E. Yerien for the
determination of quantum yield for the perfluorobutylation of 2-mercaptoethanol. Also
thanks are given to Dr. Victoria Cooke for carrying out theoretical calculations.
Financial agencies UBACyT, Conicet and ANPCyT are acknowledged.
References
[1] L. Marzo, S. K. Pagire, O. Reiser, and B. König, Angew. Chemie Int. Ed., 2018, 57,
10034-10072.
[2] (a) M. H. Shaw, J. Twilton, and D. W. C. MacMillan, J. Org. Chem., 2016, 81,
6898-6929. (b) D. E. Yerien, S. Barata-Vallejo, B. Camps, A. E. Cristófalo, M. E. Cano,
M. L. Uhrig and A. Postigo, Catal. Sci. Technol., 2017, 7, 2274-2282. (c) D. E. Yerien,
M. V. Cooke, M. C. García Vior, S. Barata-Vallejo, A. Postigo Org. Biomol. Chem.,
2019, 17, 3741-3746.
[3] T. C. Sherwood, N. Li, A. N. Yazdani, and T. G. Murali Dhar, J. Org. Chem., 2018,
83, 3000−3012.
[4] H. S. Han, E. Hye Oh, Y.-S. Jung, and S. B. Han, Org. Lett., 2018, 20, 1698-1702.
[5] J. Kong, J. Jiang, Y. Huang, Y.-G. Lou, X.-F. Li and C.-Y. He, Asian J. Org. Chem.,
2017, 6, 1737-1740.
[6] P. Chandu, K. G. Ghosh, and D. Sureshkumar, J. Org. Chem,. 2019, 84, 8771-8781.
[7] S. J. S. Düsel and B. König, J. Org. Chem., 2018, 83, 2802-2807.
[8] D. P. Curran, Synthesis, 1988, 417–439.
[9] M. Majek, F. Filace and A. J. von Wangelin, Beilstein J. Org. Chem., 2014, 10,
981–989.
[10] K. Matsuzaki, T. Hiromura, E. Tokunaga, and N. Shibata, ChemistryOpen, 2017, 6,
226 – 230.
[11] Sebastián Barata-Vallejo, Sergio M. Bonesi and Al Postigo, RSC Adv., 2015, 5,
62498-62518.
[12] A. Studer and D. P. Curran, Nat. Chem., 2014, 6, 765–773.
[13] D. Liu, M.-J. Jiao, X.-Z. Wang, and P.-F. Xu, Org. Lett., 2019, 21, 4745-4749.
[14] M. A. Cismesia and T. P. Yoon, Chem. Sci., 2015, 6, 5426.
[15] G. M. Burnett and H. W. Melville, Chem. Rev., 1954, 54, 225–288.
[16] L. Buzzetti, G. E. M. Crisenza, and P. Melchiorre, Angew. Chem. Int. Ed., 2019,
58, 3730 – 3747.
[17] (a) S. P. Pitre, C. D. McTiernan, W. Vine, R. DiPucchio, M. Grenier, and J. C.
Scaiano, Sci Rep. 2015, 5, 16397. (b) N- J. W. Straathof, B. J. P. Tegelbeckers, V.
Hessel, X. Wang and T. Noël, Chem. Sci., 2014, 5, 4768-4773.
[18] (a) D. E. Yerien, S. Barata-Vallejo, E. Mora Flores, A. Postigo, Cat. Sci. & Tech.
2020, doi.: 10.1039/d0cy00921k (b) R. C. McAtee, E. J. McClain, and C. R. J.
Stephenson, Trends in Chemistry, 2019, 1, 111-125.
[19] D. A. Armstrong, W. L. Waltz, A. Rauk, Can. J. Chem., 2006, 84, 1614–1619.
[20] Y. Shen, J. Cornella, F. Juliá-Hernández and R. Martin, ACS Catal., 2017, 7, 409–
412.
[21] K. Matsuzaki, T. Hiromura , H. Amii and N. Shibata, Molecules, 2017, 22, 1130.
[22] M. S. Lowry, J. I. Goldsmith, J. D. Slinker, R. Rohl, R. A. Pascal, Jr., G. G.
Malliaras, and S. Bernhard, Chem. Mater., 2005, 17, 5712-5719.
[23] H. J. Kuhn, S. E. Braslavsky, and R. Schmidt, Pure Appl. Chem., 2004, 76, 2105–
2146.
[24] G. H. Searle and G. S. Bull, D. A. House, J. Chem. Ed., 1989, 66, 605-608.
[25] S. Fukuzumi, K. Ohkubo, J. Ortiz, A. M. Gutiérrez, F. Fernández-Lázaro, and Á.
Sastre-Santos, J. Phys. Chem. A, 2008, 112, 10744–10752.

Graphical abstract

Highlights:
•radical nucleophilic substitutions of thiols with perfluoroalkyl groups
•radical chains in photocatalyzed reactions
•oxidative and reductive quenching cycles
•photocatalytic perfluoroalkylation reaction mechanism
•radical propagation chains in photocatalyzed reactions