Photochemical Reductive C–C Coupling with a Guanidine Electron Donor

Article abstract of DOI:10.1002/ejoc.201600978

The metal-free photoinduced reductive C–C coupling reactions of a number of substituted benzyl halides (15 examples) with the organic electron-donor 2,3,5,6-tetrakis(tetramethylguanidino)pyridine are evaluated. Depending on the substituents at the benzyl group, a C–C coupling product yield in the range 50–95 % is achieved. The photochemical benzyl-radical formation by homolytic N–C bond cleavage of the initially formed benzyl-pyridinium salts is the rate-determining step of these reactions. Electron-withdrawing as well as -donating substituents at the phenyl group increase the reaction rate. Quantum chemical computations did not reveal any correlation between either the enthalpy or Gibbs free energy of the N–C bond cleavage step and the experimentally determined first-order rate constants. Instead, the structural difference between the excited state generated by irradiation and the electronic ground state of the pyridinium ions could be used to rationalize the differences in the reaction rates.

Full text of DOI:10.1002/ejoc.201600978

DOI: 10.1002/ejoc.201600978
Full Paper
Metal-Free C–C Coupling
Photochemical Reductive C–C Coupling with a Guanidine
Electron Donor
Sven Wiesner,[a] Petra Walter,[a] Arne Wagner,[a] Elisabeth Kaifer,[a] and Hans-Jörg Himmel*[a]
Abstract: The metal-free photoinduced reductive C–C cou- ing substituents at the phenyl group increase the reaction rate.
pling reactions of a number of substituted benzyl halides (15 Quantum chemical computations did not reveal any correlation
examples) with the organic electron-donor 2,3,5,6-tetrakis(tetra- between either the enthalpy or Gibbs free energy of the N–C
methylguanidino)pyridine are evaluated. Depending on the bond cleavage step and the experimentally determined first-
substituents at the benzyl group, a C–C coupling product yield order rate constants. Instead, the structural difference between
in the range 50–95 % is achieved. The photochemical benzyl- the excited state generated by irradiation and the electronic
radical formation by homolytic N–C bond cleavage of the ini- ground state of the pyridinium ions could be used to rationalize
tially formed benzyl-pyridinium salts is the rate-determining the differences in the reaction rates.
step of these reactions. Electron-withdrawing as well as -donat-
Introduction
The construction of advanced molecular structures from small
building blocks requires a repertoire of element–element cou-
pling strategies. A number of C–C coupling strategies have
been established, which mostly rely on transition-metal cataly-
sis (e.g., Heck, Suzuki, Negishi, Sonogashira, Kumada, and Stille
Scheme 1. The perylene diimide (PDI) 1 used in metal-free photoinduced
C–C coupling reactions and the super-electron-donor 2 used for metal-free
photoinduced reductive cleavage of C–N and S–N bonds.
reactions, or direct dehydrative coupling),[1–3] but the evalua-
tion of new strategies, especially direct C–H activation[4] and
metal-free coupling reactions,[2,5,6] remains a topic of intense
research.
thermally stable, easy to handle and exhibits good solubility in
Photoredox catalysts are attractive because they offer the
many common organic solvents.[11] Based on the results of sev-
possibility of driving reactions under mild conditions.[7,8] In the
eral experiments and on quantum chemical calculations, a reac-
past, metal complexes (especially ruthenium and iridium poly-
tion pathway as sketched in Scheme 2 was suggested for the
pyridyl complexes) were applied as photosensitizers in combi-
reductive coupling of an alkyl bromide.[12] Herein, we examine
nation with redox-active compounds. Recent work highlights
in detail the reductive C–C coupling reaction for a number of
the importance of organic compounds for photoredox proc-
esses. For example, the perylene diimide derivative 1 (see
Scheme 1) represents a metal-free photosensitizer for C–C cou-
pling reactions between aryl halides and pyrroles.[9] Metal-free
reductive cleavage of C–N and S–N bonds by photoactivated
electron transfer with organic electron donors such as com-
pound 2 was reported by Murphy el al.[10]
We showed that the guanidine electron-donor 2,3,5,6-
tetrakis(tetramethylguanidino)pyridine[11] (3) could be used in
metal-free photoinduced reductive C–C coupling reactions.
Compound 3 is easily synthesised in three steps in high yields
starting from inexpensive, commercially available chemicals,
and is, like other guanidino-functionalised aromatics (GFAs),
[a] Anorganisch-Chemisches Institut, Ruprecht-Karls-Universität Heidelberg,
Im Neuenheimer Feld 270, 69120 Heidelberg, Germany
E-mail: hans-jorg.himmel@aci.uni-heidelberg.de
http://www.uni-heidelberg.de/fakultaeten/chemgeo/aci/himmel/
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejoc.201600978.
Scheme 2. Simplified reaction pathway for the reductive C–C coupling of
alkyl bromides with 2,3,5,6-tetrakis(tetramethylguanidino)pyridine (3).
Eur. J. Org. Chem. 0000, 0–0
1
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
substituted benzyl halides, and show that the yield and the group of the benzyl unit also experiences a low-field shift with
reaction rate varies greatly with different substituents at the respect to its position in the benzyl bromide. On the other
phenyl ring. In combination with quantum chemical calcula- hand, the aromatic protons of the benzyl unit are generally not
tions, a clearer picture of the factors that control the kinetics of significantly shifted.
photoreactions with the organic electron donor 3 is obtained.
In quartz glass cuvettes, solutions of one of the pyridinium
salts in CH3CN [(2.50 0.05) × 10–5
] were irradiated with a
M
mercury medium-pressure lamp and UV/Vis spectra were re-
corded as a function of irradiation time. The distance between
the light source and the cuvette and the light power (150 W)
were maintained in all experiments to ensure comparable light
exposure. Experiments with different concentrations showed
that the reaction rate is influenced by self-absorption effects for
Results and Discussion
The first step of the C–C coupling reaction with 3 is the forma-
tion of a pyridinium ion (Scheme 2).[12] We prepared a series
of pyridinium salts by reaction between equimolar amounts of
substituted benzyl halide and 3 in Et2O at room temperature
under exclusion of light (Scheme 3). The product precipitated
from the Et2O solution and was washed three times with 10 mL
portions of Et2O. All pyridinium salts were characterized by NMR
and UV/Vis spectroscopy, and by mass spectrometry (see Exp.
Section and the Supporting Information). The UV/Vis spectra of
all the salts were similar, and exhibit a relatively broad and in-
tense band around 435 6 nm that is responsible for their in-
tense yellow colour. Previous experiments with laser light
showed that the electronic excitation responsible for this band
initiates the photochemistry.[12] In addition, the pyridinium salts
give rise to bands at 340, 259 and 219 (shoulder) nm. In the 1H
NMR spectra, the pyridinium ion formation leads to a high-field
shift of the aromatic proton and a low-field shift of the methyl
protons of the guanidino groups. Depending on the benzyl
concentrations above 1 × 10–4
concentrations (less than 1 × 10–4
M
. On the other hand, for low
), the experimental data ar-
M
gue for a first-order kinetics and the derived first-order rate
constant values do not change with concentration.
Representative spectra for [3-Bn(F5)]Br, [3-Bn(m-OCF3)]Br and
[3-Bn(p-CF3)]Br are displayed in Figure 1. The UV/Vis spectra of
all other studied compounds are included in the Supporting
Information. In all cases, the band at ca. 435 nm, characteristic
for the pyridinium salt, decreased with increasing irradiation
time; instead, bands at 513 and 370 nm, typical for the dication
3
2+, appeared. Two isosbestic points at λ = 475 and 400 nm
were visible in all three cases, indicating clear conversion. The
extinction coefficient at 513 nm was then plotted as a function
of the irradiation time for further kinetic analysis. These plots
group, the aromatic proton exhibits a chemical shift of δ = are shown in Figure 2. It should be noted that the selection of
6.05–5.95 ppm, and the methyl protons give rise to two signals compounds presented in each of the two plots was made
in the region δ = 2.60–2.70 ppm. The singlet of the methylene purely for clarity, and does not group the compounds with re-
Scheme 3. Survey of the 15 benzyl derivatives studied in this work. The UV/Vis kinetic studies of the photoinduced C–C coupling reactions in CH3CN solution
started with the salts [3-Bn(R)]+Br–.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
2
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
spect to their properties. The data points were fitted by expo- in CH3CN then two equivalents of the benzyl bromide were
nential functions under the assumption that all compounds fol- added. The reaction mixture was stirred for 10 min at room
low a first-order rate law.[12] The rate constants determined in temperature to allow formation of the pyridinium salt, before
this way are compiled in Table 1. It can be seen that the the solutions were irradiated with a mercury medium-pressure
previously studied pyridinium salt of the unsubstituted lamp operating at 150 W. Alternatively, one could start with an
benzyl, [3-Bn]+, exhibits the smallest rate constant of equimolar mixture of pyridinium salt and benzyl bromide (see
(0.15 0.01) min–1.[13] Substitution with an electron-donating Scheme 4). The results were similar for both procedures; there-
or electron-withdrawing substituent leads to an increase of the fore, for practical reasons, the direct route starting with 3 and
reaction rate. The only exception is [3-Bn(p-NO2)]+, which does two equivalents of the substituted benzyl bromide was chosen.
not show any sign of C–C coupling upon light exposure (see By following the previously established C–C coupling proto-
the Supporting Information). Prolonged irradiation of all pyrid- col,[12] higher concentrations than in the UV/Vis experiments
inium-benzyl compounds led to bleaching, indicating photo- were used [c(3) = 0.5 c(benzyl bromide) = 4.7 × 10–3
M]. Due to
degradation of the produced 32+. The highest rate (k = self-absorption effects of the intensely coloured solutions,
1.80 0.07) min–1, more than one magnitude higher than that much higher irradiation times had to be used. Probes for GC–
of unsubstituted benzyl, was determined for [3-Bn(p-CF3)]+. MS measurements were taken out of the reaction mixture after
Generally, introduction of a substituent in the para-position
leads to higher rates than introduction of the same substituent
in the meta- or ortho-position, but there are some excep-
tions. In the case of the CF3 substituent, the rate constants
decrease in the order kpara [(1.80 0.07) min–1] > kmeta
[(1.00 0.03) min–1] >> kortho [(0.30 0.03) min–1].
defined irradiation times. Figure 3 plots the conversion from
the GC–MS measurements for the example of reductive C–C
coupling of p-CN substituted benzyl bromide, Bn(p-CN)-Br. In
some cases, for example for Bn(p-CN)-Br, the GC–MS measure-
ments give evidence for the formation of an isomeric coupling
product (equal mass, but slightly reduced retention time) in a
The yield of coupling product was then followed in GC–MS yield of 4–18 % (see Scheme 5). From the fragmentation pattern
experiments. In these experiments, compound 3 was dissolved in the mass spectrometric data, this isomeric product can be
Figure 1. UV/Vis kinetic measurements for the photoinduced C–C coupling reaction starting with (a) [3-Bn(F5)]Br (c = 2.48 × 10–5
M); (b) [3-Bn(p-CF3)]Br (c =
2.48 × 10–5
M
), and (c) [3-Bn(m-OCF3)]Br (c = 2.55 × 10–5
M) in CH3CN solutions.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
3
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
identified as a C(sp2)–C(sp3) coupled species (see Scheme 5).
In the mass spectrometric data recorded for the C(sp3)–C(sp3)
coupling product, [M]+ and [M]2+ signals were detected, the
latter being most intense. By contrast, the isomeric C(sp2)–
C(sp3) coupling product gives rise to signals at [M]+ and
[M – CH3]+, followed by signals of products from further frag-
mentation initiated by methyl group loss. The C(sp2)–C(sp3)
coupling product should arise from a rearrangement of the
benzyl radical through an H-shift, or, more likely, the attack of
a benzyl radical on an aromatic C atom. Very recently, Stephan
et al. reported the formation of C(sp2)–C(sp3) coupling products
with benzyl fluoride and an electrophilic phosphonium cation
as catalyst.[14] The relative yield of this isomeric coupling prod-
uct is likely to be concentration dependent. However, the self-
absorption effects hamper a detailed analysis of this issue;
therefore, we set aside measurements for different concentra-
tions. In a few experiments, the formation of benzaldehydes
was observed, despite strict handling of the compounds under
inert gas atmosphere. In Table 2, the results from all measure-
Figure 2. Plot of the kinetic data for the photochemical reductive C–C cou-
pling reaction of benzyl bromide and substituted benzyl bromides with
2,3,5,6-tetrakis(tetramethylguanidino)pyridine (3).
Table 1. Rate constants k determined from UV/Vis kinetic studies with CH3CN
solutions and concentrations of (2.50 0.05) × 10–5 –1, as well as calculated
M
enthalpy and free Gibbs energy changes for N–C bond cleavage of the pyrid-
inium-benzyl cations. The compounds are ordered with respect to their rate
constant.
k [min–1
]
BP86/def2-SVP
B3LYP/def2-TZVP//
BP86/def2-SVP[a]
ΔH
ΔG
ΔH
ΔG
[3-Bn(p-CF3)]+
[3-Bn(m-CF3)]+
[3-Bn(p-CN)]+
[3-Bn(p-SMe)]+
[3-Bn(m-OCF3)]+
[3-Bn(F5)]+
1.80 0.07
1.00 0.03
0.90 0.04
0.81 0.06
0.68 0.03
0.58 0.02
0.56 0.01
0.50 0. 01
0.46 0.01
0.39 0.05
0.30 0.01
0.23 0.01
152.2
160.5
147.8
159.2
165.2
149.8
148.8
158.1
161.8
157.8
157.1
160.3
163.2
162.3
147.3
86.6
92.4
83.5
94.7
94.4
84.7
86.5
84.5
97.5
92.3
94.6
94.6
100.7
98.1
83.1
105.1
119.1
108.5
121.1
120.3
104.7
103.0
121.2
122.7
119.4
115.6
121.5
124.5
124.2
108.1
39.5
54.9
44.3
56.6
49.5
39.6
38.8
47.7
58.5
54.0
51.4
55.7
62.0
60.0
43.9
[3-Bn(o-CF3,o′-F)]+
[3-Bn(p-OCF3)]+
[3-Bn(m-F)]+
[3-CH2Naph]+
[3-Bn(o-CF3)]+
[3-Bn(m-OMe)]+
[3-Bn(m-Me,m′-Me)]+ 0.22 0.01
[3-Bn]+
0.15 0.01
[3-Bn(p-NO2)]+
[b]
[a] B3LYP/def2-TZVP single-point calculation on the structure optimised with
BP86/def2-SVP. [b] No photoinduced C–C coupling reaction observed.
Figure 3. Conversion (GC–MS yield) in dependence of the irradiation time for
(a) 3 + 2 equiv. [(p-CN)Bn]Br, and (b) 3 + 2 equiv. [(m-CF3)Bn]Br.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
4
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 4. GC–MS reactions used to determine the yield of C–C coupling product. The pyridinium salt [3-Bn(R)]+Br– was either formed in situ or first isolated
and then applied.
Scheme 5. Benzyl reagents for which the formation of an isomer of the C(sp3)–C(sp3) coupling product was observed (the percentage of this isomeric product
are given underneath each structure), and possible C(sp3)–C(sp2) and C(sp2)–C(sp2) coupling products.
ments are compiled. The highest total yield of C–C coupling structurally characterised by X-ray diffraction (see the Support-
product (≥ 90 %) was obtained for Bn(p-CN)-Br and Bn(p-CF3)- ing Information for details).
Br, and the lowest yield (ca. 45 %) for Bn(m-OMe)-Br, Bn(m-
Subsequently, we analysed the role of the halide on the reac-
tion kinetics by treating benzyl chloride, benzyl bromide and
benzyl tosylate with 3. As visible from the immediately precipi-
OCF3)-Br and Bn(p-SMe)-Br. For Bn(o-CF3,o′-F)-Br and Bn(p-CN)-
Br, the C(sp3)–C(sp3) coupling products crystallised and were
tating product, the formation of [3-Bn]Tos was very fast in Et2O
solution at room temperature, similar to formation of [3-Bn]Br.
Table 2. C–C coupling yields from GC–MS measurements with CH3CN solu-
tions and concentrations of 4.7 × 10–3
M M for the benzyl
for 3 and 9.4 × 10–3
By contrast, the formation of [3-Bn]Cl was extremely slow in
Et2O at room temperature. The solvent was exchanged by tetra-
hydrofuran (THF) and the temperature raised to 50 °C to obtain
quantitative formation of [3-Bn]Cl. This result could be rational-
ised by the larger bond strength of the benzyl chloride bond
compared with the benzyl bromide or benzyl tosylate bond.
The kinetics of the reductive C–C coupling reaction was then
followed by UV/Vis spectroscopy in CH3CN solution for all three
benzyl reagents. Figure 4 shows the conversion as a function
of the irradiation time. It can be seen that the anion influences
the kinetics. The rate constant k, as estimated again by
exponential curve fitting, follows the order bromide
(0.15 0.01 min–1) > tosylate (0.05 0.02 min–1) > chloride
(0.03 0.01 min–1). At first glance, it is surprising that the reac-
tion is so much faster with bromide compared with tosylate.
Given that the bromide is a particularly good nucleophile in
CH3CN solution, it might assist the stabilisation of the radical
monocationic 3+·. In summary, the results show a significant
anion influence, but this influence on the rate constant is
smaller than the influence of substitution.
bromide. The time t is the time after which no change in yield was detected.
The total yield of C–C coupling product, C(sp3)–C(sp3) or C(sp3)–C(sp2), is
given in the last column. Formation of benzaldehydes was observed in some
cases, despite handling under inert gas atmosphere.
t [h][a] Yields (%)
Reagent
Product
Isomer
Aldehyde
Total[b]
Bn(p-CF3)-Br
Bn(m-F)-Br
Bn(p-CN)-Br
Bn(CH2Naph)-Br
Bn-Br
3.0
4.0
1.5
3.0
2.0
6.0
4.0
4.0
4.0
6.0
2.0
4.0
2.0
4.0
77
76
75
62
60
55
52
50
50
51
47
46
44
42
13
–
18
4
–
–
–
–
10
–
15
–
–
–
–
22
–
–
–
–
18
–
–
90
76
93
66
60
55
52
50
60
51
62
46
45
46
Bn(m-CF3)-Br
Bn(F5)-Br
Bn(m-Me,m′-Me)Br
Bn(o-CF3,o′-F)-Br
Bn(o-CF3)-Br
Bn(p-OCF3)-Br
Bn(m-OMe)-Br
Bn(p-SMe)-Br
Bn(m-OCF3)-Br
10
31
–
1
4
[a] Note: time t gives no information about the reaction rate, which was
analysed by UV/Vis spectroscopy in a different concentration regime. [b] Total
yield of coupling product (product + isomer).
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
5
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
SMe)Ph-(CH2)2-Ph(p-SMe) was obtained. The formation of cross-
coupling products supports the suggested reaction pathway.
Scheme 6. Cross-coupling experiments with benzyl bromide and Bn(p-SMe)-
Br in a 2:1 molar ratio in CH3CN. The cross-coupling compound is the main
product of this reaction.
Quantum Chemical Computations
The results compiled in Table 1 show that both electron-donat-
ing and electron-withdrawing substituents speed up the reac-
tion. There is clearly no correlation between the rate constant
and the electron density on the benzyl radical. DFT calculations
with the BP86 or the B3LYP functional were carried out to ob-
tain a more detailed insight into the factors that govern the
reaction rate. To keep the calculations as simple as possible,
we did not consider the influence of the anion. Although the
experiments showed that this influence is significant, it is still
smaller than the influence of substitution. First, the enthalpy
and Gibbs free energy change for N–C bond cleavage of the
pyridinium cation [3-Bn(R)]+ to give 3+· and the benzyl radical
were calculated (see Table 1). It was previously shown that the
C–H bond dissociation enthalpies for substituted toluene radi-
cals cannot be estimated from Hammett parameters,[15] and
both electron-donating and electron-withdrawing substituents
cause C–H bond weakening. In this respect, the results obtained
with toluene radicals are comparable with those obtained for
the pyridinium-benzyl ions. The highest ΔG value was calcu-
lated for [3-Bn(m-Me,m′-Me)]+ (60 kJ mol–1), and this compound
indeed exhibits a low k value. However, no clear correlation
between ΔG and k could be found. The smallest Gibbs free
energy change was calculated for [3-Bn(p-NO2)]+ (44 kJ mol–1
according to B3LYP/def2-TZVP single-point calculations on the
structures optimised with BP86/def2-SVP), which is the only
pyridinium compound that completely resisted a photochemi-
cal C–C coupling reaction.
Figure 4. Plot of the kinetic data for the photochemical reductive C–C cou-
pling reaction of benzyl bromide, benzyl chloride and benzyl tosylate (c =
2.50 0.04 × 10–5
M) with 2,3,5,6-tetrakis(tetramethylguanidino)pyridine (3).
The possibility of C–C cross-coupling reactions was then
tested. For these experiments, benzyl bromide (Bn-Br) with a
low rate constant for homocoupling [(0.15 0.01) min–1 under
the conditions used in the UV/Vis experiments, see Table 1],
was combined with Bn(p-SMe)-Br with a high rate constant
[(0.81 0.06) min–1]. The idea was to favour formation of the
cross-coupling product by applying Bn-Br in excess. In a first
experiment, a CH3CN solution of Bn-Br and Bn(p-SMe)-Br in a
2:1 molar ratio was prepared, and then one equivalent of 3
was added. After 90 min of irradiation, product formation was
analysed by GC–MS. The three C(sp3)–C(sp3) coupling products
Ph-(CH2)2-Ph, (p-SMe)Ph-(CH2)2-Ph and (p-SMe)Ph-(CH2)2-
Ph(p-SMe) were obtained in a 1:2:1 molar ratio. Under these
conditions, no C(sp3)–C(sp2) coupling product was detected.
Hence, the cross-coupling product was the main product (see
Scheme 6). In a second experiment, two equivalents of Bn-Br
was first mixed with one equivalent of 3. The solution was
stirred for 30 min, before one equivalent of Bn(p-SMe)-Br was
added. In this way the Bn(p-SMe)-Br and Bn-Br peaks in the GC–
MS could be used as internal standard. The analysis after 2.5 h
of irradiation shows a yield of heterocoupling product of 52 %
with respect to Bn(p-SMe)-Br. A 37:50:13 molar ratio of the three
As discussed previously,[12] excitation of an electron from the
coupling products Ph-(CH2)2-Ph, (p-SMe)Ph-(CH2)2-Ph and (p- HOMO of [3-Bn(R)]+ into its LUMO (S0→S1) initiates the photo-
Scheme 7. The quantum yield for C–C coupling is determined by the rate constants k′, k′–1 {relaxation into the electronic ground state of the pyridinium-
benzyl ion [3-Bn(R)]+} and k′′, which could all contribute to the total rate k. The results indicate that differences in the rates k′–1 explain the differences in the
total coupling rates k.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
6
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
chemistry. The absorbed photon energy (433 nm or 276 kJ mol–1) so that differences in the N–C bond energy should not affect
is clearly sufficient for N–C bond cleavage of all studied the rate. Indeed, no correlation was found between the experi-
compounds. As visualised in Scheme 7, the photoactivated
{[3-Bn(R)]+}* cation could either relax back to its electronic
ground state (with a rate constant k′–1) or react to 3+· and Bn(R)·
radicals (with a rate constant k′′). The vertical excitation ener-
gies (ΔEvert.) were calculated at the B3LYP/def2-TZVP level of
theory for the three pyridinium salts [3-Bn]+, [3-Bn(m-F)]+ and
[3-Bn(p-CF3)]+ and are between 304 and 310 kJ mol–1. In addi-
tion to similar energies, the oscillator strengths of the S0→S1
transitions (between 0.57 and 0.62) show no distinct differences
that can explain the observed trend in the rate constants.
In all cases {with [3-Bn(p-NO2)]+ as only exception, see
below}, the excitation of an electron can be described as a π–
π* transition (as visualised in Figure 5) that is localised on the
aromatic system and the imino nitrogen atoms of GFA 3. The
total reaction rate k for C–C coupling could clearly depend on
all three rates k′, k′–1 and k′′. However, the rate k′ seems to be
quite similar (comparable oscillator strengths for the electronic
excitations), and k′′ might also be comparable because the elec-
tronic structure of the excited states look similar. The excitation
energy is, in all cases, much higher than the N–C bond energy,
mentally derived total rate k and the calculated N–C bond en-
ergy. Then, factors that decrease k′–1 lead to an increase of the
overall reaction rate k. Clearly, k′–1 depends on the structural
difference between the molecule in the excited state, {[3-
Bn(R)]+}*, and in its ground state, and a large structural devia-
tion should lead to a low k′–1 value. Therefore, we compared
the structures optimised at the B3LYP/def2-SVP level of theory
of the first excited state and the ground state of [3-Bn]+ (dis-
playing the lowest rate) and [3-Bn(p-CF3)]+ (displaying the high-
est rate). The root mean square deviation (RMSD) between the
ground- and excited-state structures was determined by using
the Kabsch algorithm.[16] The superpositions of the excited-
state structures and the ground-state structures are illustrated
in Figure 6. The sphere dimensions reflect the relative RMSD
distribution in the molecule (i.e., the normalised RMSD with
respect to the largest intramolecular deviation) and the colour
code from green to red corresponds to absolute deviations be-
tween 0.0 and 0.7 Å. The structural deviation is much larger for
[3-Bn(p-CF3)]+ (RMSD = 0.627 Å) than for [3-Bn]+ (RMSD =
0.229 Å). This implies that k′–1 for [3-Bn(p-CF3)]+ is smaller than
for [3-Bn]+. Consequently, the C–C coupling rate should be
higher for [3-Bn(p-CF3)]+, in agreement with the experimental
results. To test another example, we calculated the structure of
the first excited electronic state of [3-Bn(m-F)]+. Again, the ex-
cited-state structure deviates significantly from the ground-
state structure, and a RMSD of 0.536 Å was derived. Hence, the
RMSD values follow the order [3-Bn]+ < [3-Bn(m-F)]+ < [3-Bn-
(p-CF3)]+, in line with the k value order. These results clearly
indicate that the reaction rate k for C–C coupling could be cor-
related with the structural deviation between ground state and
excited state of the pyridinium-benzyl ion. The trend in the
structural changes upon electronic excitation between [3-Bn]+,
[3-Bn(m-F)]+ and [3-Bn(p-CF3)]+ are also reflected in the higher
relaxation energies ΔErelax. = ΔEvert. – ΔEopt. (where ΔEopt. is the
energy of the optimised excited state), which increase in the
Figure 5. Visualisation of the isodensity surfaces of the molecular orbitals
involved in the electronic transition from the ground state (left) to the first
excited state of [3-Bn]+ (right), which can be described as a GFA-centred π–
π* transition. The isodensities correspond to 80 % of the electronic density,
and identical representations are obtained for the pyridinium compounds
[3-Bn(m-F)]+ and [3-Bn(p-CF3)]+.
Figure 6. Superpositions of excited-state and ground-state structures of the three cations [3-Bn]+, [3-Bn(m-F)]+ and [3-Bn(p-CF3)]+ from B3LYP/def2-SVP
calculations arranged in order of increasing RMSD, given below each formula. The sphere dimensions reflect the relative RMSD distribution in the molecule
(i.e., the normalised RMSD with respect to the largest intramolecular deviation) and the colour code from green to red corresponds to absolute deviations
between 0.0 and 0.7 Å.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
7
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 8. Simplified representation of possible reaction pathways following the initial absorption of a photon (ΔEvert.) by the pyridinium-benzyl ions
[3-Bn(R)]+. From the initial vibrationally excited S1 state, either cleavage of the N–C bond or an intersystem crossing to the T1 state can occur. In light of the
energy difference between the S1 and T1 states, and the qualitative correlation between the observed reaction rates and the structural relaxation energies
ΔErelax., an ISC event is likely not required for the formation of the coupling product.
order [3-Bn]+ (18 kJ mol–1) < [3-Bn(m-F)]+ (33 kJ mol–1) < [3-Bn- inium ions) not localised on the aromatic core but on the NO2
(p-CF3)]+ (60 kJ mol–1) and show that the electronic excitations substituent (see the Supporting Information). The fact that the
initially lead to different vibrationally excited states.
electronic structure of the first excited state differs from that of
the other pyridinium ions could be responsible for the missing
reactivity of [3-Bn(p-NO2)]+, despite of the smallest change in
Gibbs free energy for homolytic N–C bond cleavage. Hence, the
calculations explain, at least qualitatively, the differences in the
reaction rates.
Although the proposed reaction mechanism is consistent
with the experimental results, we cannot rule out the possibility
that other photochemical processes, as shown in Scheme 8,
occur in solution. An intersystem crossing (ISC) from the first
excited singlet states may lead to excited triplet states with
significantly longer lifetimes and a subsequent relaxation to the
triplet ground states; both of these states could cause the for-
mation of the C–C coupling products. To get a first estimate of
the importance of this process, we optimised the triplet
ground-state structures of [3-Bn]+, [3-Bn(m-F)]+ and [3-Bn-
(p-CF3)]+, which were found to be 205, 189 and 189 kJ mol–1,
respectively, higher in energy (ΔES–T) than the singlet struc-
tures, and therefore in principal exhibit enough energy for N–
C bond cleavage. No major structural differences between the
optimised S0 and T0 states were found for any of the three
compounds, but the N–C bond lengths were even shorter in
the triplet structures. Hence, it seems to be unlikely that the
observed reactivity comes from the triplet ground states.
Conclusions
The organic electron-donor 2,3,5,6-tetrakis(tetramethyl-
guanidino)pyridine (3) was applied in metal-free photoinduced
reductive C–C coupling reactions of a number of benzyl halides
(15 examples). The total yield of C–C coupling product (deter-
mined by GC/MS) varies between 45 and 93 %. By mixing
benzyl bromides with different reaction rates together and ap-
plying an excess of the slower reacting component, it was pos-
sible to obtain the cross-coupling product as the main product.
The first step of the reaction between 3 and a substituted
benzyl bromide Bn(R)-Br is formation of a pyridinium ion [3-
The relative energies of the T1 states obtained from the Bn(R)]+, and the subsequent photoinduced homolytic cleavage
vertical excitation energies for the T0→T1 transitions {ΔES–T
of the N–C bond is the rate-determining step of the reaction.
+
ΔEvert.–T = 325, 294 and 290 kJ mol–1 for [3-Bn]+, [3-Bn(m-F)]+ The kinetics of the reaction was followed by UV/Vis experi-
and [3-Bn(p-CF3)]+, respectively} show that, in case of [3-Bn]+, ments. Rate constants ranging from 0.15 0.01 min–1 (for un-
the vibrationally excited T1 state is 19 kJ mol–1 higher in energy substituted benzyl bromide) to 1.80 0.07 min–1 (for p-CF3-sub-
than the respective S1 state. This argues against an ISC event stituted benzyl bromide) were determined by exponential curve
in this compound and indicates that the photochemical C–C fitting for these first-order reactions. Substitution with electron-
coupling is possible from the first excited singlet state of the donating or electron-withdrawing groups generally leads to
pyridinium-benzyl cations.
faster reaction. The experiments show in addition that the reac-
tion rate depends on the leaving group X in benzyl-X reactants,
but the effect on the reaction rate is smaller than that associ-
ated with different substituents.
The proposed mechanism based solely on singlet excitations
even provides an explanation for the lack of reactivity of [3-
Bn(p-NO2)]+. In contrast to the other pyridinium compounds,
the p-NO2 derivative has a very low-lying first excited state
Quantum chemical calculations showed that the reaction ki-
(ΔEvert. = 220 kJ mol–1), which is (in contrast to the other pyrid- netics are not correlated with the N–C bond strength of the salt
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
8
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
bromide derivative, reaction times, analytical data and yields are
given in the Supporting Information.
[3-Bn(R)]Br. A more detailed analysis indicated that the struc-
tural variation of the first excited state of [3-Bn(R)]+ with respect
to the structure of the electronic ground state correlates with
the reaction rate. Large structural changes upon photoexcita-
tion of the cation both disfavour relaxation to the ground state
and favour formation of the C–C coupling product. Inspection
of the excited states also offers an explanation for the failure of
the coupling reaction with Bn(p-NO2)-Br, since they show that
the electron density in the first excited state of [3-Bn(p-NO2)]+
is mainly centred on the nitro group.
General Procedure for the Irradiation Experiments
UV/Vis Kinetic Measurements: In a Young quartz glass cuvette, a
(2.50 0.05) × 10–5
M acetonitrile solution of the respective pyridin-
ium salt of 3 was irradiated with a 150 W Hg medium-pressure lamp
under an inert gas atmosphere. After defined intervals of irradiation
(10 to 60 seconds depending on the used pyridinium salt), UV/Vis
spectra were recorded as a function of irradiation time to follow
the formation of 32+ at λ = 513 nm. A plateau in the formation of
32+ indicated the end of the photoreaction. The obtained UV/Vis
spectra for the determination of the rate constant k are given in
Figure 1 and in the Supporting Information.
The results of this work provide the basis for the application
of compound 3 in metal-free C–C coupling reactions.
GC/MS Measurements: Compound 3 (50 mg, 0.094 mmol) was dis-
solved in acetonitrile (20 mL; c = 4.7 × 10–3
M) and the respective
Experimental Details
benzyl bromide derivative (2 equiv., 0.188 mmol) was added. The
reaction mixture was stirred at room temperature under exclusion
of visible light to allow for the quantitative formation of the respec-
tive pyridinium salt. The reaction mixture was then irradiated under
water cooling, and GC/MS samples were taken after defined time
intervals to identify and determine the yield of the C–C coupling
products. The reaction yields are compiled in Table 2 and further
information is given in the Supporting Information.
General Procedures: All reactions were carried out under inert gas
atmosphere using standard Schlenk techniques. All solvents were
dried with an MBraun Solvent Purification System and stored over
molecular sieves prior to their use. Compound 3 [2,3,5,6-tetrakis-
(tetramethylguanidino)pyridine] was synthesised as described pre-
viously.[11] All benzyl bromide derivatives (purity 96–99 %) were
purchased from Sigma–Aldrich and used as delivered. UV/Vis meas-
urements were carried out with a Cary 5000 spectrometer. NMR
spectra were recorded with a BRUKER Avance II 600, a BRUKER
Avance II 400 or a BRUKER Avance DPX AC200 spectrometer. Elec-
trospray ionisation (ESI) mass spectra were recorded with a Finnigan
LCQ Quadrupole Ion trap. An Agilent 6890 gas chromatograph was
used for GC/MS experiments. Irradiation was achieved with a 150 W
medium-pressure Hg lamp (Heraeus TQ 150).
X-ray Crystal Structure Data: CCDC 1486231 {for [Ph(p-CN)]2-
(CH2)2} and 1486232 {for [Ph(o-CF3,o′-F)]2(CH2)2} contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre.
Acknowledgments
Details of the Quantum Chemical Computations: Quantum
chemical computations were performed with the Orca 3.0.3 suite
of programs or by using Turbomole.[16,17] For geometry optimisa-
tions and single-point calculations, a combination of the DFT func-
tionals BP86[18] or B3LYP[19] and the basis sets def2-SVP and def2-
TZVP was used and no symmetry constraints were imposed on the
structures.[20] Whenever possible, RI-J and RICOSX approxima-
tions[21] in conjunction with the appropriate auxiliary basis sets[22]
were used to accelerate the computations. Second-order derivatives
were computed to confirm that all optimised structures are station-
ary points on their respective potential energy surface. All given
electronic energies include zero-point-energy (ZPE) corrections that
were taken from the respective frequency computations. Vertical
excitation energies, oscillator strengths and geometry optimisations
of the excited states were carried out within the TD-DFT approach
at the B3LYP/def2-TZVP level of theory. Molecular orbitals were visu-
alised with IBOView.[23]
The authors gratefully acknowledge continuous financial sup-
port by the Deutsche Forschungsgemeinschaft (DFG). Parts of
this work were performed on the computational resource
bwUniCluster funded by the Ministry of Science, Research and
the Arts Baden-Württemberg and the Universities of the Feder-
ate State of Baden-Württemberg, Germany, within the frame-
work program bwHPC.
Keywords: Redox reactions · Photochemistry · Radicals ·
C–C coupling · Reaction mechanisms
[1] a) A. de Maijere, F. Diederich, Metal-Catalyzed Cross-Coupling Reactions,
2nd ed., Wiley-VCH, Weinheim, Germany, 2004; b) J. P. Corbet, G. Mig-
nani, Chem. Rev. 2006, 106, 2651–2710; c) G. Evano, N. Blanchard, M.
Toumi, Chem. Rev. 2008, 108, 3054–3131; d) J. Magano, J. R. Dunetz,
Chem. Rev. 2011, 111, 2177–2250.
[2] J. J. Mousseau, A. B. Charette, Acc. Chem. Res. 2013, 46, 412–424.
[3] R. Kumar, E. V. Van der Eycken, Chem. Soc. Rev. 2013, 42, 1121–1146.
[4] a) D. A. Colby, R. G. Bergman, J. A. Ellman, Chem. Rev. 2010, 110, 624–
655; b) T. W. Lyons, M. S. Sandford, Chem. Rev. 2010, 110, 1147–1169.
[5] S. Murarka, S. Wertz, A. Studer, Chimia 2012, 66, 413–417.
[6] S. Roscales, A. G. Csák, Chem. Soc. Rev. 2014, 43, 8215–8225.
[7] C. K. Prier, D. A. Rankic, D. W. C. MacMillan, Chem. Rev. 2013, 113, 5322–
5363.
General Procedure for the Synthesis of N-Benzyl Pyridinium
Salts of 3
Note: The synthesised products are all light-sensitive (especially in
solution), therefore the synthesis must be carried out with exclusion
of visible light.
Compound 3 (50 mg, 1 equiv., 0.094 mmol) was dissolved in diethyl
ether (15 mL) at room temperature. Subsequently, benzyl bromide
derivative (0.95 equiv.) was added to the pale-yellow solution and
the obtained bright-yellow reaction was stirred at room tempera-
ture under the exclusion of light, resulting in the formation of a
yellow precipitate. After filtration under inert gas atmosphere, the
crude product was washed with diethyl ether (3 × 15 mL) and dried
under reduced pressure. The amounts of the respective benzyl
[8] M. Peña-López, A. Rosas-Hernández, M. Beller, Angew. Chem. Int. Ed.
2015, 54, 5006–5008; Angew. Chem. 2015, 127, 5090–5092.
[9] I. Ghosh, T. Gosh, J. I. Bardagi, B. König, Science 2014, 346, 725–728.
[10] S. O′Sullivan, E. Doni, T. Tuttle, J. A. Murphy, Angew. Chem. Int. Ed. 2014,
53, 474–478; Angew. Chem. 2014, 126, 484–488.
[11] S. Stang, A. Lebkücher, P. Walter, E. Kaifer, H.-J. Himmel, Eur. J. Inorg.
Chem. 2012, 4833–4845.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
9
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[12]
[13]
[18] a) J. P. Perdew, Phys. Rev. B 1986, 33, 8822–8824; b) A. D. Becke, Phys.
Rev. A 1988, 38, 3098–3100.
[19] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) C. Lee, W. Yang,
R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[20] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[21] F. Neese, F. Wennmohs, A. Hansen, U. Becker, Chem. Phys. 2009, 356, 98–
109.
[22] F. Weigend, Phys. Chem. Chem. Phys. 2006, 8, 1057–1065.
[23] G. Knizia, J. E. M. N. Klein, Angew. Chem. Int. Ed. 2015, 54, 5518–5522;
Angew. Chem. 2015, 127, 5609–5613.
S. Stang, E. Kaifer, H.-J. Himmel, Chem. Eur. J. 2014, 20, 5288–5297.
Please note that this value deviates slightly from the value of
(0.23 0.03) min–1 given in ref.[12], due to changes in the experimental
setup concerning the cooling of the reaction solution during the irradia-
tion.
[14]
J. Zhu, M. Pérez, D. W. Stephan, Angew. Chem. Int. Ed. 2016, 55, 8448–
8451.
[15] N. K. Singh, P. L. A. Popelier, P. J. O'Malley, Chem. Phys. Lett. 2006, 426,
219–221.
[16] a) W. Kabsch, Acta Crystallogr., Sect. A 1976, 32, 922–923; b) W. Kabsch,
Acta Crystallogr., Sect. A 1978, 34, 827–828.
Received: August 9, 2016
Published Online: ■
[17] F. Neese, WIREs Comput. Mol. Sci. 2012, 2, 73–78.
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
10
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Metal-Free C–C Coupling
S. Wiesner, P. Walter, A. Wagner,
E. Kaifer, H.-J. Himmel* .................. 1–11
Photochemical Reductive C–C Cou-
pling with a Guanidine Electron Do-
nor
The yields and rate constants for the electron-withdrawing and -donating
photochemical reductive C–C coupling substituents and quantum-chemical
of a number of benzyl halides with the calculations are used for a detailed
electron-donor 2,3,5,6-tetrakis(tetra- analysis of the mechanism of metal-
methylguanidino)pyridine are studied. free C–C coupling.
The results for benzyl halides with
DOI: 10.1002/ejoc.201600978
Eur. J. Org. Chem. 0000, 0–0
www.eurjoc.org
11
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim