Metal-free C-C coupling reactions with tetraguanidino-functionalized pyridines and light

Article abstract of DOI:10.1002/chem.201304987

Herein we report on metal-free C-C coupling reactions mediated by the pyridine derivative 2,3,6,7-tetrakis(tetramethylguanidino)pyridine under the action of visible light. The rate-determining step is the homolytic N-C bond cleavage of the initially formed N-alkyl pyridinium ion upon excitation with visible light. The released alkyl radicals subsequently dimerize to the C-C coupling product. 2,3,6,7-Tetrakis(tetramethylguanidino)pyridine, which is a strong electron donor (E_1/2(CH₂Cl₂)=-0.76 V vs. ferrocene) is oxidized to the dication. For alkyl=benzyl and allyl, relatively high first-order rate constants of 0.23±0.03 and 0.13±0.03 s ^-1 were determined. Regeneration of neutral 2,3,6,7- tetrakis(tetramethylguanidino)-pyridine by reduction allows to drive the process in a cycle. Light on C-C coupling: Tetraguanidino-substituted pyridines are used for metal-free C-C coupling reactions with visible light (see figure).

Full text of DOI:10.1002/chem.201304987

DOI: 10.1002/chem.201304987
Full Paper
&
Photochemistry
Metal-Free CꢀC Coupling Reactions with Tetraguanidino-
Functionalized Pyridines and Light
Simone Stang, Elisabeth Kaifer, and Hans-Jçrg Himmel*[a]
Abstract: Herein we report on metal-free CꢀC coupling reac-
tions mediated by the pyridine derivative 2,3,6,7-tetrakis-
(tetramethylguanidino)pyridine under the action of visible
light. The rate-determining step is the homolytic NꢀC bond
cleavage of the initially formed N-alkyl pyridinium ion upon
excitation with visible light. The released alkyl radicals subse-
quently dimerize to the CꢀC coupling product. 2,3,6,7-Tetra-
kis(tetramethylguanidino)pyridine, which is a strong electron
donor (E1/2(CH2Cl2)=ꢀ0.76 V vs. ferrocene) is oxidized to the
dication. For alkyl=benzyl and allyl, relatively high first-
order rate constants of 0.23ꢁ0.03 and 0.13ꢁ0.03 sꢀ1 were
determined. Regeneration of neutral 2,3,6,7-tetrakis(tetrame-
thylguanidino)-pyridine by reduction allows to drive the pro-
cess in a cycle.
Introduction
Photoinduced reactions involving pyridinium salts were inten-
sively studied in the past. Hence, photoexcitation of a suitable
electron donor molecule (sensitizer) was shown to lead to one-
electron reduction of N-alkoxy-pyridinium acceptors. A prereq-
uisite for such an intermolecular electron transfer reaction is
the formation of a charge-transfer complex, in which the
donor and acceptor molecules generally establish p–p interac-
tions. Gould et al. studied the mechanism for NꢀO bond cleav-
age of the resulting N-alkoxy-pyridinyl radical.[1] Bending of the
O atom from the alkoxy group out of the pyridinyl ring plane
leads to mixing between s* and p* orbitals, which is essential
for a low barrier to NꢀO bond cleavage. The effect of electron-
withdrawing and donating groups on the barrier was also
studied and it was shown that the barrier is lowered by elec-
tron-donating groups attached to the radical.[1] Such photore-
actions have found important applications. The alkoxy radicals
produced under light were used, for example, to initiate radical
polymerization reactions.[2] Furthermore, the chain-amplified
photochemical fragmentation of N-alkoxypyridinium salts was
applied in organic synthesis.[3]
Scheme 1. Example for a photoreaction involving an N-alkyl-pyridinium
charge-transfer salt. The example shows at the same time the strength of
the NꢀC bond, which remains intact.
such as that in Scheme 1.[6] However, in this case the pyridini-
um-methyl bond remains intact, proving again its stability. In
two fundamental publications, Katritzky, Eyler et al. studied the
collisionally-activated dissociation of a series of N-alkylpyridini-
um ions in the gas phase.[7] The products were either pyridine
and alkyl cations or pyridinium ions and alkenes (see
Scheme 2). They compared their experimental results with
Compared with N-alkoxy-pyridinyl radicals, N-alkyl-pyridinyl
radicals are more stable,[4] and CV experiments prove the rever-
sible reduction of several N-alkyl-pyridinium salts.[5] Conse-
quently, up to date such compounds have found less applica-
tions. Kochi et al. reported photoreactions with pyridinium tet-
raalkylborate charge-transfer salts (some also react thermally)
[a] S. Stang, E. Kaifer, Prof. H.-J. Himmel
Anorganisch-Chemisches Institut
Ruprecht-Karls-Universitꢀt Heidelberg
Im Neuenheimer Feld 270, 69120 Heidelberg (Germany)
Fax: (+49)6221-545707
E-mail: hans-jorg.himmel@aci.uni-heidelberg.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304987.
Scheme 2. Collisionally activated gas-phase fragmentation of pyridinium
ions.
Chem. Eur. J. 2014, 20, 5288 – 5297
5288
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
AM1 (Austin Model 1) calculations. With relevance to the work
reported herein, they found that the reaction enthalpies for
alkyl cation formation decrease in the order R=Me
(see Figure 1a). The bond distances within the cationic unit are
in line with the resonance structures shown in Scheme 4a, indi-
cating delocalization of the positive charge into the guanidino
groups. Alkylation of 1 leads to a decrease of the HOMO–
LUMO gap. Consequently, the N-alkyl pyridinium salts are in-
tensively yellow colored. A strong band centered at 433 nm in
the UV/Vis spectrum of [1-CH2Ph]+ can be assigned to the
HOMO!LUMO transition (see Figure 2a and Figure S1 in the
Supporting Information for spectra of salts containing the
cation [1-CH3]+), and this transition is clearly involved in the
photochemistry. In 1, a band at a significantly higher energy of
360 nm was assigned to the
(410 kJmolꢀ1
)
>
H2C=CH-CH2 (213 kJmolꢀ1)
>
PhCH2
(172 kJmolꢀ1). In the case of R=PhCH2, the initially formed
+
PhCH2 cation could subsequently rearrange to the more
stable tropylium cation, leading to a further decrease of the re-
action enthalpy to 126 kJmolꢀ1
.
We recently reported the synthesis of the pyridine derivative
2,3,5,6-tetrakis(1,1,3,3-tetramethylguanidino)pyridine, (see
1
Scheme 3), which turned out to be a strong organic electron
HOMO!LUMO
excitation.[8]
Using density functional theory
(DFT) calculations (B3LYP/6-
311G**), the HOMO–LUMO gap
was estimated to be 4.31 eV
(288 nm) in
1
and 3.38 eV
(367 nm), 3.48 eV (356 nm) and
3.35 eV (370 nm) in [1-Me]+,
[1-CH2Ph]+ and [1-CH2CH=CH2]+,
respectively. Hence, the calcula-
tions predict correctly a decrease
Scheme 3. Metal-free photocoupling reaction of alkyl halides (R=benzyl or allyl in this work) with 2,3,5,6-tetra-
kis(1,1,3,3-tetramethylguanidino)-pyridine, 1, as discussed in this work. No reaction was observed for the related
electron donor 1,2,4,5-tetrakis(1,1,3,3-tetramethylguanidino)-benzene, 2.
donor with E1/2(CH2Cl2)=ꢀ0.76 V
vs. Fc/Fc+ (Fc=ferrocene) for
two-electron oxidation.[8] Oxida-
tion is fully reversible. Herein we
show that visible-light irradiation
of solutions of 1-alkyl-2,3,5,6-tet-
rakis(1,1,3,3-tetramethylguanidi-
no)-pyridinium salts, formed
from 1 and alkyl halides) leads
to CꢀC coupling reactions. Such
reactions are not feasible for
other pyridinium salts, and
might pave the way for the
Scheme 4. a) Resonance structures for [1-R]+ showing the delocalization of the positive charge into the guanidino
broader usage of 1 in metal-free
photochemical CꢀC coupling re-
actions. The related compound 2
(1,2,4,5-tetrakis(1,1,3,3-tetrame-
groups. b) Lewis structure of the trication [1-R]3+
.
thylguanidino)benzene, see Scheme 3),[9] which is similar to
1 in its redox potential, does not mediate CꢀC coupling reac-
tions. In this work we focus on some aspects of the mecha-
nism of this new photoreaction.
of the HOMO–LUMO gap upon alkylation (73 nm according to
experiment versus 68 nm according to the calculations for
benzyl), and only the absolute values deviate. The HOMO
energy decreases upon alkylation of 1, reducing the electron
donor capacity. Hence the CV curves show a shift by 0.23 V to-
wards higher potentials for the first (two-electron) oxidation
waves upon alkylation (with benzyl). Nevertheless, [1-CH2Ph]+
still is a strong electron donor, and [1-CH2Ph]Br can readily be
oxidized, for example, with I2 to give the salt
[1-CH2Ph]BrI1.5(I3)0.5. The structures of the trications [1-CH3]3+
and [1-CH2Ph]3+ as derived from an X-ray diffraction analysis
are displayed in Figure 1b and c. In the trication [1-CH3]3+, the
short distances N1ꢀC1 and N10ꢀC5 (1.276(3) and 1.287(3) ꢁ)
together with the long N1ꢀC7 and N10ꢀC22 bond distances
argue for localization of two positive charges mainly on the
two guanidino groups next to the N ring atom. The third posi-
Results and Discussion
Alkylation of 1
The first step in the CꢀC coupling reaction with 1 is the forma-
tion of an alkyl-pyridinium salt. Therefore we started our work
with the preparation of the N-benzyl, N-allyl or N-methyl pyri-
dinium salts of compound 1 by reaction between 1 and one
equivalent of alkyl (benzyl, allyl, or methyl) halide or triflate (tri-
fluoromethane-sulfonate, O3SCF3). The structure of the salt
[1-Me]I in the solid state was determined by X-ray diffraction
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5289
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 2. a) Comparison between the UV/Vis spectra of 1 and
[1-CH2Ph]+Brꢀ. b) Illustration of the isodensity surfaces of HOMO and LUMO
in [1-CH2Ph]+ (B3LYP/6-311G** calculations). The H atoms were omitted for
clarity.
of [1-R]+ were stored for some time under sun light, a color
change from yellow to red signaled a light-induced redox pro-
cess. A small number of red-colored crystals of [1-Me]Tf3 and
[1-CH2Ph]Br3 were isolated from [1-Me]Tf and [1-CH2Ph]Br solu-
tions exposed to sun light for several days (see Figure S2 in
the Supporting Information and the Experimental Section).
These results were the first hints of an interesting photochem-
istry.
Figure 1. Structures of a) The [1-Me]+ cation in the salt [1-Me]I, b) the
[1-Me]3+ trication in the salt [1-Me](F3CSO3)3, and c) the [1-CH2C6H5]3+ trica-
tion in the salt [1-CH2C6H5]BrI1.5(I3)0.5 (hydrogens omitted). Vibrational ellip-
soids drawn at the 50% probability level. Selected bond distances (in ꢁ) for
[1-Me]+: N1ꢀC1 1.377(10), N1ꢀC7 1.345(12), N2ꢀC7 1.360(11), N3ꢀC7
1.362(13), N4ꢀC2 1.410(11), N4ꢀC12 1.306(11), N5ꢀC12 1.375(11), N6ꢀC12
1.372(11), N7ꢀC4 1.393(11), N7ꢀC17 1.278(13), N8ꢀC17 1.392(13), N9ꢀC17
1.397(11), N10ꢀC5 1.335(11), N10ꢀC22 1.337(11), N11ꢀC22 1.354(11),
N12ꢀC22 1.345(12), N13ꢀC1 1.325(11), N13ꢀC5 1.389(10), N13ꢀC6 1.493(10),
C1ꢀC2 1.422(12), C2ꢀC3 1.388(12), C3ꢀC4 1.419(12), C4ꢀC5 1.401(11). Select-
ed bond distances (in ꢁ) for [1-Me]3+: N1ꢀC1 1.276(3), N1ꢀC7 1.385(3),
N2ꢀC7 1.323(3), N3ꢀC7 1.334(3), N4ꢀC2 1.300(3), N4ꢀC12 1.390(3), N5ꢀC12
1.326(3), N6ꢀC12 1.329(3), N7ꢀC4 1.336(3), N7ꢀC17 1.351(3), N8ꢀC17
1.345(3), N9ꢀC17 1.344(3), N10ꢀC5 1.287(3), N10ꢀC22 1.378(3), N11ꢀC22
1.346(3), N12ꢀC22 1.333(3), N13ꢀC1 1.381(3), N13ꢀC5 1.377(3), N13ꢀC6
1.478(3), C1ꢀC2 1.503(3), C2ꢀC3 1.405(3), C3ꢀC4 1.370(3), C4ꢀC5 1.500(3).
Selected bond distances (in ꢁ) for [1-CH2C6H5]3+: N1ꢀC1 1.281(6), N1ꢀC6
1.378(6), N2ꢀC6 1.336(6), N3ꢀC6 1.344(6), N4ꢀC2 1.317(6), N4ꢀC11 1.377(6),
N5ꢀC11 1.338(6), N6ꢀC11 1.327(6), N7ꢀC4 1.316(6), N7ꢀC16 1.373(6), N8ꢀC16
1.327(6), N9ꢀC16 1.349(6), N10ꢀC5 1.294(6), N10ꢀC21 1.380(6), N11ꢀC21
1.332(6), N12ꢀC21 1.336(6), N13ꢀC1 1.370(6), N13ꢀC5 1.378(5), N13ꢀC26
1.474(6), C1ꢀC2 1.505(6), C2ꢀC3 1.382(6), C3ꢀC4 1.394(6), C4ꢀC5 1.495(6).
The isodensity surfaces of the HOMO and LUMO orbitals of
[1-CH2Ph]+ are visualized in Figure 2b (B3LYP/6-311G** calcula-
tion). The pyridinium ring orbitals, but also the guanidino
group orbitals contribute to the HOMO and LUMO. In addition,
it can be seen that in the LUMO the benzyl group is also in-
volved. This mixing between the p* orbital of the pyridine ring
with guanidino and benzyl orbitals is of importance for the ob-
served photochemistry. It is also worth mentioning that both
the UV/Vis spectra (measured at various concentrations) as
well as the crystal structure data show no sign for the forma-
tion of a charge-transfer complex, in which the alkylated
1 would be electron donor and acceptor at the same time (see
below).
Photochemistry
tive charge is delocalized in the “other half” (1,3-bisguanidino-
allyl) of the molecule (see the Lewis structure in Scheme 4b). A
similar electronic situation is found for [1-CH2Ph]3+. The dihe-
dral angle between the “CN2 plane” of the C(NMe2)2 group of
each guanidino substituent and the plane of the central het-
erocycle remains large, most likely for steric reasons. Solutions
of [1-R]3+ are, like solutions of 12+, red-colored. When solutions
Solutions of the pyridinium salts in CH3CN were then irradiated
with a 150 W medium-pressure Hg lamp, either in a quartz
glass cuvette or in a Schlenk flask. The behavior was very dif-
ferent for [1-CH2Ph]+ and [1-CH3]+. In the case of [1-CH2Ph]+,
a fast photoreaction was observed, while [1-CH3]+ reacts very
slowly. After 15 min of irradiation of a 1.76·10ꢀ5 m solution of
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5290
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
very small. The isolation of a few number of [1-CH2Ph]Br3 crys-
tals upon expose of solutions of [1-CH2Ph]Br to sunlight for
several days is therefore the result of an equilibrium
([1-CH2Ph]++12+$[1-CH2Ph]3++1) which lies on the left side,
but leads to formation of some [1-CH2Ph]3+ if only a small
amount of 12+ is present. The absorption intensity at 513 nm
was plotted as a function of irradiation time, and the data
could be fitted satisfactorily by an exponential function (see
Figure 3c), in line with a first-order kinetics with a rate con-
stant k of approximately (0.23ꢁ0.03) minꢀ1. The reaction was
repeated with different concentrations of [1-CH2Ph]+Brꢀ. An in-
crease in the concentration from 1.76·10ꢀ5 m to 2.47·10ꢀ5 m did
not affect the k value. On the other hand, a further increase to
3.36·10ꢀ5 m led to a slightly reduced reaction rate. In this case
the reaction was completed in ca. 20 min. These results can be
explained by the reduction of light intensity by absorption.
Prolonged irradiation times led in all cases to a secondary reac-
tion and bleaching of the solution. The absorptions at 513
(broad) and 369 nm vanished and absorptions at 241, 270, and
341 nm (the latter with a shoulder at 370 nm) appeared
(Figure 4). However, the absence of isosbestic points in this
Figure 3. a) Photo of a solution of [1-CH2C6H5]+Brꢀ in CH3CN (2.47·10ꢀ5 m)
before (left) and after irradiation with a Hg high pressure lamp (150 W) for
15 min. b) UV/Vis spectra of a [1-CH2C6H5]+Brꢀ solution in CH3CN for differ-
ent irradiation times. c) Plot of the absorption intensity at 513 nm as a func-
tion of the irradiation time together with an exponential fit.
Figure 4. Effect of prolonged irradiation of a solution of [1-CH2C6H5]+Brꢀ in
CH3CN (2.47·10ꢀ5 m) with broadband light. The dication 12+, which is formed
after 30 min of photolysis, is consumed after 480 min irradiation.
second photoreaction signaled a more complex process that
might lead to product mixtures. The results demonstrate that
prolonged photolysis times, at least with broadband light,
have to be avoided.
[1-CH2Ph]+Brꢀ in CH3CN in a quartz glass cuvette, the solution
had changed color from yellow to red (see photo in Figure 3a),
and the spectroscopic data signaled completion of the reac-
tion. UV/Vis spectra recorded after different irradiation times
are shown in Figure 3b. The absorptions due to [1-CH2Ph]+ at
433 and 340 nm completely vanished, and new absorptions at
513 (broad) and 369 nm appeared. The spectra clearly showed
that all of the [1-CH2Ph]+ cations were consumed after 15 min.
Isosbestic points were located at 475 and 400 nm, indicating
formation of a single product with absorptions in the visible
region. The photoproduct can be identified unambiguously as
Figure 5 shows the UV/Vis spectra recorded under similar
conditions for a 1.84·10ꢀ5 m solution of [1-CH3]+Iꢀ in CH3CN.
The concentration lies in the low-concentration region which
should allow the determination of the rate constant. It can be
seen that the photoreaction proceeds significantly slower, and
even after 360 min of irradiation with broadband light about
50% of the [1-CH3]+ ions remained. After 30 min of photolysis,
the appearance of new absorptions indicated formation of 12+
in small quantities. However, as already observed in the case of
the reaction with [1-CH2Ph]+Brꢀ, longer irradiation times did
not lead to an increase of the dication concentration, since it is
itself consumed in a slow photoreaction.
the dication 12+ [8] The trication [1-CH2Ph]3+, which exhibits
.
bands at 492 and 358 nm with a different intensity ratio, can
be ruled out. Due to the large half-width of the bands, it could
in principal be formed as side product, but the presence of
clear isosbestic points shows that its concentration can only be
Subsequently, we repeated the experiments with
a 2.7·10ꢀ5 m solution of the allyl-pyridinium salt [1-CH2-CH=
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5291
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 5. UV/Vis spectra recorded for [1-Me]+Iꢀ solutions (1.84·10ꢀ5 m) in
CH3CN in dependence of the photolysis time.
CH2]+Brꢀ in quartz glass cuvettes. From the experience with
[1-CH2Ph]+Brꢀ we kept the concentration low to avoid deceler-
ation of the reaction rate due to self-absorption effects (the
UV/Vis spectra collected in our experiments are shown in Fig-
ure S3 in the Supporting Information). Again, the presence of
isosbestic points indicated a clean photoreaction in the first
minutes of irradiation, which leads to the same product,
namely the dication 12+. After 20 min of irradiation, all [1-CH2-
CH=CH2]+ ions had reacted. A plot of the absorption intensity
versus irradiation time (see Figure S4 in the Supporting Infor-
mation) argues for a first-order kinetics, in agreement with the
results obtained with the benzyl compound. However, the re-
action rate was in the case of [1-CH2-CH=CH2]+Brꢀ (0.13ꢁ
Figure 6. Plot of the GC/MS measurements and photos before and after irra-
diation of a [1-CH2C6H5]+Brꢀ solution in CH3CN (4.7·10ꢀ3 m) for a period of
120 min.
benzyl). In Figure 6, the consumption of PhCH2Br and the gen-
eration of the coupling product PhCH2CH2Ph (measured with
GC/MS) are plotted together as a function of the irradiation
time. From this plot it can be seen that prolonged irradiation
times did not decrease the amount of PhCH2CH2Ph, meaning
that the product is not involved in the second slower photore-
action in which 12+ is consumed. Up to an irradiation time of
ca. 150 min, the only species detected with GC/MS were the
coupling product PhCH2CH2Ph and the reactant PhCH2Br. Neu-
tral 1, [12+](Brꢀ)2 and [1-CH2Ph]+Brꢀ got stuck on the column.
After 150 min of irradiation, 60% of the PhCH2Br reactant mol-
ecules had reacted to the desired product, leaving 40% of the
initial PhCH2Br molecules in solution. The UV/Vis spectra
showed already at this stage complete consumption of all
[1-CH2Ph]+ ions. Longer photolysis times did not improve the
yield of the couple product, and only led to a slow decrease in
the number of reactant molecules. At the same time, the for-
mation of small amounts of benzaldehyde was detected via
GC/MS. Since we excluded H2O or O2 rigorously, we think that
the oxygen stems from reactions of the radicals produced by
irradiation (see below) with the glass walls of the reaction
vessel. In addition, another side reaction must occur to explain
the presence of significant amounts of unreacted PhCH2Br.
Since the alkylation of 1 proceeds quantitatively at room tem-
perature, this side reaction must cause the loss of some of the
formed neutral molecules of 1. This point is addressed below.
The same results as with a mixture of [1-CH2Ph]+ and
PhCH2Br were obtained when mixtures of 1 and two equiva-
lents of PhCH2Br were irradiated. On the other hand, in experi-
ments in which 1 was replaced by 2, which is comparable in
0.03 minꢀ1
)
slightly lower than for [1-CH2Ph]+Brꢀ (0.23ꢁ
0.03 minꢀ1). In full agreement with the other experiments, pro-
longed irradiation times led to a decrease of the absorptions
due to 12+ (Figure S5 in the Supporting Information).
Then we carried out the photoreaction with [1-CH2Ph]+Brꢀ
in Schlenk flasks and with higher concentrations of the reac-
tant. Under these conditions, longer irradiation times of 2 h
were necessary to complete the reaction. Using GC/MS, the
second species, besides 12+, generated under the action of the
light was identified as the carbon–carbon coupling product
PhCH2CH2Ph. An overall reaction equation, which is in agree-
ment with the experimental findings, is given in Scheme 5. Ac-
Scheme 5. Photoreaction of [1-R]+ without and with addition of RBr
(R=benzyl).
cording to this equation, it should be possible to increase the
yield of 1,2-diphenylethane by addition of one equivalent of
benzyl bromide, which reacts with 1 to give further photoac-
tive [1-CH2Ph]+. This was indeed the case. Broadband photoly-
sis of an equimolar mixture of PhCH2Br and [1-CH2Ph]+ gave
1,2-diphenylethane in a yield of up to 60% (with respect to
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5292
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
its reduction potential,[9] no reaction occurred. This result clear-
ly shows that fast electron transfer from the organic donor to
the alkyl is only possible if a bond is established (resembling in
this aspect the “inner-sphere” mechanism for electron transfer
in transition metal chemistry). In further experiments, we irradi-
ated a solution containing [1-CH2Ph]+ and allyl bromide
(C3H5Br) in an equimolar ratio. The CꢀC coupling products
were 1,2-diphenylethane (PhCH2CH2Ph) and 4-phenyl-1-butene
(PhCH2CH2CH=CH2), which were formed in a 2:1 molar ratio.
1,5-Hexadiene (H2C=C(H)CH2CH2CH=CH2) was not found. Simi-
larly, irradiation of mixtures of [1-CH2CH=CH2]+ and PhCH2Br
only produced 1,5-hexadiene and 4-phenyl-1-butene, but no
1,2-diphenylethane. All these results clearly show that irradia-
tion leads to the cleavage of the pyridine NꢀC alkyl bond in
the first place. They also indicate that cross coupling reactions
(and maybe more effective ring-closing reactions) are possible.
likely because they are subjected to fast reactions. The first-
order kinetics indicates that the cleavage of the NꢀC bond by
light is the rate-determining step.
We carried out quantum chemical calculations (B3LYP/
6-311G**) to compare the gas-phase NꢀC bond fragmentation
of the N-benzyl pyridinium ion with that of the cation
[1-CH2Ph]+. According to our calculations, the formation of
alkyl cations by heterolytic NꢀC bond cleavage is associated
with a change of the Gibbs free energy (298 K, 1 bar) of
+156 kJmolꢀ1. By contrast, the homolytic NꢀC bond cleavage
leading
to
alkyl
radicals
is
strongly
disfavored
(DG=+327 kJmolꢀ1). As mentioned in the Introduction, colli-
sionally induced NꢀC fragmentation of N-alkyl pyridinium ions
in the gas phase leads to formation of alkyl cations,[7] a result
which is in good agreement with our calculations. In the case
of [1-CH2Ph]+, the situation changes drastically (see
Scheme 7). Hence, the pathway leading to the benzyl radical
(homolytic NꢀC bond fragmentation) is now preferred. It is as-
sociated with a relatively low DG value of +73 kJmolꢀ1. On
the other hand, alkyl cation formation is highly endergonic
(DG=+306 kJmolꢀ1). Is the energy of the excited state of
[1-R]+ higher than the energy needed for homolytic NꢀC bond
fragmentation? To obtain a preliminary answer to this ques-
tion, we assumed the lowest-energy triplet state to be similar
to the open-shell first-excited singlet state. Indeed, the SOMOs
of the triplet state agree nicely with the HOMO and LUMO of
the singlet ground-state (Figure S6 in the Supporting Informa-
tion). Under this assumption, the energy difference between
the excited state and the ground state can be estimated by
calculating the singlet–triplet energy difference. The triplet
state of [1-CH2Ph]+ exhibits a Gibbs free energy which is
172 kJmolꢀ1 higher than that of the singlet ground state. This
implies that NꢀC bond fragmentation after photoexcitation is
exergonic by approximately 100 kJmolꢀ1. This estimate takes
into account relaxation of the excited state prior to bond frag-
mentation. If one purely considers the light energy pumped
into the [1-CH2Ph]+ cation (433 nm or 276 kJmolꢀ1), it can be
seen that the energy balance even allows for a barrier to frag-
mentation in the order of 100 kJmolꢀ1 (if the energy is not in-
gested otherwise). In summary, these calculations lend strong
support to the proposed pathway (see Scheme 6 and
Scheme 7).
In another experiment, we irradiated
a
solution
of [1-CH2Ph]+Brꢀ with the 465 and 488 nm lines of an Ar+ ion
laser. In both cases the photoreaction took place, but it was
faster with the 465 nm line, which is close in energy to the ab-
sorption maximum (at 433 nm) of the band assigned to the
HOMO–LUMO excitation in the UV/Vis spectrum. This result
argues for photodissociation of [1-CH2Ph]+ directly upon
photoexcitation of an electron into the LUMO. An alternative
mechanism, which would involve excitation into the
LUMO+1 or higher energetic orbitals and electron transfer to
another [1-CH2Ph]+ ion to produce a neutral [1-CH2Ph] radical
C
species which subsequently dissociates, could be excluded on
this basis. This alternative mechanism would resemble the re-
duction of N-alkoxy-pyridinium salts after electron transfer
from an excited electron donor molecule,[1] with the specialty
that [1-CH2Ph]+ acts simultaneously as electron donor and ac-
ceptor component. Such a mechanism is presumably only fea-
sible if the donor and the acceptor component could form
a charge-transfer complex, which is clearly not the case with
[1-CH2Ph]+. The proposed radical pathway sketched in
HOMO!LUMO excitations generally involve mainly the p-
system of the pyridine ring in the case of N-alkyl pyridinium
ions (p!p* transition).[1] However, the p-system is orthogonal
to the s-bond between the pyridine N and the attached R. In
the case of N-alkoxy-pyridinyl radicals, the resulting high barri-
er for NꢀO bond cleavage is reduced or vanishes completely
(barrierless fragmentation) by bending of the NꢀC bond out of
the pyridinyl ring plane.[1] This bending allows mixing between
p and s orbitals. Our quantum chemical calculations (B3LYP/6-
311G**) did not find evidence for a bending of the NꢀC bond
out of the ring plane in the singlet or triplet state of the cation
[1-CH2Ph]+. On the other hand, the frontier orbitals of the
[1-CH2Ph]+ ion are clearly not purely p orbitals, but involve
also the guanidino groups. Moreover, the illustration of the iso-
density surface of the LUMO (see Figure 2b and Figure S6 in
Scheme 6. Pathway of the photoreaction between [1-R]+ and RX (e.g.,
R=CH2Ph).
Scheme 6 also involves disproportionation of the radical cation
2+
1+ into 1 and 1. This disproportionation is supported by CV
C
experiments showing the first and second electron oxidation
from 1 to occur at the same potential.[8,10] Unfortunately, we
were not able to trace the radicals by EPR spectroscopy, most
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5293
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 8) could be cleaved by
irradiation.[14] In other work, the
CoꢀC bond of benzyl complexes
was cleaved homolytically either
thermally or by photolysis to
give benzyl radicals which di-
merize to 1,2-diphenylethane.[15]
A higher yield of 1,2-diphenyl-
ethane and a cleaner reaction
was achieved by addition of an
equivalent of benzyl bromide.
Recently the reduction of benzyl
bromide to give 1,2-diphenyl-
ethane via
a
mononuclear
cobalt-benzyl complex, anchored
to semiconductor (TiO2) nano-
crystallites, has also been ach-
ieved photocatalytically.[16] The
TiO2 material trapped and stored
the electrons needed for reduc-
Scheme 7. NꢀC bond cleavage reactions showing the preference of N-alkyl pyridinium ions for alkyl cation forma-
tion. By contrast, 1-alkyl cations prefer to form alkyl radicals. The energies are DG values (1 bar, 198 K) for
R=benzyl from B3LYP/6-311G** calculations.
the Supporting Information) also shows contributions from the
benzyl group. Overall, the mixing between the pyridine p orbi-
tals with guanidino and benzyl orbitals leads to a symmetry re-
duction and a lower barrier.
tion and enabled a multielectron transfer chemistry. Other
metal complexes which were used to produce alkyl radicals
[17]
from alkyl halides are Mn2(CO)10 and [CpFe(CO)2]2 (see Equa-
tions (b) and (c) in Scheme 8).[18] The reduction of 12+ allows
the regeneration of 1,[8,19] so that the metal-free photocoupling
process with 1 can be driven in a cycle. This cycle is visualized
in Figure 7. At the present stage, the regeneration in the last
step of this cycle must be carried out separately, and conse-
quently stoichiometric amounts of the photosensitizer and re-
ducing agent 1 are necessary. In ongoing research in our labo-
ratory we are testing cheap reducing agents which could be
added directly to the reaction mixture, and which might allow
the development of metal-free photocatalytic coupling reac-
tions with 1 as photocatalyst.
To obtain information about possible side reactions, which
could explain that some of the added alkyl bromide did not
react, we made several test experiments. In one of these ex-
periments, we irradiated solutions of 1 in CH3CN without addi-
tion of any alkyl bromide, and indeed observed the red-color-
ing of the solution after some time. The UV/Vis spectra (Fig-
ure S7 in the Supporting Information) clearly showed relatively
slow photoinduced oxidation of 1 to give 12+ (at least four
times slower than the photoinduced NꢀC bond fragmentation
of [1-CH2Ph]+Brꢀ). Such a photoinduced oxidation of 1 was
not observed in THF solutions. Unfortunately, 12+(Brꢀ)2 turned
out to be poorly soluble in THF. Therefore, when the photore-
action between [1-CH2Ph]+ and PhCH2Br was repeated in THF
solution, clouding of the solution by the formed solid 12+(Brꢀ)2
particles reduced the transmittance of the reaction mixture
and consequently led to a deceleration of the photoreaction.
Nevertheless, these results show that an undesired side reac-
tion involving the solvent can be avoided. In a continuous-
flow reactor the solution could be filtered so that the forma-
tion of solid particles should not represent a problem, and the
reaction rate should increase considerably.[11,12]
There are some reports in the literature of alkyl radical for-
mation upon irradiation of alkyl halide solutions, however
without evaluation of the mechanism (see the critical discus-
sion in Ref. [13]). Under our experimental conditions, irradiation
of alkyl halide solutions had no measurable effect. Also, the ad-
dition of pyridine to an alkyl halide solution did not initiate
any photoreaction. Only if the pyridine derivative 1 was added,
a photoreaction occurred. Photochemical generation of alkyl
radicals from alkyl halides is also possible with metal com-
plexes, especially cobalt complexes. A photolabile metal–alkyl
complex is formed in the first step. Hence, it was shown that
the CoꢀC bond in alkyl-cobaloximes (see Equation (a) in
Scheme 8. Some metal complexes that were used in the past to generate
alkyl radicals.
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5294
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tests for cross-coupling reactions. In ongoing experiments we
examine the suitability of 1 in photoinitiated ring-closing reac-
tions. In summary, we think that 1 is an extremely interesting
reagent for metal-free reductive photocoupling reactions. The
present fundamental study stimulates more research in this
field.
Experimental Section
All reactions were carried out under inert gas atmosphere using
standard Schlenk techniques. The solvents were rigorously dried
prior to their use. However, we recognized that the presence of
water traces has no measurable influence on the photochemistry
reported herein. Although 2,3,5,6-tetrakis(1,1,3,3-tetramethylguani-
dino)pyridine (1) is a strong Brønsted base,[8] the basicity is signifi-
cantly reduced in the [1-R]+, ion which is rapidly formed upon alky-
lation. It especially turned out to be not necessary to dry the ap-
plied alkyl halides. Benzyl bromide (98%), allyl bromide (99%),
methyl iodide (99.5%), and methyl triflate (98%) were purchased
from Sigma–Aldrich and used as delivered. Compound 1 was syn-
thesized as described previously.[8] Elemental analyses were carried
out at the Microanalytical Laboratory of the University of Heidel-
berg. UV/Vis measurements were carried out on a Cary 5000 spec-
trophotometer. Infrared spectra were recorded using a BIORAD Ex-
calibur FTS 3000. NMR spectra were taken on a BRUKER Avance III
600, a BRUKER Avance II 400 or on a BRUKER Avance DPX AC200
spectrometer. Bruker ApexQe FT-ICR (ESI) and JEOL JMS-700 (HR-EI)
machines were used for mass spectrometry (MS). Irradiation was
achieved with a 150 W medium pressure Hg lamp (Heraeus TQ
150) as well as with the 465 or 488 nm lines of an Ar+ ion laser (at
0.3 W, Coherent).
Figure 7. Cycle for CꢀC photocoupling reactions with 1.
Conclusion
In this work we report metal-free photocoupling reactions of
alkyl halides with the compound 2,3,5,6-tetrakis(1,1,3,3-tetra-
methylguanidino)pyridine (1), which is both photosensitizer
and reducing agent. The mechanism of this reaction was stud-
ied, and shown to follow a radical pathway. In the first step, an
N-alkyl-pyridinium salt is formed. Test experiments with related
guanidine electron donors show that the formation of the NꢀC
bond is essential. Then, irradiation with visible light leads in
the rate-determining step to homolytic cleavage of the NꢀC
bond. The guanidino groups turn the pyridine into a strong
electron donor and weaken the NꢀC bond. In quantum chemi-
cal calculations the pyridine derivative 1 was compared with
the parent pyridine, and the differences were highlighted
which make such reactions possible with 1, but not with other
known pyridines. The reaction rate was determined for the N-
benzyl-pyridinium and N-allyl-pyridinium ions (k=0.23ꢁ0.03
and 0.13ꢁ0.03 sꢀ1). In the case of the N-methyl-pyridinium
ion, the reaction was much slower. The resulting alkyl radicals
1-methyl-2,3,5,6-tetrakis(1,1,3,3-tetramethylguanidino)pyri-
dinium iodide, [1-Me]I
Compound 1 (51.0 mg, 0.096 mmol), was dissolved in 16 mL Et2O.
Then 7 mL of CH3I (16.0 mg, 0.113 mmol) were added. The reaction
mixture was stirred at room temperature for a period of 3 h.
During this time, a deep-yellow colored solid precipitated. The sol-
vent was removed under vacuum and the remaining solid washed
three times with 3 mL portions of Et2O. Subsequently the product
was dried under vacuum yielding 64.4 mg (0.095 mmol, 99%).
Crystals suitable for an X-ray diffraction analysis were obtained
1
from CH2Cl2/Et2O=2:1 solvent mixtures. H NMR (400 MHz, CD3CN):
d=6.03 (s, 1H, Hpy), 3.77 (s, 3H, pyridine N-Me), 2.73 (s, 24H,
NMe2), 2.62 ppm (s, 24H, NMe2); 13C NMR (100 MHz, MeCN), HSQC,
HMBC: d=163.77 (s, Cimino,guanidino), 160.50 (s, Cimino,guanidino’), 145.82 (s,
Cpy (bonded to guanidino’)), 131.11 (s, Cpy (bonded to guanidino)),
123.91 (d, Cpy), 39.85 (q, NMe2), 39.74 (q, NMe2), 35.83 ppm (q, Me
(bonded to pyridine N)); UV/Vis (CH3CN, c=1.97·10ꢀ5 molLꢀ1, d=
1.0 cm): l (nm, e in Lmolꢀ1 cmꢀ1)=425 (2.0·104), 333 (1.4·104), 246
(4.3·104); HRMS (ESI+): m/z calcd (%): 546.44632; found: 546.44662
([1-Me]+, 100%); elemental analysis calcd (%) for C26H52N13I
(673.35): C 46.35, H 7.78, N 27.03; found C 46.45, H 7.67, N 26.92;
crystal data for C26H52IN13: Mr=673.71, 0.30ꢂ0.20ꢂ0.15 mm3, tri-
dimerize to give the CꢀC coupling product. The radical mono-
2+
cation 1+ disproportionates to give 1 and 1. If the photore-
C
action is conducted at low concentration with a 1:2 molar ratio
of 1 and the alkyl halide, 1 is completely consumed and the
only products are the CꢀC coupling product and a salt of the
dication 12+. If the reaction is repeated with higher concentra-
tions, a yield of the photocoupling product to 60% in CH3CN
solutions was obtained, and the analysis disclosed ways to fur-
ther increase the yield. Regeneration of neutral 1 by reduction
of the dication formed in the course of the reaction offers the
possibility to drive the process in a cycle. Ongoing experiments
aim at the reduction of 12+ directly in the reaction mixture,
and at the development of metal-free photocatalytic coupling
reactions with compound 1. We also briefly discussed first
¯
clinic, space group P1, a=14.697(3), b=16.453(3), c=17.291(4) ꢁ,
a=114.78(3)8, b=97.10(3)8, g=111.63(3)8, V=3332(2) ꢁ3, Z=4,
d
calcd =1.343 Mg·mꢀ3, MoKa radiation (graphite monochromated,
l=0.71073 ꢁ), T=100 K, qrange 2.28 to 27.528. Reflections mea-
sured; 26948, indep. 15062, Rint =0.0646. Final R indices [I>2s(I)]:
R1 =0.0605, wR2 =0.1900.
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5295
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
40.24 ppm (q, NMe2); HRMS (ESI+): m/z: 622.4 ([1-CH2C6H5]+,
1-methyl-2,3,5,6-tetrakis(1,1,3,3-tetramethylguanidino)pyri-
dinium triflate, [1-Me](O3SCF3)
100%), 311.7 ([1-CH2C6H5]2+
,
5.15%); UV/Vis (CH3CN, c=
(nm,
in Lmolꢀ1 cmꢀ1)=433
2.48·10ꢀ5 molLꢀ1
,
d=0.5 cm):
l
e
(1.6·104), 340 (1.3·104), 261 (2.4·104), 217 (4.9·104); elemental analysis
calcd (%) for C32H56N13Br·2H2O (701.40): C 52.02, H 8.19, N 24.65;
found: C 52.59, H 8.12, N 24.24.
12.8 mL of methyl triflate (18.5 mg, 0.113 mmol) was slowly added
at a temperature of 08C to a solution of 50.0 mg of 1 (0.094 mmol)
in 16 mL Et2O. The reaction mixture turned yellow and a deeply-
yellow solid precipitated. After stirring the mixture at room tem-
perature for a period of 1 h, the solvent was removed and the re-
maining solid washed 3 times with 3 mL portions of Et2O. Then the
solid was dried under vacuum to obtain 49.5 mg (0.071 mmol,
[1-CH2C6H5]BrI1.5(I3)0.5
10 mg of [1-CH2C6H5]Br (0.017 mmol) were dissolved in 2 mL
CH3CN. Then 8.6 mg of I2 (0.034 mmol) were added. The reaction
mixture turned deep-red, and red-colored crystals precipitated
after addition of 2 mL Et2O. 1H NMR (200 MHz, CD3CN): d=7.45–
7.39 (m, 5H, Hphenyl), 5.51(s, 2H, CH2), 4.94 (s, 1H, HPy), 3.04 (s, 24H,
NMe2), 2.92 ppm (s, 24H, NMe2); 13C NMR (150 MHz, CD3CN): d=
166.38 (s, Cimino), 165.52 (s, Cimino), 150.76 (s, Cpy (bonded to guanidi-
no)), 149.68 (s, Cphenyl (bonded to CH2)), 136.13 (s, Cpy (bonded to
guanidino)), 129.54 (d, Cphenyl,para), 129.53 (d, Cphenyl,meta), 128.28 (d,
Cphenyl,ortho), 96.90 (d, Cpy), 51.39 (t, (CH2)benzyl) 42.23 (q, NMe2),
1
76%) of [1-Me](O3SCF3). H NMR (400 MHz, CD2Cl2): d=5.95 (s, 1H,
Hpy.), 3.79 (s, 3H, Me (bonded to pyridine)), 2.73 (s, 24H, NMe2),
2.62 ppm (s, 24H, NMe2); 13C NMR (100 MHz, CD2Cl2): d=162.94 (s,
Cimino), 160.15 (s, Cimino), 145.30 (s, Cpy (bonded to guanidino)),
131.35 (s, Cpy (bonded to guanidino)), 123.86 (d, Cpy), 39.88 (q,
NMe2), 39.81 (q, NMe2), 35.83 ppm (q, Me (bonded to pyridine N));
HRMS (ESI+): m/z calcd: 546.44632 found: 546.44685 ([LMe]+,
100%); UV/Vis (CH3CN, c=1.95·10ꢀ5 molLꢀ1, d=0.5 cm): l (nm, e in
Lmolꢀ1 cmꢀ1):426 (2.3·104), 334 (1.2·104), 250 (2.6·104) (shoulder); el-
emental analysis calcd (%) for C27H52N13SO3F3 (695.40): C 46.60,
H 7.53, N 26.17; found: C 45.97, H 7.22, N 25.37.
42.04 ppm (q, NMe2); HRMS (ESI+): m/z: 207.8 ([1-CH2C6H5]3+
50.9%), 265.8 ([1]2+, 100%); UV/Vis (CH3CN, c=1.29·10ꢀ5 molLꢀ1
,
,
d=1 cm): l (nm, e in Lmolꢀ1 cmꢀ1)=492 (0.5·104), 358 (2.9·104), 291
(5.6·104), 242 (2.5·104); Crystal data for C32H56BrI3N13: Mr=1083.51,
1-allyl-2,3,5,6-tetrakis(1,1,3,3-tetramethylguanidino)pyridini-
um bromide, [1-CH2CH=CH2]Br
3
¯
0.25ꢂ0.20ꢂ0.20 mm , triclinic, space group P1, a=11.498(2), b=
12.040(2), c=16.808(3) ꢁ, a=89.12(3)8, b=70.07(3)8, g=76.34(3)8,
ꢁ, V=2120.2(7) ꢁ3, Z=2,
d , MoKa radiation
calc =1.697 Mg·mꢀ3
8.1 mL of allyl bromide (11.4 mg, 0.094 mmol) were added to a solu-
tion of 50.0 mg of 1 (0.094 mmol) in 16 mL Et2O. The reaction mix-
ture was stirred for a period of 16 h under exclusion of light at
room temperature, resulting in the precipitation of a deeply-yellow
colored solid. The solvent was removed, the solid residue washed
three times with portions of 5 mL Et2O and the product subse-
quently dried under vacuum. Yield: 38.9 mg (0.06 mmol, 64%).
1H NMR (600 MHz, MeCN): d=6.01–6.08 (m, 1H, (CH)allyl), 6.00 (s,
(graphite monochromated, l=0.71073 ꢁ), T=100 K, qrange 2.32 to
33.008. Reflections measured: 51044, indep: 15952, Rint =0.0597;
final R indices [I>2s(I)]: R1 =0.0568, wR2 =0.1458.
[1-Me](O3SCF3)3 and [1-CH2C6H5]Br3
The compounds [1-Me](O3SCF3)3 and [1-CH2C6H5]Br3 crystallized in
small amounts from solutions of [1-Me](O3SCF3) and [1-CH2C6H5]Br
kept in the sun light for several days.
2
3
1H, Hpy), 5.19–5.24 (dd, J=21.4 Hz, J=13.8 Hz, 2H, (CH2)allyl), 5.08
(d, J=5.8 Hz, 2H, N-(CH2)allyl), 2.73 (s, 24H, NMe2), 2.62 ppm (s, 24H,
NMe2); 13C NMR (150 MHz, MeCN): d=163.32 (s, Cimino), 160.62 (s,
Cimino), 145.19 (s, Cpy (bonded to guanidino)), 133.29 (d, (CH)allyl),
130.57 (d, Cpy (bonded to guanidino)), 124.32 (d, Cpy.), 119.02 (t,
(CH2)allyl), 50.49 (t, N-(CH2)allyl)), 39.89 (q, NMe2), 39.80 ppm (q,
NMe2); HRMS (ESI+): m/z calcd: 572.46197; found: 572.46256
([1-CH2CH=CH2]+, 100%); UV/Vis (CH3CN, c=2.7·10ꢀ5 molLꢀ1, d=
1.0 cm): l (nm, e in Lmolꢀ1 cmꢀ1)=432 (1.5·104), 336 (1.4·104),
ca. 254 (2.5·104), 219 (4.0·104); elemental analysis calcd (%) for
C28H54N13Br (651.38): C 51.52, H 8.34, N 27.90; found: C 50.78,
H 8.34, N 27.16;
[1-Me](O3SCF3)3
Crystal data for C29H52F9N13O9S3: Mr=994.02, 0.40ꢂ0.35ꢂ0.30 mm3,
orthorhombic, space group Pbca, a=18.629(4), b=16.001(3), c=
28.986(6) ꢁ, V=8640(3) ꢁ3, Z=8, dcalc =1.528 Mgmꢀ3, Mo radiation
(graphite monochromated, l=0.71073 ꢁ), T=100 K, qrange 1.41 to
27.918; Reflections measured: 20662, indep: 10314, Rint =0.0337;
final R indices [I>2s(I)]: R1 =0.0473, wR2 =0.1168.
[1-CH2C6H5]Br3
1-benzyl-2,3,5,6-tetrakis(1,1,3,3-tetramethylguanidino)pyri-
dinium bromide, [1-CH2C6H5]Br
Crystal data for C32H56Br3N13: Mr=862.63, 0.30ꢂ0.15ꢂ0.15 mm3, tri-
¯
clinic, space group P1, a=10.853(2), b=11.957(2), c=17.633(4) ꢁ,
10.7 mL C6H5CH2Br (16.1 mg, 0.094 mmol) were added to a solution
of 50.0 mg of 1 (0.094 mmol) in 16 mL Et2O. The reaction mixture
was stirred for a period of 4 h at room temperature, leading to pre-
cipitation of a deeply yellow-colored solid. After partial removal of
the solvent the solution was cooled to 08C for 10 min before the
remaining solvent was removed. The solid was washed three time
with portions of 5 mL Et2O and subsequently dried under vacuum.
Yield: 53.5 mg (0.076 mmol, 81%). 1H NMR (600 MHz, MeCN): d=
7.48 (d, 2H, Hphenyl,ortho), 7.32 (t, 2H, Hphenyl,meta), 7.24(t, 1H, Hphenyl,para),
6.01 (s, 1H, Hpy), 5.69 (s, 2H, (CH2)benzyl), 2.64 (s, 24H, NMe2),
2.62 ppm (s, 24H, NMe2); 13C NMR (100 MHz, MeCN): d=163.59 (s,
Cimino,guanidino), 161.23 (s, Cimino,guanidino’), 145.75 (s, Cpy (bonded to gua-
nidino’)), 138.81 (s, Cphenyl (bonded to CH2), 131.86 (s, Cpy (bonded
to guanidino)), 130.57 (d, Cphenyl,ortho), 129.45 (d, Cphenyl,meta), 128.69
(d, Cphenyl,para) 125.06 (d, Cpy), 51.95 (t, (CH2)benzyl), 40.27 (q, NMe2),
a=76.95(3)8, b=77.13(3)8, g=78.93(3)8, V=2148.8(7) ꢁ3, Z=2,
d
calc =1.333 Mgmꢀ3, MoKa radiation (graphite monochromated, l=
0.71073 ꢁ), T=100 K, qrange 1.94 to 27.538; Reflections measured:
18427, indep: 9840, Rint =0.0663; final R indices [I>2s(I)]: R1 =
0.0602, wR2 =0.1288.
X-ray crystallographic study
Suitable crystals were taken directly out of the mother liquor, im-
mersed in perfluorinated polyether oil and fixed on top of a glass
capillary. Measurements were made with a Nonius-Kappa CCD dif-
fractometer with low-temperature unit using graphite-monochro-
mated MoKa radiation. The temperature was set to 100 K. The data
collected were processed using the standard Nonius software.[20]
All calculations were performed using the SHELXT-PLUS software
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5296
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[10] A GFA radical cation was so far only formed in a transition metal com-
plex. See: C. Trumm, O. Hꢃbner, E. Kaifer, H.-J. Himmel, Eur. J. Inorg.
Chem. 2010, 3102–3108.
[11] A. Albini, M. Fagnoni, Photochemically-Generated Intermediates in Syn-
thesis, Wiley-VCH, Weinheim, 2013, chapt. 1.
[12] See also discussion in: A. C. Hernandez-Perez, S. K. Collins, Angew.
Chem. 2013, 125, 12928–12932; Angew. Chem. Int. Ed. 2013, 52, 12696–
12700.
[13] A. Albini, M. Fagnoni, Photochemically-Generated Intermediates in Syn-
thesis, Wiley-VCH, Weinheim, 2013, chapt. 2.
[14] a) A. Ghosez, T. Gçbel, B. Giese, Chem. Ber. 1988, 121, 1807–1811;
b) B. P. Branchaud, M. S. Meier, J. Org. Chem. 1989, 54, 1320–1326.
[15] a) S. Fukuzumi, K. Ishikawa, T. Tanaka, Chem. Lett. 1985, 1355–1358;
b) K. Ishikawa, S. Fukuzumi, T. Tanaka, Bull. Chem. Soc. Jpn. 1987, 60,
563–570; c) S. Fukuzumi, K. Ishikawa, T. Tanaka, Organometallics 1987,
6, 358–365.
[16] S. Ardo, D. Achey, A. J. Morris, M. Abrahamsson, G. J. Meyer, J. Am.
Chem. Soc. 2011, 133, 16572–16580.
[17] a) B. Giese, G. Thoma, Helv. Chim. Acta 1991, 74, 1135–1142; b) G. K.
Friestad, J. Qin, J. Am. Chem. Soc. 2001, 123, 9922–9923.
[18] a) G. Thoma, B. Giese, Tetrahedron Lett. 1989, 30, 2907–2910; b) B.
Giese, G. Thoma, Helv. Chim. Acta 1991, 74, 1143–1155.
package. Structures were solved by direct methods with the
SHELXS-97 program and refined with the SHELXL-97 program.[21,22]
Graphical handing of the structural data during solution and refine-
ment was performed with XPMA.[23] Atomic coordinates and aniso-
tropic thermal parameters of non-hydrogen atoms were refined by
full-matrix least-squares calculations. CCDC 976076 ([1-Me]I),
976080 ([1-CH2C6H5]BrI1.5(I3)0.5), 976079 ([1-Me](O3SCF3)I(I3)), 976078
([1-Me](O3SCF3)3), and 976077 ([1-CH2C6H5]Br3) contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
Quantum chemical calculations
The GAUSSIAN 09 program was used for all calculations.[24] Density
functional theory (DFT) calculations were carried out with the
B3LYP functional[25] in combination with a 6–311G** basis set.[26]
The calculated structures were similar to those determined by X-
ray diffraction (see Table S1 for a comparison in the case of
[1-Me]+). To check the influence of the size of the basis set, the
6–31G* basis set was also applied (Table S2). Moreover, we repeat-
ed the calculations with the pure DFT functional BP86 (Table S3).
The results were similar.
[19] For chemical reduction of redox-active GFAs, see: A. Peters, H. Herr-
mann, M. Magg, E. Kaifer, H.-J. Himmel, Eur. J. Inorg. Chem. 2012, 1620–
1631.
[20] DENZO-SMN, Data processing software, Nonius, 1998, http://www.no-
nius.com.
[21] a) G. M. Sheldrick, SHELXS-97, Program for Crystal Structure Solution, Uni-
versity of Gçttingen, Germany, 1997, http://shelx.uni-ac.gwdg.de/
SHELX/index.html; b) G. M. Sheldrick, SHELXL-97, Program for Crystal
Structure Refinement, http://shelx.uni-ac.gwdg.de/SHELX/index.html.
[22] International Tables for X-ray Crystallography, vol. 4, Kynoch Press, Bir-
mingham, 1974.
Acknowledgements
The authors gratefully acknowledge continuous financial sup-
port from the Deutsche Forschungsgemeinschaft (DFG).
[23] L. Zsolnai, G. Huttner, XPMA, University of Heidelberg, Germany, 1994;
http://www.uni-heidelberg.de/institute/fak12/AC/huttner/soft-ware/soft-
ware.html.
Keywords: CꢀC coupling
radicals · reduction
· guanidine · photoreactions ·
[24] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian Inc., Wallingford CT, 2009.
[25] a) P. J. Stephens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Chem.
Phys. B 1994, 98, 11623–11627; b) A. D. Becke, J. Chem. Phys. 1993, 98,
5648–5652; c) C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–
789.
[26] a) R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J. Chem. Phys. 1980,
72, 650–654; b) A. D. McLean, G. S. Chandler, J. Chem. Phys. 1980, 72,
5639–5648.
[1] a) E. D. Lorance, W. H. Kramer, I. R. Gould, J. Am. Chem. Soc. 2002, 124,
15225–15238; b) E. D. Lorance, W. H. Kramer, I. R. Gould, J. Am. Chem.
Soc. 2004, 126, 14071–14078; c) E. D. Lorance, K. Hendrickson, I. R.
Gould, J. Org. Chem. 2005, 70, 2014–2020.
[2] a) W. Schnabel, Macromol. Rapid Commun. 2000, 21, 628–642; b) I. R.
Gould, D. Shukla, D. Giesen, S. Farid, Helv. Chim. Acta 2001, 84, 2796–
2812.
[3] a) D. Shukla, W. G. Ahearn, S. Farid, J. Org. Chem. 2005, 70, 6809–6819;
b) D. Shukla, S. P. Adiga, W. G. Ahearn, J. P. Dinnocenzo, J. Org. Chem.
2013, 78, 1955–1964.
[4] The properties and chemistry of N-alkyl pyridinyl radicals was conse-
quently studied in detail. See, for example, studies on pyridinyl diradi-
cals: a) E. M. Kosower, Y. Ikegami, J. Am. Chem. Soc. 1967, 89, 461–462;
b) Y. Ikegami, T. Muramatsu, K. Hanaya, S. Onodera, N. Nakayama, E. M.
Kosower, J. Am. Chem. Soc. 1987, 109, 2876–2880; c) Y. Ikegami, T. Mur-
amatsu, K. Hanaya, J. Am. Chem. Soc. 1989, 111, 5782–5787.
[5] K. Y. Lee, J. K. Kochi, J. Chem. Soc. Perkin Trans. 1 1992, 1011–1017.
[6] D. Zhu, J. K. Kochi, Organometallics 1999, 18, 161–172.
[7] a) A. R. Katritzky, C. H. Watson, Z. Dega-Szafran, J. R. Eyler, J. Am. Chem.
Soc. 1990, 112, 2471–2478; b) A. R. Katritzky, C. H. Watson, Z. Dega-Sza-
fran, J. R. Eyler, J. Am. Chem. Soc. 1990, 112, 2479–2484.
[8] S. Stang, A. Lebkꢃcher, P. Walter, E, Kaifer, H.-J. Himmel, Eur. J. Inorg.
Chem. 2012, 4833–4845.
[9] a) A. Peters, E. Kaifer, H.-J. Himmel, Eur. J. Org. Chem. 2008, 5907–5914;
b) A. Peters, C. Trumm, M. Reinmuth, D. Emeljanenko, E. Kaifer, H.-J.
Himmel, Eur. J. Inorg. Chem. 2009, 3791–3800.
Received: December 20, 2013
Published online on March 27, 2014
Chem. Eur. J. 2014, 20, 5288 – 5297
www.chemeurj.org
5297
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim