Comparative Study of the Catalytic Amination of Benzylic C-H Bonds Promoted by Ru(TPP)(py)2 and Ru(TPP)(CO)

Article abstract of DOI:10.1002/ejic.201500656

A combined experimental and DFT-based theoretical analysis elucidated the influence of the axial ligand L on the catalytic activity of Ru(porphyrin)L complexes in promoting the amination of benzylic C-H bonds by organic azides (RN₃). Experimental data indicated that the catalytic activity of Ru(TPP)(CO) (1) (TPP = dianion of tetraphenylporphyrin) is comparable to that of Ru(TPP)(py)₂ (2) (py = pyridine). DFT modelling disclosed that 2 can be regarded as a precatalyst that becomes active after the endergonic loss of one pyridine ligand to give the unsaturated species [Ru](py) (11) {[Ru] = Ru(porphine)}. This complex would react with RN₃ to give the mono-imido singlet complex [Ru](py)(NR)_S (6_S), which can be easily transformed into the triplet isomer 6_T having diradical character at the imido N atom. The subsequent formation of the benzylic amine PhCH₂NHR occurs through a radical homolytic activation of one C-H bond of the toluene substrate (PhCH₃). Conversely, by staying on the singlet potential-energy surface, 6_S can undergo dissociation of the pyridine ligand to form [Ru](NR). This complex can activate another RN₃ molecule to form the bis-imido compound [Ru](NR)₂, which is also catalytically active. At this point, the mechanism becomes independent of the nature of the original ligand L coordinated to [Ru]. The catalytic activity of Ru(TPP)(py)₂ (2) (TPP = dianion of tetraphenylporphyrin, py = pyridine) in the amination of C-H bonds by organic azides (RN₃) was experimentally compared to that of Ru(TPP)(CO) (1). A computational analysis corroborates similar radical mechanisms with a triplet imido intermediate, in spite of the electronic differences of the apical ligands.

Full text of DOI:10.1002/ejic.201500656

FULL PAPER
DOI:10.1002/ejic.201500656
Comparative Study of the Catalytic Amination of
Benzylic C–H Bonds Promoted by Ru(TPP)(py)2 and
Ru(TPP)(CO)
Gabriele Manca,*[a] Carlo Mealli,[a] Daniela Maria Carminati,[b]
Daniela Intrieri,[b] and Emma Gallo*[b]
Keywords: Ruthenium / Porphyrinoids / Amination / Density functional calculations / Reaction mechanisms
A combined experimental and DFT-based theoretical analy-
can be easily transformed into the triplet isomer 6T having
sis elucidated the influence of the axial ligand L on the cata- diradical character at the imido N atom. The subsequent for-
lytic activity of Ru(porphyrin)L complexes in promoting the
amination of benzylic C–H bonds by organic azides (RN3).
Experimental data indicated that the catalytic activity of
Ru(TPP)(CO) (1) (TPP = dianion of tetraphenylporphyrin) is
comparable to that of Ru(TPP)(py)2 (2) (py = pyridine). DFT
modelling disclosed that 2 can be regarded as a precatalyst
that becomes active after the endergonic loss of one pyridine
ligand to give the unsaturated species [Ru](py) (11) {[Ru] =
Ru(porphine)}. This complex would react with RN3 to give
the mono-imido singlet complex [Ru](py)(NR)S (6S), which
mation of the benzylic amine PhCH2NHR occurs through a
radical homolytic activation of one C–H bond of the toluene
substrate (PhCH3). Conversely, by staying on the singlet po-
tential-energy surface, 6S can undergo dissociation of the
pyridine ligand to form [Ru](NR). This complex can activate
another RN3 molecule to form the bis-imido compound
[Ru](NR)2, which is also catalytically active. At this point, the
mechanism becomes independent of the nature of the origi-
nal ligand L coordinated to [Ru].
Metal porphyrin complexes efficiently promote amin-
ation reactions of saturated and unsaturated hydrocarbons
by aryl azides,[12–29] and they show very good activity in
the amination of activated sp3 C–H bonds. Ruthenium(II)
porphyrins[19,21,30] are active catalysts in the synthesis of all-
ylic and benzylic amines. They are also efficient in the amin-
ation of benzylic C–H bonds in the α or β positions with
respect to an ester group to yield the corresponding and
biologically relevant α- and β-amino esters.[23,31–33]
One of the most extensively used amination catalyst is
the five-coordinate complex Ru(TPP)(CO) (1) (TPP = di-
anion of tetraphenylporphyrin), which has the π-acceptor
CO ligand in the axial position. The vacant coordination
site in the trans position is suitable for azide activation. Ki-
netic[20] and DFT[27] studies indicated that the reaction of
Ru(TPP)CO with ArN3 first affords the octahedral mono-
imido complex Ru(TPP)(NAr)(CO) in a singlet state, which
can be transformed into a slightly more stable triplet with
the two unpaired spins largely localised at the NAr ligand.
Such a diradical can promote the activation of different or-
ganic substrates,[27] in particular the homolysis of a C–H
bond of a hydrocarbon with consequent formation of the
corresponding amine[27a] by the so-called rebound mecha-
nism.[34] Conversely, if the species Ru(TPP)(NAr)(CO) re-
mains a singlet, the bis-imido complex Ru(TPP)(NAr)2
(7)[18,20] is formed through CO departure and another azide
activation. Complex 7 can also behave as a diradical species
for a similar type of amination chemistry.
Introduction
The catalytic syntheses of nitrogen-containing com-
pounds are important for their potential use as precursors
of biological and pharmaceutical compounds.[1,2] In re-
sponse to the demand for sustainable chemistry, the em-
ployment of organic azides (RN3) as nitrogen sources for
the synthesis of aza compounds[3–10] is constantly growing.
In fact, this class of aminating reagents shows high eco-
compatibility due to the formation of the benign molecular
nitrogen as the only by-product of the nitrene (“RN”)
transfer reaction to an organic molecule. Among available
organic azides, the aryl azides are very interesting nitrogen
sources because of their convenient reactivity/stability rela-
tionship and their easy syntheses from the corresponding
amines. More extensive employment of these aminating
agents is also favoured by their commercial availability,
thanks to an efficient and safe procedure to obtain aryl
azides in bulk amounts.[11]
[a] Istituto di Chimica dei Composti OrganoMetallici,
ICCOM-CNR,
Via Madonna del Piano 10, 50019 Sesto Fiorentino, Italy
E-mail: gabriele.manca@iccom.cnr.it
www.iccom.cnr.it
[b] Department of Chemistry, University of Milan,
Via Golgi 19, 20133 Milan, Italy
E-mail: emma.gallo@unimi.it
www.unimi.it
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201500656.
Eur. J. Inorg. Chem. 2015, 4885–4893
4885
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Both CO and NR ligands, placed trans to the vacant co- Table 1. Synthesis of compounds 3a–3h catalysed by 2.[a]
ordination site at which the azide is activated, combine π-
acceptor and σ-donor abilities. To analyse the catalytic in-
fluence of the electronic features of the axial ligand, we here
report the combined experimental and computational study
of the catalytic behaviour of Ru(TPP)(py)2 (2). Although
the physicochemical properties of 2 were already elucidated
by a variety of experimental methods, such as electrochem-
istry,[35,36] UV/Vis absorption, emission, resonance Ra-
man[37–41] and picosecond transient absorption,[42,43] the use
of this complex to catalyse C–H amination by organic az-
ides has not yet been reported.
Results and Discussion
Experimental Studies on the Catalytic Activity of
Ru(TPP)(py)2 (2)
The bis-pyridine RuII complex 2 was prepared in accord-
ance with a reported synthetic procedure[40] and used to
catalyse the reaction between aryl azides (ArN3) and vari-
ous benzylic substrates of general formula RЈRЈЈRЈЈЈC–H
to yield the corresponding aminated compounds 3a–3h
(Table 1). 3,5-Bis(trifluoromethyl)phenyl azide was used as
the aminating agent, with the sole exception of the synthesis
of compound 3c, for which 4-tert-butyl azide was employed.
Data reported in Table 1 indicate that complex 2 is a
good catalyst of C–H amination, and its catalytic efficiency
[a] Experimental conditions: 6.8ϫ10–3 mmol of the catalyst (2%
with respect to ArN3) in 15.0 mL of refluxing hydrocarbon sub-
is comparable to that of Ru(TPP)CO (1) when used to cata-
lyse the synthesis of the same compounds.[19,33] As observed strate as the reaction solvent. [b] Time required for complete azide
conversion, determined by IR spectroscopic monitoring of the de-
in 1-catalysed reactions, 2 was more effective in activating
electron-deficient azides. Indeed, the reaction of 4-tert-butyl
azide with ethylbenzene yielded 3c in a low yield (Table 1,
entry 2), which suggested an electrophilic role of azide in
the amination reaction. Complex 2 was also active in pro-
moting the amination of the allylic C–H bond of cyclohex-
crease in N3 absorbance at about 2115 cm–1. [c] Yields based on
1
ArN3 and determined by H NMR spectroscopy (2,4-dinitrotolu-
ene as the internal standard). [d] Reaction catalysed by 1.[19] [e]
Reaction run at 150 °C. [f] Reaction run at 80 °C. [g] Reaction cata-
lysed by 1.[33] [h] 6% of 2 was used.
ene by 3,5-bis(trifluoromethyl)phenyl azide. The corre- ture inhibited the formation of side products as well as the
sponding allylic amine N-(cyclohex-2-en-1-yl)-3,5-bis(tri- activation of the azide, which was more effective at the re-
fluoromethyl)aniline (3i)[20] was formed in 0.75 h with 77% fluxing temperature of cumene (≈ 150 °C).
yield. For all the described reactions, NMR and GC-MS
Considering the biological relevance of β-amino esters,
analyses of the crude product revealed the formation of the synthesis of 3-[3,5-bis(trifluoromethyl)phenylamino]-3-
ArN=NAr and ArNH2 side products derived from partial phenylpropanoate (3h) was studied under different experi-
decomposition of the employed azide.
mental conditions. As reported in entry 7 of Table 1, the
Considering that in previous reactions catalysed by best yield of the desired compound 3h (82%) was obtained
Ru(TPP)CO (1),[19,33] slightly different experimental condi- by using a catalytic ratio 2/3,5-bis(trifluoromethyl)phenyl
tions were used, the synthesis of compound 3f was repeated azide of 1:15 in methyl hydrocinnamate as the reaction sol-
in the presence of complex 1 by using the experimental con- vent.
ditions described in Table 1 to give 90% of the benzylic
The reaction was also performed in refluxing benzene
amine in 1 h. Even though the yield was higher than that with a catalytic ratio 2/azide/hydrocarbon of 1:15:1000, but
obtained in the presence of 2 (Table 1, entry 5), the reaction compound 3h was formed in 5 h with 60% yield. Import-
time increased from 20 min to 1 h. An attempt was made antly, the reaction catalysed by 2 is faster than that per-
to enhance the selectivity of 3f by reducing the formation of formed in the presence of Ru(TPP)CO (1), for which 77%
side products. Thus, azide was slowly added to the reaction yield was obtained after 10 h with a catalytic ratio 1/azide/
mixture with a syringe pump over 1.5 h, but unfortunately hydrocarbon of 1:15:1000 in refluxing benzene. This last re-
the reaction yield did not increase. Although the reaction sult is important because it indicates that two coaxial pyr-
yield was enhanced to 72% by working at 70 °C, the time idine ligands do not hamper the efficiency of the catalyst,
required for full conversion of the azide to 3f became ex- even though the latter must vacate one coordination site to
cessively long (7.0 h). Clearly, the lower working tempera- allow the activation of the azide and the subsequent amin-
Eur. J. Inorg. Chem. 2015, 4885–4893
4886
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
ation reaction (see the conclusions of the theoretical analy- vation. Moreover, the pyridine ligand differs from CO in
sis below). The reaction time was halved by increasing the having no significant π-acceptor capability, which was com-
methyl hydrocinnamate concentration, and 70% yield of 3h putationally found to be important in singlet/triplet in-
was obtained in 2.5 h by employing a catalytic ratio 2/azide/ tersystem crossing for promotion of the radical activity. For
hydrocarbon of 1:15:2500. Data reported above indicate the these reasons, we compare in the following theoretical sec-
importance of using high methyl hydrocinnamate concen- tion some important electronic aspects of catalysts 1 and 2
trations to achieve good catalytic performance. To reduce as well as the corresponding reaction profiles of amination
the reaction costs, the excess hydrocarbon was recovered at reactions.
the end of the reaction by a simple distillation process, as
already reported by us for the same reaction catalysed by
complex 1.[33]
Computational Studies
To obtain experimental information on the strength of
the Ru–py bond in 2, we studied the process of pyridine
substitution by dimethyl sulfoxide (DMSO). The reaction
with DMSO quantitatively yielded the complex
Ru(TPP)(DMSO)2 (4), which supports the hypothesis that
the two pyridine ligands in 2 are not irreversibly coordi-
nated to the metal centre and can be displaced by another
2e-donor ligand (Scheme 1, path a). In view of this result,
we treated 2 with an equimolar amount of ArN3 in order
to substitute pyridine by an azide ligand to give RuII-
(TPP)(py)(ArN3) (5) or RuIV(TPP)(py)(NAr) (6), which is
derived from 5 by elimination of molecular nitrogen
(Scheme 1, path b). Unfortunately, we did not observe the
coordination of one ArN3 ligand to the metal centre yield-
ing 5 or the mono-imido derivative 6. The NMR analysis
of the crude reaction product revealed the presence of an
equimolar mixture of Ru(TPP)(py)2 (2) and bis-imido com-
plex Ru(TPP)(NAr)2 (7). This suggests that, even if com-
pounds 5 and 6 are momentarily formed, they are too elus-
ive to be experimentally detected, most likely because 7 is
quickly formed as the thermodynamically stable product
(Scheme 1, path c). Then, we tried to detect the formation
of the key intermediate RuIV(TPP)(py)(NAr) (6) in the reac-
tion of equimolar amounts of 2 and 7. Unfortunately, the
desired nitrene-transfer reaction from 7 to 2 did not occur,
and the formation of the elusive mono-imido intermediate
6 was not observed (Scheme 1, path d).[20,27]
General Aspects of the Amination Reaction Catalysed by
[Ru](CO) (1)
Our previous computational studies illustrated the pos-
sible mechanisms of the catalytic allylic amination of cyclo-
hexene by an RN3 molecule.[27] With [Ru](CO) (1) {[Ru] =
Ru(porphine), porphine
=
parent compound of
porphyrins} as catalyst, the overall process consists of the
two interconnected cycles (a) and (b) shown in Scheme 2.
The mechanistic proposal is supported by kinetic investi-
gations and by the isolation and characterisation of the bis-
imido complex Ru(TPP)(NR)2 (7) {R = 3,5-bis(trifluoro-
methyl)phenyl}, which is catalytically active in several hy-
drocarbon aminations.[18,20]
Scheme 2. Overall mechanism of the 1-catalysed C–H amination of
cyclohexene by an organic azide (RN3).
Cycle (a) starts with RN3 activation by complex 1 to give
the mono-imido species [Ru](CO)(NR)S (CS) via a well-de-
fined TS (not shown in Scheme 2).[27] Importantly, the em-
ployment of CH3N3 in place of the experimentally used 3,5-
bis(trifluoromethyl)phenyl azide does not change the over-
all reaction profile. Only the barrier of +26.8 kcalmol–1 is
about 25% higher than that optimised by modelling the ex-
perimentally used azide. Since such a difference is not dra-
matic, CH3N3 is indicated as RN3 in all the subsequent
modelling, including that relative to reactions promoted by
2 (see below).
Scheme 1. The reactivity of complex 2 towards DMSO, ArN3 and
complex 7.
In cycle (a) the singlet minimum CS is first obtained after
the TS, although its triplet isomer [Ru](CO)(NR)T (CT) is
In conclusion, all the collected experimental data indi- –3.7 kcalmol–1 more stable. The implied intersystem cross-
cate that the catalytic activity of the bis-pyridine complex 2 ing is fundamental to trigger radical reactivity, given that
in the amination of C–H bonds by aromatic azides is com- the two unpaired electrons are largely localised at the axial
parable to, and in some cases better than, that of Ru- imido N atom. Complex CT is able to promote the homo-
(TPP)CO (1). This result is somewhat surprising, because 2 lytic cleavage of a C–H bond of cyclohexene (C6H10) to
first needs to vacate one coordination site for azide acti- form the mono-amido [Ru](CO)(NHR)D doublet interme-
Eur. J. Inorg. Chem. 2015, 4885–4893
4887
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
·
diate together with the free cyclohexenyl radical (C6H9 ). independent of the chemical nature of the hydrocarbon sub-
Subsequently, the two radicals pair their spins to afford the strate (cyclohexene or toluene).
desired allylic amine (C6H9)NHR, first as a ligand and then
The energy barrier at the transition state (TS) {[Ru]-
as a free molecule. This process corresponds to the known (CO)(NR)T*PhCH3}TS (8TS, see Figure S1) is only slightly
rebound mechanism.[34] The overall estimated free-energy higher than that obtained with cyclohexene (+9.1 vs.
gain for cycle (a) is –44.8 kcalmol–1.[27]
+7.5 kcalmol–1). Additionally, the following step toward
If not promptly converted to CT, CS has the alternative the combined doublet [Ru](CO)(NHR)D (9D) and the free
·
possibility of losing the CO apical ligand to afford access tolyl radical PhCH2 is somewhat less exergonic (–13.0 vs.
to cycle (b) of Scheme 2. The derived five-coordinate –18.3 kcalmol–1), which suggests a possibly more difficult
[Ru](NR) species has a vacant coordination site to activate process. On the other hand, the recombination of the two
a second azide molecule, which, after another similar TS, radicals to give the diamagnetic amino complex
affords the bis-imido complex [Ru](NR)2 (ES). In principle, [Ru](CO){HN(R)CH2Ph} (10) is more exergonic than that
the latter singlet complex also has the possibility of trans- in the corresponding allylic amination (–44.4 vs.
forming into the triplet isomer ET, although such process is –38.6 kcalmol–1). The final separation of the benzylic
endergonic rather than exergonic like the CSǞCT transfor- amine (PhCH2NHR) from the metal is somewhat more hin-
mation.
dered than that of the allylic amine (C6H9)NHR (+19.2 vs.
Then, one of the two coordinated NR groups of ET may +12.9 kcalmol–1). Such an important difference is reflected
initiate radical reactivity towards the cyclohexene substrate. in the overall lower exergonicity of the amination process
Hence, a similar rebound mechanism leading to the same involving toluene compared to that involving cyclohexene
allylic amine product with an energy balance comparable (–37.2 vs. –44.8 kcalmol–1, respectively).
with that of cycle (a) is also possible in this case.
Catalytic Activity of [Ru](py)2 (2)
Benzylic amination catalysed by [Ru](py)2 (2) was com-
putationally analysed. The isolated octahedral precursor 2
was optimised (Figure S2) with two identical Ru–Npy dis-
tances of 2.09 Å, which fully match those of the available
X-ray structure.[44,45] To behave as a catalyst, 2 must first
lose one pyridine ligand to allow activation of the azide,
similar to what happens at 1. The optimised five-coordinate
complex [Ru](py) (11) (Figure 1, a) has an about 0.1 Å
shorter Ru–Npy bond with respect to that of [Ru](py)2 (2).
This is attributable to the loss of any trans influence on
vacating one of the two coordination sites. The transforma-
tion of 2 into 11 involves a high energy cost of
+22.1 kcalmol–1 (first step in Scheme 4). All attempts to
identify a TS for the dissociation failed, while a scanning
technique indicated that the process proceeds monoto-
nously uphill. Therefore, the formation of 11, corroborated
by its subsequent catalytic activity, must be attributed to the
relatively high temperature used in the catalytic experiments
(Ͼ 70 °C, see Table 1).
Comparison of the Amination of Toluene and Cyclohexene
Catalysed by [Ru](CO) (1)
The mechanism of Scheme 2 refers to the 1-catalysed all-
ylic amination of cyclohexene, whereas the catalytic activity
of 2 was mainly studied in the amination of the benzylic
substrates indicated in Table 1. Since the behaviour of 1 was
previously highlighted only toward allylic amination,[27] we
now briefly illustrate some comparative aspects of benzylic
amination promoted by the same catalyst 1.[19,20] Toluene
(PhCH3) was therefore modelled as the benzylic substrate,
and Scheme 3 shows the energy profile relative to the acti-
vation of the methylic C–H bond by the diradical [Ru]-
(CO)(NR)T (CT) complex. As discussed above, CT is formed
Figure 1. Optimised structures of (a) [Ru](py) (11) and (b) [Ru]-
(py)(RN3) (5).
Scheme 3. Energy profile for formation of the benzylic amine
PhCH2NHR starting from the mono-imido diradical [Ru](CO)-
(NR) (CT).
The vacant coordination site at 11 allows the coordina-
tion of one azide to the metal to give the adduct [Ru](py)-
Eur. J. Inorg. Chem. 2015, 4885–4893
4888
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
group has already lost its original linearity, since the angle
Nα–Nβ–Nγ is 137.2°.
In Scheme 4, the barrier at 5TS of +13.3 kcalmol–1 is
about 50% lower than that in the corresponding process
catalysed by 1 (+26.8 kcalmol–1). Therefore, after the ini-
tially difficult activation of the catalytic reactivity, the evol-
ution of the amination reaction seems to be facilitated by
the presence of pyridine. Not only is the barrier lower, but
also the following imido product [Ru](py)(NR)S (6S in Fig-
ure 3, a) is significantly more stable than the carbonylated
analogue [Ru](CO)(NR)S (CS) (–37.4 vs. –27.9 kcalmol–1).
Scheme 4. Energy profile for formation of mono-imido complex
[Ru](NR) (12) starting from [Ru](py)2 (2).
(RN3) (5) of Figure 1 (b). Note that 5 has a higher stabilisa-
tion energy than the corresponding carbonylated complex
[Ru](CO)(RN3): –10.1 versus –3.5 kcalmol–1. In this re-
spect, we recall that 1 is largely stable as an unsaturated
species, whereas its azide adduct [Ru](CO)(RN3) is stabi-
lised only by weak dispersion forces,[27] as clearly estab-
lished by usage of the DFTD functional.[46]
The greater stabilisation of [Ru](py)(RN3) (5) is reflected
in the shorter Ru–Nazide coordination bond compared to
that in [Ru](CO)(RN3) (2.17 vs. 2.31 Å) due to the stronger
CO trans influence. The significant energy gain of the ad-
duct 5 may in part compensate the energy previously lost
with the first pyridine departure from 2, and thus justifies
the observed catalytic activity of Ru(TPP)(py)2.
Figure 3. Optimised structure of the spin isomers (a) [Ru](py)-
(NR)S (6S) and (b) [Ru](py)(NR)T (6T).
Importantly, the complex 6S, analogously to CS, is found
to undergo intersystem crossing to the triplet spin isomer
6T (Figure 3, b), which is the fundamental step for promot-
ing the radical reactivity. The main difference between the
two pairs of spin isomers is that, while the CSǞCT intercon-
version is somewhat exergonic, the 6SǞ6T process is ender-
gonic (–3.7 vs. +2.0 kcalmol–1), but in any case the differ-
ences are not large enough to hamper the following radical
reactivity.
For [Ru](CO)(NR)S (CT), an orbital analysis suggested
that two singly populated MO levels of the CT triplet re-
ceive a favourable contribution from the two CO π* orbit-
als.[27] In fact, a percentage of the latter is present in the
singly occupied molecular orbitals (SOMOs), which are
mainly combinations of metal dπ (dxz, dyz) and NR pπ or-
bitals. This is not exactly the case for 6T, in which the apical
pyridine ligand is not involved in the SOMOs. Therefore,
the 6SǞ6T intersystem crossing is less facile, as quantita-
tively supported by calculations. In the case of the
carbonylated complex, the minimum-energy crossing point
(MECP)[47] was found to be null.[27] Conversely, an appro-
priate scan for the 6SǞ6T interconversion (see Figure S3)
shows that 6S lies constantly below 6T, except for very short
Ru–Npy distances (Ͻ 2.02 Å), for which the isomers become
isoenergetic. Distinct MECP values of +3.8 and
+6.6 kcalmol–1 were determined for 6S and 6T, respectively,
confirming a more hindered spin crossing. This is also con-
sistent with the computed spin-density distribution in 6T
(Figure S4), which shows smaller spin accumulation at the
N imido atom compared to CT (0.9 vs. 1.59 e2 bohr–3). The
trend is opposite at the ruthenium centre (0.72 vs.
0.31 e2 bohr–3).
Azide activation by 11 proceeds, as described before,
through the TS 5TS (Figure 2), which shows previously ob-
served structural features.[27] At this point, diatomic N2 is
about to be released, since the Nβ–Nγ distance is short as
1.14 Å, while its separation from the still-coordinated Nα
atom is already quite large (1.58 Å). In the process, the N3
Figure 2. Optimised structure of [Ru](py)(RN3)TS (5TS).
Eur. J. Inorg. Chem. 2015, 4885–4893
4889
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
·
The subtle but evident electronic differences do not seem
In the next step, the radical PhCH2 is fully separated
to preclude a radical pathway, which starts with 6T from the doublet [Ru](py)(NHR)D (14D in Figure 5, a) with
(Scheme 5) and can be related to the profile of Scheme 3. an energy gain close to that of the corresponding separation
Otherwise, the singlet 6S may lose its pyridine ligand to af- from 9D in Scheme 3 (–13.0 vs. –12.0 kcalmol–1). In 14D the
ford the intermediate [Ru](NR) (12) at the relatively small radical character at the amidic N atom (0.61 e2 bohr–3) is
energy cost of +7.4 kcalmol–1. As discussed below, the ca- almost twice that of the ruthenium atom, which suggests a
talysis may still proceed from 12 with a different pattern.
new prompt coupling with PhCH2· to yield the diamagnetic
product [Ru](py){HN(R)CH2Ph}S (15) of Figure 5 (b). The
energy balance is exergonic by –32.9 kcalmol–1.
Figure 5. Optimised structure of: a) ([Ru](py)(NHR)D (14D) and b)
[Ru](py){HN(R)CH2Ph} (15).
Finally, the separation of the benzylic amine
PhCH2NHR from 15 is endergonic by +21.1 kcalmol–1, an
energy cost similar to that calculated for removing one pyr-
idine ligand from the starting complex [Ru](py)2 (2). The
departure of the benzylic amine restores the catalyst
[Ru](py) (11), which is available for a new catalytic cycle.
Concerning the overall free-energy balance of the process,
the benzylic amination described in Equation (1) catalysed
by the pyridine complex 11 is more exergonic than that pro-
moted by the corresponding [Ru](CO) (1) catalyst (–43.3 vs.
–37.2 kcalmol–1, respectively). However, whereas 1 is di-
rectly available for azide activation, 11 must be first gener-
ated from the bis-pyridine precursor 2, which has a high
energy cost of +22.1 kcalmol–1.
Scheme 5. Energy profile for formation of the benzylic amine
PhCH2NHR starting from 6T.
As occurs at CT, the radical activation of the organic sub-
strate is already enhanced at the TS 13TS (Figure 4). The
corresponding barrier appears somewhat higher than that
of 8TS for the process promoted by 1 in Scheme 3 (+12.7
vs. +9.1 kcalmol–1), but in both cases one methylic H atom
of toluene already lies halfway between the methylic carbon
atom and the N acceptor, given that the C–H and N–H
distances in 13TS are 1.34 and 1.28 Å, respectively. At this
point, the spin densities at the Ru and Nimido atoms are
0.36, 0.98 e2 bohr–3, respectively, and the exocyclic carbon
atom of the tolyl ring has also large spin localisation
(0.60 e2 bohr–3).
RN3 + PhCH3 ǞPhCH2NHR + N2
(1)
Benzylic Amination Catalysed by [Ru](NR)2
Cycle (a) of Scheme 2 for the allylic amination promoted
by catalyst 1 is not the only possible process. In particular,
the bis-imido species [Ru](NR)2 (ES) can independently be-
have as a catalyst for the comparable amination process
shown in cycle (b) of Scheme 2.[27] Species ES is obtained
as a side product of cycle (a) when the mono-imido com-
plex CS is not promptly transformed into the triplet isomer
CT. In fact, the singlet CS has a chance of losing the weakly
bound CO ligand to form the five-coordinate complex
[Ru](NR), which allows access to cycle (b). Although no
mono-imido species of the type CS or CT has been ever
isolated, their computational study accounts for a series of
experimental facts. Another azide molecule can be an-
chored and activated at the vacant site of [Ru](NR), similar
to what happens at complex 1. The subsequent N2 loss gen-
Figure 4. Optimised structure of {[Ru](py)(NR)T*(PhCH3)TS
(13TS).
}
Eur. J. Inorg. Chem. 2015, 4885–4893
4890
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
erates the bis-imido singlet [Ru](NR)2 (ES), which can be
converted to the triplet isomer ET. Such an intersystem
crossing is significantly endergonic, at variance with the
CSǞCT process, but the cost of +16.1 kcalmol–1 may not
be an insurmountable barrier in view of the high working
temperatures. In ET,[27] the two unpaired spins are initially
shared by the two NR groups, but the C–H bond activation
still occurs only at one of them (Scheme 2) to provide the
doublet [Ru](NR)(NHR)D, in which the surviving unpaired
electron is largely localised at the amido ligand. The subse-
quent spin pairing of the amido complex with the pre-
·
viously separated allylic radical C6H9 is again consistent
with the rebound mechanism affording the allyl amine
(C6H9)NHR. The overall process is exergonic by
–40.6 kcalmol–1.
Figure 7. Optimised structure of [Ru](NR){HN(R)CH2Ph} (17).
Our present experimental and computational studies on
the catalytic activity of 2 in the benzylic amination of tolu-
ene also support the alternative formation of the diamag-
netic bis-imido complex [Ru](NR)2 (ES), from which a cycle
of type (b) may be initiated. Scheme 4 shows the energy
profile for the transformation of 2 into the five-coordinate
singlet [Ru](NR) (12), which provides access to cycle (b) of
Scheme 2, which was already described for the allylic amin-
ation.[27]
clusion is that the catalytic process, initiated by 1 or 2,
eventually becomes a unique type that involves the same
bis-imido species [Ru](NR)2. (ES).
Conclusions
To obtain a comparable quantitative overview for the
The present work has illustrated the experimental cata-
benzylic amination of PhCH3, corresponding DFT calcula- lytic activity of Ru(TPP)(py)2 (2) in promoting the amin-
tions were carried out starting from triplet isomer [Ru]- ation of benzylic C–H bonds by aryl azides. Complex 2
(NR)2 (ET). It was found that the initial adduct showed comparable behaviour to the already studied
[ET*PhCH3]TS (16TS) (Figure 6) is again a key TS with a Ru(TPP)(CO) (1), which suggests that the dependence of
ΔG barrier of +10.4 kcalmol–1, which is smaller than that the catalytic activity on the nature of different axial ligands
for the amination of cyclohexene (+14.0 kcalmol–1).[48] On is of secondary importance. Experimental results were cor-
the other hand, some evident energy differences emerge af- roborated by DFT analyses, which contributed to ratio-
·
ter 16TS: (1) the separation of the organic radical (PhCH2
nalising key aspects of the catalytic mechanism, also by
·
or C6H9 ) from the amido imido doublet [Ru](NR)- elaborating a previously developed model of the 1-catalysed
(NHR)D is exergonic by –16.4 and –26.8 kcalmol–1 for tolu- allylic amination of cyclohexene.
ene and cyclohexene, respectively; (2) the rebound mecha-
The loss of one pyridine ligand from the precatalyst 2
nism to form the complex [Ru](NR){HN(R)CH2Ph} (17) affords the active catalyst [Ru](py) (11) with a rather ender-
(Figure 7) is more exergonic than in the case of cyclohexene gonic balance, which seems to disfavour the catalytic pro-
amination (–36.0 vs. –27.8 kcalmol–1); (3) the energy cost cess, at least in the beginning. On the other hand, this step
to release the amine product is approximately halved for is essential to allow coordination of the azide reactant to
benzylamine (+10.5 vs. +21.1 kcalmol–1).
the metal centre and its subsequent activation with eco-
friendly release of N2. It was found that this process occurs
through a TS with common features in all the studied cases.
The subsequent mono-imido complexes {e.g., [Ru](py)(NR)
(6S) or [Ru](CO)(NR) (CS)} transform, more or less easily,
into their triplet isomers (6T or CT), which essentially pro-
mote the radical activation of a C–H bond of the studied
benzylic substrate. Two subsequent spin couplings allow the
formation of the benzylic amine product, according to the
known rebound mechanism.[34] Alternatively, the mono-
imido singlet CS or 6S may lose the apical ligand trans to
the NR ligand, which is CO or pyridine. The vacated coor-
dination site allows another azide activation, which in any
case leads to the same bis-imido complex [Ru](NR)2. This
complex may also exist as a singlet (ES) or a triplet (ET),
Figure 6. Optimised structure of [ET*PhCH3]TS (16TS).
The somewhat contrasting trends do not clearly allow but only the latter supports the radical activation of the
one to establish which amination process (allylic vs. benz- benzylic substrate and affords benzylic amine through the
ylic) is more efficient. Conversely, the most important con- known rebound mechanism. Importantly, the process cata-
Eur. J. Inorg. Chem. 2015, 4885–4893
4891
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
centro di calcolo elettronico dell’Italia Nord-Orientale (ISCRA-CI-
NECA) (HP grant “HP10BEG2NO”) and the Centro Ricerche En-
ergia e Ambiente (CREA), Colle Val d’Elsa, Siena, Italy, for com-
putational resources.
lysed by ES and ET does not depend on whether the bis-
imido catalyst was generated from precursor 1 or 2.
Experimental Section
[1] E. Dux, Chem. Rev. (Deddington, U. K.) 2009, 19, 11–15.
[2] J. C. Lewis, R. G. Bergman, J. A. Ellman, Acc. Chem. Res.
2008, 41, 1013–1025.
[3] Organic Azides: Syntheses and Applications (Eds.: S. Bräse, K.
Banert), Wiley, Chichester, UK, 2010.
General Conditions: All reactions were performed in a nitrogen at-
mosphere by employing standard Schlenk techniques and vacuum-
line manipulations. Benzene, toluene, cyclohexene, cumene, pyr-
idine and dichloromethane were purified by distillation under ni-
trogen in the presence of CaH2 or Na. All the other starting materi-
als were commercial products used after degasification.
Solvents and Reagents: 3,5-Bis(trifluoromethyl)phenyl azide,[49] 4-
tert-butylphenyl azide,[50] TPPH2,[51] Ru(TPP)(CO) (1),[52]
Ru(TPP)(CO)(MeOH),[36] Ru(TPP)(py)2 (2),[41] Ru(TPP)(NAr)2 (7)
{Ar = 3,5-bis(trifluoromethyl)phenyl}[18] and methyl hydrocinna-
mate[53] were synthesised by methods reported in the literature or
using minor modifications thereof.
[4] S. Bräse, C. Gil, K. Knepper, V. Zimmermann, Angew. Chem.
Int. Ed. 2005, 44, 5188–5240; Angew. Chem. 2005, 117, 5320.
[5] S. Cenini, E. Gallo, A. Caselli, F. Ragaini, S. Fantauzzi, C.
Piangiolino, Coord. Chem. Rev. 2006, 250, 1234–1253.
[6] S. Cenini, F. Ragaini, E. Gallo, A. Caselli, Curr. Org. Chem.
2011, 15, 1578–1592.
[7] S. Chiba, Synlett 2012, 23, 21–44.
[8] T. G. Driver, Org. Biomol. Chem. 2010, 8, 3831–3846.
[9] T. Katsuki, Chem. Lett. 2005, 34, 1304–1309.
[10] M. Minozzi, D. Nanni, P. Spagnolo, Chem. Eur. J. 2009, 15,
7830–7840.
Instruments: NMR spectra were recorded at 300 K (unless other-
1
wise specified) and 300 or 400 MHz for H. IR, UV/Vis and mass
[11] M. Y. Weber, G. Yilmaz, G. Wille, Chim. Oggi-Chem. Today
2011, 29, 8–10.
[12] S. Cenini, E. Gallo, A. Penoni, F. Ragaini, S. Tollari, Chem.
Commun. 2000, 2265–2266.
[13] S. Fantauzzi, A. Caselli, E. Gallo, Dalton Trans. 2009, 5434–
5443.
[14] F. Ragaini, A. Penoni, E. Gallo, S. Tollari, C. L. Gotti, M.
Lapadula, E. Mangioni, S. Cenini, Chem. Eur. J. 2003, 9, 249–
259.
[15] A. Caselli, E. Gallo, S. Fantauzzi, S. Morlacchi, F. Ragaini, S.
Cenini, Eur. J. Inorg. Chem. 2008, 3009–3019.
[16] T. W.-S. Chow, G.-Q. Chen, Y. Liu, C.-Y. Zhou, C.-M. Che,
Pure Appl. Chem. 2012, 84, 1685–1704.
[17] S. Fantauzzi, E. Gallo, A. Caselli, C. Piangiolino, F. Ragaini,
S. Cenini, Eur. J. Org. Chem. 2007, 6053–6059.
[18] S. Fantauzzi, E. Gallo, A. Caselli, F. Ragaini, N. Casati, P.
Macchi, S. Cenini, Chem. Commun. 2009, 3952–3954.
[19] D. Intrieri, A. Caselli, F. Ragaini, S. Cenini, E. Gallo, J. Por-
phyrins Phthalocyanines 2010, 14, 732–740.
[20] D. Intrieri, A. Caselli, F. Ragaini, P. Macchi, N. Casati, E.
Gallo, Eur. J. Inorg. Chem. 2012, 569–580.
[21] D. Intrieri, M. Mariani, A. Caselli, F. Ragaini, E. Gallo, Chem.
Eur. J. 2012, 18, 10487–10490.
[22] Y. Liu, C.-M. Che, Chem. Eur. J. 2010, 16, 10494–10501.
[23] H.-J. Lu, V. Subbarayan, J.-R. Tao, X. P. Zhang, Organometal-
lics 2010, 29, 389–393.
[24] H. Lu, H. Jiang, L. Wojtas, X. P. Zhang, Angew. Chem. Int.
Ed. 2010, 49, 10192–10196; Angew. Chem. 2010, 122, 10390.
[25] H. Lu, J. Tao, J. E. Jones, L. Wojtas, X. P. Zhang, Org. Lett.
2010, 12, 1248–1251.
[26] H. Lu, X. P. Zhang, Chem. Soc. Rev. 2011, 40, 1899–1909.
[27] a) G. Manca, E. Gallo, D. Intrieri, C. Mealli, ACS Catal. 2014,
4, 823–832; b) P. Zardi, A. Pozzoli, F. Ferretti, G. Manca, C.
Mealli, E. Gallo, Dalton Trans. 2015, 44, 10479–10489.
[28] C. Piangiolino, E. Gallo, A. Caselli, S. Fantauzzi, F. Ragaini,
S. Cenini, Eur. J. Org. Chem. 2007, 743–750.
[29] P. Zardi, D. Intrieri, A. Caselli, E. Gallo, J. Organomet. Chem.
2012, 716, 269–274.
[30] C.-M. Che, V. K.-Y. Lo, C.-Y. Zhou, J.-S. Huang, Chem. Soc.
Rev. 2011, 40, 1950–1975.
spectra were recorded in the analytical laboratories of Milan Uni-
versity.
Computational Details: All calculations were carried out with the
Gaussian 09 package[54] at the B97D-DFT[46] level of theory. The
TPP ligand (C44H28N4) was replaced by porphine (C20H14N4),
which has H atoms in place of Ph groups at the meso positions, in
all DFT calculations. In addition, the simple CH3N3 was used in
place of the experimentally used 3,5-bis(trifluoromethyl)phenyl az-
ide reactant. All optimised structures were validated as minima
and/or transition states by calculation of vibrational frequencies.
The transition states along the reaction pathways were also vali-
dated through Intrinsic Reaction Coordinate (IRC) calculations.[55]
All calculations were based on the CPCM[56] model for the benzene
solvent, the same as used in the experiments. The effective Stutt-
gart/Dresden core potential (SDD)[57] was adopted for the ruth-
enium centre, while for all the other atomic species the basis set
was 6-31G, with addition of the polarisation functions (d,p). The
coordinates of all the optimised structures are reported in the Sup-
porting Information.
Synthesis of Ru(TPP)(dmso)2 (4):[58] Ru(TPP)(py)2 (13.7 mg,
1.54ϫ10–2 mmol) was suspended in DMSO (2.0 mL), and the re-
sulting solution was heated at 110 °C for 4.0 h. The solution was
evaporated to dryness and the crystalline violet solid was dried in
vacuo. Analytical data were in accord with those reported in the
literature.
General Procedures for Catalytic Reactions: In a typical run, the
aryl azide and the ruthenium catalyst (6.0 mg, 6.8ϫ10–3 mmol)
were dissolved in the hydrocarbon (15 mL). The resulting mixture
was heated in a preheated oil bath until complete consumption of
the azide. The catalytic reaction was monitored by IR spectroscopy
by measuring the characteristic N3 absorbance at about 2115 cm–1.
The reaction was considered finished when the absorbance value
of the azide was less than 0.01 (by using a 0.1 mm-thick cell). The
1
solvent was evaporated to dryness and the residue analysed by H
NMR analysis with 2,4-dinitrotoluene as the internal standard.
[31] A. Caselli, E. Gallo, F. Ragaini, F. Ricatto, G. Abbiati, S. Cen-
ini, Inorg. Chim. Acta 2006, 359, 2924–2932.
[32] H. Lu, Y. Hu, H. Jiang, L. Wojtas, X. P. Zhang, Org. Lett.
2012, 14, 5158–5161.
Analytical data of 3a,[19] 3b,[19] 3c,[59] 3d,[19] 3e,[19] 3f,[19] 3g,[33] 3h[33]
and 3i[60] were in accord with those reported in the literature.
[33] P. Zardi, A. Caselli, P. Macchi, F. Ferretti, E. Gallo, Organome-
tallics 2014, 33, 2210–2218.
[34] J. T. Groves, J. Chem. Educ. 1985, 62, 928–931.
[35] G. M. Brown, F. R. Hopf, J. A. Ferguson, T. J. Meyer, D. G.
Whitten, J. Am. Chem. Soc. 1973, 95, 5939–5942.
Acknowledgments
C. M. and G. M. acknowledge the Italian SuperComputing Re-
search Allocation-Consorzio Interuniversitario per la gestione del
Eur. J. Inorg. Chem. 2015, 4885–4893
4892
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[36]
[37]
[50] J. B. Gerken, M. L. Rigsby, R. E. Ruther, R. J. Pérez-Rodríg-
uez, I. A. Guzei, R. J. Hamers, S. S. Stahl, Inorg. Chem. 2013,
52, 2796–2798.
[51] A. D. Adler, F. R. Longo, J. D. Finarelli, J. Goldmacher, J. Ass-
our, L. Korsakoff, J. Org. Chem. 1967, 32, 476.
[52] J. P. Collman, A. O. Chong, G. B. Jameson, R. T. Oakley, E.
Rose, E. R. Schmittou, J. A. Ibers, J. Am. Chem. Soc. 1981,
103, 516–533.
J. P. Collman, C. E. Barnes, P. N. Swepston, J. A. Ibers, J. Am.
Chem. Soc. 1984, 106, 3500–3510.
J. P. Collman, S. T. Harford, S. Franzen, J.-C. Marchon, P.
Maldivi, A. P. Shreve, W. H. Woodruff, Inorg. Chem. 1999, 38,
2085–2092.
S. C. Jeoung, D. Kim, D. W. Cho, M. Yoon, K.-H. Ahn, J.
Phys. Chem. 1996, 100, 8867–8874.
D. Kim, Y. O. Su, T. G. Spiro, Inorg. Chem. 1986, 25, 3993–
3997.
I. R. Paeng, K. Nakamoto, J. Am. Chem. Soc. 1990, 112, 3289–
3297.
S. E. Vitols, R. Kumble, M. E. Blackwood, J. S. Roman, T. G.
Spiro, J. Phys. Chem. 1996, 100, 4180–4187.
M. Barley, D. Dolphin, B. R. James, C. Kirmaier, D. Holten,
J. Am. Chem. Soc. 1984, 106, 3937–3943.
[38]
[39]
[40]
[41]
[42]
[53] A. W. Ingersoll, Org. Synth. 1929, 9, 42.
[54] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, 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. Starov-
erov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell,
J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M.
Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Ad-
amo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Mar-
tin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09,
revision B.01, Gaussian, Inc., Wallingford CT, 2009.
[55] K. Fukui, Acc. Chem. Res. 1981, 14, 363–368.
[56] a) V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102, 1995–
2001; b) M. Cossi, N. Rega, G. Scalmani, V. Barone, J. Comput.
Chem. 2003, 24, 669–681.
[43] C. D. Tait, D. Holten, M. Barley, D. Dolphin, B. R. James, J.
Am. Chem. Soc. 1985, 107, 1930–1934.
[44] a) J. P. Collman, J. I. Brauman, J. P. Fitzgerald, J. W. Spara-
pany, J. A. Ibers, J. Am. Chem. Soc. 1988, 110, 3486–3495; b)
Y. Li, P. W. Hong Chan, N.-Y. Zhu, C.-M. Che, H.-L. Kwong,
Organometallics 2004, 23, 54–66.
[45] a) F. R. Hopf, T. P. O’Brien, W. R. Scheidt, D. G. Whitten, J.
Am. Chem. Soc. 1975, 97, 277–281; b) F. Malvolti, P. Le Maux,
L. Toupet, M. E. Smith, W. Y. Man, P. J. Low, E. Galardon, G.
Simmonneaux, F. Paul, Inorg. Chem. 2010, 49, 9101–9103.
[46] S. Grimme, J. Comput. Chem. 2006, 27, 1787–1799.
[47] a) J. N. Harvey, R. Poli, K. M. Smith, Coord. Chem. Rev. 2003,
238–239, 347–361; b) J. N. Harvey, M. Aschi, J. Chem. Soc.
Faraday Trans. 2003, 124, 129–143.
[48] A very reliable comparison of the barriers is not possible, since
for technical reasons, the energy of the cyclohexene TS was
estimated only as ΔE (electronic energy). In these terms, the
barrier is higher for cyclohexene than toluene (+3.0 vs.
+0.2 kcalmol–1), but the difference should remain also in terms
of the free energy ΔG, under the assumption that the entropic
component computed for toluene (–TΔS) is +10 kcalmol–1.
[49] M. Tanno, S. Sueyoshi, S. Kamiya, Chem. Pharm. Bull. 1982,
30, 3125–3132.
[57] M. Dolg, H. Stoll, H. Preuss, R. M. Pitzer, J. Phys. Chem.
1993, 97, 5852–5859.
[58] E. Gallo, A. Caselli, F. Ragaini, S. Fantauzzi, N. Masciocchi,
A. Sironi, S. Cenini, Inorg. Chem. 2005, 44, 2039–2049.
[59] S. Zhou, S. Fleischer, H. Jiao, K. Junge, M. Beller, Adv. Synth.
Catal. 2014, 356, 3451–3455.
[60] F. Ragaini, S. Cenini, F. Turra, A. Caselli, Tetrahedron 2004,
60, 4989–4994.
Received: June 15, 2015
Published Online: September 18, 2015
Eur. J. Inorg. Chem. 2015, 4885–4893
4893
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim