Indoles from Alkynes and Aryl Azides: Scope and Theoretical Assessment of Ruthenium Porphyrin-Catalyzed Reactions

A Journal of

Accepted Article

Title: Indoles from Alkynes and Aryl Azides. Scope and Theoretical

Assessment of Ruthenium Porphyrin-Catalyzed Reactions

Authors: Emma Gallo, Daniela Intrieri, Daniela Maria Carminati, Paolo

Zardi, Caterina Damiano, Gabriele Manca, and Carlo Mealli

This manuscript has been accepted after peer review and appears as an

Accepted Article online prior to editing, proofing, and formal publication

of the final Version of Record (VoR). This work is currently citable by

using the Digital Object Identifier (DOI) given below. The VoR will be

published online in Early View as soon as possible and may be different

to this Accepted Article as a result of editing. Readers should obtain

the VoR from the journal website shown below when it is published

to ensure accuracy of information. The authors are responsible for the

content of this Accepted Article.

To be cited as: Chem. Eur. J. 10.1002/chem.201904224

Link to VoR: http://dx.doi.org/10.1002/chem.201904224

Supported by

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

Indoles from Alkynes and Aryl Azides. Scope and Theoretical

Assessment of Ruthenium Porphyrin-Catalyzed Reactions

Daniela Intrieri,^[a]Daniela Maria Carminati,^[b]Paolo Zardi,^[c]Caterina Damiano,^[a]Gabriele Manca,*^[d]

Emma Gallo*^[a]and Carlo Mealli^[d]

Abstract: A symbiotic experimental/computational study analyzed

the Ru(TPP)(NAr)₂-catalyzed one-pot formation of indoles from

alkynes and aryl azides. Thirty different C₃-substituted indoles were

synthesized and best performances, in term of yields and

regioselectivities, were observed by reacting ArC≡CH alkynes with

3,5-(EWG)₂C₆H₃N₃azides, whilst the reaction was less efficient by

using electron-rich aryl azides. A DFT analysis describes the

reaction mechanism in terms of the energy costs and

orbital/electronic evolutions; the limited reactivity of electron-rich

azides was also justified. In summary, PhC≡CH alkyne interacts with

one NAr imido ligand of Ru(TPP)(NAr)₂to give a residually dangling

C(Ph) group which, by coupling with a C(H) unit of the N-aryl

substituent, forms a 5+6 bicyclic molecule. In the process, two

subsequent spin changes allow inverting the conformation of the sp²

C(Ph) atom and its consequent electrophilic-like attack to the

aromatic ring. The bicycle isomerizes to indole via a two-steps outer

sphere H-migration. Eventually, a ‘Ru(TPP)(NAr)’ mono-imido active

catalyst is reformed after each azide/alkyne reaction.

is responsible for anti-hypertensive, anti-inflammatory, anti-

tumor, anti-migraine, anti-cholesterol and many other

pharmacological activities (Figure 1).^[2]

Figure 1. Selected examples of bioactive and pharmaceutical indole-

containing compounds.

Introduction

The indole ring represents one of the most active

pharmacophores in Nature, which confers to indole-containing

Since the publication of the pioneering Fischer’s synthesis of

indoles,^[3]a great attention has always been devoted to design

efficient strategies to synthesize^[4]and functionalize^[5]indoles.

Although many metal-free procedures have successfully been

applied,^[6]transition metal catalyzed-methodologies have

emerged as more attractive methods due their capability in

controlling the selectivity and conditions of the processes.^[7]In

order to optimize the performance of the methodology without

affecting the process eco-sustainability, reaction components

are rigorously selected, and either ortho-disubstituted or mono-

functionalized aromatic compounds have been found to

represent valuable starting materials. Even if ortho-

disubstituted^[8]species are reagents of choice in several

synthetic procedures (Figure 2a), their use often requires time-

consuming pre-functionalization steps thus, the employment of

mono-substituted aromatic reactants is constantly increasing.

Mono-substituted species can be efficiently transformed into

molecules

a plethora of biological and/or pharmaceutical

features.^[1]The molecular skeleton of several naturally-occurring

and clinically employed molecules shows an indole motif, which

[a]

PhD D. Intrieri, Dr. C. Damiano, Prof. E. Gallo

Department of Chemistry

University of Milan

Via Golgi 19, I-20133 Milan

E-mail: emma.gallo@unimi.it

PhD D. M. Carminati.

Department of Chemistry

University of Rochester

416 Hutchison Hall, NY 14627-0216

PhD P. Zardi

Department of Chemical Sciences

University of Padua

Via F. Marzolo 1, I-35131 Padua

PhD G. Manca, Dr. C. Mealli

Istituto di Chimica dei Composti OrganoMetallici

ICCOM-CNR

[b]

9]

indoles through C-H bond activation reactions^[4a,and many

synthetic procedures, which employ N-aryl enamines^[10], N-aryl

imines,^[11]nitroarenes,^[12]nitrosoarenes,^[13]anilines,^[14]anilides^[15]

as the starting materials, have been reported up to now (Figure

2b).

Via Madonna del Piano 10, I-50019 Sesto Fiorentino

E-mail: gabriele.manca@iccom.cnr.it

Supporting information for this article is given via a link at the end of

the document.

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

Scheme 1. General scheme of ruthenium porphyrin-catalyzed synthesis of

indoles from aryl azides and alkynes.

In addition to the study of the reaction scope, a mechanistic

study, based on the availability of many experimental and

theoretical data, added rich information on the evolution of the

reaction mechanism that involves both the inner and outer

sphere of the catalyst. Thus, the paper clarifies various key

aspects of the intriguing reactivity also providing important

suggestions for the process optimization.

Figure 2. Selection of ortho-disubstituted and mono-substituted aromatic

reagents.

Results and Discussion

Study of the Reaction Scope

Indoles can be also obtained by ring-closure reactions of aryl

azides (ArN₃),^[16]which are a sustainable and atom-efficient

class of aminating agents thanks to the formation of benign N₂

as the only by-product of the annulation process. The catalytic

performance of this class of molecules is due to the formation of

a very reactive nitrene intermediate ‘ArN’ which is responsible

for amination reactions.^[16c]Figure 3 shows the most relevant

azide molecules which have been employed for the catalytic

intramolecular synthesis of indoles.

As previously communicated,^[17]the comparison of the catalytic

efficiency of different ruthenium porphyrins in promoting the

model reaction of 3,5-bis(trifluoromethyl) phenyl azide (1a) with

phenylacetylene (2a) pointed out that ruthenium bis-imido

Ru(TPP)(NAr)₂(5) (TPP = dianion of tetraphenyl porphyrin, Ar =

3,5-(CF₃)₂C₆H₃) complex^[19]performed the best. The optimization

of the experimental conditions allowed us obtaining differently

C₃-substituted indoles in high yields and short reaction times.

The catalytic activity of 5 was then further investigated and new

obtained data are here discussed together with those already

reported in our communication^[17]for providing a more complete

scenario on the applicability of the synthetic methodology.

The efficiency of complex 5 was first investigated by reacting

several substituted azides with phenylacetylene (2a), and

resulted data are showed in Table 1. The indole formation

occurs by the activation of the C-H bond^[17]which is placed onto

the ortho position with respect to the N₃functionality of the

starting azide. The presence of electron withdrawing groups

(EWG_s) on meta positions of azide had a positive effect on the

reaction productivity while, when electron-rich 3-^tBu-phenyl

azide, 4-^tBu-phenyl azide and 3,4,5-trimethoxyphenyl azide were

reacted with 2a, the formation of corresponding indoles was not

observed (vide infra for the rationalization of this effect).

Figure 3. Selection of aryl azides employed for the synthesis of indoles.

During our ongoing studies on the use of aryl azides as nitrene

sources, we discovered that azides are very reactive towards

alkynes in forming indoles rather than triazoles in the presence

of ruthenium porphyrin complexes. In our previous

communication,^[17]we reported the one-pot synthesis of

differently substituted indoles (Scheme 1) thus, in view of the

interest in developing sustainable methods for achieving such

valuable fine-chemicals, the catalytic activity of ruthenium

porphyrins was further investigated. Note that to the best of our

knowledge, only one other example was recently published on

the reaction of sulfonyl azide with isocyanides yielding 3-imine

indoles by an intermolecular mechanism.^[18]

Considering that the ring-closing reaction yielding indoles

involves one of the two C-H bonds adjacent to the N₃group,

isomers 3 and 4 (Scheme 1 and Table 1) can both be formed in

a ratio which depends on the electronic nature of azide

substituents. Clearly, the use of 3,5-disubstited azides, with

R¹= R³, resulted in the formation of only one regiosomer due to

the chemical equivalence of the two aromatic C-H bonds. The

reaction of either 3,5-(CF₃)₂C₆H₃N₃(1a) or 3,5-(NO₂)₂C₆H₃N₃

(1b) with 2a afforded corresponding 3aa and 3ba indoles in very

good yields (86% and 82%, respectively) and short reaction

times (1.0 h and 0.5 h, respectively).

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

bond was always favored. The high directing capability of CF₃

permitted the formation of 3ga indole in 82% yield and 4ga

isomer was only formed in small amounts (5% yield), as

revealed by GC-MS and ¹⁹F NMR analyses. The worse

regioselectivity, which was observed when 1f azide reacted with

2a to form a 3fa/4fa mixture, was ascribed to the mesomeric

effect of the nitro group.

Table 1. Complex 5-catalyzed synthesis of indoles 3 and 4 from 2a and aryl

azides 1a-g.^a,b

Then, in order to enlarge the reaction scope, the most reactive

and selective 3,5-disubstituted aryl azides 1a and 1b were

reacted with terminal aromatic alkynes 2b-n (Table 2).

Table 2. Synthesis of indoles of type 3 from the 5-catalyzed reaction of aryl

azides (1a-b) with alkynes (2b-n).^a,b

[a] Catalyst 5 (0.01 mmol) in 10.0 mL of refluxing benzene under N₂, mol ratio

5/azide/2a = 1:50:250. [b] Isolated yield. [c] Ref 17. [d] Mol ratio 5/azide/2a =

1:50:1000. [e] NMR yield calculated in the 3/4 mixture.

The high regioselectivity of the reaction was proved by the

formation of C₃-substituted isomer as the sole reaction product,

not traces of C₂‑substituted indole were obtained. It is important

to underline that when 1b azide was employed as the reactant,

the corresponding 3ba indole precipitated from the reaction

mixture and the desired product was isolated in a pure form by a

simple filtration at the end of the catalytic reaction. This

convenient work-up of the reaction rendered the procedure more

sustainable because any other expensive and time-consuming

purification procedures, such as flash chromatography, were not

required. The reaction of 3,5-(Cl)₂C₆H₃N₃(1c) with 2a formed

3ca indole in a 60% yield pointing out that the reaction efficiency

decreases by decreasing the electron withdrawal power of azide

substituents.

Both 3 and 4 indole isomers were obtained either when two

different substituents were present onto the 3 and 5 positions of

the azide or 3,4-disubstituted aryl azides were employed as

starting materials, due to the presence of two chemically

inequivalent C-H aromatic bonds nearby the N₃moiety. The

reaction of 3-(CF₃)-5-(Cl)C₆H₃N₃(1d) or 3-(NO₂)-4-(Me)C₆H₃N₃

(1e) azides with 2a yielded indoles 3da and 3ea as the most

abundant isomers suggesting that the reaction pathway favors

the formation of indoles which bear the most electron

withdrawing group on the 4 position of the heterocycle. The 4-

substituted indoles 3fa and 3ga were also obtained, as principal

isomers, when either 3-(NO₂)C₆H₄N₃(1f) or 3-(CF₃)C₆H₄N₃(1g)

azides were employed as the reagents. Even if, an electron

withdrawing R³substituent (CF₃or NO₂group) can activate C-H

bonds which are located either in ortho or para position referring

to R³, achieved data indicated that the cleavage of an ortho C-H

[a] Catalyst 5 (0.01 mmol) in 10.0 mL of refluxing benzene under N₂, mol ratio

5/azide/alkyne = 1:50:250. [b] Isolated yield. [c] Ref 17.

Achieved results suggest a modest dependence of the reaction

productivity on electronic characteristics of the employed alkyne.

Even if, similar indole yields were observed by using both

electron-rich and electron-poor alkynes in the reaction with 1a or

1b, in many cases shortest reaction times (see Experimental

Section for details) were achieved by employing alkyne reagents

bearing an electron donating group (EDG) onto the para

aromatic position (e.g. compare 3ae, 3af, 3be with 3ab, 3ad,

3bh, Table 2).

Conversely to the observed low electronic effect, the steric

hindrance of alkynes influenced the reaction productivity to a

greater extent. Even if mono-substituted alkynes were efficiently

converted into corresponding indole independently for the

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

position where the substituent was placed (3bi and 3bk, Table

2), the presence highly substituted aryl alkyne moiety was

responsible for a decrease of the reaction yield; the reaction of

1b with the sterically hindered 2,4,5-(Me)₃C₆H₂CCH (2n)

species yielded the corresponding indole 3bn in the low 37%

yield (Table 2). The procedure did not tolerate the presence of

reactive functional groups such as an aldehyde and the reaction

of 1b with 2l yielded indole 3bl in a very low yield (17%) due to

the formation of several side-products which were not identified.

It is important to remind that all the indoles (3bb, 3bd, 3bh, 3be,

3bg, 3bi, 3bj, 3bk, 3bl, 3bn, Table 2) which were obtained by

using 3,5-(NO₂)₂C₆H₃N₃(1b) as the precursor, were directly

collected from the reaction mixture by a simple filtration.

leading the C₂-iminated product. Thus, in order to favor the

conversion of the azide into corresponding indole with respect to

the C₂-imination side-reaction, an alkyne excess was always

employed.

In summary, the study of the reaction scope disclosed that the

procedure is highly effective for producing indoles from electron-

poor azides and terminal aromatic alkynes whilst, the reaction

efficiency dropped by using either electron-rich azides or internal

as well as aliphatic alkynes. Then, in order to examine the power

and limitations of the synthetic procedure, the catalytic

mechanism was investigated by both experimental and

computational analyses.

Finally, we studied the reactivity of internal alkynes PhCCPh

(2o) and PhCCMe (2p) towards 1a or 1b azides but

unfortunately, the synthesis of indoles 3ao, 3ap and 3bo (Table

3) occurred in low yields and long reaction times due to the

steric hindrance of starting alkynes (see Experimental Section).

Even if a modest productivity was always registered, only the

isomer which bears the more sterically hindered phenyl group on

the C₃position of the heterocycle, was isolated. Similar results

were obtained by reacting CO₂EtCCH (2q) with 1a where the

desired indole 3aq was isolated after 13 hours in the very low

yield of 13%. The methodology was ineffective when aliphatic

alkynes, such as trimethylsilyl acetylene or 1 heptyne, were the

selected reagents.

Mechanistic Investigation

In order to study the reaction mechanism, some experiments

were performed and reported in our previous communication.^[17]

A summary of those data is listed in the following (see SI for

experimental details): i) Indole isn’t formed by a rearrangement

process of a triazole molecule, which can be obtained by the

free-catalysts reaction of alkyne with azide. In fact, no reaction

was observed by treating catalyst 5 with a 4:1 mixture of 1-(3,5-

bis(trifluoromethyl)phenyl)-5-phenyl-1H-1,2,3-triazole and 1-(3,5-

bis(trifluoromethyl)phenyl)-4-phenyl-1H-1,2,3-triazole to exclude

a possible metal-catalyzed transformation of triazole into indole

by N₂release. ii) The imido “NAr” moiety of Ru(TPP)(NAr)₂(5)

can be directly transferred to the alkyne substrate, as supported

by the stoichiometric reaction between 5 and phenylacetylene

(2a) which yielded indole 3aa (Table 1) in 50% yield. iii) The

analysis of the reaction between phenylacetylene-d (PhC≡CD)

and 1a revealed that the C₂position of the obtained indole was

occupied by a deuterium atom to indicate that the C-D alkyne

bond is not broken during the annulation reaction. As a

consequence, a hydrogen-transfer (HAT) reaction should occur

from the aromatic C-H bond of azide to indole nitrogen atom to

form the NH bond. iv) The cleavage of the aromatic C-H bond

does not represent the rate-determining step of the catalytic

cycle, as suggested by the lack of an appreciable isotopic effect

(k_H/k_D= 1.1) in the reaction of 2a with an equimolar amount of

1a and its deuterated 3,5-(CF₃)₂C₆D₃N₃derivative; v) The

participation of radical mechanisms can be excluded since 3aa

was formed in comparable rates both in in the absence and in

the presence of the radical trapping agent TEMPO (2,2,6,6-

tetramethylpiperidin-1-yl)oxy).

Table 3. Complex 5-catalyzed synthesis of indoles from alkynes 2o-q and

aryl azides 1a or 1b.^a,b

On the basis of experimental results described above and

literature data, which report the formation of unstable

antiaromatic 1H-azirines^{[6, 21]}from the reaction of nitrenes with

acetylenes as well as the transformation of 2H-azirines into

[a] Catalyst 5 (0.01 mmol) in 10.0 mL of refluxing benzene under N₂, mol

ratio 5/azide/alkyne = 1:50:250. [b] Isolated yield. [c] Ref 17.

Considering the good affinity of ruthenium for oxygen-containing

species, the reduced catalytic performances, observed when

alkynes 2l (Table 2) and 2q (Table 3) were employed, can be

due to the coordination of the alkyne C=O group to ruthenium

atom, which causes the catalyst deactivation. This hypothesis

was supported by the decomposition of Ru(TPP)(NAr)₂(5) that

was observed when 5 was reacted with the sole 2l by using the

experimental conditions reported in Table 2.

indoles,^[22]we hypothesized^[17]the initial formation of

a

metastable three-membered ring. The potential mechanism to

transform N-aryl-1H-azirine into the desired indole 3aa, possibly

involving a ring-opening reaction, is presented in Scheme 2.

Unfortunately, no spectroscopic evidence supports the formation

of the mentioned azirine intermediate, hence any hint on the

feasibility of the process may emerge from an in silico analysis.

The corresponding details, reported in a section of the SI, seem

to exclude such a possibility

It should be noted that, as previously reported,^[20]ruthenium

porphyrin catalysts can promote the insertion of the nitrene

functionality of an aryl azide into the C₂-H bond of an indole

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

(a)

kinetic order with

respect to azide 1a

Scheme 2. Proposed pathway of the indole formation by a ring-opening

reaction of a putative N-aryl 1H-azirine.

-

time (s)

4,0E-03

3,0E-03

2,0E-03

1,0E-03

0,0E+00

(b)

kinetic order with respect

to Ru(TPP)(NAr)₂(5)

Kinetic Study.

The kinetic study of the 5-catalyzed indole formation was

undertaken by using the model reaction between

3,5-bis(trifluoromethyl)phenyl azide (1a) and phenylacetylene

(2a). First, the consumption of aryl azide was followed by IR

spectroscopy by monitoring the decrease of the intensity of the

1a absorption (A) at 2116 cm^-1in the reaction where the

5/azide/alkyne ratio of 2:50:250 was employed (see SI for

experimental details). Since the reaction was run in the

presence of a large 2a excess (pseudo zero-order in alkyne), the

linear fit of ln(A/Ao) (A = IR absorbance at each measure; Ao =

4,0E-04

9,0E-04

1,4E-03

1,9E-03

2,4E-03

[Ru(TPP)(NAr)₂] (mol/L)

1,0

0,8

0,6

0,4

0,2

0,0

(c)

kinetic order with

respect to alkyne 2a

initial IR absorbance) versus time indicated

a first-order

0,03

0,05

0,07

0,09

0,11

0,13

0,15

0,17

dependence of the reaction with respect to azide 1a

concentration (Figure 4a) and that no relevant catalyst

deactivation occurred during the reaction.

[phenylacetylene] (mol/L)

Figure 4. Kinetic dependence of the reaction rate with respect to 1a (a), 5 (b)

and 2a (c) concentrations.

The kinetic order of the reaction rate with respect to catalyst 5

concentration was investigated by repeating the experiment

described above in the presence of different amounts of the

catalyst. A first-order dependence on azide concentration was

always observed for each kinetic run (see SI) and the values of

the observed kinetic constants (k_obs) show a linear dependence

on the catalyst concentration (Figure 4b), indicating that the

reaction is also first-order in 5. This rules out any mechanistic

proposal in which either the porphyrin dimer formation or any

reaction involving more than one "Ru(TPP)" unit is involved.

The kinetic order with respect to the alkyne 2a was finally

investigated by recording the rate of a series of reactions run at

a constant catalyst concentration and a variable concentration of

the alkyne, which was always large enough to ensure that a

pseudo zero-order rate with respect to this reagent would be

observed. The observed first-order kinetic constants again

varies linearly with the alkyne concentration (Figure 4c, see SI

for a plot of all individual reaction profiles), proving that the

reaction rate is also first order in 2a.

Thus, the kinetic equation describing the reactions can be

formulated as -d[azide]/dt = k'_obs[Ru(TPP)(NAr)₂][azide][alkyne].

Since a trimolecular reaction can be excluded, a kinetic equation

of this kind implies that two of the reaction components must be

involved in a reversible interaction during the cycle. Considering

that 1a and 2a do not form indole 3aa in a free-metal reaction^[17]

and that triazoles cannot be transformed into indole in the

presence of the ruthenium catalyst (see above), kinetic data

indicates that either azide or alkyne should reversibly coordinate

the ruthenium catalytically active species A to give either

species B or D, as shown by the alternative path a or path b in

Scheme 3. It is important to note that the observed first-order

with respect to the catalyst 5 excludes a possible reaction

between species B and D, in which both reagents are activated

by the metal coordination. In fact, in this latter case a bimetallic

pathway should result in a second-order dependence with

respect to the catalyst concentration. The B or D adducts cannot

be experimentally detected because the reaction equilibrium

should be shifted to the left, towards the side of the azide and

alkyne reagents.

The two alternative mechanisms of the reaction between

phenylacetylene (but also another alkyne) and substituted aryl

azides (path a and path b of Scheme 3) differ for the inverted

order of the interaction between the ruthenium species and the

two involved reagents, with no distinction emerging from

kinetics.

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

emerge, such as a spin state change of the system, which

suitably rearranges the stereochemistry to allow the C-C

coupling, hence the formation of the bicyclic molecule (see

below).

Questionable reliability of path b in Scheme 3. The mono-

imido complex A is common to the two foreshadowed pathways.

Path b suggests the dihapto coordination of the alkyne to the

metal to give complex D, which would couple with one azide

reactant, allowing the ring-closing reaction to indole. Although

complex D could be fully optimized (see Figure S18), the

feasibility of path b could not be further corroborated, as

discussed in the SI.

Details of the more probable catalytic formation of indole

through path a. A definitely more realistic computational

evidence emerged on the evolution of the system through path a

of Scheme 3. In fact, all the important intermediates, transition

states (TS) and products were fully characterized, thus allowing

the determination of reasonable profiles from both the

stereochemical and energetic viewpoints. It emerged that the

process evolves thanks to the attainment of intermediates of

variable spin multiplicity.

Scheme 3. Ru-catalyzed indole formation.

The experimental Ru(TPP)(NAr)₂catalyst 5 plus all the reactants

sum up to > 100 atoms, hence all the computational models,

characterized by the letter m, have been slightly simplified. The

used porphine ligand has H atoms in place of the meso-aryl

substituents of the experimentally used porphyrin and one of the

two imido ligands is as simple as NMe in the bis-imido model

[Ru](NMe)(NAr) ([Ru] = Ru(porphine), Ar = 3,5-(CF₃)₂C₆H₃),

whose optimized ground state structure (singlet 5m_S) is shown in

Figure 5.

The very first step of Scheme 3 is the general stoichiometric

reaction between complex 5 and phenylacetylene 2a giving

indole 3aa, as a sacrificial side-product. Due to the presence of

two 3,5-(CF₃)₂C₆H₃N imido moieties as axial ligands of catalyst

5, the formation of active species A is inevitably accompanied by

the release of one equivalent of 3aa, independently from the

employed azide reactant. This side-reaction can be neglected in

view of the very low amount of 3aa (stoichiometric ratio with

respect to 5), which can be easily separated from the target

indole corresponding to the employed azide and alkyne

reactants. Similarly, the formation of 1a-derived indole is not a

drawback in all the indole syntheses here reported.

A detailed analysis of the reaction between 5 and 2a is highly

educational because implies the mechanistic underpinnings of

the whole catalytic cycle. The mentioned step generates the

mono-imido species A that is potentially the active catalyst for

either the cycle a or b, depending on the first formed

intermediate B or D (Scheme 3).

Analogously to other similar mechanisms of nitrene transfer

reactions mediated by ruthenium imido complexes,^[23]

intermediate B can easily lose N₂to form bis-imido complex C,

which could positively interact with alkyne yielding the desired

indole (Scheme 3, path a). Conversely, the dihapto coordination

of alkyne 2a to A would lead to intermediate D that may interact

with an azide also giving the final indole product (Scheme 3,

path b). In order to shed some light on the mechanism of the

indole formation, a systematic DFT study was performed.

Figure 5. Optimized structure of bis-imido singlet complex [Ru](NMe)(NAr)

(5m_S).

Its subsequent reaction with PhCCH (2a) and the following

steps were computationally monitored up to the release of the

indole product 3aa. At this point, the residual mono-imido

fragment [Ru](NMe) (Am) becomes the fundamental catalyst, as

it activates a new 1a azide molecule and regenerates the

precursor 5m_Sthrough a series of subsequently described steps.

The reliability of 5m_Sas a model catalyst was corroborated by a

satisfactory comparison with a known experimental structure.^[19b]

In our model, the 0.05 Å difference between the two Ru-N axial

Electronic Underpinnings of the indole formation

mechanism based on DFT calculations

The potential pathways presented in Scheme 3 represent a

useful guideline for the study of the reaction mechanism. Many

of the proposed steps are corroborated by computing the

evolution of the stereochemical, electronic and energetic

features. Unprecedented aspects of the chemical reactivity

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

[24]

distances suggests a possible steric hindrance between the aryl

substituent of the imido moiety and the porphine macrocycle.

Also remarkable are the bent and eclipsed N-C bonds (N₁-C₃

and N₁-C₅, Figure 5) of the two imido ligands, whose plane is

staggered with respect to any Ru-N_eqlinkage. Any deviation

from the given arrangement plays a key role in the tuning of the

reactivity.

triatomic CO₂

where eight delocalized  electrons determine

four half bonds, normally depicted as two consecutive double

ones.^[25]Also, the known R-N=C=N-R carbodiimide, which is

additionally relevant for its reactivity with group 4 and 5 imido

complexes,^[26]compares with 5m_Sfor the similarly bent NR

groups.^[27]

Also due to the delocalization of the  electrons in 5m_S, the in

plane  donor power of any imido sp²hybrid is limited, but still

appropriate for the interaction with the incoming alkyne (see

below). On the other hand, bis-imido complexes are also known

to promote diradical reactivity, as in the case of the allylic

amination and aziridination of unsaturated alkenes by azides.^[23c,

Basic  electron interactions in the bis-imido complex. The

MO description of the delocalized MeN=Ru=NAr _and _||

systems in Scheme 4 is helpful to interpret the overall electronic

evolution.

28]

For instance, the radical mechanism can be in principle

R

associated to the 5m_S5m_Tintersystem crossing, which

promotes one electron from the _n.b. lone pair to the higher

_||^*a.b. level (see Scheme 4). This process requires the not small

energy cost of ~+16 kcal mol^-1, which remains almost constant

for various imido ligands. It was reported the diradical triplet

promotes the C–H homolysis of an allylic substrate, while a

subsequent rebound mechanism restores the diamagnetism of

the system.^[29]

In the present case, the diradical mechanism can be safely

excluded for any missing evidence of homolysis and the

mentioned inactivity of TEMPO radical scavenger, when added

to the reaction mixture (see above). It must be anticipated that

the observed indole formation, first analyzed over the singlet

Potential Energy Surface (PES), requires at some point an

access to the triplet state for stereochemical and electronic

reasons to be later illustrated.

R

N

a.b.



N

a.b.

N



½

R

+

R

½

R

N

(2-)



½

N

Ru

N

(VI)

Ru

N

^N

N



½



½



^n.b.

n.b.

R

N

R

(2-)

R

N

R

N



b.

N

R

b.



R

Structural aspects of the imido-alkyne coupling. It is debated

in literature whether an alkyne can be prone to either a

nucleophile or an electrophilic attack.^[30]In principle, the former

mode is difficult for the unavailability of suitable vacant C p_

orbitals, which appear only in the high lying CC * levels.

These may be affected only by a rather strong base, which may

otherwise favor the PhCCH deprotonation. In contrast, both the

NR ligands of 5m_Sare weak donors with a more pronounced

effect at the NAr ligand for the presence of the two electron

withdrawing CF₃aromatic substituents.

Scheme 4. Delocalized _and _||sets of interactions in a bis-imido [Ru]

diamagnetic complex with the implications for the structural formulas

appearing in the box ( bonds in red).

None of the three sets of double levels is exactly degenerate

due to the bending of the two imido ligands in one plane. Overall

the diagram depicts the basic NR(p_)/Ru(d_) interactions, while

the role of the porphine  levels is in first approximation

excluded. A combination of delocalized imido lone pairs (_n.b.

and _||n.b.) lies in between the bonding and antibonding levels,

namely _b.+_||b. and _a.b.+_||a.b., respectively. The MO

diagram contains a total of 8 electrons, four of which are

attributed to the Ru(II) filled d_xz+d_yzorbitals, while the other four

belong to the two NR imido ligands. Such a picture also

emerged in our previous studies of the azide activation by [Ru],

where the departing N₂leaves 2e^-at the coordinated NR

grouping.^[23c]The available Ru d_electrons preferentially localize

at the trans-axial N atoms for electronegativity reasons, thus

formally implying an oxidized Ru(VI) ion and two dianionic imido

ligands. The >60% Ru character of the highest _||a.b. and _a.b.

vacant levels corroborates the existence of two Ru=N trans-axial

double bonds, as indicated by the first structural formula in the

box of Scheme 4. In actuality, the latter picture is better

described by four half  bonds with 1e^-in each case, associated

to two dative NRu  linkages (red arrows). Incidentally, the

MO  picture of Scheme 4 recalls that of the much simpler linear

Examples of the coupling between an imido ligand with an

alkyne were reported for early Transition Metal (TM)

complexes,^[31]while corresponding reactions involving late TM

systems have been less investigated. The computationally

ascertained energy profile of Scheme

5 starts from the

separated species [Ru](NMe)(NAr) (5m_S) and phenylacetylene

(2a), which jointly represent the initial zero energy point.

The optimized structure of the weak adduct 6m_S, as the first

encountered minimum, is shown at the top left side of Figure

6.^[32]At its right side, the transition state 7m_S-TSwas first located

via a relaxed scan technique that progressively shortened

N₁···C₁separation from 3.88 Å to the bonding value.

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

related ruthenium salen complex.^[34]For comparison, an

analogous path over the triplet PES is unlikely in view of the

~+20.0 kcal mol^-1higher barrier of 7m_T-TSin Figure S19. The

final singlet point 7m_S(0°)(Figure 6) is characterized by a

C₅N₂N₁C₃dihedral angle of 0° that implies the co-planarity of the

trans imido and amido axial ligands. The -16.4 kcal mol^-1energy

gain in 7m_S(0°)vs. 7m_S-TSis mainly due to the exergonicity of the

newly formed N₁-C₁ bond of 1.42 Å, which does not exclude

some  delocalization over the adjacent C₁=C₂linkage of

1.33 Å.^[35]The singlet minimum 7m_S(0°)also exists as the 4.3 kcal

mol^-1more stable triplet isomer 7m_T(0°)in Figure S20, which

minimally differs for conformation and geometry. Under these

circumstances, no reaction coordinate was identified for

detecting the intersystem crossing barrier. However, 7m_T(0°)

further converts into the another triplet isomer, namely 7m_T(90°)in

Figure 6, which lies lower by only -0.6 kcal mol^-1. This occurs

with an evident structural rearrangement with the 90° rotation of

the trans-axial amido and imido ligands. About halfway (~40°),

the Transition State 7m_T-TS(40°)was identified (see Figure S21)

with the flat energy barrier of +0.3 kcal mol^-1. The dangling chain

Ru-N₁-C₁(H)=C₂Ph in all the structures of Figure 6 is indicative of

a terminal vinyldenic anion, likely formed upon the in plane lone

pair transfer from the N₁atom to the terminal C₂one. Similar

features were previously detected by some of us (C.M. and G.M)

in an experimental/theoretical study of an alkyne rearrangement

over a Co-Co linkage.^[36]Another similar example of a vinylidenic

dangling chain has been proposed for a catalytic metallo-nitrene

alkyne metathesis.^[37]In any case, the terminal Ph substituent

leans toward the aryl moiety at the N atom with potential steric

problems for the five-membered ring closure (C₂-C₄linkage,

Figure 6). More importantly, the C₂sp²conformation appears

inadequate for the coupling due to the improper orientation of

the C₂ hybrid. The orbital/electronic underpinnings of the

imido-alkyne coupling are described in the following section.

Scheme 5. Energy profile for the 5m_S+2a coupling, characterized by an

intersystem crossing and the 90° rotation of the new amido ligand

(7m_T(0°)7m_T(90°)transformation).

Orbital/electron rearrangements in the imido-alkyne

coupling preceding the ring closure. Figure 7 shows the

ensemble of the three 1_||, 2_||and 1_frontier MOs, which in

both 7m_S(0°)and 7m_T(0°)have maximum width of ~0.80 eV. The

two electrons available to the levels in question are consistent

with both the singlet and the triplet state and, since the latter is

more stable by only -4.3 kcal mol^-1, intersystem crossing cannot

to be excluded.

Figure 6. The 6m_Sadduct and the transition state 7m_S-_TSupon N₁···C₁

coupling; the subsequent singlet complex 7m_S(0°)and its triplet isomer 7m_T(90°)

with orthogonal imido/amido ligands.

The unique imaginary frequency at -383.5 cm^-1of the fully

optimized species 7m_S-_TSis indicative of the N₁-C₁stretching

with the assignment further corroborated by an IRC analysis.

Although in 7m_S-TSthe N₁-C₁separation is still as large as

1.98 Å,^[33]the lost linearity of the alkyne moiety (H₁-C₁-C₂and

C₁-C₂-C_ipsoangles of 147.0° and 157.0°, respectively) implies an

electron density transfer from the N₁atom into a _||* level. The

latter point is also confirmed by the 0.04 Å elongation of the

original C₁C₂triple bond consistently with the computed 90 cm^-1

red shift of the stretching frequency.

Figure 7. The close frontier levels of 7m_S(0°)and 7m_T(0),where 1_||and 2_||are

either the HOMO and LUMO or the two SOMOs, respectively. In the triplet,

1_may also compete for electrons.

The total energy barrier of +18.3 kcal mol^-1to reach 7m_T-TSmay

be indicatively compared with the +22.9 kcal mol^-1value found

for the coupling of a 3-hexyne with the apical nitride ligand of a

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

In the 7m_S(0°), 1_||and 2_||correspond to the HOMO and LUMO

of the system with an energy gap of +0.40 eV, while in triplet

7m_T(0°), the same levels are the most probable SOMOs, which

are separated by only 0.23 eV. On the other hand, also the

slightly higher 1_level may compete for population. To analyze

the latter aspects, a more sophisticated treatment, such as one

of the various multireference methods or the broken symmetry

approach, would be more appropriate. However, due to the

complexity of the system, we limited the investigation to the

basic DFT method in association with perturbation theory

arguments,^[38]which were extracted from the simplified

interaction and Walsh diagrams of Scheme 6.^[39]Here, the left

side diagram for 7m_S(0°)highlights the three orbital interactions of

the alkyne _||and _||* combinations with either the _||n.b. and

_||a.b. level (shown in Scheme 4) giving rise to the 1_||and 2_||

MO, respectively.

Figure 8. Structural formula of 7m_S(0°)with separated charges at the N₁and C₂

atoms.

The two electrons, which previously allowed two consecutive in

plane Ru-N  half bonds, are now localized below the porphine

macrocycle ensuring the MeN₂=Ru  linkage (solid line).

Conversely, the _distribution is substantially unaffected with

respect to 5m_S. In this manner, 6 and 7 electrons are counted for

the N₁(amido) and N₂(imido) atoms, respectively, with the metal

remaining Ru(VI).

The eventual closeness of the 1_||, 2_||and 1_levels allows a

spontaneous formation of the triplet isomer 7m_T(0°), which

promptly interconverts into 7m_T(90°)(Scheme 5), where the 0.8

eV width of the three frontier MOs is maintained. The achieved

orthogonality of the imido and amido ligands is a cornerstone of

the process, as it will be clarified later. This implies a reshuffling

of the _||/_combinations, as highlighted from the comparison of

Figures 7 and 9. The latter shows that the C₂sp² hybrid is now

localized at the lowest of the three levels (1_||), while 1_lies

higher by +0.5 eV. The highest 2_||and the lower 1_MOs

feature MeN₂=Ru d_/p_antibonding character.

Scheme 6. a) General interaction diagram applicable to the singlet an triplet

models of type 7m. b) Evolution of the frontier levels in the 6m_S7m_S(0°)

transformation (total energy in red).

Essentially, the two populated _||and _||n.b. levels cause a

significant 4e^-repulsion, mitigated by the mixing of _||* into 1_||.

This cancels out the in-plane C₁p_orbital from 1_||but enhances

the C₂one (first drawing of Figure 7). A similar _||* mixing occurs

in the antibonding _||/_||a.b. combination (a 2c/2e^-interaction),

enhancing the weight of C₂p_in 2_||, also shown in Figure 7.

Conversely, 1_maintains the original _a.b. character of

Scheme 4 due to its symmetry properties.

The Walsh diagram, which monitors the frontier MOs upon the

N₁C₁coupling, shows how the 4e^-repulsion destabilizes 1_||up

to the region of the vacant levels 2_||and 1_, with a reduced

steepness of the curve after the TS. Then, the system stabilizes

(red curve) thanks to the N₁-C₁ bond formation. At the 7m_S(0°)

minimum, 1_||becomes a C₂lone pair of a vinylidene carbanion

(Figure 8), similarly to another case already investigated by

some of us.^[40]This corresponds to the transfer of the in plane

lone pair at the dianionic imido ligand of 5m_S(box of Scheme 4)

to the C₂atom, hence the separation of the negative charges in

7m_S(0°), as shown in Figure 8.

Figure 9. The three modified frontier MOs of 7m_T(90°)as determined by the

orthogonality of the amido and imido ligands.

In the staggered conformation of the trans-axial ligands of

7m_T(90°), the plane of the d_xzand d_yzorbitals also contains one of

the N_equat-Ru-N_equatvectors, thus suggesting an involvement of

the porphine’s  system in the overall  electron delocalization.

The result is confirmed by the computed spin densities of 0.28,

0.22 and 0.23 e²bohr^-3for the Ru, N₁and N₂atoms,

respectively, plus a residual 0.25 e²bohr^-3value for the non-

innocent porphine.

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

Structural and electronic aspects of the five-membered ring

closure. In all the species examined up to now (see Figure 6),

the C₂sp²conformation is unsuited for a direct C₂···C₄coupling,

hence the Ph alkyne’s substituent must rotate by ~180° about

the C₂=C₁double bond. However, all of our attempts of

validating either a singlet or a triplet intermediate of this type

failed, since the corresponding model immediately converged to

either the ring-closed isomer 8m_Sor 8m_T; the former is shown in

Figure 10. None of the latter species has ever been

experimentally detected, but the in silico result indicates how the

reaction proceeds through a complex of the closed bicyclic

molecule 3aa_iso, i.e., an isomer of the final indole product 3aa. At

variance with the latter, 3aa_isois not planar due to the C₄sp³

hybridization and the still present C₄-H₄bond. The lack of

experimental evidence for any species of type 8 suggests that

3aa_isopromptly separates from the metal and undergoes a

subsequent outer sphere isomerization with the H₄atom transfer

from C₄to N₁, as described in a following section.

Figure 11. Singlet/triplet relaxed energy scans for the C₂-C₄coupling, which

associates to the torsion about the C₁=C₂linkage.

To explain the orbital/electronic trends of Figure 11 it must be

recalled that the minimum 7m_T(90°)(Scheme 5) has the unpaired

electrons in the orthogonal MOs of Figure 9, namely 1_||and 1_.

The single population of the latter suggests an allylic-type radical

in the N₁=C₁=C₂region, that presumably favors the rotation of

the phenyl group and stereochemical inversion of the C₂sp²

center. Near the CP crossing point, the reorienting C₂ hybrid in

1_||starts “feeling” the C₄one, which derives from a  bonding

MO of the aryl ring. The level in question transforms into a C₄

sp³lone pair thanks to the pinning down of the C₄-H₄linkage, as

indicated by the structure in Figure 10. The resulting C₂-C₄*

combination (Figure 12) is swept away from the frontier region,

leaving only the two orthogonal and close MeN₂=Ru *

combinations of Figure 9. Hence, the two available electrons are

either paired or unpaired at the levels 1_and 2_||, even if the

singlet 8m_Sis somewhat favored after the CP. The feature

persists even after the separation of the [Ru](NMe) (Am_S)

fragment from the bicyclic molecule, already suggesting that the

spin multiplicity is intrinsic in the metal fragment (Am_Sis -5.0

kcal mol^-1more stable than Am_T).

Figure 10. The optimized spin isomer 8m_Sthat already features a 6+5 bicyclic

molecule.

The in silico detection of both 8m_Sand 8m_T(Figure S22) shows

a weaker Ru-N₁bonding in the former (2.27 vs. 2.20 Å) as well

as a pronounced tilt of the coordinated five membered ring with

respect to the octahedron’s main axis (~30° vs. 8°). Thus, the

loss of the bicyclic molecule more likely occurs at the singlet

complex, which is also energetically favored by -4.8 kcal mol^-1.

In this case, another intersystem crossing must occur after that

of the 7m_S(0°)7m_T(0°)transformation. This is confirmed by the

plot in Figure 11, which, by using a scan technique, monitors the

shortening of the C₂···C₄distance and indicates the Minimum

Energy Crossing Point (MECP).^[41]Both the species 7m_S(0°)and

7m_T(0°)were fixed as starting points of the singlet (black) and

triplet (red) curves, respectively. Perhaps, the constantly small

energy difference seems to confirm the competitiveness of the

two spin states in the process. The torsion about the C₁=C₂

linkage eventually reaches the values of 176° and 171° in 8m_S

and 8m_T, respectively. The crossing point (CP) occurs at the

C₂···C₄distance of ~2.37 Å with a torsion of ~145°. Then, the

singlet species immediately stabilizes, while the triplet inverts its

trend only at ~2.20 Å and never returns to be the lower isomer.

Based on these data, the MECP is estimated to be as low as

+3.4 kcal mol^-1.

Figure 12. The C₄-C₂* destabilizing level upon the ring closure in either 8m_S

or 8m_T.

The formation of the C₂-C₄ bond is consistent with a typical

chemical substitution at the aromatic ring of the original ArN

imido grouping. In principle, the process may have alternative

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

nucleophilic or electrophilic character, depending on the

inductive and/or  delocalization effects of the aryl substituents,

which determine a different regioselectivity.^[42]In the case of the

present aryl ring with EWG substituents at para and ortho

positions with respect to the C₄atom plus the ortho amido EDG

one, the evolution of the process is hardly predictable, even if

some MO consideration support the electrophilic substitution.

While the C₂vinylidenic anion proposed for the singlet 7m_S(0°)of

Figure 8 would suggest a nucleophilic attack, the picture is

significantly modified at the triplet 7m_T(90°). In fact, the half

vacated 1_||level already implies a partial electrophilicity of the

C₂atom that could be already attacked by the C₄sp³lone pair to

give the 3aa_isobicyclic molecule in 8m_T. The effect is definitely

magnified when 1_||is totally vacated in the singlet ground state.

The scission of 3aa_isofrom either 8m_Sor 8m_Tis similarly

endergonic (+10.1 and +12.3 kcal mol^-1, respectively), mainly

due to the active dispersion forces, in spite of the weak Ru-N₁

linkage of 2.27 in the singlet complex. It may be discharged

that 3aa_isomay separate as a triplet for its +28 kcal mol^-1higher

energy, that excludes the final indole formation to be radical. It

remains to establish whether the process of the 3aa_iso3aa

isomerization already starts when the bicycle is still anchored to

the metal or is an outer sphere process. The latter point will be

described in more details in a separate section.

redistribution implies an inner Ru(VI)Ru(IV) reduction with

respect to the starting bis-imido complex (5m_S). The mentioned

electron relationships are illustrated in Scheme 7b, that shows

how one imido dianion first transforms into an amido and then

into an uncharged imino ligand upon the C₂-C₄coupling.

Aspects of the isomerization of the 3aa_isobicyclic molecule

to the indole 3aa. The process most likely occurs on the singlet

PES involving the C₄C₃N₁stepwise transfer of the H₄atom

throughout the bicyclic molecule. It remains to establish at which

point the latter detaches from the metal. Recall that in the

starting 8m_Scomplex (Figure 10) the large Ru-N₁distance of

2.27 Å and the 29° bias of the five-membered ring already

suggested a poor coordination in spite of the +10.1 kcal mol^-1

cost for the scission, which is contrasted by the dispersion

forces. To frame better the picture, we first examined the outer

sphere isomerization, as shown in Scheme 8.

Electronic/orbital aspects of the [Ru](NMe) (Am) catalyst.

Upon the departure of the formed bicyclic molecule, the residual

five-coordinated metal fragment is either the singlet Am_Sor the

triplet Am_T. In any case, there remain six  electrons, which

differently affect the nature of the MeN₂=Ru multiple bonding. As

qualitatively shown in Scheme 7a, there are 2+2 orthogonal

bonding and antibonding levels with one electron pair being

available to the two * combinations. Depending on the _||^*-_*

gap, the two electrons may be either coupled or split,

determining the Am_Sor the Am_Tspin isomers, respectively.

Scheme 8. Free energy profile for the outer sphere 3aa_iso3aa isomerization.

The first intermediate 3aa_iso’ is attained through the TS₁barrier

of +19.0 kcal mol^-1due to disfavoring H₄bridging position in

between the C₄and C₃atoms.^[43]Not all of the lost energy is

recovered at 3aa_iso’, which lies +2.4 kcal mol^-1above 3aa_iso

,

while the major energy gain occurs at the subsequent step.

Here, the N₁imino lone pair, which was originally engaged in

metal coordination, is used for the final N₁-H₄linkage. The

encountered TS₂barrier of +13.6 kcal mol^-1is again attributable

to the H₄bridging position in between the C₃and N₁atoms. After

TS₂, the system gains as much as -54.0 kcal mol^-1thanks to the

achieved planarity and aromaticity of the indole. Overall, the

outer sphere process of Scheme 8 is as exergonic as -37.9 kcal

mol^-1.

Scheme 7. a)  electron distribution in the Am–type compounds. b) Evolution

of the highest ^*electron pair for Am along the pathway leading first to 8m_Sor

8m_T.

Alternatively, we also considered that the isomerization of 3aa_iso

starts while the bicyclic molecule still anchored to the metal. In

fact, by excluding the dissociation of 8m_S, the intermediate

complex 9m could be fully optimized as shown in Figure S23. In

the latter, the 3aa_iso’ isomer acts as a donor to the metal by still

using the N₁lone pair without any major penalization of the

Ru-N₁strength nor the stereochemistry of the bicyclic molecule.

This is corroborated by the very similar binding energies of 8m_S

and 9m, with the latter being disfavored by only +1.3 kcal mol^-1.

A consistent picture also applies to the immediate precursors

8m_Sor 8m_T. In any case, an overall MeN=Ru double bond

exists, since the two highest electrons either allow a full double

bond (vacant _*) or two half  bonds (two SOMOs). The two

available ^electrons are those of the filled C₂ hybrid at the

intermediate 7m_S(0°), which were in turn derived from one pristine

alkyne CC  linkage. Remarkably, the described electron

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

Implicitly, 9m must dissociate to free the N₁lone pair for allowing

the 3aa formation. In this respect, it must be mentioned that the

scission of 9m is more expensive than that of 8m_S(+13 vs.

+10.1 kcal mol^-1). From this data, it is hard to predict which of the

two alternative pathways of the isomerization is more likely.

favoured starting point B but also features the highest TS barrier

of +20.0 kcal mol^-1possibly in line with its chemical inertness.

On the other hand, the profiles of the Cl and CF₃substituted

compounds may be consistent with their reactivity in producing

corresponding indoles with different yields (Table 1). The adduct

B has the largest stability in the CF₃case, implying a sufficient

lifetime to continue the reaction. From this viewpoint, the +7.0

kcal mol^-1higher energy of the B chlorine analogue seems to

disfavour the process, which is however not inhibited because of

the rather low TS barrier to be passed.

The exergonicity of all the reactions in Scheme 9 may seem

inconsistent with the lack of the reactivity for the X = Me. Some

weak points in the profile of the latter species have already been

illustrated and it cannot be excluded that the initially disfavouring

thermal effects may influence the kinetic behaviours.

Influence of the electronic characteristics of aryl azide on

the catalytic mechanism. Amongst our experiments, the

formation of indole was investigated for the reaction between

PhCCH and selected azides of formula 3,5-X₂C₆H₃N₃(X = CF₃,

Cl and Me). Even if the most efficient process occurs when

X = CF₃, the catalysis also proceeds if X = Cl with the formation

of the corresponding indole with 60% yield (vs. 86% yield

observed when X = CF₃) (see Table 1). Conversely, for X = Me

the reaction did not occur at all. To shed some light on the

different behaviors, the already illustrated reactivity pathways

were computationally tested also for azides bearing the Cl and

Me EDG substituents besides the CF₃EWG ones. In any case,

the calculations indicated quite similar trends and no evidence

emerged that the EDG groups significantly alter the features of

the five-membered ring closure. In view of the similar computed

mechanistic behaviors for differently substituted azides, we

looked for some alternative explanation of the experimental

reactivity. With reference to the path a of Scheme 3, we focused

on the formation of the bis-imido species C from the five-

coordinated fragment Am and variously substituted azides. As

previously shown,^[17]the azide must first anchor to the metal to

be activated over the singlet PES via the N₂loss. In Scheme 9,

the free energy profiles of the species with Me, Cl and CF₃

substituents outline a potential exergonic access to the bis-imido

complex C (or 5m_Sin the case of X = CF₃). To shed some light

on the missed C formation in the case of X = Me, one must

focus on the stability of the initial adduct B and the relative

difficulty in attaining the subsequent TS.

Conclusions

The Ru(TPP)(NAr)₂(5) complex showed a very good catalytic

activity in promoting the synthesis of C₃-substituted indoles (up

to 30 examples) by the one-pot reaction of aryl azides with

alkynes. The study of the reaction scope revealed that the

reaction is very efficient and completely regioselective for

terminal alkynes and aryl azides bearing on the 3 and 5 aromatic

positions the same electron withdrawing group. It should be

noted that when the 3,5-(NO₂)₂C₆H₃N₃azide was reacted with

different alkynes, the corresponding indoles precipitated from

the reaction mixture in a pure form. Thus, a convenient filtration,

instead of less sustainable and more expensive purification

procedures (e.g. chromatographic purification), was used to

isolate the desired compounds.

The joint experimental/theoretical study rationalized the catalytic

indole formation with the key mechanistic aspects being

supported by orbital and electronic arguments. In particular, it

emerged how the stereochemistry of the system is dynamically

affected by an interplay of singlet and triplet ground states that

eventually allows the conformational inversion of the sp²C (Ph)

atom. Also, thanks to the vacancy of the latter an electrophilic

attack to one aryl’s position allows the C-C coupling. Then, the

formed bicyclic molecular skeleton detaches from the metal with

a subsequent outer sphere isomerization affording the desired

indole. At this point, the separated five-coordinated mono-imido

species A (Scheme 3) represents the actual reaction catalyst

that repetitively activates a new azide molecule.^[23c]The role of

the EWG vs. the EDG substituents at the aryl ring of the azide

has been compared to explain the correlation between the

catalytic performance and the stability and lifetime of the initial

ruthenium/azide adduct complex (B type). For convenience,

Scheme 10 summarizes the overall mechanism with the

stepwise energies. A more standard reaction profile is presented

in Scheme S1 (SI).

Scheme 9. Comparative energy underpinnings of the bis-imido complex C

formation, which is inhibited in presence of methyl EDG aromatic substituents.

By assuming constant dispersion and entropy terms, the

estimated energy order of the B compounds (namely -10.3, -3.3

and -1.1 kcal mol^-1for X = CF₃, Cl and Me, respectively) is

mainly due to the enthalpy effects introduced by the different

substituents. The Me-containing species has not only the least

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

optimized species, includes pairwise sets of coordinates, when available.

Also, the coordinates of some key models such as 5m_S’, 5m_T’, 7m_S’,

7m_T’, 8m_S’ and 8m_T’, with 3,5-(CF₃)₂C₆H₃substituents at both the

N-imido groups, are given in the SI to support the consistency with the

used simpler models. Qualitative MO arguments were developed with the

CACAO package^[39]after verifying the gross consistency between EHMO

and DFT results.

General. All reactions were carried out under nitrogen atmosphere

employing standard Schlenk techniques and vacuum-line manipulations.

Anhydrous benzene was purified by using standard procedures and

stored under inert atmosphere. Organic azides^[50]and Ru(TPP)(NAr)₂

(5)^[19b]were synthesized by methods reported in literature.

Phenylacetylene (2a) was filtered over activated alumina, distilled and

stored under nitrogen. All the other starting materials were commercial

products and used as received. NMR spectra were recorded at room

temperature either on Bruker Avance 300-DRX, operating at 300 MHz for

¹H, at 75 MHz for ¹³C and at 282 MHz for ¹⁹F, or on a Bruker Avance

400-DRX spectrometers, operating at 400 MHz for ¹H NMR, 100 MHz for

¹³C and 376 MHz for ¹⁹F. Chemical shifts (ppm) are reported relative to

TMS. The ¹H NMR signals of the compounds described in the following

have been attributed by COSY and NOESY techniques. Assignments of

the resonance in ¹³C NMR were made by using the APT pulse sequence

and HSQC and HMBC techniques. GC-MS analyses were performed on

a Shimadzu QP5050A equipped with Supelco SLB -5 ms capillary

column (L 30m × I.D. 0.25 µm film thickness). GC analyses were

performed on a Shimadzu GC-2010 equipped with a Supelco SLB -5 ms

capillary column (L 10m × I.D. 0.1 µm film thickness). Infrared spectra

were recorded on a Varian Scimitar FTS 1000 spectrophotometer. UV-

Vis spectra were recorded on an Agilent 8453 instrument. Mass spectra

were recorded in the analytical laboratories of Milan University.

Microanalyses were performed on a Perkin Elmer 2400 CHN Elemental

Analyzer instrument.

Scheme 10. Suggested catalytic cycle of the azide/alkyne coupling leading to

indole. G values (kcal mol^-1) are reported for each step together with the

associated barrier (in brackets).

The process starts with the reaction of bis-imido complex 5m_S

with an alkyne to give the singlet intermediate 7m_Sfeaturing a

dangling chain. The terminal C(Ph) atom of 7m_Shas an unsuited

sp²conformation for the coupling with the aryl ring, hence the

system undergoes a stereochemical rearrangement. This is

allowed

by

a

double

intersystem

crossing

(singlettripletsinglet), whose implications for the mechanism

were underlined. In this manner, the bicyclic skeleton 3aa_isois

formed as a ligand in 8m_Sthrough an aromatic-like electrophilic

substitution. Notice that such a step implies an inner redox

process, namely Ru(VI)Ru(IV) reduction and amidoimino

oxidation (see Scheme 7b). Finally, the 3aa_isobicyclic molecule

dissociates from the metal and undergoes two subsequent H

migration steps to give the final indole 3aa. Importantly, the

separation of the bicyclic molecule affords the five-coordinated

metal fragment [Ru](NMe) (Am) as the active catalyst, which

activates a new azide molecule and restores the bis-imido

precursor 5m. Overall, the indole formation is energetically

highly favored (i.e., -99.9 kcal mol^-1) with the largest barrier of

+19.0 kcal mol^-1. Notice that the latter is found for the indole

outer sphere isomerization, while in the metal-catalyzed cycle, a

very similar barrier of +18.3 kcal mol^-1was estimated to reach

7m_S-TS(Scheme 5). Based on the energy closeness, no precise

information arises about the actual rate determining step, for

which not even the kinetics offers useful indications.

Procedure for catalytic reactions. Method A: Alkyne (2.5 mmol) and

aryl azide (0.50 mmol) were added to a benzene (10 mL) solution of

complex 5 (11.7 mg, 0.010 mmol) and the resulting mixture was refluxed

until the complete conversion of aryl azide, which was monitored by IR

spectroscopy (ν_N=N= 2200-2000 cm^-1). The solvent was evaporated to

dryness and the crude purified by flash chromatography. Method B: the

procedure was identical to that described in method A except for the

purification step. After the complete conversion of aryl azide, desired

products were collected in a filter in a pure form. Method C: the

procedure was identical to that described in method A, except for the

amount (1.1 mL, 10.0 mmol) of employed phenylacetylene (2a). At the

end of the reaction, the 2a excess was recovered in 90-95% yield (GC

analysis, biphenyl as the internal standard) by vacuum distillation of the

obtained mixture.

Synthesis of indole 3aa from Ru(TPP)(NAr)₂(5) and phenylacetylene

(2a). Phenylacetylene (2a) (27 L, 0.24 mmol) was added to a benzene

(30.0 mL) solution of complex 5 (57.4 mg, 0.049 mmol) and the resulting

solution was refluxed until the complete disappearance of 5 (TLC

monitoring, Al₂O₃, n-hexane/CH₂Cl₂= 9:1). GC-MS analysis revealed the

formation of compound 3aa that was isolated by flash chromatography

(SiO₂, n-hexane/AcOEt = 9:1) (50% yield, considering the transfer of only

one nitrene functionality from 5 to 2a).

Experimental Section

Computational Details. . All of the models were optimized at the B97D-

DFT level of theory^[44]by using the Gaussian 09 package.^[45]Minima

and/or transition states were validated by vibrational frequencies. The

CPCM model^[46]was used to mimic the experimentally used benzene

solvent. The effective Stuttgart/Dresden core potential (SDD)^[47]was

adopted for ruthenium with the 6-31G+(d, p) basis set for all the other

atoms.^[48]To confirm the acceptability of the latter original approach,

several optimizations were also carried out with the triple zeta basis set

def2-TZVP.^[49]The structural and energy divergences appear minor as it

can be verified in the Supporting Information, which besides all the

Isotope tracing experiment using phenylacetylene-d₁(2a-d₁).

Compound 1a (86 L, 0.50 mmol) and 2a-d₁(99 atom % D) (275 L, 2.5

mmol) were added to benzene (10.0 mL) solution of complex 5 (11.7 mg,

0.010 mmol). The resulting mixture was refluxed for 1.0 h until the

complete aryl azide conversion. ¹H NMR (400 MHz, CDCl₃) spectrum of

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

the purified indole showed the absence of the signal attributed to C₂-H

atom of the indole ring (Figure S2).

S. Cacchi, G. Fabrizi, Chem. Rev. 2011, 111, PR215-PR283; h) M. G.

Ciulla, S. Zimmermann, K. Kumar, Org. Biomol. Chem. 2019, 17, 413-

431.

[5]

a) J.-B. Chen, Y.-X. Jia, Org. Biomol. Chem. 2017, 15, 3550-3567; b) S.

Cacchi, G. Fabrizi, Chem. Rev. 2005, 105, 2873-2920; c) Y. Yang, Z.

Shi, Chem. Commun. 2018, 54, 1676-1685; d) Q. Q. Cheng, Y. Yu, J.

Yedoyan, M. P. Doyle, ChemCatChem 2018, 10, 488-496; e) J. A.

Leitch, Y. Bhonoah, C. G. Frost, ACS Catal. 2017, 7, 5618-5627; f) T.

A. Shah, P. B. De, S. Pradhan, T. Punniyamurthy, Chem. Commun.

2019, 55, 572-587; g) J. Kalepu, P. Gandeepan, L. Ackermann, L. T.

Pilarski, Chem. Sci. 2018, 9, 4203-4216.

Kinetic Isotope Effect (KIE) experiment. Compound 2a (275 L, 2.5

mmol) and an equimolar mixture of 1a (43 L, 0.25 mmol) and 1a-d₃(94

atom % D) (for details see supporting information) (44 L, 0.25 mmol)

were added to a benzene (10.0 mL) solution of complex 5 (12.0 mg,

0.010 mmol). The solution was refluxed for 15 minutes until the azide

conversion of 35% was reached. The solvent was evaporated to dryness

and the crude was purified by flash chromatography (SiO₂, n-

hexane/AcOEt = 9:1) to give a mixture of 3aa and 3aa-d₂(the labile D of

the N_(indole)-D bond was replaced by H during the purification). The k_H/k_D

ratio of 1.1 was determined by ¹H NMR spectroscopy (300 MHz, CDCl₃)

(Figure S3).

[6]

[7]

[8]

B. C. G. Soderberg, Curr. Org. Chem. 2000, 4, 727-764.

S. W. Youn, T. Y. Ko, Asian J. Org. Chem. 2018, 7, 1467-1487.

a) R. C. Larock, E. K. Yum, J. Am. Chem. Soc. 1991, 113, 6689-6690;

b) D. F. Taber, P. K. Tirunahari, Tetrahedron 2011, 67, 7195-7210; c) L.

Ackermann, L. T. Kaspar, C. J. Gschrei, Chem. Commun. 2004, 2824-

2825; d) J. Barluenga, M. A. Fernández, F. Aznar, C. Valdés, Chem.

Eur. J. 2005, 11, 2276-2283; e) J. Barluenga, F. J. Fañanás, R. Sanz,

Y. Fernández, Chem. Eur. J. 2002, 8, 2034-2046; f) X.-S. Ning, X.

Liang, K.-F. Hu, C.-Z. Yao, J.-P. Qu, Y.-B. Kang, Adv. Synth. Catal.

2018, 360, 1590-1594.

Synthesis of 3aa in the presence of TEMPO. Compounds 1a (86 L,

0.50 mmol), 2a (275 L, 2.5 mmol) and TEMPO (21.0 mg, 0.13 mmol)

were added to a benzene (10.0 mL) solution of complex 5 (11.7 mg,

0.010 mmol) and the resulting solution was refluxed for 1.3 h until the

complete consumption of aryl azide 1a. The solvent was evaporated to

dryness and the crude purified by flash chromatography (SiO₂, n-

hexane/AcOEt = 9:1) (125.0 mg of 3aa, 76%).

[9]

a) N. Yoshikai, Y. Wei, Asian J. Org. Chem. 2013, 2, 466-478; b) B. J.

Stokes, T. G. Driver, Eur. J. Org. Chem. 2011, 4071-4088; c) F. Collet,

R. H. Dodd, P. Dauban, Chem. Commun. 2009, 5061-5074.

General kinetic procedure. Compounds 2a and 1a were added to a

benzene (5.0 mL) solution of complex 5 under nitrogen. The flask was

capped with a rubber septum and immediately placed in an oil bath

preheated to 75°C. The aryl azide consumption was followed by IR

spectroscopy (≈2114 cm^-1) by withdrawing samples of the solution (0.1

mL) at regular time intervals. The apparent first order rate constants were

[10] a) S. Würtz, S. Rakshit, J. J. Neumann, T. Dröge, F. Glorius, Angew.

Chem. Int. Ed. 2008, 47, 7230-7233; b) X.-L. Lian, Z.-H. Ren, Y.-Y.

Wang, Z.-H. Guan, Org. Lett. 2014, 16, 3360-3363; c) J. J. Neumann,

S. Rakshit, T. Dröge, S. Würtz, F. Glorius, Chem. Eur. J. 2011, 17,

7298-7303; d) Z. He, W. Liu, Z. Li, Chem. Asian J. 2011, 6, 1340-1343;

e) R. Bernini, G. Fabrizi, A. Sferrazza, S. Cacchi, Angew. Chem. Int.

Ed. 2009, 48, 8078-8081; f) W.-Q. Liu, T. Lei, Z.-Q. Song, X.-L. Yang,

C.-J. Wu, X. Jiang, B. Chen, C.-H. Tung, L.-Z. Wu, Org. Lett. 2017, 19,

3251-3254; g) X. Mao, T. Tong, S. Fan, L. Fang, J. Wu, X. Wang, H.

Kang, X. Lv, Chem. Commun. 2017, 53, 4718-4721; h) S. Tang, X.

Gao, A. Lei, Chem. Commun. 2017, 53, 3354-3356.

fitted to the equation –d[ArN₃]/dt

= k_app[5][1a][2a]. The catalyst

concentration was calculated by the exact amount of catalyst weighed in

each run and was considered to remain constant during the reaction. See

SI for other experimental details.

[11] a) Y. Wei, I. Deb, N. Yoshikai, J. Am. Chem. Soc. 2012, 134, 9098-

9101; b) Z. Shi, F. Glorius, Angew. Chem. Int. Ed. 2012, 51, 9220-

9222; c) P. S. Naidu, S. Kolita, M. Sharma, P. J. Bhuyan, J. Org. Chem.

2015, 80, 6381-6390.

Acknowledgements

We are grateful to prof. Fabio Ragaini of University of Milan for

his valuable suggestions regarding the kinetic study.

G.M. and C.M. acknowledge the ISCRA-CINECA (HP grant

HP10C2Q178 and HP10CFMSCC) for the computational

resources.

[12] a) F. Ragaini, A. Rapetti, E. Visentin, M. Monzani, A. Caselli, S. Cenini,

J. Org. Chem. 2006, 71, 3748-3753; b) F. Ragaini, F. Ventriglia, M.

Hagar, S. Fantauzzi, S. Cenini, Eur. J. Org. Chem. 2009, 2009, 2185-

2189; c) A. Penoni, K. M. Nicholas, Chem. Commun. 2002, 484-485.

[13] a) B. Liu, C. Song, C. Sun, S. Zhou, J. Zhu, J. Am. Chem. Soc. 2013,

135, 16625-16631; b) C. Wang, Y. Huang, Org. Lett. 2013, 15, 5294-

5297; c) A. Penoni, G. Palmisano, Y.-L. Zhao, K. N. Houk, J. Volkman,

K. M. Nicholas, J. Am. Chem. Soc. 2008, 131, 653-661; d) F. Tibiletti,

M. Simonetti, K. M. Nicholas, G. Palmisano, M. Parravicini, F. Imbesi,

S. Tollari, A. Penoni, Tetrahedron 2010, 66, 1280-1288; e) S.

Muthusamy, A. Balasubramani, E. Suresh, Org. Biomol. Chem. 2018,

16, 756-764; f) S. Muthusamy, A. Balasubramani, E. Suresh, Adv.

Synth. Catal. 2017, 359, 786-790; g) G. Ieronimo, G. Palmisano, A.

Maspero, A. Marzorati, L. Scapinello, N. Masciocchi, G. Cravotto, A.

Barge, M. Simonetti, K. L. Ameta, K. M. Nicholas, A. Penoni, Org.

Biomol. Chem. 2018, 16, 6853-6859.

Keywords: indole • azide • ruthenium • porphyrin • DFT • energy

profiles

[1]

[2]

a) S. Lancianesi, A. Palmieri, M. Petrini, Chem. Rev. 2014, 114, 7108-

7149; b) L. C. Klein-Junior, C. d. Santos Passos, A. P. Moraes, V. G.

Wakui, E. L. Konrath, A. Nurisso, P.-A. Carrupt, C. M. Alves de Oliveira,

L. Kato, A. T. Henriques, Curr. Top. Med. Chem. 2014, 14, 1056-1075.

a) N. Chadha, O. Silakari, Eur. J. Med. Chem. 2017, 134, 159-184; b)

S. Thokchom Prasanta, S. Okram Mukherjee, Mini-Rev. Med. Chem.

2018, 18, 9-25; c) S. Suzen, Curr. Org. Chem. 2017, 21, 2068-2076; d)

D. Goyal, A. Kaur, B. Goyal, ChemMedChem 2018, 13, 1275-1299.

E. Fischer, H. Hutz, Ber. 1895, 28, 585-587.

[14] a) M. Peña-López, H. Neumann, M. Beller, Chem. Eur. J. 2014, 20,

1818-1824; b) Z. Shi, C. Zhang, S. Li, D. Pan, S. Ding, Y. Cui, N. Jiao,

Angew. Chem. Int. Ed. 2009, 48, 4572-4576; c) L. Ackermann, A. V.

Lygin, Org. Lett. 2012, 14, 764-767; d) Y. Liang, N. Jiao, Angew. Chem.

Int. Ed. 2016, 55, 4035-4039; e) J. Wang, M. Wang, K. Chen, S. Zha,

C. Song, J. Zhu, Org. Lett. 2016, 18, 1178-1181; f) J. Su, K. Zhang, M.

Zhuang, F. Ma, W.-Q. Zhang, H. Sun, G. Zhang, Y. Jian, Z. Gao, Asian

J. Org. Chem. 2019, 8, 482-486.

[3]

[4]

a) J. Yamaguchi, A. D. Yamaguchi, K. Itami, Angew. Chem. Int. Ed.

2012, 51, 8960-9009; b) G. R. Humphrey, J. T. Kuethe, Chem. Rev.

2006, 106, 2875-2911; c) K. Krüger, A. Tillack, M. Beller, Adv. Synth.

Catal. 2008, 350, 2153-2167; d) M. Inman, C. J. Moody, Chem. Sci.

2013, 4, 29-41; e) G. W. Gribble, J. Chem. Soc., Perkin Trans. 1 2000,

1045-1075; f) M. Bandini, Org. Biomol. Chem. 2013, 11, 5206-5212; g)

[15] a) M. P. Huestis, L. Chan, D. R. Stuart, K. Fagnou, Angew. Chem. Int.

Ed. 2011, 50, 1338-1341; b) D. R. Stuart, P. Alsabeh, M. Kuhn, K.

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

Fagnou, J. Am. Chem. Soc. 2010, 132, 18326-18339; c) P. Patel, G.

Borah, Eur. J. Org. Chem. 2017, 2017, 2272-2279.

Kawakita, E. P. Beaumier, Y. Kakiuchi, H. Tsurugi, I. A. Tonks, K.

Mashima, J. Am. Chem. Soc. 2019, 141, 4194-4198.

[16] a) S. Cenini, F. Ragaini, E. Gallo, A. Caselli, Curr. Org. Chem. 2011,

15, 1578-1592; b) T. G. Driver, Org. Biomol. Chem. 2010, 8, 3831-

3846; c) J. G. Harrison, O. Gutierrez, N. Jana, T. G. Driver, D. J.

Tantillo, J. Am. Chem. Soc. 2016, 138, 487-490; d) N. Jana, T. G.

Driver, Org. Biomol. Chem. 2015, 13, 9720-9741; e) C. Kong, N. Jana,

C. Jones, T. G. Driver, J. Am. Chem. Soc. 2016, 138, 13271-13280; f)

S. Parisien-Collette, C. Cruche, X. Abel-Snape, S. K. Collins, Green

Chem. 2017, 19, 4798-4803; g) B. Lu, Y. Luo, L. Liu, L. Ye, Y. Wang, L.

Zhang, Angew. Chem. Int. Ed. 2011, 50, 8358-8362; h) X.-R. Song, R.

Li, T. Yang, J. Bai, R. Yang, X. Chen, H. Ding, Q. Xiao, Y.-M. Liang,

Tetrahedron Lett. 2018, 59, 3763-3766.

[32] The components are structurally unaffected by the interaction, but their

almost null free energy balance of +0.5 kcal mol^-1implies an attraction

due to the dispersion forces, which is counterbalanced by the ~+10 kcal

mol^-1entropy term. For these reasons, no Transition State could be

detected for 6m_S.

[33] The -383.5 cm^-1value of the unique imaginary frequency corroborated

N1-C1 as the proper reaction coordinate, given its opposite oscillations

from the equilibrium.

[34] W.-L. Man, J. Xie, P.-K. Lo, W. W. Y. Lam, S.-M. Yiu, K.-C. Lau, T.-C.

Lau, Angew. Chem. Int. Ed. 2014, 126, 8603-8606.

[35] A reduced bond order with respect to the original triple bond is

corroborated by the computed IR 391 cm^-1red shift

[17] P. Zardi, A. Savoldelli, D. M. Carminati, A. Caselli, F. Ragaini, E. Gallo,

ACS Catal. 2014, 4, 3820-3823.

[36] M. Harris, S. A. Brusey, A. Moore, Y. Ortin, H. Müller-Bunz, G. Manca,

C. Mealli, M. J. McGlinchey, ChemPlusChem 2016, 81, 292-306.

[37] a) A. R. Thornton, S. B. Blakey, J. Am. Chem. Soc. 2008, 130, 5020-

5021; b) A. R. Thornton, V. I. Martin, S. B. Blakey, J. Am. Chem. Soc.

2009, 131, 2434-2435.

[18] Z.-Y. Gu, J.-H. Li, S.-Y. Wang, S.-J. Ji, Chem. Commun. 2017, 53,

11173-11176.

[19] a) D. Intrieri, A. Caselli, F. Ragaini, P. Macchi, N. Casati, E. Gallo, Eur.

J. Inorg. Chem. 2012, 569-580; b) S. Fantauzzi, E. Gallo, A. Caselli, F.

Ragaini, N. Casati, P. Macchi, S. Cenini, Chem. Commun. 2009, 3952-

3954.

[38] R. Hoffmann, Angew. Chem. Int. Ed. 1982, 21, 711-724.

[39] C. Mealli, D. M. Proserpio, J. Chem. Educ. 1990, 67, 399.

[40] E. Peréz-Carreño, P. Paoli, A. Ienco, C. Mealli, Eur. J. Inorg. Chem.

1999, 1999, 1315-1324.

[20] J. Wei, W. Xiao, C.-Y. Zhou, C.-M. Che, Chem. Commun. 2014, 50,

3373-3376.

[21] a) K. Banert, S. Bochmann, M. Hagedorn, F. Richter, Tetrahedron Lett.

2013, 54, 6185-6188; b) D. J. Anderson, T. L. Gilchrist, C. W. Rees,

Chem. Commun. 1969, 147; c) G. Dequirez, V. Pons, P. Dauban,

Angew. Chem. Int. Ed. 2012, 51, 7384-7395.

[41] a) J. N. Harvey, R. Poli, K. M. Smith, Coord. Chem. Rev. 2003, 238-

239, 347-361; b) J. N. Harvey, M. Aschi, Faraday Discuss. 2003, 124,

129-143.

[42] M. B. Smith, J. March in Advanced Organic Chemistry: Reactions,

Mechanisms, and Structure (6th ed.), New York: Wiley-Interscience,

2007, ISBN 978-0-471-72091

[43] The total barrier starting from 8m_Sis +29.1 kcal mol^-1, given the initial

cost of the 3aa_iso’ scission from the metal.

[22] a) C. Wu, J. Li, B. Yan, Dalton Trans. 2014, 43, 5364-5374; b) D. F.

Taber, W. Tian, J. Am. Chem. Soc. 2006, 128, 1058-1059; c) S. Jana,

M. D. Clements, B. K. Sharp, N. Zheng, Org. Lett. 2010, 12, 3736-3739.

[23] a) P. Zardi, A. Pozzoli, F. Ferretti, G. Manca, C. Mealli, E. Gallo, Dalton

Trans. 2015, 44, 10479-10489; b) G. Manca, C. Mealli, D. M. Carminati,

D. Intrieri, E. Gallo, Eur. J. Inorg. Chem. 2015, 2015, 4885-4893; c) G.

Manca, E. Gallo, D. Intrieri, C. Mealli, ACS Catal. 2014, 823-832.

[24] T. A. Albright, J. K. Burdett and M.-H. Whangbo in Orbital Interactions

in Chemistry, Second Edition, 2013, 834 pp, John Wiley & Sons.

[25] At variance with the metal complex, the central carbon atom is involved

in the  delocalization by using two p_rather than d_orthogonal orbitals,

but in any case the four axial  linkages are supported by one bonding

electron each (four half axial linkages in total).

[44] a) G. Stefan, J. Chem. Phys 2006, 124, 034108; b) A. D. Becke, J.

Chem. Phys 1993, 98, 5648-5652.

[45] 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. 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. Jr. Montgomery, J. E. Peralta, F.

Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N.

Staroverov, 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. 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,

ꢁ. Farkas, J. B. Foresman, J. V.Ortiz, J.Cioslowski, D. J. Fox Gaussian

0.9, revision B.01; Gaussian, Inc.: Wallingford, CT, 2010

[26] a) T.-G. Ong, G. P. A. Yap, D. S. Richeson, J. Am. Chem. Soc. 2003,

125, 8100-8101; b) F. Montilla, A. Pastor, A. n. Galindo, J. Organom.

Chem. 2004, 689, 993-996.

[27] a) T. Peddarao, A. Baishya, M. K. Barman, A. Kumar, S. Nembenna,

New J. Chem. 2016, 40, 7627-7636; b) M. Findlater, N. J. Hill, A. H.

Cowley, Dalton Trans. 2008, 4419-4423.

[28] a) P. F. Kuijpers, J. I. van der Vlugt, S. Schneider, B. de Bruin, Chem.

Eur. J. 2017, 23, 13819-13829; b) M. Goswami, V. Lyaskovskyy, S. r.

R. Domingos, W. J. Buma, S. Woutersen, O. Troeppner, I. Ivanoviꢀ-

Burmazoviꢀ, H. Lu, X. Cui, X. P. Zhang, E. J. Reijerse, S. DeBeer, M.

M. van Schooneveld, F. F. Pfaff, K. Ray, B. de Bruin, J. Am. Chem.

Soc. 2015, 137, 5468-5479; c) A. I. O. Suarez, H. Jiang, X. P. Zhang,

B. de Bruin, Dalton Trans. 2011, 40, 5697-5705.

[46] 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.

[47] M. Dolg, H. Stoll, H. Preuss, R. M. Pitzer, J. Phys. Chem. 1993, 97,

5852-5859.

[48] G. A. Petersson, A. Bennett, T. G. Tensfeldt, M. A. Al-Laham, W. A.

Shirley, J. Mantzaris, J. Chem. Phys 1988, 89, 2193-2218.

[49] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys 1994, 100, 5829-

5835.

[29] a) J. T. Groves, J. Chem. Educ. 1985, 62, 928; b) D. Balcells, C.

Raynaud, R. H. Crabtree, O. Eisenstein, Chem. Commun. 2008, 744-

746.

[30] J. P. Brand, J. Waser, Chem. Soc. Rev. 2012, 41, 4165-4179.

[31] a) N. Vujkovic, B. D. Ward, A. Maisse-François, H. Wadepohl, P.

Mountford, L. H. Gade, Organometallics 2007, 26, 5522-5534; b) K.

[50] M. Tanno, S. Sueyoshi, S. Kamiya, Chem. Pharm. Bull. 1982, 30, 3125-

3132.

10.1002/chem.201904224

Chemistry - A European Journal

FULL PAPER

Entry for the Table of Contents

FULL PAPER

The Ru(TPP)(NAr)₂–catalyzed one-

pot formation of indoles from alkynes

and aryl azides was analyzed by a

synergic experimental/computational

study. The reaction was effective for

the synthesis of differently C₃-

Daniela Intrieri, Daniela Maria Carminati,

Paolo Zardi, Caterina Damiano, Gabriele

Manca,* Emma Gallo,* Carlo Mealli

Page No. – Page No.

Indoles from Alkynes and Aryl

Azides. Scope and Theoretical

Assessment of Ruthenium Porphyrin-

Catalyzed Reactions

substituted indoles and the DFT

investigation elucidates the energy

costs as well as associated orbital and

electronic evolutions of the reaction.

Article Doi

Indoles from Alkynes and Aryl Azides: Scope and Theoretical Assessment of Ruthenium Porphyrin-Catalyzed Reactions

DOI: 10.1002/chem.201904224

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201904224

Full text of DOI:10.1002/chem.201904224

Products guided by the article

R&D Labs maybe for 1131611-99-8

Relevant to this article

Hot Product