Experimental and computational studies on the C-H amination mechanism of tetrahydrocarbazoles via hydroperoxides

Article abstract of DOI:10.1002/chem.201405376

The acid-catalyzed reactions of photochemically generated tetrahydrocarbazole peroxides with anilines have been studied experimentally and computationally to identify the underlying reaction mechanism. The kinetic data indicate a reaction order of one in

Full text of DOI:10.1002/chem.201405376

DOI: 10.1002/chem.201405376
Full Paper
&
Reaction Mechanisms
Experimental and Computational Studies on the CꢀH Amination
Mechanism of Tetrahydrocarbazoles via Hydroperoxides
Naeem Gulzar,[a] Kevin Mark Jones,[a] Hannelore Konnerth,[b] Martin Breugst,*[b] and
Martin Klussmann*[a]
Abstract: The acid-catalyzed reactions of photochemically
generated tetrahydrocarbazole peroxides with anilines have
been studied experimentally and computationally to identify
the underlying reaction mechanism. The kinetic data indi-
cate a reaction order of one in the hydroperoxide and zero
in the aniline. Computational investigations using density
functional theory support the experimental findings and pre-
dict an initial tautomerization between an imine and enam-
ine substructure of the primarily generated tetrahydrocarba-
zole peroxide to be the rate controlling step. The enamine
tautomer then loses hydrogen peroxide upon protonation,
generating a stabilized allylic carbocation that is reversibly
trapped by solvent or aniline to form the isolated products.
Introduction
Oxidative coupling reactions for the functionalization of CꢀH
bonds have emerged as an important synthetic tool in recent
years.[1] They help to save time, steps, and materials as two
substrates can be directly coupled without the need for pre-
functionalization. However, some of these reactions require
harsh conditions and their overall sustainability is further di-
minished if stoichiometric amounts of synthetic oxidants are
required. This problem can be overcome by using catalytic
conditions and environmentally benign molecular oxygen as
the terminal oxidant.[1a,2]
We recently reported a Brønsted-acid-catalyzed aerobic ami-
nation of tetrahydrocarbazole and indole derivatives by using
the strategy of CꢀH functionalization via Intermediate Perox-
ideS (CHIPS, Scheme 1a).[3] By treating the starting materials,
for example, tetrahydrocarbazoles 1, with oxygen in the pres-
ence of visible light and a photosensitizer, hydroperoxides 2
are generated. These can be subjected to acid catalysis in the
presence of anilines and some other N-nucleophiles, giving the
substitution products 3. The hydroperoxides can be conven-
iently isolated prior to substitution, but they can also be direct-
ly subjected to the second step in a one-pot procedure. The
reactions do not require elevated temperatures, stoichiometric
Scheme 1. a) Amination of tetrahydrocarbazole derivatives 1 by the strategy
of CꢀH functionalization via intermediate peroxides (CHIPS); b) common
proposal for the observed shift from the 4a to the 1 position during nucleo-
philic substitution reactions with activated tetrahydrocarbazoles.
amounts of synthetic oxidants, or expensive metal catalysts,
but simply visible light, a sensitizer, and a simple Brønsted
acid. The latter can either be trifluoroacetic acid in catalytic
amounts or acetic acid as solvent, depending on the nature of
the substrates.[3a] The reaction could be applied to the synthe-
sis of pharmaceutically active compounds such as
4
(Scheme 1), which has antiviral properties in the nanomolar
range.[4]
[a] Dr. N. Gulzar, Dr. K. M. Jones, Dr. M. Klussmann
Max-Planck-Institut fꢀr Kohlenforschung
The interesting shift occurring in the transformation of 2 to
3, which substitutes the hydroperoxide at position 4a to install
the N-nucleophile at position 1 (Scheme 1a), has been ob-
served in a couple of related cases before. Substitution reac-
tions with tetrahydrocarbazoles bearing chloride, bromide, or
acetate leaving groups, for example, resulted in the same shift
from position 4a to 1,[5] while direct substitutions are also
known.[5c] The generally proposed rationale for the shift occur-
ring in these substitution reactions is an intermediate imine–
Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim an der Ruhr (Germany)
E-mail: klusi@mpi-muelheim.mpg.de
[b] H. Konnerth, Dr. M. Breugst
Universitꢁt zu Kçln
Department fꢀr Chemie
Greinstraße 4, 50939 Kçln (Germany)
E-mail: mbreugst@uni-koeln.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201405376.
Chem. Eur. J. 2015, 21, 1 – 11
1
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
enamine tautomerization that activates position 1 (Scheme 1b).
However, detailed mechanistic studies are lacking.
most appropriate to determine the reaction order with respect
to both substrates 5 and 6.
1
Another aspect of this reaction that is of interest to us is the
utilization of a peroxide moiety as leaving group in acid-cata-
lyzed reactions. Organic hydroperoxides can rearrange by OꢀO
bond cleavage in the presence of acid, a well-known example
is the key step in the industrial synthesis of phenol and ace-
tone by the Hock process using cumyl hydroperoxide.[6] On the
other hand, acid-catalyzed nucleophilic substitution reactions
of hydroperoxides by CꢀO bond cleavage are also known,
albeit with a limited number of substrates.[7] The reaction of
singlet oxygen with indole derivatives is well-known to give
rise to peroxides such as 2.[8] Methods that can utilize hydro-
peroxides selectively in substitution reactions would thus
enable CꢀH functionalizations in two simple steps; by photo-
catalyzed aerobic oxygenation and subsequent acid-catalyzed
substitution.
We used H NMR spectroscopy to determine formation of 7
from experiments that were quenched by base to stop the re-
action at fixed intervals. Each reaction was repeated two to
three times to ensure reproducibility. A polynomial curve of
third order was fitted to the average conversion versus time
profile, the first derivative of which then delivered the reaction
rate profile (see the Supporting Information for further details).
To determine the reaction order in aniline 6, a series of ex-
periments with constant concentrations of peroxide 5 and
varying concentrations of 6 was conducted, with 6 being in
excess over 5. The initial reaction rates were found to be
slightly irregular but nearly constant at different concentrations
of 6 (Figure 1).
We have experienced that a detailed understanding of the
reaction mechanism is extremely valuable in method develop-
ment.[7i] Because we want to further extend the CHIPS concept
into true one-pot CꢀH functionalization reactions, we started
a detailed mechanistic investigation of the reaction shown in
Scheme 1a. Herein, the results of a combined experimental
and computational study are presented.
Results and Discussion
Experimental studies
The reaction of choice for the mechanistic studies was the cou-
pling of hydroperoxide 5 with p-cyanoaniline in acetic acid to
give the substituted tetrahydrocarbazole 7 (Scheme 2a). This
reaction proceeds cleanly with high yields, which is also the
case when the conditions are changed to methanol as solvent
and trifluoroacetic acid as catalyst.[3a]
Figure 1. Determining the reaction order in aniline 6 for the reaction shown
in Scheme 2a, plotting initial rates versus [6]0. All reactions: 0.05m 5, AcOH
solvent, 26.58C.
As an approximation, the reaction order in 6 appears to be
zero. The irregularity of the kinetic behavior could be due to
the fact that the product precipitates during the reaction and
that larger excesses of the weakly basic aniline nucleophiles
reduce the final product yield, as was observed before.[3a]
The reaction order in peroxide 5 was determined in the
same way by varying its concentration while keeping all other
parameters constant (Figure 2). For this set of experiments, the
peroxide was used equimolar or in excess compared to the
aniline.
The rate is clearly increasing with an increasing concentra-
tion of peroxide 5, and from a linear regression analysis a reac-
tion order in 5 of approximately 1.0 could be derived.
Scheme 2. a) Model reaction for this mechanistic study; b) possible key in-
termediates 8 and 9 for SN2’- and SN1-like pathways, respectively.
A graphical evaluation of this kinetic data, by using the reac-
tion progress kinetic analysis method,[9] leads to the same con-
clusions (see the Supporting Information). Similar results were
also obtained from kinetic studies using reaction calorimetry
for reactions with trifluoroacetic acid as catalyst in methanol
and DMSO, respectively. These experiments revealed approxi-
mate reaction orders of 0.7 and 1.0, respectively, in 5, zero in 6
and 1.0 in acid (see the Supporting Information). Although the
rate behavior appears to be more complex, these findings
clearly support an approximate rate expression for the reaction
Possible scenarios for the acid-catalyzed substitution reac-
tion, including the shift from position 4a to position 1, are
attack of the aniline on the hydroperoxide 8 (and its protonat-
ed form, respectively) or the cation 9, which can be viewed as
SN2’- and SN1-like mechanisms, respectively (Scheme 2b). To
distinguish between these two cases, kinetic studies appeared
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
2
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 4. Reversible formation of substitution products with solvent and
anilines.
structure (Scheme 4). When subjected to reaction conditions,
most of 10b converted to 11, and when 11 was dissolved in
acetic acid, it partially converted to 10b. Likewise, we found
that the coupling product with p-nitroaniline, 12, could be par-
tially converted to 7 when treated with 6, and vice versa. A
competition experiment of 5 with both anilines gave a mixture
of products 7 and 12 in a 1:1 ratio. All of these results suggest
that the postulated intermediate cation 9 can be trapped by
solvent and aniline nucleophiles alike, that all of these prod-
ucts are formed reversibly and that the trapping products with
anilines are thermodynamically favored (Scheme 4; see the
Supporting Information for details).
Figure 2. Determining the reaction order in peroxide 5 for the reaction
shown in Scheme 2a; plotting initial rates versus [5]0. All reactions: 0.05m 6,
AcOH solvent, 26.58C.
in acetic acid as given in Equation (1) and accordingly an SN1-
like mechanism via carbocation 9 (Scheme 2b).
r ꢁ k ꢂ ½5ꢃ
ð1Þ
The next question of interest to us was how the postulated
cationic intermediate 9 was formed. For this purpose, the reac-
tion was performed in deuterated acetic acid. A deuterium in-
corporation of approximately 28% was detected at position 1
of the coupling product 7-d (Scheme 3, see the Supporting In-
formation for further details).
To detect H2O2, which would be liberated according to the
proposed mechanism of Scheme 2b, we analyzed a reaction
mixture after two hours, at which point full conversion of the
peroxide had been attained. On adding an acidic solution of
K2Cr2O7, the mixture turned blue, indicating the formation of
CrO(O2)2.[10] We had shown previously that this method had
a negligible cross sensitivity with an organic hydroperoxide,[7i]
which was also confirmed for 5.
Based on all of the above-mentioned observations, we con-
clude that the acid-catalyzed tautomerization of imine 5 to en-
amine 8 is followed by CꢀO bond cleavage to give resonance-
stabilized carbocation 9 and hydrogen peroxide. The added
nucleophile 6 or the solvent can trap 9 to form the coupling
product 7 and the less-stable adduct 10a, both of which exist
in equilibrium with 9 (Scheme 5).
Scheme 3. Deuterium incorporation during the reaction in AcOD and postu-
lated enamine tautomer 8.
All attempts to observe 8 or 9 by studying the reaction of 5
in the presence of acid separately only led to decomposition
This partial incorporation of deuterium at position 1 sug-
gests an acid-catalyzed equilibrium between imine 5 and its
enamine tautomer 8 prior to formation of the cation 9. At-
tempts to detect the enamine by NMR spectroscopy were un-
successful. This and the relatively low incorporation of deuteri-
um suggest that 8 is highly reactive and reacts to afford the
coupling product before complete equilibrium H–D exchange
is achieved.
During 1H NMR analysis of the reaction mixture, we observed
indications for the intermediacy of acetoxy-substituted tetrahy-
drocarbazole 10a, but the product could not be isolated for
characterization. We could, however, isolate the corresponding
acetate 10b from 6-bromo tetrahydrocarbazole and confirm its
Scheme 5. Proposed reaction mechanism, based on experimental studies.
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
3
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
into mostly unidentified byproducts. As the kinetic experi-
ments were also not entirely conclusive and the negative reac-
tion order in 6 potentially indicates a more complex mecha-
nism, we looked for additional insight from computational
studies.
Computational studies
Figure 3. Putative isomeric tetrahydrocarbazole peroxides and their relative
stabilities [in kcalmolꢀ1, M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-
31+G(d,p)/IEFPCM(MeOH)].
As the computational approach also allows the study of inter-
mediates that are too unstable to be observed experimentally,
we analyzed the putative reaction mechanisms employing
density functional theory [M06-2X-D3/def2-TZVPP/IEFPCM//
M06-2X-D3/6-1+G(d,p)/IEFPCM(MeOH)]. Although most ex-
periments presented herein were performed in acetic acid as
solvent, it has been shown previously that similar results are
obtained in methanolic solution with trifluoroacetic acid as
Brønsted acid.[3a] Since calculations using methanol and acetic
acid as solvents gave very similar results, we chose to use
methanol as the solvent and trifluoroacetic acid as Brønsted
acid for the computational studies. In methanol, direct solvent
participation especially in proton-transfer reactions, is some-
what less likely due to the lower acidity. Furthermore, the com-
putational investigations employed 4-nitroaniline (13) instead
of the cyanoaniline 6. Previous studies have already shown
that both nucleophiles are similarly reactive and should react
through the same reaction mechanism.[3a] Further validation
for comparable reactivities comes from almost equal reaction
free energies for the combinations of differently substituted
anilines (e.g., 6 or 13) and tetrahydrocarbazole peroxide 5
(Scheme 6; see the Supporting Information for more details).
The computational result of comparable stabilities of 7 and 12
agrees nicely with the results from the competition experiment
between 6 and 13.
stable conformers. The peroxide 5, isolated from the photooxy-
genation reaction, with an imine substructure is more stable
than the isomer with an enamine substructure 8, which can be
rationalized by the stabilization of the annulated benzene ring.
Aromaticity within the indole is restored in 14, which renders
14 the most stable peroxide (DDG=ꢀ11.5 kcalmolꢀ1). The cor-
responding imines anti-15 and syn-15 were calculated to be
slightly more stable or isoenergetic comparted to the initially
formed peroxide 5. As neither 14, anti-15 nor syn-15 could be
observed under the experimental conditions, we have to con-
clude that their formation is kinetically hindered or that they
are rapidly consumed under the reaction conditions. As treat-
ment of 5 with acid in the absence of any nucleophiles does
not result in clean formation of 14 or 15 but in decomposition
to a mixture of mostly unknown products, the former explana-
tion is more likely.
As the tautomerization of 5 to 8 could be an important step
along the reaction mechanism, we had a closer look at this iso-
merization. In principle, three different pathways have to be
considered for this isomerization; a direct, intramolecular
proton transfer, protonation and deprotonation, or a trifluoro-
acetic acid-catalyzed proton shuttle (Scheme 7). Intramolecular
processes can be ruled out as previous investigations revealed
barriers of >60 kcalmolꢀ1 for these transformations.[11] Accord-
ing to our calculations, the initial N-protonation of 5 yielding
5ꢀH_c occurs readily without a significant barrier, but we were
unable to locate any transition states for a deprotonation at
the carbon terminus (!8). Instead, all estimated transition
states converged to the structure for a proton shuttle transi-
tion state TS1 (DG° =21.9 kcalmolꢀ1). Protonation at the nitro-
gen atom is almost complete and the deprotonation of the
carbon atom lacks behind, which renders TS1 highly asynchro-
nous. The intrinsic reaction coordinate (IRC) starting from TS1
confirmed that 5 is transformed to 8 via this transition state.
Based on these results, we conclude that a proton shuttle
mechanism via trifluoroacetic acid is most likely for the tauto-
merization of 5 to 8. The high barrier and the endergonic for-
Scheme 6. Overall thermodynamics for the reactions of the differently sub-
stituted anilines 6 and 13 and tetrahydrocarbazole peroxides (5) [in kcal
molꢀ1, M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/
IEFPCM(MeOH)].
Prior to the analysis of the complete reaction mechanism,
we investigated whether the initially formed tetrahydrocarba-
zole peroxide could undergo isomerization reactions. The ex-
perimental studies suggested that peroxide 5 tautomerizes to
the enamine isomer 8 (see above). Additional isomerizations
(!14, 15; Figure 3) could also occur via acid-catalyzed dissoci-
ation–association reactions or through [3,3]-sigmatropic rear-
rangements of the protonated peroxides. Figure 3 summarizes
the relative thermodynamic stabilities (DDG) of the most
1
mation of 8 are in line with the failure to detect 8 by H NMR
spectroscopy, but its presence was supported indirectly by the
reaction in AcOD (Scheme 3).
In the first step of the proposed mechanism, protonation of
tetrahydrocarbazole peroxide takes place. Therefore, we ana-
lyzed the stabilities of all possible protonated species (i.e., pro-
tonation at O1, O2, and the nitrogen atoms). As we cannot
rule out the participation of the peroxides 14 and 15 based on
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
4
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 7. Different mechanisms for the tautomerization of the imine 5 to
the enamine 8 catalyzed by trifluoroacetic acid [in kcalmolꢀ1, M06-2X-D3/
def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/IEFPCM(MeOH)].
experimental studies alone, we decided to keep them for the
computational analysis for the sake of completeness. These re-
sults are summarized in Figure 4.
Figure 4. Relative stabilities of differently protonated tetrahydrocarbazole
peroxides [in kcalmolꢀ1, M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-
31+G(d,p)/IEFPCM(MeOH)]; [a] could not be located, see main text.
In general, the nitrogen atom is the most basic center of the
tetrahydrocarbazole system and consequently, protonation at
this atom leads to the most stable cationic structures. These N-
protonated peroxides do not lie on the reaction coordinate for
the transformation under investigation, and protonation at the
oxygen atom is needed for the reaction to proceed. However,
protonation of either oxygen is significantly higher in energy
(DDG>14.3 kcalmolꢀ1). As a result, the concentration of O-pro-
tonated peroxides in solution should be rather small. Interest-
ingly, no stable protonated adduct 8ꢀH_a could be located on
the potential energy surface. Protonation of the oxygen atom
of peroxide 8 resulted in an immediate CꢀO bond fission with
formation of a stabilized carbocation (!9) and hydrogen per-
oxide.
Subsequently, hydrogen peroxide is released from the pro-
tonated species and a carbocation is formed. Different possibil-
ities have been evaluated and are summarized in Scheme 8,
and the different transition states are depicted in Figure 5.
Three different allyl cations (9, 16a and 16b) could be formed
from the corresponding protonated peroxides. Our calculations
indicate that the aza-allyl cations 16a and 16b are significantly
higher in energy than the analogue 9 (DDG=37.6 and
Figure 5. Transition states for the dissociation of hydrogen peroxide, relative
activation free energies (in kcalmolꢀ1) and selected bond lengths (in ꢁ)
[M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/IEFPCM(MeOH)].
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
therefore a much earlier transi-
tion state. The high energies of
both intermediates and transi-
tion states render the formation
of 16a or 16b very unlikely.
In principle, each cationic in-
termediate could now be at-
tacked
by
a
nucleophile
(Scheme 9). The unstable aza-
allyl cations 16a and 16b (if
formed at all) are so reactive
that they will undergo combina-
tions with p-nitroaniline (13)
without significant barriers (!
17aꢀH or anti-17dꢀH). In con-
trast, transition states could be
obtained for the nucleophilic
attack of 13 on both ends of the
less reactive isomer 9. The attack
at the indole substructure
through TS3a yields the ammo-
nium ion 17bꢀH in an ender-
gonic reaction. The attack at the
cyclohexane substructure results
not only in the most stable am-
monium ion 12ꢀH but also pro-
ceeds through the more favora-
Scheme 8. Different pathways for hydrogen peroxide dissociation from protonated tetrahydrocarbazole peroxides
with respect to the initially generated peroxide 5 [in kcalmolꢀ1, M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-
31+G(d,p)/IEFPCM(MeOH)].
35.3 kcalmolꢀ1, respectively). Whereas a protonation of the
oxygen atom of the enamine tautomer 8 directly leads to the
carbocation 9 in a concerted
ble transition state TS3b (Scheme 9, Figure 6). In both transi-
tion states, stabilizing p–p interactions between the aromatic
fashion without the formation of
the protonated peroxide 8ꢀH_a,
transition states could be locat-
ed for hydrogen peroxide cleav-
age in the other isomers. A very
small barrier of 0.3 kcalmolꢀ1 be-
tween 14ꢀH_a and TS2c has
been calculated on the M06-2X-
D3/6-31+G(d,p)/IEFPCM poten-
tial energy surface, whereas neg-
ative barriers were obtained
from high-level M06-2X single-
point calculations. Thus, the con-
version of 14ꢀH_a to 9 via TS2c
occurs without a significant bar-
rier. While
9 can be formed
easily, transition states for the
formation of the aza-allyl cations
16a and 16b are considerably
higher (DG° =51.2 and 54.5 kcal
molꢀ1
,
respectively; Scheme 8
and Figure 5). The breaking CꢀO
bond is significantly longer in
TS2a and TS2d than in TS2c re-
flecting the highly exergonic re-
action from 14ꢀH_a to
9
Scheme 9. Different pathways for the nucleophilic attack of 4-nitroaniline (13) on the allyl cations 14 [free ener-
(DDG=ꢀ17.4 kcalmolꢀ1)
and gies in kcalmolꢀ1 with respect to 05, M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/IEFPCM(MeOH)].
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
rings render the aryl groups almost parallel. The shorter CꢀN
bond length in TS3a (Figure 6) can again be rationalized with
the endergonicity of this step and thereby with a later transi-
tion state. Attempts to locate the analogous SN2 or SN2’ transi-
tion states (Scheme 2b) starting from different transition state
guesses failed, which further supports the experimental evi-
dence for an SN1-type mechanism.
Deprotonation of the ammonium ions affords the final sub-
stitution products 12 and 17. Among those, the experimentally
isolated isomer 12 is by far the most stable (Scheme 9). As 12
is the only isomer in which the indole substructure has been
reinstalled, it is not surprising that 12 is the only product that
is formed in an overall exergonic reaction.
Figure 6. Transition states for the nucleophilic attack of 13 on the allyl
cation 9, relative activation free energies (in kcalmolꢀ1) and selected bond
lengths (in ꢁ) [M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/
IEFPCM(MeOH)].
Scheme 10. Calculated Gibbs free energy profiles for the formation of the substitution product 12 from the peroxide 5 and the aniline 13 [in kcalmolꢀ1, M06-
2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/IEFPCM(MeOH)].
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
7
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Based on the computational analysis presented above, two
pathways are conceivable for the CꢀH functionalization and
the corresponding Gibbs free energy profiles are shown in
Scheme 10. The lowest energy pathway (black lines in
Scheme 10) involves the initial tautomerization to the enamine
isomer 8, protonation and elimination of hydrogen peroxide
and subsequent nucleophilic attack of aniline. The rate-limiting
step for this pathway was calculated to be the initial isomeriza-
tion (TS1), which is in line with the experimentally determined
rate law (first order in peroxide, zeroth order in aniline, first
order in acid if used in catalytic amounts in DMSO or metha-
nol). The calculated activation free energy of 21.9 kcalmolꢀ1 is
in good qualitative agreement with the experimental reaction
times of two to four hours. As an alternative, the initially gen-
erated peroxide 5 could isomerize to the thermodynamically
most stable peroxide 14. In this case, the dissociation of hydro-
gen peroxide would be the rate-limiting step. However, as this
would require an activation free energy of more than 38 kcal
molꢀ1 and there is no experimental evidence for an involve-
ment of 14, this pathway has to be discarded.
diastereomer isolated in the experimental studies is preferred
both thermodynamically and kinetically (see the Supporting In-
formation for a detailed discussion).
Especially in reactions with low conversion to the desired
products, we could observe and characterize the spirocyclic in-
dolinone by-product 21. This could be formed from the tetra-
hydrocarbazole alcohol 20, which in turn might be generated
either by reduction of the corresponding peroxide 5 or by the
attack of water on the allyl cation 9 (Scheme 12). The former is
probably more likely, as the formation of 16a is highly ender-
gonic. The alcohol 20 is subsequently protonated at the nitro-
gen atom and the resulting b-hydroxy iminium ion 20ꢀH un-
dergoes a ring contraction through TS3. This step requires
a substantial activation free energy (DG° =33.1 kcalmolꢀ1). The
cationic ketone 21ꢀH is subsequently deprotonated to yield
the spiro compound 21 in an overall almost thermoneutral re-
action. The rate-limiting step for this reaction is significantly
higher than those for the formation of the amino product 12
(see above) and therefore, the spiro compound will not be
formed in significant amount.
As mentioned above, the cationic intermediate 9 can be re-
versibly trapped by various nucleophiles and we have addi-
tionally calculated the stabilities of these intermediates
(Scheme 11). All products are formed in exergonic reactions
and all reaction free energies fall in a very small energy range
of only 1 kcalmolꢀ1. When free energies are considered, only
the methanol adduct 19 was calculated to be slightly more fa-
vorable than the amino product 12. However, 12 is slightly
preferred (DDH=1.1 kcalmolꢀ1) when reaction enthalpies are
compared, which is in line with the experimental preference of
12 over 19. Additionally, product 12 precipitates during the re-
action, further driving the equilibrium to its side.
The reversible formation of 12 also explains the diastereose-
lectivities we had previously observed in products bearing
a substituent in the annulated cyclohexane ring.[3a] The major
Scheme 12. Formation of a spiro-indolinone 21 and the rate-limiting transi-
tion state TS4 as a potential side reaction [in kcalmolꢀ1, M06-2X-D3/def2-
TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/IEFPCM(MeOH)].
Conclusions
In summary, our combined experimental and computational in-
vestigations of the reaction of hydroperoxide 5 with aniline 6
in acetic acid support an acid-mediated nucleophilic substitu-
tion of hydrogen peroxide via an SN1-like mechanism. The ob-
served shift of the leaving group at position 4a to the new Cꢀ
N bond at position 1 is rationalized by an acid-mediated tauto-
Scheme 11. Thermodynamic stabilities of different products obtained from
nucleophilic attack on the intermediate carbocation 9 [free energies in kcal
molꢀ1 M06-2X-D3/def2-TZVPP/IEFPCM//M06-2X-D3/6-31+G(d,p)/
IEFPCM(MeOH)].
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
8
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
mann-averaged ensembles. All DFT calculations were performed
with Gaussian 09.[20]
merization of the imine form of the hydroperoxide to an en-
amine, which generates a stabilized allylic cation after loss of
hydrogen peroxide. Computational investigations indicate that
the rate-limiting step is this initial tautomerization, which is in
excellent agreement with the reaction orders from kinetic
studies and the incorporation of deuterium during the reaction
in deuterated acetic acid. These investigations could be rele-
vant for other reactions with tetrahydrocarbazole derivatives
bearing a leaving group at position 4a. They also support the
strategy of CꢀH functionalization via intermediate peroxides
(CHIPS) and could be helpful in future extensions towards
other substrates.
Acknowledgements
Financial support from the DFG (KL 2221/3-1 and Heisenberg
scholarship to M.K., KL 2221/4-1), the Fonds der chemischen
Industrie (Liebig scholarship to M.B.) and the MPI fꢂr Kohlen-
forschung is gratefully acknowledged. We thank Philipp
Schulze for support with the analysis of H2O2. This work used
the Cologne High Efficiency Operating Platform for Sciences
(CHEOPS).
Experimental Section
Keywords: computational chemistry
· kinetics · nitrogen
heterocycles · peroxides · reaction mechanisms
Experimental Details
Unless otherwise indicated, all reagents and solvents were pur-
chased from commercial distributors and used as received. The hy-
droperoxide 5 and the products 7 and 12 were prepared according
[1] a) N. Gulzar, B. Schweitzer-Chaput, M. Klussmann, Catal. Sci. Technol.
2014, 4, 2778–2796; b) S. A. Girard, T. Knauber, C.-J. Li, Angew. Chem.
2014, 126, 76–103; Angew. Chem. Int. Ed. 2014, 53, 74–100; c) C.
Zheng, S.-L. You, RSC Adv. 2014, 4, 6173–6214; d) R. Rohlmann, O. G.
MancheÇo, Synlett 2013, 24, 6–10; e) C. S. Yeung, V. M. Dong, Chem.
Rev. 2011, 111, 1215–1292; f) C. Liu, H. Zhang, W. Shi, A. Lei, Chem. Rev.
2011, 111, 1780–1824; g) M. Klussmann, D. Sureshkumar, Synthesis
2011, 353–369; h) C. J. Scheuermann, Chem. Asian J. 2010, 5, 436–451;
i) A. R. Dick, M. S. Sanford, Tetrahedron 2006, 62, 2439–2463.
1
to published procedures.[3a,8e] For H and 13C NMR spectra of new
compounds and further details of kinetic and other experiments,
see the Supporting Information.
Computational Details
[2] a) Z. Shi, C. Zhang, C. Tang, N. Jiao, Chem. Soc. Rev. 2012, 41, 3381–
3430; b) A. N. Campbell, S. S. Stahl, Acc. Chem. Res. 2012, 45, 851–863.
[3] a) N. Gulzar, M. Klussmann, Org. Biomol. Chem. 2013, 11, 4516–4520;
b) N. Gulzar, M. Klussmann, J. Visualized Exp. 2014, 88, e51504 (DOI:
51510.53791/51504).
[4] a) S. D. Boggs, K. S. Gudmundsson, L. D. A. Richardson, P. R. Sebahar, Tet-
rahydrocarbazole derivatives and their pharmaceutical use, WO 2004/
110999A1, 2004; b) K. S. Gudmundsson, HCV Inhibitors, WO 2006/
121467A2, 2006.
[5] a) F. Ying-Hsiueh Chen, E. Leete, Tetrahedron Lett. 1963, 4, 2013–2017;
b) R. J. Owellen, J. Org. Chem. 1974, 39, 69–72; c) R. J. Owellen, C. A.
Hartke, J. Org. Chem. 1976, 41, 102–104; d) C. Chan, C. Li, F. Zhang, S. J.
Danishefsky, Tetrahedron Lett. 2006, 47, 4839–4841; e) K. Higuchi, M.
Tayu, T. Kawasaki, Chem. Commun. 2011, 47, 6728–6730; f) R. E. Ziegler,
S.-J. Tan, T.-S. Kam, J. A. Porco, Angew. Chem. 2012, 124, 9482–9485;
Angew. Chem. Int. Ed. 2012, 51, 9348–9351; g) H. Zaimoku, T. Hatta, T.
Taniguchi, H. Ishibashi, Org. Lett. 2012, 14, 6088–6091.
For all intermediates of the CꢀH functionalization of tetrahydrocar-
bazole, the conformational space was explored using the OPLS-
2005 force field[12] and a modified Monte Carlo search routine im-
plemented in MacroModel version 9.8.[13] An energy cut-off of
20 kcalmolꢀ1 was used for the conformational analysis, and struc-
tures with heavy-atom RMSDs less than 1 ꢁ after the initial force
field optimization were considered to be the same conformer. The
remaining structures where subsequently optimized with the
meta-hybrid M06–2X functional,[14] Grimme’s D3 dispersion correc-
tion (zero-damping),[15] and the double-z basis set 6-31+G(d,p).
Solvation by methanol was accounted for by using the integral
equation formalism polarizable continuum model (IEFPCM) for all
calculations (optimizations, frequencies, and single-point ener-
gies).[16] Polarizable continuum models do not have a large impact
on the calculated frequencies, but might be necessary for the loca-
tion of transition states in some cases.[17] Vibrational analysis veri-
fied that each structure was a minimum or a transition state. Fol-
lowing the intrinsic reaction coordinates (IRC) confirmed that all
transition states connected the corresponding reactants and prod-
ucts on the potential energy surface. Thermal corrections were cal-
culated from unscaled harmonic vibrational frequencies at the
same level of theory for a standard state of 1 molLꢀ1 and 298.15 K.
Entropic contributions to the reported free energies were calculat-
ed from partition functions evaluated with Truhlar’s quasiharmonic
approximation.[17] This method uses the same approximations as
the usual harmonic oscillator approximation except that all vibra-
tional frequencies lower than 100 cmꢀ1 are set equal to 100 cmꢀ1
to correct for the breakdown of the harmonic oscillator approxima-
tion for low frequencies. Electronic energies were subsequently ob-
tained from single-point calculations employing the M06-2X func-
tional with the large triple-z def2-TZVPP basis set,[18] IEFPCM for
methanol, and Grimme’s D3 dispersion correction. An ultrafine grid
corresponding to 99 radial shells and 590 angular points was used
throughout this investigation for numerical integration of the den-
sity.[19] The reported free energies refer to the lowest-energy con-
former and are almost identical to those obtained from Boltz-
[6] a) H. Hock, S. Lang, Ber. Dtsch. Chem. Ges. 1944, 77, 257–264; b) H.
Hock, H. Kropf, Angew. Chem. 1957, 69, 313–321.
[7] See for example: a) A. G. Davies, R. V. Foster, R. Nery, J. Chem. Soc. 1954,
2204–2209; b) C. W. Jefford, J. C. Rossier, S. Kohmoto, J. Boukouvalas,
Helv. Chim. Acta 1985, 68, 1804–1814; c) I. Saito, H. Nakagawa, Y. H.
Kuo, K. Obata, T. Matsuura, J. Am. Chem. Soc. 1985, 107, 5279–5280;
d) L. Liguori, H.-R. Bjorsvik, F. Fontana, D. Bosco, L. Galimberti, F. Minisci,
J. Org. Chem. 1999, 64, 8812–8815; e) P. H. Dussault, H.-J. Lee, X. Liu, J.
Chem. Soc. Perkin Trans. 1 2000, 3006–3013; f) ꢃ. Pintꢄr, A. Sud, D. Sure-
shkumar, M. Klussmann, Angew. Chem. 2010, 122, 5124–5128; Angew.
Chem. Int. Ed. 2010, 49, 5004–5007; g) ꢃ. Pintꢄr, M. Klussmann, Adv.
Synth. Catal. 2012, 354, 701–711; h) C. G. Piscopo, S. Buhler, G. Sartori,
R. Maggi, Catal. Sci. Technol. 2012, 2, 2449–2452; i) B. Schweitzer-
Chaput, A. Sud, ꢃ. Pinter, S. Dehn, P. Schulze, M. Klussmann, Angew.
Chem. 2013, 125, 13470–13474; Angew. Chem. Int. Ed. 2013, 52, 13228–
13232.
[8] For reviews, see: a) M. V. George, V. Bhat, Chem. Rev. 1979, 79, 447–
478; b) M. R. Iesce, F. Cermola, F. Temussi, Curr. Org. Chem. 2005, 9,
109–139; for tetrahydrocarbazole, see: c) R. J. S. Beer, L. McGrath, A.
Robertson, J. Chem. Soc. 1950, 2118–2126; d) B. Witkop, J. B. Patrick, M.
Rosenblum, J. Am. Chem. Soc. 1951, 73, 2641–2647; e) C. A. Mateo, A.
Urrutia, J. G. Rodrꢅguez, I. Fonseca, F. H. Cano, J. Org. Chem. 1996, 61,
810–812.
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
9
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[9] a) D. G. Blackmond, Angew. Chem. 2005, 117, 4374–4393; Angew. Chem.
Int. Ed. 2005, 44, 4302–4320; b) J. S. Mathew, M. Klussmann, H. Iwa-
mura, F. Valera, A. Futran, E. A. C. Emanuelsson, D. G. Blackmond, J. Org.
Chem. 2006, 71, 4711–4722.
[18] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[19] S. E. Wheeler, K. N. Houk, J. Chem. Theory Comput. 2010, 6, 395–404.
[20] 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. Na-
katsuji, 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. Ha-
segawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Ra-
ghavachari, 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. Zakr-
zewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Dan-
iels, ꢇ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussi-
an 09, revision D.01; Gaussian, Inc., Wallingford CT, 2009.
[10] D. F. Evans, J. Chem. Soc. 1957, 0, 4013–4018.
[11] a) J.-F. Lin, C.-C. Wu, M.-H. Lien, J. Phys. Chem. 1995, 99, 16903–16908;
b) C.-C. Wu, M.-H. Lien, J. Phys. Chem. 1996, 100, 594–600; c) J. Andrꢄs,
L. R. Domingo, M. T. Picher, V. S. Safont, Int. J. Quantum Chem. 1998, 66,
9–24.
[12] J. L. Banks, H. S. Beard, Y. Cao, A. E. Cho, W. Damm, R. Farid, A. K. Felts,
T. A. Halgren, D. T. Mainz, J. R. Maple, R. Murphy, D. M. Philipp, M. P. Re-
pasky, L. Y. Zhang, B. J. Berne, R. A. Friesner, E. Gallicchio, R. M. Levy, J.
Comput. Chem. 2005, 26, 1752–1780.
[13] MacroModel, V. 10.2, Schrçdinger, LLC, New York, NY, 2013.
[14] Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241.
[15] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104.
[16] E. Cancꢆs, B. Mennucci, J. Tomasi, J. Chem. Phys. 1997, 107, 3032–3041.
[17] R. F. Ribeiro, A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B
2011, 115, 14556–14562.
Received: September 23, 2014
Published online on && &&, 0000
&
&
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
10
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
Substitute a hydroperoxide: Tetrahy-
drocarbazoles can undergo a CꢀH ami-
nation via intermediate hydroperoxides
and acid-catalyzed substitution with ani-
lines. The mechanism of this reaction
was studied by kinetics and DFT calcula-
tions and an initial imine–enamine tau-
tomerization was found to be the rate-
determining step.
&
Reaction Mechanisms
N. Gulzar, K. M. Jones, H. Konnerth,
M. Breugst,* M. Klussmann*
&& – &&
Experimental and Computational
Studies on the CꢀH Amination
Mechanism of Tetrahydrocarbazoles
via Hydroperoxides
Chem. Eur. J. 2015, 21, 1 – 11
www.chemeurj.org
11
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ