Mechanistic Investigations of the Catalytic Formation of Lactams from Amines and Water with Liberation of H2

Article abstract of DOI:10.1021/jacs.5b01750

The mechanism of the unique lactam formation from amines and water with concomitant H₂ liberation with no added oxidant, catalyzed by a well-defined acridine-based ruthenium pincer complex was investigated in detail by both experiment and DFT calculations. The results show that a dearomatized form of the initial complex is the active catalyst. Furthermore, reversible imine formation was shown to be part of the catalytic cycle. Water is not only the oxygen atom source but also acts as a cocatalyst for the H₂ liberation, enabled by conformational flexibility of the acridine-based pincer ligand. (Figure Presented).

Full text of DOI:10.1021/jacs.5b01750

Article
pubs.acs.org/JACS
Mechanistic Investigations of the Catalytic Formation of Lactams
from Amines and Water with Liberation of H2
Urs Gellrich,†,‡ Julia R. Khusnutdinova,†,§,‡ Gregory M. Leitus,∥ and David Milstein*,†
∥
†Department of Organic Chemistry and Department of Chemical Research Support, Weizmann Institute of Science, Rehovot,
76100, Israel
S
* Supporting Information
ABSTRACT: The mechanism of the unique lactam formation
from amines and water with concomitant H2 liberation with no
added oxidant, catalyzed by a well-defined acridine-based
ruthenium pincer complex was investigated in detail by both
experiment and DFT calculations. The results show that a
dearomatized form of the initial complex is the active catalyst.
Furthermore, reversible imine formation was shown to be part
of the catalytic cycle. Water is not only the oxygen atom source
but also acts as a cocatalyst for the H2 liberation, enabled by
conformational flexibility of the acridine-based pincer ligand.
acridine-based Ru pincer complex 1 in the presence of a
catalytic amount of NaOH. Complex 1 was previously
developed in our group as a catalyst for alcohol amination
with NH3 as well as the reverse reaction, deamination of
primary amines, via a “hydrogen borrowing” pathway.14−16
Importantly, the lactam formation shown in Scheme 1 is
conceptually different from “hydrogen borrowing” method-
ology, and the overall reaction is accompanied by evolution of 2
equiv of H2 gas. Water plays a crucial role in this transformation
acting as a source of oxygen atom for formal oxygenation of a
CH2 group. In principle, such transformation resembles water
splitting where organic substrate (amine in this case) is used as
an acceptor of O atom. The only other examples of such CH2
group oxygenation by water were reported recently, and they
are limited to alcohol conversion to carboxylate or CO2 (in case
of MeOH) in the presence of a stoichiometric base.17,18
Notably, the reaction in Scheme 1 is accomplished with only a
catalytic amount of a base in combination with 1 or even under
base-free conditions using a modified catalyst 2 (vide infra).
We previously proposed that the amine-to-lactam conversion
involves intermediate imine formation via H2 liberation
catalyzed by 1 (Scheme 2). Water attack would yield a
hemiaminal, and a second dehydrogenation then finally results
in the lactam formation.13
INTRODUCTION
■
Amide formation is one of the most important topics in organic
synthesis since this structural element is found in many
pharmaceuticals and polymers. A great variety of methods have
been developed for amide formation, including coupling of
acylating reagents with amines, hydrolysis of nitriles, dehydro-
genative coupling of amines with alcohols, and other
methods.1−3 However, the formation of amides via direct
oxygenation of a CH2 group in the α-position to N remains
relatively rare and typically requires the use of a stoichiometric
oxidant (PhIO, peroxides, O2, RuO2/NaIO4, etc.).4−12
Recently, our group has developed a fundamentally new
method for lactam formation from cyclic amines using water as
the only reagent and as the oxygen atom source (Scheme 1).13
This transformation does not require any stoichiometric
oxidant apart from water and is accompanied by liberation of
H2 gas, thus providing an atom economical approach for the
synthesis of lactams. The lactam formation is catalyzed by the
Scheme 1. Lactam Formation from Cyclic Amines Using
Water As the Oxygen Atom Source Catalyzed by 1
Scheme 2. Proposed Pathway for Lactam Formation
Catalyzed by 1
Received: February 16, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
This hypothesis is in agreement with the experimental
finding that H2 was detected by GC/TCD after the reaction as
described in a previous publication.13 The imine intermediate
was proposed based on the H/D exchange experiments,
however, it was not observed directly under catalytic
conditions. Moreover, it was proven that water acts as a source
of the oxygen atom in the lactam product based on experiments
using H218O.13
However, the catalytic mechanism and the role of acridine-
based pincer ligand in this unprecedented amine-to-lactam
dehydrogenation remained unclear. The aim of the study
presented herein is to investigate the mechanism by combining
experimental and theoretical investigations.
Figure 1. ORTEP representation of 2 (50% probability ellipsoids).
Selected bond lengths (Å) and angles (deg): Ru1−N1 2.124(2), Ru1−
P1 2.3301(9), Ru1−P2 2.324(1), Ru1−C28 1.816(4), Ru1−H1
1.47(4), C7−C6 1.502(5), C7−C8 1.502(5); P1−Ru1−P2
160.55(3), N1−Ru1−C28 154.93(13), C6−C7−C8 108.1(3).
RESULTS AND DISCUSSION
■
Role of a Base and Transformation of the Ligand
during Catalysis. Previous studies involving complex 1
reported by our group indicate that the acridine-based pincer
ligand exhibits reactivity at C9 position of the acridine
backbone, leading to reduction of the central acridine ring.19
In particular, it was previously shown that complex 1 reacts
with H2 in the presence of a base to give a hydrido carbonyl
complex 2 (Scheme 3, Method A) in which a central acridine
ring is reduced.19 Complex 2 was previously isolated and fully
characterized by NMR, IR spectroscopy, and elemental
analysis.19
The X-ray structure of 2 reveals a distorted square pyramidal
RuII center bound to a pincer ligand through two P atoms and
an acridine N atom. It unambiguously confirms that the central
acridine ring is reduced: the bond distances C7−C6 and C7−
C8 at the cental acridine ring are typical for single C−C bonds,
and the bond angle C6−C7−C8 is characteristic of a
tetrahedral sp3 hybridized central carbon. The hydrogen atom
H1 was localized as an electron density peak (see Experimental
Details section). No ligands are present trans to the hydride,
likely due to its strong trans influence. The Ru−N bond
distance (2.124 Å) is significantly shorter than the Ru−N
distance reported for complex 1 (2.479 Å), suggesting that the
reduction of the central acridine ring leads to a stronger
interaction between the Ru center and a negatively charged N
donor.15,19 Moreover, the X-ray structure is in good agreement
with the DFT-calculated structure for 2 (vide infra).
Scheme 3. Synthesis of the Dearomatized Hydride Carbonyl
a
Complex 2
We then tested the catalytic activity of complex 2 in the
absence of a base in amine-to-lactam conversion under the
conditions that were previously reported for catalyst 1 in the
presence of catalytic NaOH, using pyrrolidine (A) as a model
substrate. When pyrrolidine was heated in a dioxane-water
mixture (1:1 v/v) in the presence of 1 mol % of 2 at 150 °C for
48 h, 2-pyrrolidone (B) was obtained as a major product in
55% yield. The major byproduct was 4-(1-pyrrolidinyl)butanoic
acid (C) formed in 10% yield, similar to the results reported
previously.13 The comparison of these results with the
analogous reaction in the presence of complex 1 (1 mol %)
and NaOH (1.5 mol %) is given in Table 1 and shows that the
conversion and the product yields are similar in these two
systems. This result suggests that complex 2 may be the
catalytically competent species under the catalytic conditions
which is also consistent with our DFT study (vide infra).
Importantly, the catalytic activity of complex 2 was observed
in the absence of NaOH suggesting that a base is not required
for amine-to-lactam conversion. The role of NaOH in the
lactam formation catalyzed by 1 is most likely to promote the
formation of the catalytically active species 2. The possible
mechanism of the formation of 2 was studied by DFT and will
be discussed in more detail in the following sections.
a
Method A (ref 17): H2 (5 bar), KOH (1 equiv), toluene, 135 °C, 48
h. Method B: benzyl alcohol (3.3 equiv), NaH (12 equiv), toluene,
RT, 2 h.
In addition, Hofmann et al. recently reported the synthesis of
Ru complexes with an analogous acridine-based pincer ligand
bearing bis(cyclohexyl)phosphine donors instead of bis-
(isopropyl)phosphine donors and their reactivity in amination
of alcohols with NH3.20 They showed that the reaction of the
analogous (AcrPNPCy)RuH(CO)Cl complex with benzyl
alcohol in the presence of a base also leads to the reduction
of the central acridine ring accompanied by benzyl alcohol
dehydrogenation.20
We hypothesized that the acridine-based ligand AcrPNPiPr
could undergo a similar reduction during the reaction with
amines or alcohols and synthesized complex 2 according to the
modified procedure reported by Hofmann et al. (Scheme 3,
Method B), by the reaction with benzyl alcohol (or ethanol) in
the presence of NaH. The same complex can also be obtained
when tBuOK is used as a base. Complex 2 was isolated in 77%
The mass balance in entries 1 and 2 of Table 1 sums up to
99−99.5%, taking into account that the formation of C
consumes 2 equiv of pyrrolidine A, thus indicating that no
other significant side reactions occur under these conditions.
The possible pathways leading to C were discussed in a
previous publication.13
yield and characterized by H, 13C, and 2D NMR. Single
crystals of 2 were grown by slow evaporation of pentane
solution and were characterized by X-ray diffraction (Figure
1).21
1
B
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
demonstrated that the trimer and monomer interconvert easily
in solution and cannot be separated.22 Therefore, this substrate
was used under standard catalytic conditions as a mixture of the
monomer and trimer in a water-dioxane solution (Scheme 6).
Table 1. Reactivity of Pyrrolidine in Water-Dioxane (1:1 v/
v) Catalyzed by 2 and 1/NaOH
Scheme 6. Catalytic Experiments Using 1-pyrroline As a
Substrate
conv. of A
B, %
C, %
entry
catalyst
2 (1 mol %)
T, °C
(%)
yield
yield
a
1
150
150
76
55
59
10
a
,
b
2
1 (1 mol %) NaOH
(1.5 mol %)
78.5
9.5
3
4
2 (1 mol %)
135
135
75
64
35
5
4
These experiments showed that independently from the
applied catalytic system (1/NaOH or 2), 2-pyrrolidone was
formed in yields similar to those obtained using pyrrolidine A
as a substrate (Table 2 vs Table 1). These results are consistent
1 (1 mol %) NaOH
(1.5 mol %)
42.5
a
Pyrrolidine (1 mmol), catalyst 2 (1 mol %) or 1 (1.0 mol %)/NaOH
(1.5 mol %), water (1.5 mL), and dioxane (1.5 mL) were heated in a
50 mL thick wall pressure tube for 48 h at 150 °C (silicon oil bath
temperature), if not indicated otherwise. % NMR yields and
conversions were determined by NMR and reported as averages of
Table 2. Conversion of Imine D to Lactam under Standard
a
Catalytic Conditions
two runs. from ref.13
b
entry
catalyst
conv. of D/D′ (%)
B, % yield
A, % yield
Interestingly, complex 2 also shows catalytic activity at a
lower temperature, 135 °C, yielding 2-pyrrolidinone in a
slightly higher yield, likely due to a slower side reaction leading
to the formation of C, which was formed in only 5% yield
(entry 3). By comparison, complex 1 in combination with
NaOH shows only moderate catalytic activity at this temper-
ature (entry 4).
Imine Reactivity under the Catalytic Conditions. We
have previously proposed that the amine-to-lactam conversion
likely occurs through an imine intermediate (Scheme 2).13
Since an imine could not be detected among the reaction
products, its formation was deduced based on an H/D
exchange experiment showing that significant H/D incorpo-
ration in the α- as well as the β-positions occurs when
pyrrolidine is heated in D2O-dioxane in the presence of 1/
NaOH at 135 °C (Scheme 4), while only a trace amount of 2-
1
2
1/NaOH
>99
>99
50
55
11
13
2
a
Reaction conditions: pyrroline D/D′ (1 mmol), catalyst 1 (1.0 mol
%)/NaOH (1.5 mol %) or 2 (1.0 mol %), and H2O-dioxane 1:1 v/v,
150 °C, 48 h. Yields and conversions were determined by NMR.
with imine D being part of the catalytic cycle. Furthermore, the
formation of pyrrolidine A was observed, likely formed via
hydrogenation of D with the H2 evolved in the conversion of
imine D to lactam B. This finding proves the assumption made
so far based on the H/D exchange experiments, that the imine
formation is reversible under the catalytic conditions and that
imine D is a viable intermediate of the lactam formation.
Possibility of the Reverse Reaction of Lactam-to-
Amine Conversion. To determine whether under anhydrous
reaction conditions and under high pressure of H2 we can
achieve the reverse transformation, hydrogenation of lactams to
cyclic amines, we examined the possibility of hydrogenation of
2-pyrrolidone to pyrrolidine catalyzed by complex 2. When 2-
pyrrolidine was heated in anhydrous dioxane in the presence of
1 mol % of 2 under 60 bar of H2, only a trace amount of
pyrrolidine was detected in the reaction mixture after 72 h at
150 °C (Scheme 7). Similarly, attempted hydrogenation of 2-
Scheme 4. Deuterium Incorporation in the α- and β-Position
of Pyrrolidine
Scheme 7. Attempted Hydrogenation of 2-Pyrrolidone to
Pyrrolidine
pyrrolidone was formed at this temperature as reported
previously.13 This is consistent with reversible dehydrogenation
of pyrrolidine to form 1-pyrroline (D) in the presence of the
Ru catalyst that occurs before the rate limiting step.
To confirm that 1-pyrroline (D) is indeed a viable
intermediate in the lactam formation, we synthesized it. 1-
pyrroline is known to exist in an equilibrium mixture with the
trimer D′ in solution (Scheme 5),22 and previous studies
pyrrolidone catalyzed by 1/NaOH in water-dioxane failed to
produce pyrrolidine.13 This suggests that the reverse reaction
likely has a high activation barrier, and the lactam formation is
practically irreversible under standard catalytic conditions even
in a closed system. This is fully consistent with the DFT studies
as discussed below.
Scheme 5. Monomer−Trimer Equilibrium in a Solution of 1-
Pyrroline
Computational Investigations. In order to gain more
insights into the elementary steps of the catalytic trans-
formation, we performed DFT calculations at the DSD-
C
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
PBEB95-D3BJ/def2-TZVPP//BP86-D3/def2-SV(P) level of
theory. In order to validate the level of theory used for the
geometry optimizations, we compared the structure of 2
derived from X-ray with the computed one, showing a good
agreement between experimental and computed bond lengths
and angles (Table 3).
acridine ring is similar to that calculated by Hofmann et al. for
the reaction of analogous complex (AcrPNPCy)RuH(CO)Cl
with MeOH.
Under catalytic conditions the pyrrolidine can displace the
chloride and be deprotonated by NaOH (either simultaneously
or consecutively). A β-hydride transfer from the deprotonated
pyrrolidine ligand to the C9 of central acridine ring with a low
barrier of 3.1 kcal/mol then leads to the active catalyst 2-mer
after dissociation of 1-pyrroline (pathway b) (see Figure 2). In
both pathways formation of the reduced complexes 4-mer and
6-mer is exothermic.
Table 3. Structural Parameters of 2 Measured by X-ray and
Computed by DFT
X-ray
BP86-D3/def2-SV(P)
Ru−N [Å]
2.124(2)
1.817(4)
2.327
2.136
1.832
Ru-CO [Å]
a
Ru−P [Å]
2.326
N−Ru−CO [deg]
P−Ru−P [deg]
154.93(13)
160.55(3)
151.01
158.81
a
Average over the two bond lengths.
The DSD-PBEB95-D3BJ/def2-TZVPP level of theory,
chosen for the single point energy calculations, was recently
shown to yield results very close to explicitly correlated coupled
cluster benchmark calculations for reaction energies and
barriers involving late transition-metal complexes with pincer
ligands.23 To take the influence of the solvent into account, the
SMD model was applied for a 1:1 mixture of dioxane/water,
with a dielectric constant of ε = 40.28, being the average weight
of dioxane and water.
Formation of 2 and Isomerization Pathways. Since 2
was shown to be the active catalyst, we first calculated a
pathway for its formation under the conditions used for the
direct synthesis (Scheme 3, Method B) and under catalytic
conditions. We assumed that in the presence of NaH, the
benzyl alcohol used in the synthesis of 2 gets deprotonated, and
the alkoxide substitutes the chloride leading to 3-mer (Scheme
8).
Figure 2. TS-2 involving hydride transfer from coordinated
pyrrolidine to the C9 position of the acridine ring, leading to imine
dissociation and formation of the active catalyst 2-mer. Bond lengths
are given in angstrom.
We then focused on possible stereoisomers of 2 and the
isomerization pathway connecting them (Scheme 9). We
Scheme 9. Isomerization Pathway Connecting Different
a
Conformers of 2
A β-hydride transfer to the C9 position of the acridine ring
with a barrier of 15.6 kcal/mol leads to the reduced species 4-
mer. Dissociation of the formed benzaldehyde yields 2-mer
(Scheme 8, pathway a). This dearomatization mechanism via
direct β-hydride transfer from the coordinated alkoxide to the
Scheme 8. Formation of 2 Using Benzyl Alcohol and NaH
(pathway a) or Pyrrolidine (pathway b) under Catalytic
a
Conditions
a
Calculated free energies (kcal/mol) in brackets with respect to 2-mer.
assumed that the dearomatized acridine-based pincer ligand
with the reduced central ring is sufficiently flexible to adopt
both meridional and facial coordination modes in Ru
complexes. In particular, some complexes with a fac-
coordinated reduced ligand were reported previously by our
group.15,19 Starting from 2-fac, 2-mer is accessible via TS-3 with
a
Calculated free energies (kcal/mol).
D
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a low barrier of 3.3 kcal/mol. This conformer is predicted to be
the most stable and indeed the one found by X-ray analysis. A
flip of the acridine ring yields 2′-mer. This ring flip shows the
highest single barrier calculated in the isomerization pathway
(21.8 kcal/mol). A change in the position of the CO and the
hydride ligand yields 2″-mer, which can directly interconvert to
2′-fac. For the most stable mer conformer 2-mer and the most
stable fac conformer 2′-fac, we also investigated the effect of
pyrrolidine coordination to these complexes.
The free energy differences and the barriers between all
conformers shown in Scheme 9 are small enough to make all
conformers of 2 accessible at a reaction temperature of 150 °C.
Therefore, all of them can serve as potential active catalysts.
H2 Liberation Pathways and Insights into the Role of
Water in H2 Elimination. Catalytic pathways starting from 2′-
fac will now be discussed in detail, focusing on a possible
explanation for the experimentally observed H2 liberation. We
first computed a pathway in which a hydride is transferred from
the C9 position of the acridine ligand to the ruthenium (TS-7),
forming the cis hydride species 7-fac. H2 elimination via TS-8
with an estimated energy of 33.7 kcal/mol yields the
ruthenium(0) species 8-fac with a free coordination site, able
to undergo an oxidative insertion in the NH bond of
pyrrolidine.24 A proton transfer from the β positon of the
heterocycle in 5′-mer and dissociation of the imine from 6′-mer
regenerates 2′-fac (Scheme 10).
the C9 position to the ruthenium (see Supporting Information
for more detail).
Interestingly, we found that after coordination of a water
molecule to 2′-fac, H2 elimination involving the hydride bound
to the ruthenium and a proton of the water molecule can take
place via a significantly lower barrier of 31.6 kcal/mol (TS-10
in Scheme 11; see also Figure 3) with respect to 2-mer. The
Scheme 11. Imine Formation via H2 Elimination from the
a
Hydroxy Complex 10-fac
a
Calculated free energies (kcal/mol) in brackets with respect to 2-mer.
Scheme 10. H2 Elimination and Imine Formation via a
a
Ruthenium(0) Intermediate
Figure 3. Transition state TS-10 for H2 elimination from the aqua
complex 10-fac. Bond lengths are given in angstrom.
calculated Mulliken atomic charge of the hydrogen bound to
water and involved in this process rises from 0.266 to 0.281 e
upon coordination of the water to 2′-fac, rendering it more
protic.
a
Calculated free energies (kcal/mol) in brackets with respect to 2-
mer.
Coordination of pyrrolidine to the hydroxy complex 10-fac
yields the complex 11-fac in which a N−H···O hydrogen bond
between the pyrrolidine ligand and the hydroxyl ligand is
already present. A proton transfer with concomitant H2O
liberation (TS-11) results in the formation of 12-fac. A similar
transition state in the alcohol-assisted coordinated amide
protonation was previously reported in the course of the
investigation of the ruthenium−acridine complex catalyzed
amination of primary alcohols.20 The vacant coordination site
in this complex renders a β-hydride elimination from the
deprotonated pyrrolidine possible, resulting after dissociation in
the formation of the imine and the regeneration of 2′-fac. The
actual barrier for the imine formation via the pathway depicted
in Scheme 11 is 31.6 kcal/mol with respect to 2-mer. The H/D
exchange in the α-CH2 group of pyrrolidine observed in D2O-
The activation energy required for the hydride transfer (TS-
7) from the ring to the ruthenium and the high energy of TS-9
and TS-2′ relative to 2-mer render this mechanistic scenario
unlikely even at high reaction temperatures. Since previous
studies on related ruthenium pincer complexes highlighted the
importance of water as proton shuttle, we investigated if water
could significantly lower the activation energy required for the
liberation of hydrogen.25,26 Transition states in which water
acts as a proton shuttle for the direct formation of 7-fac by
“bridging” Ru−H and the C9 position of the acridine with one
or two molecules of water were calculated to be even higher in
energy than the transition state for the hydride transfer from
E
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
dioxane (Scheme 4) is consistent with a reversible pyrrolidine-
to-(1-pyrroline) dehydrogenation and can be explained by H/D
exchange of the Ru−H in complex 2 (either as 2′-fac or 2-mer
conformer) or 13-fac with D2O and then reinsertion of 1-
pyrroline into Ru−D bond (Scheme 11). The possible
pathways for H/D exchange in β-CH2 position of pyrrolidine
were discussed in a previous publication.13
The possibility of inner sphere H2O-assisted hydrogen
liberation from a similar ruthenium complex was discussed
previously by Yoshizawa et al. but was found to be unlikely due
to high barriers required for the loss of CO ligand that enables
cis-arrangement of Ru−H and R−OH2 ligands.25 Suresh et al.
considered an alternative mechanism via rearrangement of a
CH2-arm-deprotonated (mer-PNN)Ru species to allow for cis-
arrangement of Ru−H and Ru−OH2.27 In most complexes
considered so far, the H2O-assisted hydrogen liberation takes
place via an outer sphere mechanism (either via external attack
or water molecule bridging the deprotonated arm) or the H2O
molecule has first to replace a ligand, while PNN ligand
remaines coordinated in a mer-fashion.25,28,29 By contrast, we
believe that the mechanism of H2 elimination through
transition state TS-10 is enabled by conformational mer/fac
Figure 4. Pathways of pyrrolidine dehydrogenation involving (a)
hydride transfer from the C9 position of the ligand (red); (b) H2
elimination from the aqua complex 10-fac (black); or (c) H2
elimination from the pyrrolidine complex 7′-fac (purple).
solvent, it is likely that the pyrrolidine as well as the imine form
hydrogen bonds to surrounding water molecules. We therefore
investigated if this affects the free energy associated with the
imine formation.
Interestingly, the explicit solvation by hydrogen bonds (A·
2H2O, Scheme 13) significantly lowers the endothermicity of
flexibility of the reduced AcrPNPiPr ligand allowing for
coordination of water cis to a hydride. This is a unique feature
of the acridine-based pincer ligand system compared to the
“classical” pyridine-based PNN and PNP ligand motifs in which
fac-coordination is not accessible, however, different modes of
H2 elimination are possible through deprotonation of the CH2
arm. In addition, such conformational flexibility likely enables
β-hydride elimination pathway leading to 10-fac.
a
Scheme 13. Free Energies of the 1-pyrroline (D) Formation
An alternative pathway, in which the H2 elimination involves
the pyrrolidine rather than water was also investigated (Scheme
12).
Scheme 12. H2 Elimination Involving the Pyrrolidine
a
Instead of Water As Proton Source
a
Calculated free energies (kcal/mol).
a
Calculated free energies (kcal/mol) in brackets with respect to 2-mer.
The coordination of pyrrolidine to 2′-fac is energetically
preferred compared to the coordination of water. However, the
barrier for H2 elimination from the pyrrolidine complex 14-fac
(37 kcal/mol) is 16 kcal/mol higher than starting from the
aqua complex 9-fac. Both findings may reflect the higher
basicity of pyrrolidine compared to water. The three
mechanistic scenarios depicted in Schemes 10, 11, and 12 are
summarized as potential free energy surface in Figure 4,
indicating that the pathway involving water (Scheme 11) is the
most favorable.
Formation of the imine is endothermic, in agreement with
the experimental finding that the imine cannot be observed
under typical reaction conditions. The barrier for the backward
reaction, yielding the pyrrolidine, is smaller, which is consistent
with the deuterium-labeling experiments, showing that the
imine formation is reversible. However, under the actual
reaction conditions using a 1:1 mixture of dioxane and water as
the imine formation (by 6 kcal/mol). The effect is less
pronounced with three surrounding water molecules. The
reaction is furthermore driven in the direction of the imine by
H2 leaving the condensed phase, which is not reflected by the
DFT calculations.
Formation of a Hemiaminal and a Lactam. The next
step involves the hemiaminal formation by a nucleophilic attack
of water at the imine. Coordination of the imine to ruthenium
may render the imine more electrophilic and therefore facilitate
this step. We were able to locate a transition state for the
nucleophilic attack of water at the double bond of the imine of
13-fac (Scheme 14, top). This transition state involves the
direct elimination of H2 according to IRC calculations.
However, the computed barrier of 48.7 kcal/mol renders this
transition state unlikely to be part of the catalytic trans-
formation. An attack of water on the liberated imine in which a
second water molecule serves as proton shuttle is more likely,
F
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
formation. This is in agreement with the deuterium-labeling
experiments which indicate that the dehydrogenation of the
pyrrolidine occurs prior to the rate limiting step. Once the
hemiaminal is formed, the dehydrogenation can take place in a
similar manner as the dehydrogen ation of the pyrrolidine.
Starting from the hydroxy complex 10-fac N-coordination of
the hemiaminal is energetically favored compared to O-
coordination (Scheme 15). On the other side, the proton
transfer process with concomitant H2O elimination involving
the OH group (TS-17) of the hemiaminal shows a lower
barrier than proton transfer from the NH group (TS-19),
which is in line with the acidity of these groups. The activation
energy for the β-hydride elimination is lower starting from the
N-coordinated complex 15-fac. Dissociation of the O- or N-
coordinated complexes 18-fac and 20-fac yields the lactam or
lactim, respectively (Scheme 14). The pathways for the lactim
and the lactam formation are summarized in Figure 5. The
highest barrier for the lactim formation would be the proton
transfer from the NH to the hydroxyl group bound to the
ruthenium (TS-19), whereas the β-hydride elimination to form
the lactam (TS-18) represents the highest barrier for the direct
lactam formation. Both mechanistic scenarios seem realistic at
the reaction temperature of 150 °C. The lactam is computed to
be 13.0 kcal/mol more stable than the lactim, which shows that
the driving force for the catalytic cycle is indeed the formation
of the lactam.
Scheme 14. Inner- and Outersphere Pathways for the
Hemiaminal Formation
a
a
Calculated free energies (kcal/mol)
having a barrier of 30.9 kcal/mol (Scheme 14, bottom). An
additional water molecule lowers this barrier to 27.0 kcal/mol.
Nevertheless, by taking into account that the imine formation
is endothermic, the hemiaminal formation represents the rate-
determining step of the catalytic transformation with an overall
barrier of 33.0 kcal via TS-15 and 33.5 kcal/mol via TS-16,
depending if one or two water molecules assist the hemiaminal
Furthermore, once the hemiaminal is formed, the barrier for
the forward reaction is significantly lower than the barrier for
the backward reaction, explaining the observed irreversibility of
the overall transformation.
a
Scheme 15. Pathways for the Lactam and Lactim Formation Starting with the Hydroxy Complex 10-fac
a
Calculated free energies (kcal/mol) in brackets with respect to 2-mer.
G
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
instruments equipped with a 30 m column (Restek 5MS, 0.32 mm
internal diameter) with a 5% phenylmethylsilicone coating (0.25 mm)
and helium as carrier gas.
Synthesis of 2. Sodium hydride (29.7 mg, 1.24 mmol, 12 equiv)
was added to a stirred solution of 1 (62.2 mg, 0.103 mmol) and benzyl
alcohol (37.0 mg, 3.3 equiv) in toluene (7 mL). The reaction mixture
was stirred for 2 h at RT and then filtered through a Celite plug. The
resulting orange toluene solution was evaporated to dryness, and the
solid residue was washed with pentane several times to remove benzyl
benzoate and excess benzyl alcohol. The resulting orange powder was
dried under high vacuum for several hours. Yield 45.5 mg (0.0796
mmol, 77%). X-ray quality crystals were obtained by slow evaporation
of a pentane solution. When ethanol (4 equiv) is used instead of
benzyl alcohol, the same product was obtained in lower yield (12.5 mg
from 30.4 mg, 44% yield). The same product 2 can be obtained if
tBuOK (2 equiv) is used as a base instead of NaH. NMR spectra of the
product match those for previously reported and characterized 2
obtained by a different method.19 31P{1H} NMR (C6D6, 500 MHz):
Figure 5. Pathways for the lactam formation (blue) and the lactim
formation (black). All energies are given with respect to 2-mer.
1
74.58 ppm (calibrated against H3PO4 external reference). H NMR
(C6D6, 500 MHz): −20.58 (t, 1H, J = 22 Hz, RuH), 0.68 (vq, 6H, JHP
= 13.5 Hz, JHH = 6.8 Hz, PCH(CH3)2), 0.98 (vq, 6H, JHP = 13.9 Hz,
JHH = 6.9 Hz, PCH(CH3)2), 1.15 (vq, 6H, JHP = 13.6 Hz, JHH = 6.8 Hz,
PCH(CH3)2); 1.22 (vq, 6H, JHP = 15.1 Hz, JHH = 7.1 Hz,
PCH(CH3)2), 1.44−1.56 (m, 2H, PCH(CH3)2), 1.90−2.03 (m, 2H,
PCH(CH3)2), 2.49−2.60 (m, 2H, PCHH), 2.92 (d, 2H, JPH = 12.5 Hz,
PCHH), 3.54 (d, 1H, JHH = 14.3 Hz, acridine C9HH), 3.72 (d, 1H, JHH
= 14.3 Hz, acridine C9HH), 6.8−7.0 (m, 4H, aryl), 7.10−7.25 (m, 2H,
overlaps with solvent peak). 13C{1H} NMR (C6D6, 500 MHz): 17.6
(s, PCH(CH3)2), 18.7 (s, PCH(CH3)2), 19.2 (s, PCH(CH3)2), 20.4 (s,
PCH(CH3)2), 24.3 (t, JCP = 7.8 Hz, PCH(CH3)2), 25.7 (t, JCP = 14.9
Hz, PCH(CH3)2), 27.9 (t, JCP = 9.2 Hz, PCH2). 36.5 (s, acridine C9),
119.3 (s, Ar), 121.7 (s, Ar, quat), 126.6 (s, Ar), 128.3 (s, Ar), 153.1 (t,
JCP = 4.5 Hz, Ar, quat), 210.1 (s, CO).
CONCLUSION
■
In summary, we have presented herein a thorough mechanistic
study of the acridine-based ruthenium pincer complex catalyzed
formation of lactams from amines using water with no added
oxidant. The conclusions drawn from the DFT calculations
agree well with the experimental findings. Complex 2 with the
reduced central ring of the acridine system was shown to be the
actual catalyst, which is in agreement with the DFT
computations. The isolated complex 2 is catalytically active in
the pyrrolidine-to-pyrrolidone conversion even in the absence
of a base, showing that catalytic NaOH that was used previously
in combination with the catalyst precursor 1 is not required for
the lactam formation, once the active catalyst is formed. The
conformer 2-mer was predicted to be most stable and was
indeed the conformer observed by X-ray diffraction. The
computations revealed that the imine formation is endothermic
and proceeds with an overall small barrier, which is in
agreement with the deuterium-labeling experiments and the
observed reactivity of the imine under catalytic conditions.
Finally, the DFT computations have shown that water might
catalyze the H2 liberation in a rare fashion, enabled by
General Procedure for Pyrrolidine-to-Pyrrolidone Conver-
sion Catalyzed by 2. A 50 mL pressure tube equipped with a
magnetic stirring bar was charged under N2 atmosphere with a
solution of 0.010 mmol of 2 in 1.5 mL of dioxane, 1.5 mL of distilled
water, and 1 mmol of pyrrolidine. The reaction mixture was stirred and
heated in a Teflon-sealed tube at 150 °C silicon bath temperature.
After 48 h, the reaction mixture was cooled down, an internal standard
(20 μL of pyridine) was added, and the reaction mixture was analyzed
by NMR. For NMR analysis, a sample of the reaction mixture (30−50
1
μL) was dissolved in 0.5−0.6 mL of D2O and analyzed by H NMR.
The crude yields were determined by 1H NMR by integration vs
pyridine internal standard; long delay time (10 s) was used to quantify
the amount of products. Characterization of the products was reported
previously.13 WARNING! Pressure develops during the reaction in a
closed system due to H2 evolution and significant solvent vapor
pressure at high temperatures. All reactions should be performed in
thick-glass pressure tubes using proper shielding. The reaction
mixtures should be cooled down in ice bath before slowly releasing
the H2 gas formed.
conformational flexibility of the reduced AcrPNPiPr ligand that
allows for its fac-coordination. The results presented in this
study might be useful for the development of other atom
economic amide formations as well as the discovery of new
catalytic reactions involving water as oxidant with H2 liberation.
EXPERIMENTAL DETAILS
■
General Specifications. All manipulations were carried out under
a nitrogen atmosphere using standard Schlenk and glovebox
techniques if not indicated otherwise. All chemicals for which
synthesis is not given were commercially available from Aldrich,
Acros, or STREM and were used as received without further
purification. 1,4-Dioxane and toluene were purified prior to use by
refluxing and distilling over Na/benzophenone under an argon
atmosphere. Distilled water was purchased from Aldrich (HPLC
grade) and was degassed prior to use by bubbling argon for at least 20
min. 4,5-Bis(di-iso-propylphosphinomethyl)acridine (AcrPNP) and
(AcrPNP)RuH(CO)Cl (1) were prepared according to the literature
procedures.14 NMR spectra were recorded on a Avance III − 300
Bruker or AV-500 Bruker Avance spectrometer. Chemical shifts are
referenced to TMS or to residual solvent resonance peaks (HDO peak
at 4.79 ppm in D2O).30 Mass spectra were recorded on Micromass
Platform LCZ 4000, using electrospray ionization (ESI) mode. GC-
MS was carried out on HP 6890 (flame ionization detector and
thermal conductivity detector) and HP 5973 (MS detector)
X-ray Structure Determinations of 2. Crystal data:
C28H41N1O1P2Ru1 orange, 0.03 × 0.12 × 0.14 mm3, orthorhombic,
P212121, a = 11.2347(18)Å, b = 12.603(3)Å, c = 18.665(4)Å from 20°
of data, T = 100(2)K, V = 2642.7(9) Å3, Z = 4, Fw = 570.63, Dc =
1.434 Mg·m−3, μ = 0.736 mm−1. Data collection and processing:
Bruker Appex2 KappaCCD diffractometer, MoKα (λ = 0.71073 Å),
graphite monochromator, 15569 reflections collected, −14 ≤ h ≤ 11,
−15 ≤ k ≤ 15, −16 ≤ l ≤ 23, frame scan width = 0.5°, scan speed 1°
per 240 s, typical peak mosaicity 0.69°, 5495 independent reflections
(R-int = 0.0297). The data were processed with Denzo-Scalepack.
Solution and refinement: structure solved by direct methods with
ShelexT. Full matrix least-squares refinement based on F2 with
SHELXL-97. 310 parameters with 0 restraints, final R1 = 0.0264 (based
on F2) for data with I > 2σ(I) and, R1 = 0.0317 on 5495 reflections,
goodness-of-fit on F2 = 1.048, largest electron density peak = 0.419
Å−3, deepest hole −0.426 Å−3. Position of hydrogen H1 was localized
as an electron density peak. Positional and isotropic temperature
H
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
parameters of H1 were varied during last 10 cycles of refinement
together with parameters of other atoms.
(10) Klobukowski, E. R.; Mueller, M. L.; Angelici, R. J.; Woo, L. K.
ACS Catal. 2011, 1, 703.
Computational Details. All geometries were optimized with the
BP86 functional and the def2-SV(P) basis set together with
corresponding core potential for ruthenium. The D3 dispersion
correction was used for the geometry optimizations. Thermodynamic
properties were obtained at the same level of theory from a frequency
calculation. All free energies are calculated under standard conditions
unless otherwise noted. Single point energies were obtained with the
dispersion corrected, spin scaled double hybrid functional DSD-
PBEB95-D3BJ together with a larger def2-TZVPP basis set. In order
to increase the efficiency of the calculations, the density fitting,
respectively, the RIJOSX approximation with the approximate basis
sets was used. To take the influence of the solvent into account, the
SMD model was applied for a 1:1 mixture of dioxane/water. The
geometry optimizations were performed with Gaussian 09, whereas
the double-hybrid single point calculations utilized the ORCA
program package.31,32
(11) Tanaka, K.; Yoshifuji, S.; Nitta, Y. Chem. Pharm. Bull. 1988, 36,
3125.
(12) Nishinaga, A.; Shimizu, T.; Matsuura, T. J. Chem. Soc., Chem.
Commun. 1979, 970.
(13) Khusnutdinova, J. R.; Ben-David, Y.; Milstein, D. J. Am. Chem.
Soc. 2014, 136, 2998.
(14) Gunanathan, C.; Milstein, D. Angew. Chem., Int. Ed. 2008, 47,
8661.
(15) Gunanathan, C.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
2009, 131, 3146.
(16) Khusnutdinova, J. R.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2013, 52, 6269.
(17) Balaraman, E.; Khaskin, E.; Leitus, G.; Milstein, D. Nat. Chem.
2013, 5, 122.
(18) Rodríguez-Lugo, R. E.; Trincado, M.; Vogt, M.; Tewes, F.;
Santiso-Quinones, G.; Grutzmacher, H. Nat. Chem. 2013, 5, 342.
̈
(19) Gunanathan, C.; Gnanaprakasam, B.; Iron, M. A.; Shimon, L. J.
W.; Milstein, D. J. Am. Chem. Soc. 2010, 132, 14763.
(20) Ye, X.; Plessow, P. N.; Brinks, M. K.; Schelwies, M.; Schaub, T.;
Rominger, F.; Paciello, R.; Limbach, M.; Hofmann, P. J. Am. Chem.
Soc. 2014, 136, 5923.
(21) The previously published X-ray structure of 2 (ref 17) was
obtained for single crystals grown under different crystallization
conditions: this structure includes benzene from crystallization and
exhibits disorder; no hydride could be located.
(22) Baker, J. D.; Heath, R. T.; Millar, J. G. J. Chem. Ecol. 1992, 18,
1595.
(23) Kesharwani, M. K.; Martin, J. M. L. Theor. Chem. Acc. 2014, 133,
1452.
(24) Since an optimization with analytical calculation of the Hessian
and a stepsize of 0.01 bohr did not converge, we estimated the barrier
for H2 elimination from an relaxed PES scan over the H−H distance
with a final stepsize of 0.01 Å.
(25) Li, J.; Shiota, Y.; Yoshizawa, K. J. Am. Chem. Soc. 2009, 131,
13584.
ASSOCIATED CONTENT
* Supporting Information
Experimental details of synthetic procedures, catalytic reactions,
and computational details. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
■
*david.milstein@weizmann.ac.il
Present Address
§Okinawa Institute of Science and Technology Graduate
University, 1919−1 Tancha, Onna-son, Kunigami-gun, Okina-
wa, Japan 904−0495
Author Contributions
‡These authors contributed equally.
Notes
(26) Iron, M. A.; Ben-Ari, E.; Cohen, R.; Milstein, D. Dalton Trans.
2009, 9433.
The authors declare no competing financial interest.
(27) Sandhya, K. S.; Suresh, C. H. Organometallics 2011, 30, 3888.
(28) Yang, X.; Hall, M. B. J. Am. Chem. Soc. 2010, 132, 120.
(29) Orlova, G.; Scheiner, S.; Kar, T. J. Phys. Chem. A 1999, 103, 514.
(30) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. J. Org. Chem. 1997,
62, 7512.
ACKNOWLEDGMENTS
■
This research was supported by the European Research Council
under the FP7 framework (ERC no. 246837), the Israel Science
Foundation and the Kimmel Center for Molecular Design.
J.R.K. thanks the Feinberg Graduated School for a Post
Doctoral Fellowship. U.G. thanks the DAAD for a Post
Doctoral Fellowship. D.M. holds the Israel Matz Professorial
Chair of Organic Chemistry.
(31) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
REFERENCES
■
(1) Rees, C. W.; Meth-Cohn, O.; Katritzki, A. R.; Moody, C. J.
Comprehensive Organic Functional Group Transformations, Vol. 5,
Synthesis: Carbon with Two Attached Heteroatoms with at Least One
Carbon-to-Heteroatom Multiple Link; Pergamon: Oxford, UK, 1995.
(2) Naota, T.; Murahashi, S. Synlett 1991, 693.
(3) Gunanathan, C.; Ben-David, Y.; Milstein, D. Science 2007, 317,
790.
(4) Moriarty, R. M.; Vaid, R. K.; Duncan, M. P.; Ochiai, M.; Inenaga,
M.; Nagao, Y. Tetrahedron Lett. 1988, 29, 6913.
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
revision D1; Gaussian, Inc: Wallingford, CT, 2009.
(32) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73.
(5) Wu, X.-F.; Sharif, M.; Pews-Davtyan, A.; Langer, P.; Ayub, K.;
Beller, M. Eur. J. Org. Chem. 2013, 2013, 2783.
(6) Wu, X.-F.; Sharif, M.; Feng, J.-B.; Neumann, H.; Pews-Davtyan,
A.; Langer, P.; Beller, M. Green Chem. 2013, 15, 1956.
(7) Wu, X.-F.; Bheeter, C. B.; Neumann, H.; Dixneuf, P. H.; Beller,
M. Chem. Commun. 2012, 48, 12237.
(8) Wang, Y.; Kobayashi, H.; Yamaguchi, K.; Mizuno, N. Chem.
Commun. 2012, 48, 2642.
(9) Xu, W.; Jiang, Y.; Fu, H. Synlett 2012, 23, 801.
I
DOI: 10.1021/jacs.5b01750
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX