ESI-MS Insights into Acceptorless Dehydrogenative Coupling of Alcohols

Article abstract of DOI:10.1021/acscatal.6b00623

Acceptorless dehydrogenative coupling (ADC) reactions catalyzed by a series of Ru and Os complexes were studied by ESI-MS. Important ethoxo, 1-ethoxyethanolate, and hydride intermediates were intercepted in the ADC of ethanol to ethyl acetate. Collision-induced dissociation (CID) experiments were applied as a structure elucidation tool and as a probe of the propensity of the reaction intermediates to evolve acetaldehyde, ethyl acetate, and H₂, relevant to the catalytic cycle. The key mechanistic step producing ethyl acetate from the 1-ethoxyethanolate intermediates was documented. Energy-dependent CID experiments demonstrated the importance of a vacant coordination site for efficient production of ethyl acetate. The versatility and potential broad applicability of ESI-MS and its tandem version with CID was further illustrated for the ADC reaction of alcohols with amines, affording amides. A mechanism related to that found for the ester synthesis is plausible, with the key step involving formation of a hemiaminaloxide intermediate.

Full text of DOI:10.1021/acscatal.6b00623

Research Article
pubs.acs.org/acscatalysis
ESI-MS Insights into Acceptorless Dehydrogenative Coupling of
Alcohols
Cristian Vicent*,† and Dmitry G. Gusev*,‡
†Serveis Centrals d’Instrumentacio
́
Cientıfica, Universitat Jaume I, 12071 Castellon
́
, Spain
́
‡Department of Chemistry and Biochemistry, Wilfrid Laurier University, Waterloo, Ontario N2L 3C5, Canada
S
* Supporting Information
ABSTRACT: Acceptorless dehydrogenative coupling (ADC)
reactions catalyzed by a series of Ru and Os complexes were
studied by ESI-MS. Important ethoxo, 1-ethoxyethanolate, and
hydride intermediates were intercepted in the ADC of ethanol to
ethyl acetate. Collision-induced dissociation (CID) experiments
were applied as a structure elucidation tool and as a probe of the
propensity of the reaction intermediates to evolve acetaldehyde,
ethyl acetate, and H2, relevant to the catalytic cycle. The key
mechanistic step producing ethyl acetate from the 1-
ethoxyethanolate intermediates was documented. Energy-
dependent CID experiments demonstrated the importance of a
vacant coordination site for efficient production of ethyl acetate. The versatility and potential broad applicability of ESI-MS and
its tandem version with CID was further illustrated for the ADC reaction of alcohols with amines, affording amides. A mechanism
related to that found for the ester synthesis is plausible, with the key step involving formation of a hemiaminaloxide intermediate.
KEYWORDS: catalytic hydrogenation, pincer complexes, acceptorless dehydrogenation, electrospray ionization,
collision induced dissociation
dehydrogenated by the catalyst to yield the final ester or amide
products, respectively.1a
INTRODUCTION
■
Acceptorless dehydrogenative coupling (ADC) of alcohols
according to Scheme 1a,b is an atom-economical and
environmentally benign method for the production of esters
and amides.1 The reverse hydrogenation (HY) process is also
feasible, since the HY/ADC reactions are mechanistically
related, and ADC catalysts have demonstrated useful activity in
ester and amide hydrogenation.2 Milstein’s group developed
the first efficient homogeneous catalysts for the syntheses of
esters from primary alcohols3 and amides from alcohols and
amines.4 Subsequently, a series of Ru and Os catalysts have
been prepared demonstrating improved reaction efficiency and
selectivity.2b,5,6
The continuing growth in HY/ADC catalyst development is
in contrast with a relatively small number of examples dedicated
to the elucidation of the reaction mechanism. Some insights
have been derived from (i) the presence or, more commonly,
absence of NMR-observable intermediates during the catalytic
reactions, (ii) stoichiometric reactions under relevant con-
ditions, and (iii) DFT calculations.7 The intermediacy of alkoxo
and dihydride species was proposed for the Milstein catalyst, in
a sequence defined by alcohol addition, followed by β-hydrogen
elimination with concomitant formation of the corresponding
aldehyde (Scheme 1c). Ultimately, the dihydride eliminates H2,
thereby regenerating the catalyst. Independently, the aldehyde
reacts with the alcohol or amine substrate to afford the
corresponding hemiacetal or hemiaminal intermediates that are
The detection of intermediates is important in a mechanistic
study; however, this task is anticipated to be anything but trivial
because of the short-lived transient nature of the intermediates
under catalytic ADC conditions (typically at T ≥ 80 °C).
Indeed, monitoring metal species formed under the operating
catalytic conditions of reactions in Scheme 1a,b remains
virtually unexplored. In this respect, electrospray ionization
mass spectrometry (ESI-MS) has become a valuable tool for
studying fast transformations8 by providing snapshots of the
dynamic ionic composition of the reaction solutions. Collision-
induced dissociation (CID) experiments further contribute to
the utility of mass spectrometry as a mechanistic tool in
organometallic chemistry.9 Convenient sample introduction
techniques allow chemical reactions to be monitored in situ
even at high operating temperatures, with minimal sample
manipulation. In this context, McIndoe’s group recently
described the pressurized sample infusion technique coupled
directly to the mass spectrometer that proved to be particularly
useful.10
Herein, we applied ESI-MS for detection of ADC reaction
intermediates under catalytic conditions. Pincer-type complexes
1−5 bearing PyNP,11 SNS,12 PNP,13 and PyNNP14 ligands
Received: March 1, 2016
Revised: April 5, 2016
© XXXX American Chemical Society
3301
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
Scheme 1. (a, b) Dehydrogenative Coupling of Alcohols To Yield Esters and Amides, (c) Schematic Representation of the
Intermediates Proposed for the Milstein Catalyst, and (d) Complexes 1−5 Studied in This Work
(Scheme 1d) are efficient for hydrogenation of a range of
substrates (esters, ketones, imines) as well as for the
acceptorless dehydrogenative coupling of alcohols. Recently,
we studied the mechanism of the ADC reactions of methanol
and ethanol catalyzed by 5 using a combination of ESI-MS
experiments with DFT calculations.14 The versatility of the ESI-
MS technique is further underscored in the present
investigation of a broader spectrum of catalysts with different
efficiencies in ADC of alcohols aimed at establishing whether a
common mechanism might be operative in all cases. Moreover,
an investigation of the related dehydrogenative reaction, the
ADC of ethanol and primary amines to yield acetamides, is also
presented.
with ethanol (or ethanol doped with NaBF4) using a T-shaped
connection directly coupled to an ESI source (see Figure S1 in
the Supporting Information).10d We also attempted monitoring
reaction intermediates by taking aliquots at different time
intervals followed by an off-line dilution with EtOH, but this
approach did not allow detecting most of the intermediates
involved; instead, oxidized species due to the formal uptake of
oxygen were dominant. Establishing the composition of all
metal-containing species was facilitated by the characteristic
isotopic pattern at natural abundance of Ru and Os, and it was
carried out by comparison of the experimental and theoretical
isotope patterns using the MassLynx program 4.1. For the CID
experiments, the cations of interest were mass-selected using
the first quadrupole (Q1) and interacted with argon in the
collision cell at variable collision energies (Elaboratory = 3−20
eV). The ionic products of fragmentation were analyzed with a
time-of-flight analyzer (for the QTOF instrument) or the
second quadrupole (for the Quattro LC QqQ instrument). The
isolation width was 1 Da, and the most abundant isotopomer
was mass-selected in the first quadrupole analyzer.
EXPERIMENTAL SECTION
■
Electrospray Ionization Mass Spectrometry (ESI-MS)
and Collision-Induced Dissociation (CID). ESI-MS studies
were conducted on a QTOF Premier or a Quattro LC (QqQ)
instrument, equipped with an orthogonal Z-spray-electrospray
interface (Waters, Manchester, UK). A capillary voltage of 3.5
kV was used in the positive ESI(+) scan mode, and the cone
voltage was adjusted to a low value (typically Uc = 5−15 V) to
control the extent of fragmentation in the source region. The
Q-TOF instrument was operated in the W mode at a resolution
of ca. 15000 (fwhm). Nitrogen as the drying and cone gas was
set to flow rates of 300 and 30 L/h, respectively. For the
Quattro LC instrument, the desolvation and nebulization gas
was nitrogen set to flow rates of 400 and 90 L/h, respectively.
As mentioned above, both ester hydrogenation and ADC of
alcohols are expected to follow the same mechanism; however,
online monitoring of the ADC reaction by ESI-MS is
operationally simpler, as it avoids the use of high H2 pressure.
Catalytic runs were monitored for 3 h using 0.01 or 0.05% of
catalyst and 1% of NaOEt with respect to ethanol. Sample
solutions were delivered to the spectrometer at different time
intervals using the pressurized sample infusion (PSI) technique.
Attempts to use continuous delivery of the reaction mixture,
which contains NaOEt (the base in all ADC reactions), caused
progressive reduction of the ion abundances for all species in
the ESI mass spectra, accompanied by deposition of nonvolatile
colorless solids in the ESI chamber. This technique has proved
to be ideal for in situ analysis of complex mixtures formed
during catalysis.10a−c Typically, a N2 positive pressure (1−3 psi)
was used to transfer sample solutions that were further diluted
RESULTS AND DISCUSSION
■
Single-Stage ESI-MS. Studying a chemical system by ESI-
MS requires that the species must be charged. Two aspects
make the catalytic chemistry of Scheme 1 particularly suitable
for an ESI-MS investigation. On the one hand, using the
catalysts with excess base (NaOEt) produced a useful
enhancement of the ion abundances of the [M + Na]+ sodium
adducts.15 On the other hand, since ethanol served as both the
substrate and the solvent, the ion suppression effects due to the
excess of substrate16 were not encountered in the present study.
Thus, high catalyst loadings were not necessary and realistic
catalytic conditions could be used for the ESI-MS reaction
monitoring.
For complexes 1−5, observed ionizations typically involved
chloride or hydride (Cl− or H−) loss. In ethanol, the most
favorable ionization process was the M−Cl bond breaking that
produced [1 − Cl]+ and [2 − Cl]+ cations as the base peaks for
1 and 2, respectively. For 3, the [3 − H]+ cation was observed
as the dominant species, whereas for 4 and 5, the [4 − Cl]+ and
[5 − Cl]+ cations accompanied by [4 − H]+ and [5 − H]+ were
detected, the last two albeit with much lower ion abundances.
The alkali ion adduct formation could be promoted by addition
of NaBF4.10d,17 The observations of the original compounds as
3302
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
sodium ion adducts is anticipated to be crucial to
unambiguously determine the intact coordination environment
at the metal site during catalysis.
Acceptorless Dehydrogenative Coupling (ADC) of
Ethanol To Yield Ethyl Acetate. In this section, we discuss
ESI-MS detection of intermediates relevant to the ADC of
ethanol while using the pressurized sample infusion (PSI)
technique.10 In the reactions investigated herein, NaOEt base in
the catalytic system produces significant ion abundances of Na+
adducts. Chemical speciation of the reaction of Scheme 2 was
In addition to peaks due to 1 (peaks at m/z 719.2, 777.2, and
793.1 assigned to [1 − Cl]+, [1 + Na]+ and [1·O + Na]+,
respectively), the ethoxo intermediate 1b was observed during
the first 30 min as [1b − H]+ at m/z 729.2, whereas 1c
(manifested as [1c − H]+ and [1c + Na]+ at m/z 773.2 and
797.1, respectively) was dominant after 1 h from the start of the
reaction (see Figure 1). Species 1d (and in general the related
dihydrides formed from complexes 2−4) was not detected as
the [1d + Na]+ adduct because of its low abundance.18 The low
abundance of 1d−4d suggests that they are most likely not
involved in a rate-limiting step along the catalytic cycle. The
presence of mixed Cl/H and Cl/1-ethoxyethanolate (C4O2H9)
ligand complexes, namely RuCl(H)(PPh3)(PyNHP) (1e) and
RuCl(C4O2H9)(PPh3)(PyNHP) (1f) (see Table S1 in the
Supporting Information for details), was also evident in the
early stages of the ADC reaction. For example, species 1e,f
(manifested as [1e + Na]+ and [1f + Na]+ at m/z 743.2 and
831.2, respectively) were evident in the ESI mass spectrum after
1 h. The observation of both groups of intermediates 1b,c and
1e,f highlights the versatility of the ESI-MS technique in
unraveling complex transformations. Formation of side
products such as acetate species or oxidized species (formally
corresponding to mass increases of 16 and hereafter formulated
as M·O species) was also observed. For instance, the acetate-
containing intermediate 1g was observed as [1g − Cl]+, [1g +
Na]+, and [1g·O + Na]+ cations at m/z 743.2, 801.1, and 817.0,
respectively, upon ESI-MS. CID mass spectra confirmed the
presence of the acetates as judged by the release of acetic acid
(Δm = 60) as a fragmentation channel. We believe that this
high reactivity of catalysts 1 and 2, in terms of the large number
of products identified from them, correlates with their high
ADC activity (see further discussion below for 4). The putative
16-electron amido intermediates 1a and 2a were not observed
by ESI-MS under the catalytic conditions. This is rationalized
by recognizing their intrinsic reactivity toward ethanol leading
to 1b and 2b,11b,12 respectively, and by taking into the
consideration the large excess of ethanol with respect to
ruthenium under the catalytic conditions.
Scheme 2. ADC of Ethanol
practically the same on catalysis by 1 or 2, in terms of both the
nature of the detected Ru species and their temporal evolution.
Formation of intermediates 1b,c (see Scheme 3 for the
Scheme 3. Reaction Intermediates in ADC of Ethanol
Catalyzed by 1
proposed structures) was evident from the ESI-MS analysis on
the basis of the m/z values and isotopic patterns, as illustrated
in Figure 1 for the catalytic reaction using 1 monitored after 1
h. Table S1 in the Supporting Information collects the m/z
values for the identified species. The analogous intermediates
2b,c, structurally related to 1b,c by replacement of the PyNHP
by the SNHS ligand, were detected by ESI-MS in the ADC
reaction catalyzed by 2.
Catalyst 1 and intermediate 1c were both detected as intact
Na+ ion adducts under the catalytic conditions. In general, the
ion abundances of Na+ adducts were strongly affected by the
intrinsic chemical composition of each intermediate. For
example, chlorine-containing species were clearly observed as
Na+ adducts due to the well-known stabilizing effect of the Ru−
Cl··Na interaction in the gas phase.10d 1-Ethoxyethanolate
intermediate 1c could be readily identified as a Na+ ion adduct,
whereas dilution with ethanol in the presence of traces of
NaBF4 was required to visualize the intact Na+ ion adduct of
the ethoxo intermediate 1b.
When dimer 3 was used as the catalyst, the initial intensity of
the ESI-MS detected dimeric species [3 − H]+ at m/z 941.2
was significantly reduced relative to a new set of mononuclear
complexes. This is clearly illustrated in Figure 2, showing the
ESI mass spectrum of the ADC reaction of ethanol catalyzed by
3. In this case, intermediates 3b,c (expected to be structurally
analogues to 1b,c) could be identified by ESI-MS as [3b − H]+
and [3c − H]+ at m/z 517.1 and 561.2, respectively. Other side
products featuring acetate and oxidized species were also
evidenced (see Figure S3 in the Supporting Information).
Dimeric species were invariably observed during the catalytic
reaction, thus suggesting the role of the dinuclear complexes as
Figure 1. ESI mass spectrum recorded via the PSI technique after 1 h, using a triple-quadrupole analyzer. Relevant ethoxo and 1-ethoxyethanolate
intermediates are [1b − H]+ at m/z 729.2 and [1c − H]+ and [1c + Na]+ at m/z 773.2 and 797.1, respectively. Peak assignments of all observed
species are described in the text.
3303
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
Figure 2. ESI mass spectra of ethanol (2 mL) in the presence of 0.05 mol % of 3 (top) and 0.05 mol % of 3 and 1 mol % of EtONa after heating 30
min under reflux (bottom), using the Q-TOF instrument. The base peak corresponds in both cases to the dimer [3 − H]+ cation (m/z 941.2).
Intermediates 3b,c were identified as [3b − H]+ and [3c − H]+ at m/z 517.1 and 561.2, respectively.
Scheme 4. Fragmentations Observed upon CID of Mass-Selected [1b + Na]+ and [1b − H]+ Cations Derived from the ESI-MS
Analysis of 1b
a continuous reservoir of the mononuclear active catalyst. The
low intensity of the mononuclear species hindered the
identification of the intrinsically low abundance Na+ ion
adducts in the present case.
formed from 4 was clearly evident and might correlate with the
reduced catalytic activity of this complex.19
Collision-induced dissociation CID experiments constitute a
powerful technique for probing the structure and reactivity of
reaction intermediates.9 For example, CID experiments on the
sodium ion adducts of 1b,c aided in their structural assignment
since liberation of the neutrals, NaOEt and NaC4O2H9,
respectively, strongly supports the structures of Scheme 3.
Moreover, the CID experiments revealed the intrinsic
propensity of the ethoxo, 1-ethoxyethanolate, and dihydride
intermediates to release neutral species relevant to the catalytic
cycle of the ADC of ethanol. A summary of the fragmentation
pathways of mass-selected [1b + Na]+ and [1b − H]+ cations is
presented in Scheme 4. CID of [1b + Na]+ resulted in
liberation of NaOEt as the dominant fragmentation pathway,
whereas [1b − H]+ eliminated EtOH as the main dissociation
process. Release of acetaldehyde was also observed from both
[1b − H]+ and [1b + Na]+, but with a lower abundance. These
Unlike complexes 1−3, which proved to be efficient in the
ADC reaction of ethanol to ethyl acetate,11,12 complex 4 gave
lower conversions.11b We decided to explore the latter to
ascertain the nature of species formed from 4 in the catalytic
solution. The chemical speciation of the ADC reaction
catalyzed by 4 was virtually identical with that observed for 1
and 3 in terms of the identity of the metal-containing species;
however, formation of the intermediates was significantly
retarded (see Figure S4 in the Supporting Information for
details) and the ESI mass spectra were less crowded. In
particular, the acetate and oxidized species that were observed
for compounds 1−3 and 5 were detected only as minor species
with 4. The enhanced stability of the reaction intermediates
3304
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
Figure 3. CID mass spectra of (a) [1b − H]+ (m/z 729.2), (b) [2b − H]+ (m/z 602.2), (c) [3b − H]+ (m/z 517.1), and (d) [4b − H]+ (480.2),
recorded at CEElab = 10 eV.
Scheme 5. Proposed Reactions Linking the Intermediates 1a−c
Figure 4. CID mass spectra of the (a) [1c − H]+ (m/z 773.2) and (b) [2c − H]+ (m/z 646.2) cations recorded at CEElab = 10 eV.
results suggest that the ethoxides are more prone to eliminate
ethanol rather than the product of ethanol dehydrogenation,
acetaldehyde. Identical trends were observed upon CID of [2b
− 4b − H]+ and [2b − 4b + Na]+ cations (see Figure 3 for
illustrative CID mass spectra).
PNN complexes.7a,f The absence of NMR-observable free
aldehyde during the catalytic reactions of 111,12 and the
dominant nondehydrogenative fragmentation pathway ob-
served in the CID experiments suggest that 1b′ may rapidly
re-form the alkoxide 1b or, alternatively, it is rapidly attacked by
the solvent, EtOH. The latter reaction possibly leads to the
unstable dihydrogen intermediate 1c′, which was not detected
by ESI-MS, ultimately leading to the experimentally detected
intermediate 1c. The reversibility of coordinated aldehyde and
alkoxo ligands in the periphery of the metal has been
emphasized theoretically and proved to be crucial to trigger
hemiacetal formation to ultimately yield the ester.7b
The intermediacy of hemiacetals in the ADC of alcohols was
envisioned by Murahashi20 and Shvo21 and later invoked by
others.3,22 Bergens demonstrated formation of a hemiacetal-
oxide complex in the reaction of RuH2((R)-BINAP)((R,R)-
dpen) with γ-butyrolactone,23 while Milstein succeeded in
selectively forming acetals24 as the reaction products during the
catalytic transformation of primary alcohols with the liberation
The dominant observation of neutral losses, EtOH and
EtONa, in the CID spectra of [1b − H]+ and [1b + Na]+,
respectively, agrees with the equilibrium between 1b and the
16-electron amido complex 1a via ethanol elimination as
depicted in Scheme 5. The release of acetaldehyde from 1b in
the gas phase, even if it is barely detectable, is indicative of
formation of the acetaldehyde intermediate 1b′ presumably
(but not necessarily)7b via β-hydrogen elimination in a 16-
electron intermediate with an uncoordinated Py group (see
Scheme 5).3 Let us note that an alternative mechanism might
be responsible for the acetaldehyde formation; e.g., recent
computational work indicated that the inner-sphere mechanism
is energetically less favorable than the outer-sphere mechanism,
via bifunctional hydrogen transfer, for a series of ruthenium
3305
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
Scheme 6. Fragmentation Pathways Observed upon CID Conditions of Mass-Selected [1c + Na]+ and [1c − H]+ Cations
Derived from ESI-MS of 1c
Scheme 7. Proposed Sequence of Reactions of 1c To Yield 1d and Ethyl Acetate, Followed by Regeneration of 1a
of H2, using an acridine-based ruthenium pincer complex.
Except for our recent report,14 formation of hemiacetals in the
catalytic ADC reactions of alcohols has not been observed,
neither as free species nor as hemiacetaloxide ligands
coordinated to a metal center.
We performed CID experiments to ascertain the intrinsic
reactivity of the 1-ethoxyethanolate intermediate 1c. The CID
spectra are exemplified in Figure 4 for [1c − H]+ and [2c −
H]+ cations. Gas-phase fragmentation reactions of the ESI-MS
detected species [1c − H]+ and [1c + Na]+ produced ethyl
acetate as a fragmentation channel. Additionally, the CID mass
spectra of [1c + Na]+ displayed liberation of sodium
ethoxyethanolate.
The fragmentation reactions depicted in Scheme 6 are also
characteristic of the related 1-ethoxyethanolate intermediates
2c−4c in the catalytic solutions of 2−4. It is worth emphasizing
that the CID spectra of [1c − H]+, [2c − H]+, [3c − H]+, and
[4c − H]+ all displayed exclusive liberation of ethyl acetate
followed by H2, thus indicating the central role of the metal
hemiacetaloxide intermediates in the ADC reaction. The
product ester formation presumably occurs via β-hydrogen
elimination from a 16-electron hemiacetaloxide intermediate,
for example, 1d′ in Scheme 7 where the hemilability of the
pincer ligand plays a major role, to yield the dihydride 1d that
subsequently eliminates H2 to produce 1a (see Scheme 7). In
this respect, a relevant observation was made during the energy-
resolved CID experiments of [1c−4c − H]+ cations, which
produced ethyl acetate at significantly lower collision energies
than the corresponding [1c−4c + Na]+ species. This difference
can be attributed to the presence of a vacant coordination site
in the 16-electron [1c−4c − H]+ cations. We should note that
the vacant coordination site in [1c−4c − H]+ would not
necessarily be trans to the hemiacetaloxide as shown in Scheme
6 but may be located cis to the hemiacetaloxide, considering the
facile mer/fac isomerization well-documented for the SNS and
PNN complexes.11,12 A vacant site cis to the hemiacetaloxide
would facilitate β-hydrogen elimination leading to ethyl acetate.
As mentioned above, intermediates 1d−4d were not observed
experimentally because of their low stability under catalytic
conditions; however, they could be evidenced in the gas phase
as the product ions from ethyl acetate liberation of 1c−4c that
subsequently liberate H2 (see Figure 4). This latter elementary
step corresponds to 1d to 1a transformation and catalyst
regeneration, as illustrated in Scheme 7.
Acceptorless Dehydrogenative Coupling (ADC) of
Ethanol and Amines To Yield Acetamides. Recently, we
reported that OsHCl(CO)[PyNNP] (5; Scheme 1) is efficient
for dehydrogenative coupling of alcohols and amines without
solvent and under relatively mild conditions.14 For example,
heating ethanol with benzylamine (1/1) in the presence of 0.05
mol % of 5 afforded N-benzylacetamide (see Scheme 8) in 96%
yield after 17 h. No initiation time was manifested, and a
constant H2 bubbling was evidenced as the reflux temperature
was reached.14 Using the pressurized sample infusion technique
coupled with ESI-MS, we undertook a study of the reaction
intermediates relevant to this process during the first 3 h. We
succeeded at intercepting a hemiaminaloxide species, 5c
(Figure 5), identified on the basis of the m/z value, isotopic
pattern, and CID experiments. This intermediate was observed
by ESI-MS, and its relative abundance remained practically
3306
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
Scheme 8. Formation of N-Benzylacetamide
CONCLUSIONS
■
Whereas ESI-MS25 and desorption electrospray ionization mass
spectrometry (DESI-MS)26 have been used for studies of
ruthenium-catalyzed hydrogenations, these techniques have
received little attention in mechanistic studies of catalytic
alcohol dehydrogenation reactions.14,27 In this paper, the ESI-
MS and CID experiments were used as structure elucidation
tools, to unravel the intrinsic reactivity of reaction intermediates
in acceptorless dehydrogenative coupling (ADC) reactions of
ethanol catalyzed by a series of pincer-type Ru and Os
complexes. A series of intermediates comprising metal ethoxo,
1-ethoxyethanolate, and hydride functional groups were
intercepted and characterized during the ADC of ethanol to
yield ethyl acetate; their compositions were ascertained via the
detection of the intact alkali ion adducts [M + Na]+. A
functional group−reactivity relationship could be established
where the ethoxides were shown to be involved in the
reversible liberation of ethanol and/or formation of acetalde-
hyde, the metal dihydrides exhibited a strong tendency toward
H2 elimination, and the key 1-ethoxyethanolate intermediates
preferentially liberated ethyl acetatethe ultimate product of
the ADC reaction. A mechanistically related ESI-MS study of
the ADC of alcohols and amines allowed us to intercept and
characterize the metal-containing hemiaminaloxide reaction
intermediates.
unchanged for 3 h. When ethanol was reacted with n-
butylamine, the corresponding hemiaminaloxide intermediate
could also be observed by ESI-MS.
Under the CID conditions, the hemiaminaloxide intermedi-
ates exhibited liberation of the corresponding acetamides as
fragmentation pathways. Figure 5b illustrates the gas-phase
reactivity of the mass-selected [5c − H]+ cation that comprises
amide formation accompanied by H2 loss. A minor
fragmentation channel that resulted in elimination of the free
amine was also evident. We further notice that no dihydride
intermediate 5d (analogues to 1d−4d) could be detected by
ESI-MS; this is in agreement with the known instability of 5d
with respect to H2 loss in solution.14 The experimental
evidence of Figure 5 suggests that hemiaminaloxide complexes
such as 5c are viable intermediates en route to amides in a way
mechanistically analogous to that proposed for the hemi-
acetaloxide intermediates leading to esters in Scheme 7. The
temporal evolution of the reaction intermediates and the CID
experiments certainly suggest that the ADC of alcohols and
amines might follow closely related mechanisms. We
hypothesize that the diverging step is the addition of the
alcohol vs amine to the coordinated aldehyde intermediate, as
illustrated in Scheme 5, to form a hemiacetaloxide vs
hemiaminaloxide intermediate that subsequently evolves to
the ester or amide product, respectively.
It is well-recognized that catalytic hydrogenation of carbonyl
compounds and the reverse alcohol dehydrogenation can
proceed in an inner- and/or outer-sphere fashion. The latter
process is unique among the organometallic mechanisms by
avoiding substrate binding to the metal center of the catalyst;
therefore, the concerted process of H2 transfer from (or to) the
metal takes place via fleeting intermediates where the substrate
is only loosely connected with the catalyst. While the power of
mass spectrometry in detecting short-lived reaction intermedi-
ates is impressive, it is nevertheless unlikely that the catalytic
outer-sphere H2 transfer reactions can be followed and
documented by MS techniques. It is conceivable that
Figure 5. (a) Pressurized sample infusion (PSI) ESI mass spectrum of a sample from the catalytic reaction of Scheme 8. (b) CID mass spectrum of
the mass-selected hemiaminaloxide [5c − H]+.
3307
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
(5) (a) Balaraman, E.; Milstein, D. Hydrogenation of polar bonds
catalysed by ruthenium-pincer complexes. Top. Organomet. Chem.
2014, 48, 19−44. (b) Chelucci, G.; Baldino, S.; Baratta, W. Acc. Chem.
Res. 2015, 48, 363−379. (c) Chelucci, G.; Baldino, S.; Baratta, W.
Coord. Chem. Rev. 2015, 300, 29−85. (d) Dub, P. A.; Ikariya, T. ACS
Catal. 2012, 2, 1718−1741.
(6) (a) Younus, H. A.; Ahmad, N.; Su, W.; Verpoort, F. Coord. Chem.
Rev. 2014, 276, 112−152. (b) Younus, H. A.; Su, W.; Ahmad, N.;
Chen, S.; Verpoort, F. Adv. Synth. Catal. 2015, 357, 283−330.
(c) John, J. M.; Bergens, S. H. Angew. Chem., Int. Ed. 2011, 50, 10377−
10380. (d) John, J. M.; Loorthuraja, R.; Antoniuk, E.; Bergens, S. H.
Catal. Sci. Technol. 2015, 5, 1181−1186. (e) Friedrich, A.; Schneider,
S. ChemCatChem 2009, 1, 72−73. (f) Schneider, S.; Meiners, J.;
Askevold, B. Eur. J. Inorg. Chem. 2012, 2012, 412−429. (g) Musa, S.;
Ackermann, L.; Gelman, D. Adv. Synth. Catal. 2013, 355, 3077−3080.
(h) Musa, S.; Fronton, S.; Vaccaro, L.; Gelman, D. Organometallics
2013, 32, 3069−3073.
(7) (a) Li, H.; Wang, X.; Huang, F.; Lu, G.; Jiang, J.; Wang, Z.-X.
Organometallics 2011, 30, 5233−5247. (b) Hasanayn, F.; Baroudi, A.
Organometallics 2013, 32, 2493−2496. (c) Montag, M.; Zhang, J.;
Milstein, D. J. Am. Chem. Soc. 2012, 134, 10325−10328. (d) Chen, X.;
Jing, Y.; Yang, X. Chem. - Eur. J. 2016, 22, 1950−1957. (e) Jiao, H.;
Junge, K.; Alberico, E.; Beller, M. J. Comput. Chem. 2016, 37, 168−176.
(f) Qu, S.; Dang, Y.; Song, C.; Wen, M.; Huang, K.-W.; Wang, Z.-X. J.
Am. Chem. Soc. 2014, 136, 4974−4991.
complexes 1−5 studied here can act as both inner- and outer-
sphere catalysts, and the two mechanisms are most likely in
competition in the hydrogenation/dehydrogenation reactions.
Additionally, solvation can play a crucial role by triggering some
of the elementary steps occurring in solution.7 It is the extreme
power of mass spectrometry in detecting gaseous species
present in vanishingly low concentrations that, ironically, makes
it challenging to claim with enough certainty that the
intermediates documented in this work represent the principal
reaction pathway and thus “the mechanism” of the ADC
reactions of ethanol.
The inner-sphere ADC mechanism is distinctly different
from the outer-sphere process by involving a coordinated
aldehyde intermediate, formed via the classical β-hydrogen
elimination. The latter requires an empty coordination site and
a certain degree of lability/hemilability of the catalyst. In this
regard, a comparative study of the catalysts of varying degrees
of hemilability, also including stable complexes with all strongly
bonded ligands, should provide important mechanistic insights.
A fitting example from our own work is the comparison of
RuHCl(CO)[HN(CH2CH2PPh2)2] and RuHCl(CO)-
[PyCH2NHCH2CH2PPh2]. These complexes gave 4700 and
4200 turnovers, respectively, in the ADC of refluxing ethanol to
ethyl acetate in 24 h under equivalent reaction conditions.11b If
the former catalyst retains an intact Ru[κ3-PNP] fragment at 78
°C, this would be an indication of the outer-sphere mechanism
being the principal pathway. There is no doubt that future
studies would furnish a definitive answer.
(8) (a) Coelho, F.; Eberlin, M. N. Angew. Chem., Int. Ed. 2011, 50,
5261. (b) Santos, L. S. Eur. J. Org. Chem. 2008, 2008, 235−253.
(c) Chen, P. Angew. Chem., Int. Ed. 2003, 42, 2832−2847. (d) Eberlin,
M. N. Eur. Mass Spectrom. 2007, 13, 19. (e) Schroder, D. Acc. Chem.
̈
Res. 2012, 45, 1521−1532. (f) Iacobucci, C.; Reale, S.; De Angelis, F.
Angew. Chem., Int. Ed. 2016, 55, 2980−2993.
(9) (a) Bohme, D. K.; Schwarz, H. Angew. Chem., Int. Ed. 2005, 44,
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.6b00623.
̈
■
2336−2354. (b) O’Hair, R. A. J. Chem. Commun. 2006, 1469−1481.
(c) Vikse, K. L.; McIndoe, J. S. Pure Appl. Chem. 2015, 87, 361−377.
(10) (a) Vikse, K. L.; Woods, M. P.; McIndoe, J. S. Organometallics
2010, 29, 6615−6618. (b) Vikse, K. L.; Ahmadi, Z.; Luo, J.; Van Der
Wal, N.; Daze, K.; Taylor, N.; McIndoe, J. S. Int. J. Mass Spectrom.
2012, 323−324, 8−13. (c) Yunker, L. P. E.; Stoddard, R. L.; McIndoe,
J. S. J. Mass Spectrom. 2014, 49, 1−8. (d) Wang, H.-Y.; Yim, W.-L.;
S
Schematic picture of the PSI ESI-MS experimental setup,
additional ESI-MS data, and peak assignments for various
samples (PDF)
Kluner, T.; Metzger, J. O. Chem. - Eur. J. 2009, 15, 10948−10959.
̈
(11) (a) Spasyuk, D.; Smith, S.; Gusev, D. G. Angew. Chem., Int. Ed.
2012, 51, 2772−2775. (b) Spasyuk, D.; Gusev, D. G. Organometallics
2012, 31, 5239−5242.
AUTHOR INFORMATION
Corresponding Authors
*E-mail for C.V.: barrera@uji.es.
*E-mail for D.G.G.: dgoussev@wlu.ca.
■
(12) Spasyuk, D.; Smith, S.; Gusev, D. G. Angew. Chem., Int. Ed.
2013, 52, 2538−2542.
(13) Bertoli, M.; Choualeb, A.; Lough, A. J.; Moore, B.; Spasyuk, D.;
Gusev, D. G. Organometallics 2011, 30, 3479−3482.
(14) Spasyuk, D.; Vicent, C.; Gusev, D. G. J. Am. Chem. Soc. 2015,
137, 3743−3746.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(15) Zheng, Q.; Liu, Y.; Chen, Q.; Hu, M.; Helmy, R.; Sherer, E. C.;
Welch, C. J.; Chen, H. J. Am. Chem. Soc. 2015, 137, 14035−14038.
We are grateful to the NSERC Canada, the Ontario
Government, and Wilfrid Laurier University for financial
support and the SCIC of the UJI for providing mass
spectrometry facilities.
(16) (a) Sorribes, I.; Wienhofer, G.; Vicent, C.; Junge, K.; Llusar, R.;
̈
Beller, M. Angew. Chem., Int. Ed. 2012, 51, 7794−7798. (b) Adrio, L.
A.; Gimeno, J.; Vicent, C. Chem. Commun. 2013, 49, 8320−8322.
(17) Wang, H.-Y.; Yim, W.-L.; Guo, Y.-L.; Metzger, J. O.
Organometallics 2012, 31, 1627−1634.
REFERENCES
■
(18) While the peak at m/z 685.2 may certainly come from
intermediate 1d via hydride loss, namely [1d − H]+, this species is not
diagnostic because 1b,c intermediates should also produce this species
via X release (X = ethoxo, 1-ethoxyethanolate, respectively). In
addition, we cannot definitely rule out that the intermediates 1d−4d
are not seen due to their inefficient Na+ adduct formation.
(19) Wassenaar, J.; Jansen, E.; Van Zeist, W. J.; Bickelhaupt, F. M.;
Siegler, M. A.; Spek, A. L.; Reek, J. N. H. Nat. Chem. 2010, 2, 417−
421.
(1) (a) Gunanathan, C.; Milstein, D. Science 2013, 341 (6143),
1229712. (b) Werkmeister, S.; Junge, K.; Beller, M. Org. Process Res.
Dev. 2014, 18, 289−302. (c) Pritchard, J.; Filonenko, G. A.; van
Putten, R.; Hensen, E. J. M.; Pidko, E. A. Chem. Soc. Rev. 2015, 44,
3808−3833.
(2) (a) Gunanathan, C.; Milstein, D. Acc. Chem. Res. 2011, 44, 588−
602. (b) Gunanathan, C.; Milstein, D. Chem. Rev. 2014, 114, 12024−
12087.
(3) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. J. Am. Chem. Soc.
(20) Murahashi, S.-I.; Naota, T.; Ito, K.; Maeda, Y.; Taki, H. J. Org.
Chem. 1987, 52, 4319−4327.
(21) Menashe, N.; Shvo, Y. Organometallics 1991, 10, 3885−3891.
2005, 127, 10840−10841.
(4) Gunanathan, C.; Ben-David, Y.; Milstein, D. Science 2007, 317
(5839), 790−792.
3308
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309

ACS Catalysis
Research Article
(22) Nielsen, M.; Junge, H.; Kammer, A.; Beller, M. Angew. Chem.,
Int. Ed. 2012, 51, 5711−5713.
(23) Takebayashi, S.; Bergens, S. H. Organometallics 2009, 28, 2349−
2351.
(24) Gunanathan, C.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
2009, 131, 3146−3147.
(25) (a) Daguenet, C.; Scopelliti, R.; Dyson, P. J. Organometallics
2004, 23, 4849−4857. (b) Kenny, J. A.; Versluis, K.; Heck, A. J. R.;
Walsgrove, T.; Wills, M. Chem. Commun. 2000, 99−100. (c) Pelagatti,
P.; Carcelli, M.; Calbiani, F.; Cassi, C.; Elviri, L.; Pelizzi, C.; Rizzotti,
U.; Rogolino, D. Organometallics 2005, 24, 5836−5844.
(26) (a) Perry, R. H.; Splendore, M.; Chien, A.; Davis, N. K.; Zare, R.
N. Angew. Chem., Int. Ed. 2011, 50, 250−254. (b) Perry, R. H.;
Brownell, K. R.; Chingin, K.; Cahill, T. J.; Waymouth, R. M.; Zare, R.
N. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 2246−2250. (c) Brownell,
K. R.; McCrory, C. C. L.; Chidsey, C. E. D.; Perry, R. H.; Zare, R. N.;
Waymouth, R. M. J. Am. Chem. Soc. 2013, 135, 14299−14305.
(27) Filonenko, G. A.; Aguila, M. J. B.; Schulpen, E. N.; van Putten,
R.; Wiecko, J.; Mueller, C.; Lefort, L.; Hensen, E. J.; Pidko, E. A. J. Am.
Chem. Soc. 2015, 137, 7620−7623.
3309
DOI: 10.1021/acscatal.6b00623
ACS Catal. 2016, 6, 3301−3309