Acceptorless dehydrogenative coupling of alcohols catalysed by ruthenium PNP complexes: Influence of catalyst structure and of hydrogen mass transfer

Article abstract of DOI:10.1016/j.jcat.2016.06.001

Base-free catalytic acceptorless dehydrogenative homo-coupling of alcohols to esters under neat conditions was investigated using a combined organometallic synthesis and kinetic modelling approach. The considered bifunctional ruthenium aliphatic PNP complexes are very active, affording TONs up to 15,000. Notably, gas mass transfer issues were identified, which allowed us to rationalize previous observations. Indeed, the reaction kinetics are limited by the rate of transfer from the liquid phase to the gas phase of the hydrogen co-produced in the reaction. Mechanistically speaking, this relates to the interconverting couple amido monohydride/amino bishydride. Overcoming this by switching into the chemical regime leads to an initial turnover frequency increase from about 2000 up to 6100?h^?1. This has a significant impact when considering assessment of novel or reported catalytic systems in this type of reaction, as overlooking of these engineering aspects can be misleading.

Full text of DOI:10.1016/j.jcat.2016.06.001

Journal of Catalysis 340 (2016) 331–343
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Acceptorless dehydrogenative coupling of alcohols catalysed by
ruthenium PNP complexes: Influence of catalyst structure and of
hydrogen mass transfer
Lei Zhang a, Guillaume Raffa a, Duc Hanh Nguyen a, Youssef Swesi b, Louis Corbel-Demailly c, Frédéric Capet a,
Xavier Trivelli d, Simon Desset a, Sébastien Paul a, Jean-François Paul a, Pascal Fongarland b,c
,
a,
Franck Dumeignil a, , Régis M. Gauvin
⇑
⇑
a Univ. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UMR 8181, UCCS – Unité de Catalyse et Chimie du Solide, F-59000 Lille, France
b Laboratoire de Génie des Procédés Catalytiques, LGPC, CNRS, CPE Lyon, Université de Lyon, F-69616 Villeurbanne, France
c IRCELyon, CNRS-Lyon 1, 2 Avenue Albert Einstein, F-69626 Villeurbanne Cedex, France
d Univ. Lille, CNRS, UMR 8576, UGSF – Unité de Glycobiologie Structurale et Fonctionnelle, F-59000 Lille, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Base-free catalytic acceptorless dehydrogenative homo-coupling of alcohols to esters under neat condi-
tions was investigated using a combined organometallic synthesis and kinetic modelling approach. The
considered bifunctional ruthenium aliphatic PNP complexes are very active, affording TONs up to
15,000. Notably, gas mass transfer issues were identified, which allowed us to rationalize previous obser-
vations. Indeed, the reaction kinetics are limited by the rate of transfer from the liquid phase to the gas
phase of the hydrogen co-produced in the reaction. Mechanistically speaking, this relates to the intercon-
verting couple amido monohydride/amino bishydride. Overcoming this by switching into the chemical
regime leads to an initial turnover frequency increase from about 2000 up to 6100 hꢀ1. This has a signif-
icant impact when considering assessment of novel or reported catalytic systems in this type of reaction,
as overlooking of these engineering aspects can be misleading.
Received 9 May 2016
Revised 31 May 2016
Accepted 1 June 2016
Keywords:
Acceptorless dehydrogenative coupling
Alcohol
Ruthenium
Pincer complexes
Kinetic modelling
Ó 2016 Elsevier Inc. All rights reserved.
1. Introduction
mentally friendly, atom-efficient catalytic process producing
simultaneously esters and hydrogen [5]. Depending on the reaction
The majority of chemicals arising from raw biomass transfor-
mations are (poly)oxygenated compounds such as alcohols, poly-
ols, sugars and carboxylic acids [1]. The overwhelming
abundance of hydroxyl group in the bio-based chemicals desig-
nates the alcohol function as a key in the production of higher-
value chemicals from bio-based feedstock. Bio-based primary alco-
hols are actually used as biofuels as well as platform molecules in
the production of higher-added value chemicals. For example, 1-
butanol is considered as an advanced biofuel and bioethanol, for
which worldwide production has reached 93 billion litres in
2014, currently accounts almost 90% of the global biofuel produc-
tion [2,3]. These alcohols have been identified as platform mole-
cules to give industrial access to hydrogen, hydrocarbons, heavier
alcohols and oxygenated derivatives [4].
conditions and the substrates nature, this reaction’s scope is not
limited to the production of symmetric esters from primary alco-
hols homocoupling: it can be extended to mixed primary-
secondary esters, lactones and polyesters, for instance (Scheme 1)
[6–8].
Catalytic systems involving conventional (‘‘innocent”) ligands,
i.e., ligands that are not involved in the alcohol activation, typically
require high reaction temperatures (about 180 °C) and few exam-
ples were reported [9]. On the other hand, Shvo’s pioneering work
demonstrated that cyclopentadienone-based ruthenium com-
plexes are capable of catalysing the ADC of benzyl alcohol to the
corresponding ester in the presence of 2 mol% ruthenium catalyst
loading and at comparatively milder temperature (145 °C). This is
enabled by the use of ligand-metal bifunctional catalyst, wherein
the redox activity is combined between the metal center and a
cyclopentadienone ligand (Scheme 2) [10]. Further on, building
on this metal-ligand cooperativity (MLC) concept, Milstein et al.
achieved multiple, major breakthroughs using specific metal/
non-innocent ligand combinations that involve the nowadays
In this context, a reaction such as primary alcohol acceptorless
dehydrogenative coupling (ADC) has gained interest as an environ-
⇑
Corresponding authors.
http://dx.doi.org/10.1016/j.jcat.2016.06.001
0021-9517/Ó 2016 Elsevier Inc. All rights reserved.

332
L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
systems is
a matter of concern when considering future
developments in the use of such catalytic systems at larger scales.
To date, only a few examples describing the base-free ADC reac-
tion were reported (Fig. 1). Milstein complexes and related PNN
systems are active using typically catalyst loadings ranging from
0.1 to 1 mol% [12,24]. Examples of the use of osmium and ruthe-
nium hydrides catalysts supported by a aliphatic PNP pincer ligand
to prepare the esters are scarce, and mixed results regarding yields
were obtained [14b–d]. Iron hydrido-borohydride PNP analogues
exhibit good catalytic activity but suffer from fast deactivation
and thus require higher catalyst loadings to reach comparable
performances [25]. Benefiting from PNN pincer-type ligand design,
Gusev osmium complexes achieved this transformation in moder-
ate to high yield [15a,15c]. In almost all the cases, the addition of a
base, typically an alkali metal alkoxide in large excess, was still
necessary to enable reaching higher catalytic activity [14a,15c].
In the present study, we focused on ruthenium complexes fea-
turing archetypal aliphatic PNP ligands based on the amino-bis(2-
phosphinoethyl) framework, with variation of phosphine sub-
stituents. We applied the as-designed catalytic systems in the
ADC reactions under base-free conditions. We further addressed
mechanistic and kinetic modelling aspects, which allowed us to
rationalize previous experimental observations from the literature.
We could thus highlight previously underestimated engineering
aspects that are definitely key to achieve a larger process efficiency
for this reaction.
Scheme 1. Alcohols acceptorless dehydrogenative coupling reactions variations
and corresponding reverse hydrogenation reactions.
2. Results and discussion
2.1. Synthesis and characterisation of (pre)catalysts
The complexes used in the present study are displayed on Fig. 2.
1a–c, 2c, 3a, 3c, 4 and 5 have been reported in the literature
[14c,21a,26,27]. Several of them (1c, 2c and 5) are now
commercially available. We describe here the synthesis and
characterisation of the missing elements of the series (2a, 2b,
3b). The main spectroscopic features of the complexes are given
Scheme 2. Selected MLC systems used for ADC of alcohols.
well-established aromatization/dearomatization process of
pyridine-based pincer complexes (Scheme 2), which is applied to
promote the ADC of alcohols, among other transformations [11].
Remarkably, such systems yield high selectivity under mild condi-
tions (0.1 [Ru] mol% loading, 115 °C).
in Table
1 for comparison. This combines both new and
previously reported data for enabling a complete overview of the
Following Milstein’s initial reports [12], several other related
complexes were further developed and have found numerous
applications [13]. Iron, ruthenium and osmium complexes sup-
ported by pincer ligands bearing a central secondary amino func-
tion (PNP [14], PNN [15] or SNS [16]) have been introduced,
mostly by Beller and Gusev. These rely on the MLC concept involv-
ing an amido-amino couple (Scheme 2) [17]. Similarly, cooperative
catalytic systems involving alkoxy-alcohol pairs developed by Gel-
man (Scheme 2) were also studied. They however require longer
reaction times (36 h) to reach satisfactory yields [18]. Among these
examples, the PNP systems have attracted a specific interest. The
remarkable catalytic performances and fair ease of access to such
catalysts has blossomed into their application in organic synthesis
[19–23] and potentially in industrial processes [19b]. Several
examples of such complexes and corresponding ligands are
commercially available from major suppliers. Most frequently,
the hydrido chloride precursors, more stable than the
catalytically-involved dihydride counterparts, are employed in
combination with a large excess of base as a catalyst activator
[14a,15,16]. One of the main drawbacks of the use of such
additives is the generation of additional base-induced side-
products, which implies the need for further additional
purification efforts prior to isolation of the targeted product.
Thus, development of efficient and robust base-free catalytic
Fig. 1. Main examples of complexes used in the base-free ADC reaction.

L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
333
infrared and elemental analysis. Their 1H NMR spectra respectively
display [Ru]–H signals at ꢀ13.52 and ꢀ13.60 ppm (triplet, JHP c.a.
17 Hz), and broad [Ru]–HBH3 peaks centered around ꢀ2.30 ppm,
indicating fast exchange on the NMR time-scale. Investigations
by 2D 1H–1H NOESY NMR on 2a and 2b revealed that N–H and
BH4 moieties are mutually in cis position (see Supplementary
Data). In both cases, a singlet was observed in 31P{1H} NMR at
77.8 (2a) and 68.9 ppm (2b), as well as in 11B{1H} NMR (Table 1).
The coordination mode of the borohydride was determined as
monohapto from infrared spectroscopy, as it comprises a charac-
teristic pattern in the m(BH) region [30].
The solid state structure of 2a was determined by single crystal
X-ray diffraction study (Fig. 3). The ruthenium center features a
distorted octahedral configuration, with the PNP ligand in a merid-
ional configuration, and with the carbonyl group in trans position
to the N–H. Apical positions are occupied by the BH4 and by the
hydride, which is in trans configuration relative to the N–H hydro-
gen in agreement with the 1H–1H NMR NOESY study. The geome-
try, bond lengths and angles are similar to those of the iron
analogue [FeH(HBH3)(CO)(La] [31].
Bishydride species 3a–b were obtained by chloride/hydride
exchange from the corresponding hydrido chloride derivatives in
yields of 92 and 69%, respectively, upon treatment with NaHBEt3
at room temperature (Scheme 3) [21a,32]. The new species 3b
features spectroscopic properties close to that of 3a, such as the
expected 1H NMR hydride signals (ꢀ6.17 and ꢀ6.14 ppm, td,
Fig. 2. Ligands and complexes used in this study.
2
2JHP = 19.2 Hz and JHH = 4.2 Hz), and gives rise to a singlet at
80.5 ppm in the 31P NMR spectrum. Regarding 3c, attempts to
purify the complex after its preparation using similar procedure
were thwarted by fast decomposition in solution into
unidentified species. Nevertheless, the compound identity could
be assessed based on similar spectroscopic properties compared
to 3a and 3b (Table 1).
During attempts to crystalize 3c in a CH2Cl2/Et2O mixture, sin-
gle crystals of the known [RuCl2(CO)(Lc)] complex were obtained,
which consist of a mixture of trans and cis isomers in a 15:1 ratio,
respectively. As only the trans-[RuCl2(CO)(Lc)] form is structurally
characterised [15b,20a], we present here the structure of the cis
analogue (Fig. 4). It features a distorted octahedral geometry
with a meridional tridentate PNP ligand. Both Cl1 and Cl1b
chloride ligands are mutually in cis position. One chloride is
apically located on the same side of the Ru-PNP plane (cis
position) with the amino hydrogen atom. Accordingly, the
carbonyl ligand occupies the remaining apical position trans to
the amino hydrogen atom.
Applying the same reaction conditions to 1d leads to the known
amido monohydride 4 instead of the targeted bishydride 3d [26].
The solid-state structure of this compound was determined using
X-ray diffraction (Fig. 5). Ruthenium features a distorted square
pyramidal geometry, with the PNP ligand in a meridional configu-
ration, the carbonyl ligand in trans position to the amido, and the
hydride ligand in the apical position. The sp2 hybridation of the
N atom is evidenced by the sum of angles (354.82°). Bond lengths
and angles are comparable to the trimethylphosphine coordinated
analogue [RuH(L)(La)] (L = CO, PMe3) [32,33]. Presumably, sponta-
neous release of H2 from (non-isolated) 3d affords compound 4.
Under 1 bar H2 in C6D6, compound 3d is formed in about 80%,
which allows its spectroscopic characterisation by multinuclear
NMR (Table 1). Spectroscopic data are indeed comparable to the
3a–c derivatives. Upon flushing with argon, 4 is fully regenerated.
As a general comment, over these series of compounds, we
noted that 2D 1H–15N NMR is a very informative spectroscopic tool
to determine the nature of nitrogen atom bonding, as this nucleus
features significantly distinct chemical shifts between amino (30–
50 ppm) and amido structures (ꢁ180 ppm, Table 1).
whole series. For sake of rigorous comparison, all the herein
reported spectroscopic data are issued from analyses of the
present study.
Complexes 1a–b and 1d were prepared in good yield (82% and
78%, respectively), using a method reported in the literature
(Scheme 3) [14c]. Indeed, heating an equimolar mixture of
[RuHCl(CO)(PPh3)3] and ligand La,b at 165 °C in diglyme results in
the stereoselective formation of complexes featuring mutually
trans N–H and Ru–H, as shown by 1H–1H 2D NOESY experiments
(see Supplementary Data). For the ligand Ld bearing tert-butyl
substituents at the phosphorous atom, a mixture of two isomeric
hydride complexes (1d1 and 1d2) in a 1:5 ratio was obtained, as
previously observed by Ding, who, however, did not discuss the
nature of the observed compounds [28]. Elemental analysis is fully
in line with the presence of isomers. 1H NMR spectrum in CDCl3 of
the minor isomer (1d1) features a characteristic Ru–H triplet at
ꢀ16.01 ppm, which thus falls in the same chemical shift order as
that of hydrides from complexes 1a–b (Table 1). Moreover, a 2D
1H–1H NOESY NMR experiment showed that the N–H and the
hydride are mutually in cis position. The major isomer (1d2) gives
rise to a significantly shielded triplet at ꢀ21.9 ppm. The amino pro-
ton resonates as an unusually deshielded broad triplet at 6.34 ppm,
to be compared with 2.74 ppm in 1d1. 2D 1H–1H NOESY NMR does
not show any spatial proximity between the amino group and the
ruthenium hydride. Moreover, the 15N chemical shift is around
57 ppm (2D 1H–15N HSQC NMR), confirming the amino form (vide
infra). Based on these spectroscopic observations, we proposed
that 1d2 is an amino hydride pentacoordinated cationic complex,
with H-bonding interaction between the N–H and the outer-
sphere chloride anion.
The new borohydride complexes 2a and 2b were prepared in
high yields (90% and 88%, respectively) by treating the respective
chloride precursors 1a and 1b with NaBH4 in a EtOH/PhMe mixture
at 65 °C, using a procedure modified compared to that used for
yielding commercially available 2c (Scheme 3) [29]. The synthesis
could be achieved using a lower amount of NaBH4, thus affording
the compound with minor boron salts side products generation.
Compounds 2a and 2b were characterised by multinuclear NMR,

334
L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
Table 1
Selected spectroscopic data.
Complex
1H{31P} (ppm)
31P{1H} (ppm)
75.8 (s)
11B{1H} (ppm)
–
15N (ppm)
56
t
(cmꢀ1
nd
)
1a
RuH: ꢀ16.30 (s)
NH: 3.43 (brt)
RuH: ꢀ16.36 (s)
NH: 3.46 (brt)
1b
66.1 (s)
–
57
56
N–H: 3178 (m), 3168 (m)
RuH: 1955 (m)
CO: 1914 (vs)
1c
RuH: ꢀ15.23 (s)
53.8 (d)
–
–
N–H: 3141 (m)
NH: 3.99 (brt)
(J = 14 Hz)
1d1: 87.0 (s)
1d2: 88.1 (s)
RuH: 1961 (m), CO: 1917 (vs), 1907 (vs)
N–H: 3201 (m), 3160 (m)
CO: 1898 (vs)
1d
RuH (1d1): ꢀ16.02 (s)
RuH (1d2): ꢀ21.9 (s)
NH (1d1): 2.74 (brt)
NH (1d2): 6.34 (brt)
1d1: 41
1d2: 57
2a
2b
2c
RuH: ꢀ13.52 (s)
RuHBH3: ꢀ2.30 (br)
NH: 3.88 (brt)
77.8 (s)
68.9 (s)
56.9 (s)
ꢀ34
ꢀ34
ꢀ33
48
48
51
NH: 3196, (m)
CO: 1920 (vs)
BH4: 2350 (m, broad), 2304 (w)
RuH: 1874 (m)
NH: 3198 (m)
RuH: ꢀ13.60 (s)
RuHBH3: ꢀ2.29 (brd)
NH: 3.88 (brt)
CO: 1924 (vs)
BH4: 2342 (m, broad), 2296 (w)
RuH: 1863 (m, broad)
NH: 3198 (w)
CO: 1915 (vs)
BH4: 2360 (s), 2328 (m) and 2284 (m)
RuH: 1844 (m)
RuH: ꢀ12.35 (s)
RuHBH3: ꢀ2.24 (brd)
NH: 4.32 (brt)
3a
3b
3c
3d
RuH: ꢀ6.40
92.2 (s)
80.5 (s)
66.5 (s)
108.9 (s)
–
–
–
–
32
32
nd
30
NH: 3172 (w)
CO: 1881 (vs)
(d, JHH = 5.0 Hz)
RuH: ꢀ6.26
(d, JHH = 5.0 Hz)
NH: 2.15 (br)
RuH: ꢀ6.17
NH: 3203 (w)
CO: 1898 (vs)
(d, JHH = 4.2 Hz)
RuH: ꢀ6.14
(d, JHH = 4.2 Hz)
NH: 2.42 (brt)
RuH: ꢀ5.80
nd
nd
(d, JHH = 5.3 Hz)
RuH: ꢀ5.78
(d, JHH = 5.3 Hz)
NH: 4.79 (brt)
RuH: ꢀ6.12 (s)
RuH: ꢀ5.86 (s)
NH: 2.42 (brt)
4a
4d
RuH: ꢀ18.71 (brs)
RuH: ꢀ20.87 (brs)
94.0 (s)
110.0 (s)
–
–
193
177
CO: 1883 (vs)
CO: 1869 (vs)
As reported, the chloride derivatives 1 were not active under
base-free conditions (Table 2, Entry 1). Upon addition of a base
[14a,15b], (1.3 mol% EtONa, that is around 52 equivalent to Ru),
these complexes led to very active catalytic systems with TOF0
up to 2600 hꢀ1, producing butyl butyrate in good to excellent
yields (Table 2, Entries 2–4). However, the use of the alkoxide
base had a detrimental effect on selectivity. It induced the
formation of heavier alcohols via the Guerbet reaction (see
Supplementary Data) detected as mixed esters [34]. Though minor,
these side-reactions lessen the interest of this transformation, as
they imply the use of additional purification steps to separate these
compounds upon upscaling. On the other hand, the borohydride
and dihydride analogues 2–3 enabled n-butanol conversion under
base-free conditions, here again achieving good to excellent yields,
except for 3b for which a 70% yield was obtained. Moreover, at any
reaction time, for 2a–c and 3a–c, the selectivity was larger than
99%, the reaction mixture being free of any detectable side product.
Complexes 2a–c exhibited activities comparable to those of com-
plexes 1a–c with TOF0 ranging from 2220 hꢀ1 and 2770 hꢀ1
(Table 2, Entries 5–7). Complexes 3a–c were slightly less active
with TOF0 ranging from 1930 to 2090 hꢀ1 (Table 2, Entries 8–10).
Complex 3a gave a nearly quantitative yield while Gusev et al.
reported modest results (21% yield, 6 h) with a higher catalyst
loading (1000 ppm). n-Butanol substrate quality as well as its
Scheme 3. Preparation of complexes 1a–d, 2a–b and 3a–c. (i) NaBH4, EtOH/PhMe,
65 °C, 4 h. (ii) NaHBEt3 (1 equiv.), THF or PhMe, 16 h, r.t.
2.2. Catalytic investigations
Having at hand this series of complexes, we proceeded to the
systematic investigation of their catalytic performances in the
benchmark n-butanol acceptorless dehydrogenative coupling reac-
tion using a 240 ppm catalyst loading, that is, with a substrate-to-
catalyst ratio of about 4150 (Table 2 and Fig. 6). For this purpose,
we monitored the composition of the reaction mixture over time.

L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
335
Fig. 5. ORTEP view of the solid state structure of 4. Ellipsoids are given at the 50%
probability level. Selected bond lengths (Å) and angles (deg): P1–Ru1 = 2.3253(4),
P2–Ru1 = 2.3232(4), N1–Ru1 = 2.0069(14), Ru1–H = 1.48(3), C1–Ru1 = 1.8470(17),
C1–O1 = 1.166(2), P1–Ru1–H = 93.8(10), P2–Ru1–P1 = 165.806(16), P2–Ru1–
H = 91.8(10), N1–Ru1–P1 = 82.81(4), N1–Ru1–P2 = 83.03(4), N1–Ru1–H = 109.7
(10), C1–Ru1–N1 = 169.68(7), C1–Ru1–P1 = 96.82(5), C1–Ru1–P2 = 96.94(5), C1–
Ru1–H = 80.6(10) C11–N1–Ru1 = 122.86(11), C12–N1–C11 = 108.72(13), C12–N1–
Ru1 = 123.24(11).
Fig. 3. ORTEP view of the solid state structure of 2a. Ellipsoids are given at the 50%
probability level. Selected bond lengths (Å) and angles (deg): Ru1–P1 = 2.3160(3),
Ru1–P2 = 2.3299(3), Ru1–N1 = 2.1863(9), Ru1–H1d = 1.876(18), Ru1–H = 1.561(18),
Ru1–C17 = 1.8389(12),
O1–C17 = 1.1565(14),
N1–H1 = 0.890(15),
P1–Ru1–
P2 = 163.300(10), P1–Ru1–H1d = 92.9(5), P1–Ru1–H = 85.7(7), P2–Ru1–H1d = 96.6
(5), P2–Ru1–H = 85.0(7), N1–Ru1–P1 = 83.07(3), N1–Ru1–P2 = 82.65(2), N1–Ru1–
H1d = 93.7(5), N1–Ru1–H = 87.2(6), C17–Ru1–P1 = 97.00(4), C17–Ru1–P2 = 96.57
(4), C17–Ru1–N1 = 175.62(4), C17–Ru1–H1d = 90.7(5), C17–Ru1–H = 88.4(6), H1d–
Ru1–H = 178.3(8), Ru1–N1–H1 = 104.5(10), C9–N1–Ru1 = 113.01(7), C9–N1–
C8 = 110.68(8), C9–N1–H1 = 106.6(9).
Table 2
Catalytic ADC reaction of n-butanol.
b
Entry
[Ru]
Base (mol%)
TOF0 (hꢀ1
)
TON
Yield (%)a,b
1
2
3
4
5
6
7
8
9
1a
1a
1b
1c
2a
2b
2c
3a
3b
3c
4d
5
–
0
2470
2480
2590
2770
2220
2570
1940
1930
2090
0
0
0
1.3
1.3
1.3
–
–
–
–
–
–
–
3670
4000
3540
4040
4000
3830
4040
2920
3630
0
88
96
85
97
96
92
97
70
10
11
12
87
0
–
290
790
19 (85c)
Reaction conditions: [Ru] (0.027 mmol; 240 ppm); n-butanol (10 mL; 109 mmol),
130 °C, 5 h, argon stream.
a
Butyl butyrate yield determined by 1H NMR and GC-FID/MS.
Two runs average. TOF0 and yield are given as an average of the two best runs.
Reaction time: 23 h.
b
c
Fig. 4. ORTEP view of the solid state structure of cis-[RuCl2(CO)(Lc)]. Ellipsoids are
given at the 50% probability level. Selected bond lengths (Å) and angles (deg): Ru1–
P2 = 2.3495(3), Ru1–P1 = 2.3372(3), Ru1–N1 = 2.1795(11), Ru1–Cl1 = 2.4087(3),
Ru1–Cl2b = 2.410(10), Ru1–C29b = 1.79(3), O1b–C29b = 1.16(4), N1–H1 = 0.869
(18), P2–Ru1–Cl1 = 86.554(12), P2–Ru1–Cl2b = 100.13(17), P1–Ru1–Cl2b = 96.42
above). While amido monohydride complexes are commonly
proposed as the active species in this reaction, no n-butanol
conversion was observed in the presence of complex 4d (Table 2,
Entry 11). Treatment of 4d with ethanol (1 equiv.) at room
temperature or at 80 °C in C6D6 did not give rise neither to the
formation of any new complex nor to alcohol-derived reaction
products. Interestingly, this complex was successfully used in
primary alcohol oxidation into the corresponding acid, though
with a significantly lesser activity than derivatives featuring less
bulky substituents on the phosphorous centers, which tends to
indicate that too high steric protection is detrimental to the
catalytic activity [26,33]. Finally, in order to compare our results
with those of other commercially available ruthenium catalysts
active in the absence of a base, we probed Milstein’s catalyst’s
(17),
P1–Ru1–Cl1 = 92.513(12),
Cl2b–Ru1–Cl1 = 174.056(14),
P1–Ru1–
P2 = 163.438(12), N1–Ru1–Cl1 = 86.39(3), N1–Ru1–P2 = 82.13(3), N1–Ru1–
Cl2b = 176.4(2), N1–Ru1–P1 = 81.31(3), C29b–Ru1–Cl1 = 177.4(7), C29b–Ru1–
P2 = 91.9(7), C29b–Ru1–Cl2b = 87.2(8), C29b–Ru1–P1 = 89.6(6), C29b–Ru1–
N1 = 95.5(8),
Ru1–N1–H1 = 103.6(12),
C14–N1–Ru1 = 113.61(8),
C15–N1–
Ru1 = 114.74(8), C15–N1–C14 = 111.80(10).
purification procedure could be at the origin of this discrepancy
since the authors reported a high yield (91%, 4 h) for the isoamyl
alcohol ADC reaction [14c]. Complex 3b bearing cyclohexyl
substituents was less robust with a catalyst deactivation after
2 h. Erratic results were obtained with complex 3c, in line with
the difficulties encountered previously during its purification (see

336
L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
Table 3
Substrate scope for ADC catalysed by 2c.
Entry Substrate
TOF0 (hꢀ1
)
TON
Yield (%)a
1
2
3
4
2570b
2380
3230
3070
3800 (3 h)b
3800 (3 h)
3800 (3 h)
4200 (3 h)
92(3 h)b
92 (3 h)
92 (3 h)
94 (3 h)
5
6
1280
1640
3600 (6 h)
3860 (5 h)
85(6 h)
90 (5 h)
Reaction conditions: 2c (0.027 mmol; 240 ppm), alcohol (109 mmol), 130 °C, argon
stream.
a
Ester yield determined GC-FID/MS and 1H NMR.
Two runs average.
b
Fig. 6. Butyl butyrate yield over time for catalysts 1a (with and without base), 2a
and 3a (see Table 2 for experimental details).
(Table 3). Heavier linear alcohols were transformed at slightly
higher rates than n-butanol (Table 3, Entries 3–4), since the reac-
tion mixture temperature is not limited by the alcohols boiling
point (i.e., 130 °C for higher alcohol vs. 118 °C for n-butanol). Steric
hindrance had a detrimental effect on the reaction rate, as b-
substituted alcohols are converted at a slower rate (TOF0 ranging
from 1280 hꢀ1 to 1640 hꢀ1, Table 3, Entries 5–6).
We tentatively probed the reaction temperature influence on
conversion mediated by 2c. However, at lower temperatures, solu-
bility issues (or catalytic initiation) hampered efficient catalysis to
take place [19a,35]. Seemingly sluggish reaction is observed at
80 °C (7% conversion after 5 h, 290 TON). Better results were
obtained at 100 °C, namely 33% conversion after 5 h (1370 TON).
Subsequent investigations were focused on the effect of the cat-
alyst loading in order to probe the system’s limitation, as reported
in Table 4. Interestingly, lowering catalyst 2c concentration from
500 to 100 ppm did not impact the final, nearly quantitative yield
(Table 4, Entries 1–3), but gave rise to a more active catalytic sys-
tem with a higher TOF0 (4650 hꢀ1, Table 4, Entry 3). Using lower
catalytic loadings (60 and 20 ppm, Table 4, Entries 4–5) led to an
even more initially active (up to TOF0 = 5600 hꢀ1), but less robust,
catalytic system with catalytic deactivation after 5 h
(TON = 14,900; butyl butyrate yield = 35%, Table 4, Entry 5), which
shows the limitations of this system.
(5) performance under similar reaction conditions. This compound
proved to be less active: after an induction period of about 2 h,
conversion rose with a TOF0 of 290 hꢀ1 and with a maximal yield
of 85% within 23 h (Table 2, Entry 12).
Regarding the influence of the pre-catalysts structure, we
noticed that the hydrido chloride systems combined with a base
lead to highly active systems, but are plagued by selectivity issues.
On the other hand, the bishydrido compounds appear as more sen-
sitive than the other pre-catalysts, and may achieve lower yields.
When combined to the issues encountered during isolation
attempts, this indicates that the sensitivity of bishydride species
complicates their handling and use as catalysts. On the other hand,
the borohydride derivatives, which may be considered as BH3-
protected bishydrides, are more robust and provide activities and
productivities similar to or better than that of 1a–c/base systems.
Moreover, catalyst activation was fast and no induction period
was observed. Mechanistically, BH3 release could occur upon ther-
mal treatment or could be promoted by a nucleophile such as the
alcohol substrate. This was recently investigated by NMR experi-
ments and DFT calculations by Guan in the case of the iron deriva-
tive: the trans-[FeH2(CO)(La)] complex was formed from the
borohydride species either in situ upon heating or in the presence
of phosphine as borane scavenger [35]. Moreover, activation of 2c
was reported to proceed with a 10–15 min induction time in fatty
methyl esters hydrogenation at 135 °C that may also be partially
attributed, in our opinion, to its low solubility [19a]. When
comparing the 1/base and 2 catalytic systems, it is most likely
that the side products generated during activation (NaCl for 1a–
c, BH3 and derivatives for 2a–c) are not innocent toward species
involved in catalysis, as, for instance, chloride sources proved to
be detrimental to catalytic activity [36]. This may explain the
slightly better results obtained with the hydrido-borohydride
derivatives.
The catalytic behaviour of iPr-substituted 2a and Cy-substituted
2b was also probed at low catalytic loading in order to further
investigate the phosphorous substituent effect. At 60 ppm, while
2a gives good conversion (87%, Table 4, Entry 6), 2b affords a lower
yield in butyl butyrate (71%, Table 4, Entry 8). Initial activities are
comparable for both species, with TOF0 around 4400 hꢀ1. With a
20 ppm catalyst loading, productivity of the system drops in both
cases (Table 4, Entries 7 and 9). Interestingly, the initial activity
When comparing the influence of the phosphorus substituent,
we can observe only a slight influence, putting aside the extreme
case of the tert-butyl derivatives. The observed marginal differ-
ences are difficult to rationalize, and the ranking of the (pre)-
catalysts within the series is not identical. For instance, in the
bishydride series (3a–c), that is, starting with the (postulated)
actual species involved in the catalytic cycle, P-substituents (iPr,
Cy, Ph) have no obvious effect on the initial activity (Table 2;
Entries 8–10). When considering robustness, based on final yield,
the trend is iPr > Ph > Cy.
Table 4
Influence of the catalyst loading on ADC reaction of n-butanol.
Entry [Ru] [Ru] loading (ppm) TOF0 (hꢀ1
)
TONa
Yield (%)a,b
a
1
2
3
4
5
6
7
8
9
2c
2c
2c
2c
2c
2a
2a
2b
2b
500
240
100
60
20
60
20
60
20
2130
2570
4650
4900
5600
4300
6100
4450
4600
2000 (2 h) 99 (2 h)
3800 (3 h) 94 (3 h)
10,000 (3 h) 97 (3 h)
14,500 (5 h) 87 (5 h)
14,900 (5 h) 35 (5 h)
14,000 (5 h) 84 (5 h)
13,700 (5 h) 30 (5 h)
12,000 (5 h) 71 (5 h)
10,400 (5 h) 19 (5 h)
For the next step of the study, we selected commercially avail-
able complex 2c for the scope, temperature and catalyst concentra-
tion studies. All the tested primary alcohols were transformed into
the corresponding ester in high yield with an excellent selectivity
Reaction conditions: 1-butanol (109 mmol), 130 °C, argon stream.
a
Two runs average.
b
Butyl butyrate yield determined by 1H NMR and GC-FID/MS.

L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
337
is significantly increased in the case of 2a, with an initial turnover
frequency of 6100 hꢀ1, higher than that of 2b (4600 hꢀ1). When
discussing the ranking of the 2a–c pre-catalysts, one may consider
the initial turnover frequency as a good indicator, although it is
disconnected from long-term robustness (as this value is measured
within the first stage of the catalytic runs, where no induction per-
iod is observed, and as catalysts decomposition occurs within 3–
5 h). Thus, it appears that the hydrido-borohydrides activity could
be ranked as 2a (6100 hꢀ1) > 2c (5600 hꢀ1) > 2b (4600 hꢀ1). How-
ever, even if those represent the averages from results of 2 separate
runs, these conclusions should be examined with caution, as the
range of initial activities for different catalysts remains rather nar-
row, especially for 2a and 2c. As a last comment, it is noteworthy
that the robustness of the systems follows the order: 2c P 2a > 2b.
Interestingly, such a strong, beneficial catalyst concentration
effect was previously reported in the literature by Gusev and
coworkers and was at this stage not clearly understood [37]. A
hypothesis of these authors is based on the potential formation
of inactive, polynuclear aggregates at high catalyst concentrations.
However, other physico-chemical phenomena could be at the ori-
gin of these observations. Indeed, even if the reaction proceeds effi-
ciently under standard batch conditions, it is not
thermodynamically favored (EtOH: DrG0 = 14.5 kJ molꢀ1) and the
equilibrium constant depends on the hydrogen concentration in
the considered reaction mixture [38]. Also to be considered is the
issue of Ru complex regeneration after dehydrogenation steps (that
is, to close the catalytic cycle) that could be linked to hydrogen
concentration in the liquid phase. During the reaction, the pro-
duced hydrogen is transferred from the liquid phase to the gas
phase through the liquid–gas interface. If the hydrogen mass trans-
fer is not efficient enough, hydrogen will temporarily accumulate
in the liquid phase during the first stage of the reaction. Higher cat-
alyst loadings are naturally expected to generate larger hydrogen
amounts of which the release from the liquid media could be lim-
ited by the mass transfer efficiency of the reactor.
In order to better understand the relationship between hydro-
gen concentration in the liquid phase and the catalytic system per-
formances, we studied the effect of the stirring rate, as it is well-
known to strongly influence the liquid–gas mass transfer efficiency
and, consequently, hydrogen accumulation. Building on this
hypothesis, we studied complex 2c’s catalytic activity (60 ppm,
80 g n-butanol scale) at various stirring rates in a 250 mL double-
walled reactor fitted with a Teflon impeller and Teflon baffles
allowing a good thermal and stirring control (Table 5). In all cases,
butyl butyrate was produced in a high yield (90%) with an excel-
lent selectivity. The catalyst apparent initial activity, ranging from
2880 hꢀ1 at 300 rpm (rpm: rotations per minute) to 5440 hꢀ1 at
1800 rpm, increased with the stirring rate, demonstrating that
the liquid–gas exchange is indeed largely influenced by the level
of hydrogen accumulation in the liquid phase (Fig. 7).
2.3. Kinetic modelling of the ADC process
In order to check our hypothesis that hydrogen accumulation in
the liquid phase is actually correlated to the catalytic system per-
formances, and since this hydrogen concentration was not directly
measurable, we have developed a model including chemical steps
and hydrogen liquid–gas mass transfer to simulate the experi-
ments, from Table 5, which were performed under semi-batch con-
ditions. The objective of this methodology is to give a quantitative
description of all the phenomena lumping together chemistry and
physical aspects. Thus, a simplified approach was proposed, based
on the commonly proposed catalytic cycle (Scheme 4) [14a]. It
involves the formation of an amido monohydride ruthenium(II)
Cat-H from the trans-dihydride ruthenium(II) complex Cat-H2
through hydrogen release (connected to the above-described
compounds of type 4 and 3, respectively). Then, an alcohol
dehydrogenation step gives back Cat-H2 and produces aldehyde.
The latter is quickly trapped by an alcohol molecule, giving a
hemiacetal intermediate. Finally, through another connected
catalytic loop, hemiacetal dehydrogenation by Cat-H releases the
ester and regenerates the trans-dihydride Cat-H2.
Table 5
Influence of the stirring rate on n-butanol conversion.
Entry
Stirring rate (rpm)a
TOF0 (hꢀ1
)
TON
Yield (%)b
1
2
3
4
300
800
1200
1800
2880
4320
5120
5440
15,880 (20 h)
15,880 (5 h)
14,690 (5 h)
15,880 (5 h)
92 (20 h)
92 (5 h)
90 (5 h)
92 (5 h)
Reaction conditions: 2c (0,65 mmol; 60 ppm), n-butanol (80 g; 1.09 mol), 130 °C,
argon stream.
Based on this well-accepted mechanism, a homogeneous cat-
alytic kinetic model was proposed, by considering pseudo first-
order reaction kinetics for each of the reaction steps. All the species
framed in Scheme 4 were considered in the calculations. As the
substrate scope (Table 3) revealed a detrimental steric bulk effect
on reactivity, the rate of the two different dehydrogenation steps
for n-butanol and the corresponding hemiacetal are considered
as different (designated as r1 and r3, respectively). In our case, just
as in previous reports neither aldehyde nor hemiacetal intermedi-
ates have been observed in the reaction mixture [14a,15b]. These
are thus reactive species that do not accumulate in the reactor.
The acetalization step (rate r2) that transforms the aldehyde into
the hemiacetal is therefore considered as fast, possibly being a
metal-catalysed process [25]. Obviously, the amino-bishydride
Cat-H2 to amido-hydride Cat-H reaction that corresponds to the
regeneration of the catalyst is identical for both dehydrogenation
steps (rate r4). The reaction rate of each step in Scheme 4 can be
expressed as follows:
a
rpm: rotations per minute.
Butyl butyrate yield determined by GC-FID/MS.
b
7000
6000
5000
4000
3000
2000
1000
0
r1 ¼ k1 ꢂ C1 ꢂ C2
r2 ¼ k2 ꢂ C1 ꢂ C3
r3 ¼ k3 ꢂ C8 ꢂ C2
r4 ¼ k4 ꢂ C4
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
0
300
600
900
1200 1500 1800
Impeller rotation per minute
r5 ¼ kLa ꢂ C6
Fig. 7. Calculated evolution of instantaneous rates in time.

338
L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
Scheme 4. Reaction scheme used for kinetic modelling.
C1, C2, C3, C4, C6, C7 and C8 represent the concentrations of the fol-
lowing species: primary alcohol, active form of catalyst (Cat-H),
aldehyde intermediate, inactive form of catalyst (Cat-H2), hydrogen
in liquid and gas phase, and hemiacetal intermediate, respectively.
k1–k4 and kLa are rate constants for each step. The liquid–gas mass
transfer of H2 (r5) is taken into account by using the double-film
theory, and assuming that the mass-transfer resistance is only sig-
nificant on the liquid side, and not on the gas side. We have also
assumed that hydrogen concentration in the interface was negligi-
ble since the argon flow rate was high enough to lead to a negligible
Within this correlation, kLa is strongly correlated with the
impeller rotation rate N (power 2.45). Since the reaction is per-
formed under neat conditions, physical properties of the liquid
are expected to change during the reaction (viscosity, density, sur-
face tension, diffusivity. . .), between reaction’s start and end (for
instance, by a factor 2 in mixture diffusivity). Thus, as the mass-
transfer coefficient kLa depends on the reaction mixture physical
properties, it was then recalculated as a function of time by taking
into account the evolution of density, diffusivity and viscosity
using values estimated from the Prosim(c) database. Density was
found as being mostly constant with the reaction advancement,
allowing us to consider a volume remaining constant with time.
Using the rate equations, the mass balance equations for all the
components of the reaction network are thus given by the follow-
ing expressions:
H2 partial pressure in the gas phase. With this model, the influence
of the concentration of H2 in the liquid phase is supposed to impact
the ratio between the active Cat-H and inactive Cat-H2 forms of the
catalyst. According to the mass balance equations, which are pre-
sented below (Eqs. (11)–(18)), it is not possible to determine a sim-
ple expression that would correlate hydrogen concentration and the
ratio between the active and the inactive form of the catalyst, due to
the complexity of the corresponding mathematic system. This can
only be obtained through the complete integration of the ordinary
differential equations using numerical calculations. Furthermore,
regarding the H2 liquid–gas transfer process, the kLa constant,
namely the volumetric mass-transfer coefficient, depends on
numerous parameters linked to the reactor stirrer and to the
physico-chemical properties of the presently involved fluids (like
viscosity and density). In other words, the kLa parameter does not
directly depend on the ADC process chemistry. In order to decrease
the number of kinetic constants to be determined through numer-
ical adjustment to a more reasonable but still reliable number, kLa
has been estimated from a correlation developed for a mechanically
agitated tank of a similar volume (Eq. (6)) [39].
dC1=dt ¼ ꢀr1 ꢀ r2
dC2=dt ¼ ꢀr1 ꢀ r3 þ r4
dC3=dt ¼ r1 ꢀ r2
dC4=dt ¼ r1 þ r3 ꢀ r4
dC5=dt ¼ r3
ð11Þ
ð12Þ
ð13Þ
ð14Þ
ð15Þ
ð16Þ
ð17Þ
ð18Þ
dC6=dt ¼ r4 ꢀ r5
dC7=dt ¼ r5
dC8=dt ¼ r2 ꢀ r3
All the kinetic parameters k1–k4 have then to be determined in
order to fit the experimental concentrations of substrate and pro-
duct with the calculated ones. To do so, a non-linear regression
was performed by using MATLABÒR2014a. The subroutine
‘‘ode15s” was used in the integration of the ordinary differential
equations (mass balance), whereas the subroutine ‘‘LSQCURVEFIT”
(Trust-region reflective) was applied in the non-linear regression.
The as-calculated concentration profiles of n-butanol and butyl
butyrate were then compared with the experimental ones from
the catalytic runs described in Table 5. The Residual Sum of
Squares (RSS) between the calculated and the experimental con-
centration was used as the objective function to obtain the opti-
mum kinetic parameters (Eq. (19)):
kLa ¼ 1:5 ꢃ 10ꢀ4 ꢄ Re1:45 ꢄ Sc0:5 ꢄ We0:5 ꢄ DH =d2
ð6Þ
2
with
Re ¼ Nc2 ꢄ qL
=lL
ð7Þ
ð8Þ
Sc ¼ lL
=qLDH
2
c2
DH
We ¼
ð9Þ
2
where N is the stirring rate, d is the impeller diameter,
uid volumetric mass, is the liquid dynamic viscosity,
face tension and DH2 is the bulk diffusivity of H2 in the liquid phase.
q
r
is the liq-
is the sur-
RSS ¼ nðexperimental dataÞðCexp;i ꢀ Ccalc;i
Þ
ð19Þ
X
2
l
i¼1
The correlation has thus been rearranged as follows:
where n refers to the considered experimental data set. The param-
eter estimation was performed by following a built-in option, i.e., by
using the last estimation result as a starting point for the next one.
Table 6 shows the kinetic constants k1–k4 obtained using our
N2:45
ꢄ
q1:45 ꢄ d2:4
kLa ¼ 1:5 ꢃ 10ꢀ4
ꢄ
D0H:5
ð10Þ
l0:95
ꢄ
r0:5
2

L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
339
Table 6
Adjusted kinetic and physical parameters after modelling.
Stirring rate (rpm)
k1 (L molꢀ1 sꢀ1
⁄ 100)
)
k2 (L molꢀ1 sꢀ1
)
k3 (L molꢀ1 sꢀ1
)
k4 (sꢀ1
)
kLa initial (sꢀ1
)
kLa final (sꢀ1
)
(
r
300
800
1200
1800
0.1 (22)
0.42 (34)
1.1 (331)
4.77 (194)
0.01
0.08
0.23
0.62
0.02
0.19
0.52
1.4
r: Standard deviation.
Fig. 8. Parity plot between experimental and calculated concentrations of butanol
and butyl butyrate.
Fig. 10. Evolution of the concentration: (top) of the reactant, intermediates and
products (solid line: calculated values; dots: experimental values); (bottom) of the
active and inactive form of the catalyst calculated as a function of time (1200 rpm,
[Ru] = 0.67 mmol Lꢀ1, T = 385 K).
Fig. 9. Calculated evolution of instantaneous rates vs. time.
methodology, along with the different values used for kLa, calcu-
lated according to Eq. (10). The quality of the fit is deemed accept-
able according to the parity plot presented in Fig. 8. Note that k1–k4
represent a unique set of parameters that are not depending on the
obtained from the non-linear regression method, only k1 and k2
are to be considered as reliable values, as they feature the lowest
r
values. Nevertheless, the higher value for k3 is in line with the fact
that no hemiacetal species was observed under our conditions
experimental conditions. According to the standard deviation (r)
(despite the large standard deviation).

340
L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
3. Conclusions
Base-free catalytic acceptorless dehydrogenative homo-
coupling of alcohols to esters under neat conditions was investi-
gated. Bifunctional ruthenium aliphatic PNP complexes shown to
be very active toward transformation of various primary aliphatic
alcohols, achieving up to 15,000 TON. From the reaction conditions
optimization, severe mass transfer issues related to H2 by-product
release were identified. Namely, the apparent initial activities were
limited by hydrogen diffusion, due to insufficient liquid–gas trans-
fer of the released hydrogen from the reaction mixture. From a
mechanistic point of view, this is related to the couple amido
monohydride/amino bishydride: the conversion of the latter into
the former is the last step which is necessary to complete the cat-
alytic cycle. Inhibition of this reaction affects the apparent reaction
rate. Overcoming this by switching into the chemical regime leads
to an initial turnover frequency increase from about 2900 up to
5400 hꢀ1. This has a significant impact when considering assess-
ment of novel or reported catalytic systems in this type of reaction:
overlooking of these engineering aspects can be misleading and
therefore, greater attention should be paid to catalytic setup for,
e.g., further larger scale implementation of such processes.
Fig. 11. Calculated influence of rotational impeller speed on H2 concentration.
When instantaneous rates are plotted against time as presented
in Fig. 9, we can see that all the rates r1 to r5 are similar in terms of
magnitudes except for the beginning of the reaction for r3, as this
reaction rate depends on the primary aldehyde product. There is
no clear limiting rate in our simplified reaction network, in which
some rates are moreover very close (r1 and r2; r4 and r5).
4. Materials and methods
4.1. Instruments and reagents
All the experiments were carried out under argon atmosphere
using a glove-box or a vacuum line using standard Schlenk tech-
niques. All the ruthenium complexes and PNP ligands were stored
under argon. 1-butanol (99.9%), 1-pentanol (99%), 1-octanol
(99.9%), 1-tetradecanol (97%), 2-methyl-1-butanol (99.0%), 2-
ethyl-1-hexanol (99.6%) were purchased from Aldrich. Liquid alco-
hols were dried over activated 3A molecular sieves (20 wt%) for
48 h and outgassed by extensive argon bubbling prior to use.
Deuterated solvents were purchased from Eurisotop and dried, dis-
tilled, outgassed prior to use. Dichloromethane-d2 was dried over
CaH2, THF-d8 was dried over Na/benzophenone and benzene-d6
was dried over NaK alloy. Ligands La, Lb, Ld, [RuHCl(CO)(PPh3)3],
1c (trade name: Ru-MACHO), 2c (trade name: Ru-MACHO-BH)
and 5 (Milstein catalyst) were purchased from Strem Chemicals.
Previously reported complexes 1a, 1b, 3a and 4a were prepared
following literature procedures [14c,21a]. GC-FID/MS was carried
out on Agilent 7890A (flame ionization detector and MS detector)
instrument equipped with a Zebron ZB-Bioethanol column (30 m
Fig. 10 shows an example of the evolution of the computed con-
centrations of substrate and product as a function of time at
1200 rpm (Table 5, entry 3). We can observe a deviation for long
reaction times between the predicted and the experimental con-
centrations in n-butanol. This is attributed to the deactivation of
the Ru catalyst. We checked that, even in the presence of a signif-
icant reverse reaction (ester to alcohol), the trend obtained from
the simulation was not modified. We can also observe that an
accumulation of the bishydride form (Cat-H2) takes place during
the reaction due to an accumulation of hydrogen in the liquid
phase. The differences measured in terms of TOF by varying the
impeller rotation rate are then fully explained by the performances
of the reactor related to the liquid–gas mass transfer (expressed by
the value of kLa). The higher this value, the more efficient the liq-
uid–gas mass transfer, which means a lower accumulation of
hydrogen in the liquid phase.
This is confirmed by Fig. 11 that features the calculated H2 con-
centration in the liquid phase as a function of time for different
stirring rates. Along these lines, we can also conclude that the pro-
cess performances could be improved if the liquid–gas mass trans-
fer could be intensified. The actual (maximal) TOF0 has not been
reached under the herein considered experimental conditions, as
H2 stripping is still a limiting factor (in a semi-batch even if mini-
mal the H2 concentration is still not null, and there is therefore
room for process efficiency improvement).
An ideal case would be to suppress accumulation of H2 in the
liquid phase, consequently avoiding build-up of the Cat-H2 non-
active form. In other terms, within the catalytic cycle, efficient H2
release from reaction mixture would not hinder regeneration of
Cat-H. For implementation of an industrial process, this issue has
to be considered to optimize productivity, and for laboratory-
scale experiments, it can be responsible of severe deviation
between studies involving different experimental set-ups.
column length ꢄ 0.25 mm internal diameter ꢄ 1.00
l
m
film
thickness) with helium as a carrier gas. 1H, 13P, 13C, 15N and 11B
NMR spectra were recorded at 300 K on a Bruker Avance II 400
NMR spectrometer. 1H and 13C NMR chemical shifts are reported
in ppm (d) downfield from tetramethylsilane. 13P NMR chemical
shifts are reported in ppm (d) downfield from H3PO4. 15N NMR
chemical shift are reported in ppm (d) downfield from NH3, and
were indirectly determined from 2D 1H–15N HMBC/HSQC. 11B
NMR chemical shift are reported in ppm (d) downfield from
BF3ꢂEt2O. Common abbreviations used in the NMR experiments
are as follows: s singlet (s), doublet (d), triplet (t), virtual triplet
(vt), quartet (q), multiplet (m). IR spectra were recorded on a
Nicolet 6700 FT-IR spectrometer equipped with a Praying Mantis
mirror chamber (from Harrick Scientific) by using a DRIFT cell
equipped with KBr windows. The samples were prepared under
argon in a glove-box. Typically, 64 scans were accumulated for
each spectrum (resolution 4 cmꢀ1). Data are reported as follows:
weak (w), medium (m), strong (s) and very strong (vs).

L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
341
11
4.2. Complexes preparation and characterisation
NMR (CD2Cl2, 40 MHz):
128 MHz): d (ppm) ꢀ34.5. IR(KBr):
d
(ppm) 48.
B{1H} NMR (CD2Cl2,
m
(cmꢀ1) 682 (w), 781 (w),
831 (w), 855 (w), 890 (w), 964 (w), 979 (w), 1074 (m), 1174 (w),
4.2.1. [RuHCl(CO)(NH(C2H4PtBu2)2)] (1d)
The preparation of 1d was carried out according to the reported
method with slight modification [40]. A suspension of [RuHCl(CO)
(PPh3)3] (200 mg, 0.21 mmol) and NH(C2H4PtBu2)2 (10 wt% in THF,
0.23 mmol) was concentrated under vacuum. Then, diglyme (2 ml)
was added and the suspension heated to 165 °C for 1 h giving a yel-
low solution. The solution was left to let it cool to room tempera-
ture giving a white crystalline solid. The supernatant was carefully
removed, and the white powder washed three times with Et2O and
dried under vacuum (r.t., 10ꢀ3 mbar). Isolated yield: 96% (106 mg).
1198 (w), 1447 (m), 1876 (m), 1923 (vs), 2296 (w), 2340 (m),
2850 (s), 2925 (s), 3197 (m). Elemental analysis: Calc. for C29H58
-
BNOP2Ru: C, 57.04; H, 9.57; N, 2.29%. Found: C, 56.89; H, 9.80; N,
2.47%.
4.2.4. [RuH2(CO)(NH(C2H4PCy2)2)] (3b)
A suspension of 1b (202 mg, 0.32 mmol) in toluene (4 ml) was
treated by a solution of NaHBEt3 (1 M in toluene; 290 mg;
0.33 mmol) at room temperature. The white suspension quickly
turned yellow. The reaction media was left overnight under stirring
at room temperature giving a turbid yellow solution. The reaction
mixture volume was partially reduced under vacuum and the
remaining solution filtered through a sintered glass filter. Then,
the filtrate was concentrated under vacuum, layered with pentane
and stored in the freezer of the glove-box overnight. The cold
supernatant was quickly removed via a syringe and the white crys-
talline powder quickly washed with cold pentane. Isolated yield:
69% (132 mg). 1H NMR (C6D6, 400 MHz): d (ppm) ꢀ6.27 (td,
2
1H NMR (CDCl3, 400 MHz): d (ppm) ꢀ16.02 (t, JHP = 16.0 Hz, 1H,
2
RuH, 18% minor isomer), ꢀ21.9 (t, JHP = 18.6 Hz, 1H, RuH, 82%
v
v
major isomer), 1.31 (vt, JHP = 6.6 Hz, 18H), 1.38 (vt, JHP = 6.7 Hz,
18H), 2.13–2.20 (m, 2H), 2.21–2.30 (m, 2H), 2.34–2.44 (m, 2H),
3.43–3.40 (m, 2H), 6.34 (br, 1H, NH). 31P{1H} NMR (CDCl3,
162 MHz): d (ppm) 88.1 (s, 84%, major isomer) and 87.0 (s, 16%,
minor isomer). 13C{1H, 31P} NMR (CDCl3, 100 MHz): d (ppm) 25.4,
29.6, 30.4, 36.1, 38.2, 53.7, 206.6. 2D 1H–15N HSQC NMR (CDCl3,
40 MHz): d (ppm) 57. IR(KBr):
m
(cmꢀ1) 685 (w), 811 (w), 961
2
2
(w), 1178 (w), 1364 (w), 1389 (w), 1470 (m), 1898 (vs), 2865
(m), 2895 (m), 2955 (m), 2986 (m), 3160 (w), 3200 (w). Elemental
analysis: Calc. for C21H46ClNOP2Ru: C, 47.86; H, 8.80; N, 2.66%.
Found: C, 47.67; H, 9.08; N, 2.77%.
2JHP = 19.5 and JHH = 4.2 Hz, 1H, RuH), ꢀ6.24 (td, JHP = 19.5 and
2JHH = 4.2 Hz, 1H, RuH), 1.04–1.38 (m, 15H), 1.53–1.90 (m, 27H),
2.02–2.35 (m, 11H). 31P{1H} NMR (C6D6, 162 MHz): d (ppm) 80.5
(s). 13C{1H} NMR (C6D6, 100 MHz): d (ppm) 26.7, 26.9, 27.6, 27.8,
v
28.0, 28.2, 28.3, 29.8, 29.9, 30.2, 38.2 (vt, JCP = 10.9 Hz), 41.0 (vt,
v
v
4.2.2. trans-[RuH(HBH3)(CO)(NH(C2H4PiPr2)2)] (2a)
JCP = 11.9 Hz), 54.2, 210.3 (vt, JCP = 12.5 Hz, CO). 2D 1H–15N HSQC
A solution of NaBH4 (160 mg, 4.23 mmol) in ethanol (12 ml)
was added via a cannula to a suspension of complex 1a (195 mg,
0.41 mmol) in toluene (16 mL). Then, the suspension was heated
at 65 °C during 4 h leading to an opalescent solution. Volatiles
were removed under vacuum giving a white residue. Dichloro-
methane was added and the slightly yellow suspension filtered
off. This extraction was repeated three times. Finally, the combined
dichloromethane extracts were concentrated under vacuum giving
the targeted product as an off-white solid (169 mg, 90%). 1H NMR
NMR (C6D6, 40 MHz) d (ppm) 32. IR(KBr): m
(cmꢀ1) 684 (w), 712
(w), 737 (w), 781 (w), 828 (w), 854 (w), 889 (w), 912 (w), 945
(w), 958 (w), 1006 (w), 1031 (w), 1073 (w), 1117 (w), 1173 (w),
1195 (w), 1243 (w), 1268 (w), 1296 (w), 1310 (w), 1340 (w),
1405 (w), 1446 (w), 1505 (w), 1811 (w), 1898 (s), 2848 (m),
2919 (s), 3203 (m). Elemental analysis: Calc. for C29H55NOP2Ru:
C, 58.37; H, 9.29: N, 2.35%. Found: C, 58.18; H, 9.54; N, 2.39%.
4.2.5. [RuH2(CO)(NH(C2H4PPh2)2)] (3c)
2
(CD2Cl2, 400 MHz): d (ppm) ꢀ13.52 (t, JHP = 17.6 Hz, 1H, RuH),
This compound was prepared similarly to 3b by using Ru-
MACHO 1c (250 mg, 0.42 mmol) instead of complex 1b with a lar-
ger amount of solvent (40 mL). Isolated yield: 45% (108 mg). 1H
NMR (THF-d8, 400 MHz): d (ppm) ꢀ5.78 (m, 2H, RuH), 2.08–2.16
(m, 2H), 2.20–2.30 (m, 2H), 2.87–2.92 (m, 2H), 3.08–3.21 (m,
2H), 4.79 (brt, 1H, NH), 7.25–7.32 (m, 13H), 7.94–7.99 (m, 7H).
13P{1H} NMR (THF-d8; 162 MHz): d (ppm) 66.5 (s). This compound
was not stable in THF-d8 and slowly evolved into yet unidentified
new species. 1H NMR (THF-d8, 300 MHz): d (ppm) ꢀ5.74 (t,
2JHP = 20.3 Hz, 2H, RuH), 2.07–2.32 (m, 4H), 2.89–2.94 (m, 2H),
3.06–3.23 (m, 2H), 4.69 (brt, 1H, NH), 7.27–7.35 (m, 12H), 7.95–
8.00 (m, 8H). 13P{1H} NMR (THF-d8, 121 MHz): d (ppm) 66.7 (s).
3
ꢀ2.30 (br, 4H, RuHBH3), 1.13 (dd, JHH = 7.0 Hz and JHP = 14.7 Hz,
3
6H), 1.17 (dd, JHH = 6.9 Hz and JHP = 16.4 Hz, 6H), 1.25 (dd,
3
3JHH = 7.1 Hz and JHP = 12.3 Hz, 6H), 1.37 (dd, JHH = 7.4 Hz and
JHP = 16.2 Hz, 6H), 1.83–1.92 (m, 2H), 2.13–2.30 (m, 6H), 2.45–
2.54 (m, 2H), 3.19–3.23 (m, 2H), 3.88 (br, 1H, NH). 31P{1H} NMR
(CD2Cl2, 162 MHz):
d
(ppm) 77.8 (s). 13C{1H} NMR (CD2Cl2,
v
100 MHz): d (ppm) 17.9, 18.9, 20.7, 20.9, 24.5 (vt, JHP = 12.7 Hz),
v
v
29.2 (vt, JHP = 11.5 Hz), 29.8 (vt, JHP = 9.1 Hz), 207.0 (vt,
v
JHP = 11.8 Hz). 2D 1H–15N HSQC NMR (CD2Cl2, 40 MHz): d (ppm)
48. 11B{1H} NMR (CD2Cl2, 128 MHz): d (ppm) ꢀ34.6. IR(KBr):
m
(cmꢀ1) 697 (w), 785 (w), 832 (w), 968 (w), 1044 (w), 1069 (m),
1245 (w), 1382 (w), 1461 (m), 1864 (m), 1874 (m), 1920 (vs),
2308 (w), 2350 (m), 2870 (m), 2890 (w), 2931 (w), 2942 (w),
4.2.6. [RuH(CO)(N(C2H4PtBu2)2)] (4d)
2956 (m), 2973 (w), 3196 (m). Elemental analysis: Calc. for C17H42
-
This species was previously prepared by Prechtl and coworkers
by treating RuH(H2)[N(C2H4PtBu2)2] with a primary alcohol [26].
We describe herein an alternative procedure. A suspension of 1d
(98 mg, 0.18 mmol) in toluene (2 mL) was treated by a commercial
solution of NaHBEt3 (1 M in toluene, 181 mg, 0.20 mmol). The mix-
ture was stirred overnight at room temperature giving an opales-
cent turbid orange solution. Then, the solution was filtered
through a sintered glass filter; the filtrate was concentrated under
vacuum, layered with pentane and stored at ꢀ20 °C. The cold
supernatant was quickly removed via a syringe and the orange
powder washed with cold pentane. Isolated yield: 78% (71 mg).
BNOP2Ru: C, 45.34; H, 9.40: N, 3.11%. Found: C, 45.21; H, 9.48; N,
3.12%. Single crystals suitable for X-ray diffraction measurement
were obtained by layering a toluene solution of 2a with pentane
and stored overnight at ꢀ20 °C.
4.2.3. trans-[RuH(HBH3)(CO)(NH(C2H4PCy2)2)] (2b)
This compound was prepared similarly to 2a by using complex
1b (200 mg, 0.43 mmol) instead of complex 1a and using a lower
amount of NaBH4 (48 mg, 1.37 mmol). Isolated yield: 88%
(171 mg). 1H NMR (CD2Cl2, 400 MHz):
d
(ppm) ꢀ13.60 (t,
2JHP = 17.9 Hz, 1H, RuH), ꢀ2.33 (br, 4H, RuHBH3), 1.22–1.61 (m,
22H), 1.69–1.97 (m, 20H), 2.09–2.28 (m, 8H), 3.12–3.26 (m, 2H),
3.87 (br, 1H, NH). 31P{1H} NMR (CD2Cl2; 162 MHz): d (ppm). 68.9
(s). 13C{1H, 31P} NMR (CD2Cl2, 100 MHz): d (ppm) 26.5, 26.8, 27.0,
27.2, 27.7, 28.1, 28.4, 28.8, 28.9, 54.8, 207.4. 2D 1H–15N HSQC
1H NMR (C6D6, 400 MHz): d (ppm) ꢀ20.87 (t, JHP = 16.4 Hz, 1H,
2
v
v
RuH), 1.24 (vt, JHP = 6.4 Hz, 18H), 1.28 (vt, JHP = 6.4 Hz, 18H),
1.77–1.87 (m, 2H), 1.9–1.97 (m, 2H), 3.10–3.18 (m, 2H,), 3.42–
3.55 (m, 2H). 31P{1H} NMR (C6D6, 162 MHz): d (ppm) 110.0 (s).
2D 1H–15N HMBC NMR (40 Hz, C6D6): d (ppm) 177. Single crystals

342
L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
suitable for X-ray diffraction measurement were obtained by slow
evaporation of a benzene solution.
temperature ca. 90 °C), all the catalyst was dissolved giving a
slightly yellow solution. After 15 min, reflux conditions were
obtained (solution temperature ca. 112 °C). At this time, a first ali-
quot (0.2 mL) was taken, which corresponds to the reaction t0.
Then, aliquots (ca. 0.2 mL) were periodically sampled to monitor
the reaction progress over time. Liquid samples were diluted with
CDCl3 and analysed by 1H RMN to determine yield, turnover num-
ber and turnover number frequency.
4.3. Catalytic tests
4.3.1. TOF0 determination
For all the alcohols acceptorless dehydrogenative coupling reac-
tions, the initial turnover frequency, i.e., TOF0, was determined
from the plot of the turnover number as a function of time. The
TOF0 was calculated from the slope of the linear regression per-
formed on the initial linear part of the plot. In most of the cases,
the catalytic tests were duplicated to ensure good repeatability.
Yield, turnover number (TON) and turnover number frequency
(TOF) are given as an average of two runs. See Supplementary Data
for details.
4.4. X-ray structure determination
A single crystal of each compound was mounted under inert
perfluoropolyether wax on a Mitegen MicroLoopTM. The single-
crystal X-rays measurements were performed at 100 K under N2
stream from a Cryostream 700 device (OxfordCryosystems). Data
were collected using an Apex II CCD 4K Bruker diffractometer
(k = 0.71073 Å). The structures were solved using SHELXT and
refined by least-squares procedures on F2 using SHELXL2014
[41,42]. All the hydrogen atoms were placed in theoretical posi-
tions and refined riding on their parent atoms except for the
hydride H attached to the Ru atom, which was located from differ-
ence Fourier maps and refined isotropically. ORTEP drawings were
generated with ORTEP-3 [43]. Crystallographic data have been
deposited at the Cambridge Crystallographic Data Centre as Sup-
plementary Publication Nos. CCDC 1449421-1449423. Copies of
the data can be obtained free of charge on application to the Direc-
tor, CCDC, 12 Union Road, Cambridge CB2 1EZ, U.K. (fax: (+44)
1223 336 033; e-mail: deposit@ccdc.cam.ac.uk).
4.3.2. Typical procedure for acceptorless dehydrogenative coupling of
1-butanol conducted in Schlenk tubes
In argon filled glove-box, the selected complex (0.026 mmol;
240 ppm) was weighted in a Schlenk tube containing a stirring
bar. After connection to a Schlenk line, 1-butanol (10 mL; 8.10 g;
109 mmol) was added via a syringe under an argon stream. The
Schlenk tube was then equipped with a condenser topped by an
argon bubbler. The system was heated using an oil bath (130 °C)
and stirred magnetically under an argon stream. Aliquots (ca.
0.1 mL) were periodically sampled to monitor the reaction pro-
gress over time. The liquid samples were weighted, mixed with a
known amount of cyclohexane (internal standard) and diluted
with dichloromethane. The samples were then analysed by GC-
FID/MS for determination of conversion, yield, turnover number,
turnover frequency and mass balance analysis. Alternatively, ali-
quots were diluted with CDCl3 and analysed by 1H NMR for deter-
mination of yield, turnover number and turnover frequency. Both
analytical methods gave identical results.
4.5. Modelling experiments procedure
Four catalytic runs varying by the impeller rotation speed (300,
800, 1200, 1800 rpm) and conducted at the same temperature
(measured temperature: 115 °C) were selected (Table 5). The data
were implemented into a single data file in the MatlabÒR2014a
software.
4.3.3. Typical procedure for acceptorless dehydrogenative coupling of
1-butanol conduced in a double walled reactor.
4.3.3.1. Reactor description. A 250 mL double walled reactor (Rad-
ley) with a bottom valve and fitted with a Teflon impeller and
Teflon baffles was used. The reactor lid ports were equipped with
a reflux condenser topped with an oil bubbler, an argon inlet, an
injector port closed with a septum and a dipping thermocouple.
The reactor was heated and the reaction temperature monitored
using a Huber Ministat 230 connected to the thermocouple dipped
in the reaction mixture. Silicon oil was used as heating liquid. A
Heidolph overhead mechanical stirrer was used for the impeller
rotation and accurate control of the stirring rate.
Acknowledgements
This work was performed in partnership with the SAS PIVERT,
within the frame of the French Institute for the Energy Transition
(Institut pour la Transition Energétique (ITE) P.I.V.E.R.T. (www.in-
stitut-pivert.com) selected as an Investment for the Future
(‘‘Investissements d’Avenir”). This work was supported, as part of
the Investments for the Future, by the French Government under
the reference ANR-001-01. L. Zhang thanks the China Scholarship
Council (CSC 2011008127) for providing the financial support of
his PhD. CNRS, Chevreul Institute (FR 2638), Ministère de
l’Enseignement Supérieur et de la Recherche, Région Nord – Pas
de Calais and FEDER are acknowledged for supporting and funding
partially this work.
4.3.3.2. Typical procedure. The reactor was preheated to 100 °C
under a gentle argon flow for 4 h. In an argon-filled glove-box,
the catalyst 2c (38 mg; 0.065 mmol; 60 ppm) was weighted and
transferred to a pressure-equalized dropping funnel (custom built).
Under argon flow, a first 1-butanol portion was introduced in three
fractions via a 20 mL syringe (3 ꢄ 20 mL; 48 g) into the reactor. The
temperature set point was reduced to 80 °C and the stirring was
started with a rate of 100 rpm. The septum was replaced by the
pressure-equalized dropping funnel containing the powdered cat-
alyst. Another 1-butanol portion (16 g) was introduced into the
dropping funnel giving a white suspension. Then, the suspension
was added to the reactor. The dropping funnel was rinsed with
the last portion of 1-butanol (16 g; total: 80 g; 1.08 mol). At this
point, the reaction media was a white-yellow suspension. The
dropping funnel was replaced by a septum. The argon inlet was
switched from lid reactor to the oil bubbler topping the reflux con-
denser. The circulating heating oil was set to 130 °C and the stir-
ring increased to the desired rate. After 5 min (solution
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2016.06.001.
References
[1] (a) R.A. Sheldon, Green Chem. 16 (2014) 950–963;
(b) J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539–554;
(c) A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
[2] K.R. Szulczyk, Int. J. Energy Environ. 1 (2010) 501–512.
[3] Renewal Fuels Associated (RFA), <http://www.ethanolrfa.org/resources/
industry/statistics/#1456770302650-ba1c280f-bad8>.
[4] (a) J. Sun, Y. Wang, ACS Catal. 4 (2014) 1078–1090;
(b) J. Rass-Hansen, H. Falsig, B. Jørgensen, C.H.J. Christensen, Chem. Technol.
Biotechnol. 82 (2007) 329–333.

L. Zhang et al. / Journal of Catalysis 340 (2016) 331–343
343
[5] (a) S. Werkmeister, J. Neumann, K. Junge, M. Beller, Chem. Eur. J. 21 (2015)
12226–12250;
[21] (a) Y. Li, M. Nielsen, B. Li, P.H. Dixneuf, H. Junge, M. Beller, Green Chem. 17
(2015) 193–198;
(b) C. Gunanathan, D. Milstein, Chem. Rev. 114 (2014) 12024–12087.
[6] (a) J. Cheng, M. Zhu, C. Wang, J. Li, X. Jiang, Y. Wei, W. Tang, D. Xue, J. Xiao,
Chem. Sci. (2016), http://dx.doi.org/10.1039/c6sc00145a;
(b) D. Srimani, E. Balaraman, B. Gnanaprakasam, Y. Ben-David, D. Milstein,
Adv. Synth. Catal. 354 (2012) 2403–2406;
(c) B. Gnanaprakasam, Y. Ben-David, D. Milstein, Adv. Synth. Catal. 352 (2010)
3169–3173.
[7] (a) J. Zhang, E. Balaraman, G. Leitus, D. Milstein, Organometallics 30 (2011)
5716–5724;
(b) P. Sponholz, D. Mellmann, C. Cordes, P.G. Alsabeh, B. Li, Y. Li, M. Nielsen, H.
Junge, P.H. Dixneuf, M. Beller, ChemSusChem 7 (2014) 2419–2422;
(c) A. Monney, E. Barsch, P. Sponholz, H. Junge, R. Ludwig, M. Beller, Chem.
Commun. 50 (2014) 707–709;
(d) M. Nielsen, E. Alberico, W. Baumann, H.-J. Drexler, H. Junge, S. Gladiali, M.
Beller, Nature 495 (2013) 85–89.
[22] (a) Q. Han, X. Xiong, S. Li, Catal. Commun. 58 (2015) 85–88;
(b) N.J. Oldenhuis, V.M. Dong, Z. Guan, Tetrahedron 70 (2014) 4213–4218;
(c) N.J. Oldenhuis, V.M. Dong, Z. Guan, J. Am. Chem. Soc. 136 (2014) 12548–
12551.
[23] (a) M. Peña-López, H. Neumann, M. Beller, Chem. Commun. 51 (2015) 13082–
13085;
(b) J. Zhao, J.F. Hartwig, Organometallics 24 (2005) 2441–2446.
[8] D.M. Hunsicker, B.C. Dauphinais, S.P. Mc Ilrath, N.J. Robertson, Macromol.
Rapid. Commun. 33 (2012) 232–236.
[9] S.-I. Murahashi, T. Naota, K. Ito, Y. Maeda, H. Taki, J. Org. Chem. 52 (1987)
4319–4327.
[10] (a) Y. Blum, Y. Shvo, J. Organomet. Chem. 282 (1985) C7–C10. For recent
reviews on the Shvo catalyst:;
(b) Y. Li, H. Li, H. Junge, M. Beller, Chem. Commun. 50 (2014) 14991–14994.
[24] (a) L.-P. He, T. Chen, D. Gong, Z. Lai, K.-W. Huang, Organometallics 31 (2012)
5208–5211.
[25] (a) S. Chakraborty, P.O. Lagaditis, M. Förster, E.A. Bielinski, N. Hazari, M.C.
Holthausen, W.D. Jones, S. Schneider, ACS Catal. 4 (2014) 3994–4003;
(b) P.J. Bonitatibus Jr., S. Chakraborty, M.D. Doherty, O. Siclovan, W.D. Jones, G.
L. Soloveichik, Proc. Natl. Acad. Sci. USA 112 (2015) 1687–1692.
[26] J.-H. Choi, L.E. Heim, M. Ahrens, M.H.G. Prechtl, Dalton Trans. 43 (2014)
17248–17254.
[27] M.S. Rahman, P.D. Prince, J.W. Steed, K.K. Hii, Organometallics 21 (2002) 4927–
4933.
[28] Z. Han, L. Rong, J. Wu, L. Zhang, Z. Wang, K. Ding, Angew. Chem. Int. Ed. 51
(2012) 13041–13045.
(b) M.C. Warner, C.P. Casey, J.-E. Bäckvall, Top. Organomet. Chem. 37 (2011)
85–125;
(c) B.J. Conley, M.K. Pennington-Boggio, E. Boz, T.J. Williams, Chem. Rev. 110
(2010) 2294–2312.
[11] (a) J.R. Khusnutdinova, D. Milstein, Angew. Chem. Int. Ed. 54 (2015) 12236–
12273.
[12] J. Zhang, G. Leitus, Y. Ben-David, D. Milstein, J. Am. Chem. Soc. 127 (2005)
10840–10841.
[13] (a) C. Gunanathan, D. Milstein, Science 341 (2013) 249–260;
(b) C. Gunanathan, D. Milstein, Acc. Chem. Res. 44 (2011) 588–602.
[14] (a) M. Nielsen, H. Junge, A. Kammer, M. Beller, Angew. Chem. Int. Ed. 51 (2012)
5711–5713;
[29] (a) WO Pat. WO2012144650(A1), 2012.;
(b) WO Pat. WO201148727(A1), 2011.
[30] T.J. Marks, J.R. Kolb, Chem. Rev. 77 (1977) 263–293.
[31] E. Alberico, P. Sponholz, C. Cordes, M. Nielsen, H.-J. Drexler, W. Baumann, H.
Junge, M. Beller, Angew. Chem. Int. Ed. 52 (2013) 14162–14166.
[32] A. Friedrich, M. Drees, M. Käss, E. Herdtweck, S. Schneider, Inorg. Chem. 49
(2010) 5482–5494.
(b) M. Nielsen, A. Kammer, D. Junge, H. Cozzula, S. Gladiali, M. Beller, Angew.
Chem. Int. Ed. 50 (2011) 9593–9597;
(c) M. Bertoli, A. Choualeb, A.J. Lough, B. Moore, D. Spasyuk, D.G. Gusev,
Organometallics 30 (2011) 3479–3482;
(d) M. Bertoli, A. Choualeb, D.G. Gusev, A.J. Lough, Q. Major, B. Moore, Dalton.
Trans. 40 (2011) 8941–8949.
[15] (a) D. Spasyuk, C. Vicent, D.G. Gusev, J. Am. Chem. Soc. 137 (2015) 3743–3746;
(b) D. Spasyuk, D.G. Gusev, Organometallics 31 (2012) 5239–5242;
(c) D. Spasyuk, S. Smith, D.G. Gusev, Angew. Chem. Int. Ed. 51 (2012) 2772–
2775.
[33] L. Zhang, D.H. Nguyen, G. Raffa, X. Trivelli, F. Capet, S. Desset, S. Paul, F.
Dumeignil, R.M. Gauvin, ChemSusChem (2016), http://dx.doi.org/10.1002/
cssc.201600243.
[34] (a) The Ru-MACHO low catalytic activity in the Guerbet reaction has been
recently reported: R.L. Wingad, P.J. Gates, S.T.G.S., D.F. Wass, ACS Catal. 5
(2015)5822–5826.;
[16] D. Spasyuk, S. Smith, D.G. Gusev, Angew. Chem. Int. Ed. 52 (2013) 2538–2542.
[17] D. Gelman, S. Musa, ACS Catal. 2 (2012) 2456–2466.
[18] (a) K. Oded, S. Musa, D. Gelman, J. Blum, Catal. Commun. 20 (2012) 68–70;
(b) S. Musa, I. Shaposhnikov, S. Cohen, D. Gelman, Angew. Chem. Int. Ed. 50
(2011) 3533–3537.
(b) D. Gabriëls, W.Y. Hernández, B. Sels, P. Van Der Voort, A. Verberckmoes,
Catal. Sci. Technol. 5 (2015) 3876–3902. For a recent review on the Guerbet
reaction.
[35] (a) S. Chakraborty, H. Dai, P. Bhattacharya, N.T. Fairweather, M.S. Gibson, J.A.
Krause, H. Guan, J. Am. Chem. Soc. 136 (2014) 7869–7872;
(b) S. Qu, H. Dai, Y. Dang, C. Song, Z.-X. Wang, H. Guan, ACS Catal. 4 (2014)
4377–4388.
[19] (a) J. Neumann, C. Bornschein, H. Jiao, K. Junge, M. Beller, Eur. J. Org. Chem.
(2015) 5944–5948;
(b) J. Deutsch, J.J. Pinto, S. Doerfelt, A. Martina, A. Köckritza, Flavour. Fragr. J.
30 (2015) 101–107;
(c) N.T. Fairweather, M.S. Gibson, H. Guan, Organometallics 34 (2015) 335–
339;
(d) C. Ziebart, R. Jackstell, M. Beller, ChemCatChem 5 (2013) 3228–3231;
(e) D. Lazzari, M.C. Cassani, M. Bertola, F.C. Morenoa, D. Torrente, RSC Adv. 3
(2013) 15582–15584;
(f) T. Otsuka, A. Ishii, P.A. Dub, T. Ikariya, J. Am. Chem. Soc. 135 (2013) 9600–
9603;
[36] S. Werkmeister, K. Junge, B. Wendt, E. Alberico, H. Jiao, W. Baumann, H. Junge,
F. Gallou, M. Beller, Angew. Chem. Int. Ed. 53 (2014) 8722–8726.
[37] For a discussion about the catalyst loading effect, see reference [15b] and more
specifically discussion on page 5241 and the quoted references.
[38] B.E. Poling, J.M. Prausnitz, J.P. O’Connell, Properties of Gases and Liquids, fifth
ed., McGraw-Hill Education, New York, 2004.
[39] V. Meille, N. Pestre, P. Fongarland, C. de Bellefon, Ind. Eng. Chem. Res. 43
(2004) 924–927.
[40] The formation of the complex characterized by a triplet at ꢀ16.02 ppm (60%)
was favored by refluxing the reaction mixture in toluene for 5 hours, see
reference [28].
(g) W. Kuriyama, T. Matsumoto, O. Ogata, Y. Ino, K. Aoki, S. Tanaka, K. Ishida, T.
Kobayashi, N. Sayo, T. Saito, Org. Process Res. Dev. 16 (2012) 166–171.
[20] (a) J. Kothandaraman, M. Czaun, A. Goeppert, R. Haiges, J.-P. Jones, R.B. May, G.
K.S. Prakash, G.A. Olah, ChemSusChem 8 (2015) 1442–1451;
(b) N.M. Rezayee, C.A. Huff, M.S. Sanford, J. Am. Chem. Soc. 137 (2015) 1028–
1031.
[41] G.M. Sheldrick, Acta Cryst. A 71 (2015) 3–8.
[42] G.M. Sheldrick, Acta Cryst. C71 (2015) 3–8.
[43] L.J. Farrugia, J. Appl. Crystallogr. 30 (1997) 565.