Hydroaminomethylation in Aqueous Solvent Systems – An Efficient Pathway to Highly Functionalized Amines

Article abstract of DOI:10.1002/adsc.201701463

Tandem-catalyzed hydroaminomethylation is a useful tool for synthesizing linear amines from olefins and amines in an atom-efficient manner. To enable the coupling of highly functionalized, hydrosoluble amines with non-water-soluble olefins, this reaction must be transferred to aqueous biphasic solvent systems. In this work, we systematically evaluate reaction conditions to provide a selective hydroaminomethylation of 1-octene with diethanolamine as model substrates. Although water is both the condensation side product and the solvent, yields of 79% were achieved using a catalytic system consisting a rhodium precursor and a sulfonated diphosphine ligand. This approach was applied to other functionalized amines, proving this concept is a suitable tool for the catalytic alkylation of highly functionalized amines. (Figure presented.).

Full text of DOI:10.1002/adsc.201701463

Title: Hydroaminomethylation in Aqueous Solvent Systems – An
Efficient Pathway to Highly Functionalized Amines
Authors: Thiemo Alexander Faßbach, Fridolin Sommer, and Andreas
Vorholt
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.201701463
Link to VoR: http://dx.doi.org/10.1002/adsc.201701463

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
Hydroaminomethylation in Aqueous Solvent Systems – An Efficient
Pathway to Highly Functionalized Amines
Thiemo A. Faßbach, Fridolin O. Sommer, Andreas J. Vorholt*
Fakultät für Bio- und Chemieingenieurwesen, Lehrstuhl für Technische Chemie, Technische Universität
Dortmund, Emil-Figge-Straße 66, 44227 Dortmund (Germany)
*E-Mail: andreas.vorholt@tu-dortmund.de
Abstract: Tandem-catalyzed hydroaminomethylation is a useful tool for synthesizing linear amines from olefins
and amines in an atom-efficient manner. To enable the coupling of highly functionalized, hydrosoluble amines
with non-water-soluble olefins, this reaction must be transferred to aqueous biphasic solvent systems. In this
work, we systematically evaluate reaction conditions to provide a selective hydroaminomethylation of 1-octene
with diethanolamine as model substrates. Although water is both the condensation side product and the solvent,
yields of 79% were achieved using a catalytic system consisting a rhodium precursor and a sulfonated
diphosphine ligand. This approach was applied to other functionalized amines, proving this concept is a suitable
tool for the catalytic alkylation of highly functionalized amines.
Introduction
Linear amines are of great importance in the production of pharmaceuticals, agrochemicals
or surfactants.[1,2] There are numerous ways to synthesize amines, e.g. the Buchwald-
Hartwig amination.[3,4] Hydroaminomethylation provides another means of catalytic synthesis
of higher amines and is a term that describes a tandem catalyzed reaction consisting of a
hydroformylation and a subsequent reductive amination (Scheme 1).[5,6]
Scheme 1: General hydroaminomethylation reaction sequence.
Hydroformylation and hydrogenation, which are both transition metal catalyzed reaction
steps, are irreversible. The enamine condensation step in between is an equilibrium. This
means that high amounts of water and/or enamine hamper the reaction in a whole.
In general, rhodium catalysts are used, as they show high activities for both the
hydroformylation and the hydrogenation step under mild conditions.[7] Since the
hydroformylation step can be adjusted to ensure high selectivities to the linear aldehyde, it is
an efficient method for synthesizing anti-Markovnikov products. Identifying catalytic
conversions that enable linear amination is among the most challenging tasks in chemistry
research.[8]
The olefin and the amine are typically dissolved in the same phase, maintaining high reaction
rates, especially for the condensation reaction.[9,10] Highly polar amines, e.g. amines
functionalized with carboxylic acids or hydroxyl groups, cannot generally be coupled with
olefins in this way. Amines other than relatively short alkyl amines are thus rarely reported in
the literature. Thermomorphic solvent systems (TMS)[11] provide an elegant solution to this
issue, though these have to be developed for every single substrate combination specifically
and are often limited to amines of medium polarity.[12]
The use of aqueous solvent systems in the hydroaminomethylation could enable
hydrocarbons to be coupled to highly polar amines. The spectrum of products that result
could include pharmaceuticals or surfactants. However, the enamine condensation is a
reversible reaction that incorporates the formation of water, which poses a great challenge.
Accordingly, working in aqueous solutions that shift the equilibrium of that reaction to the
aldehyde intermediate is crucial. The first hydroaminomethylation of ammonia with olefins in
aqueous solvent systems was reported by Beller. Using sulfonated diphosphines and an
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
iridium co-catalyst, olefins up to the chain length of 1-pentene were converted to the
respective primary amine.[13] Another example for this is the hydroaminomethylation of
isoprene.[14]
.
The approach for the hydroaminomethylation of higher olefins in aqueous solvent systems
was only realized using additional surfactants, such as quaternary ammonium compounds.[15]
Weberskirch took this approach a step further, using triphenylphosphine functionalized
poly(2-oxazoline)s as ligands for the substrates 1-octene and dimethylamine to react inside
micelles with a bimetallic catalytic system.[16] The conversion of substrates with strongly
opposing polarities without additives has yet to be described in the literature.
This work presents the development of an efficient hydroaminomethylation of water-soluble
amines with non-polar olefins that focuses on a simple, viable procedure. Crucial factors for
achieving this goal include the condensation equilibrium and the sufficient mixing of the
substrates. The influence of temperature and substrate ratios on these crucial factors are
investigated systematically. Selecting a proper solvent is an important precondition for these
factors. These investigations are carried out using 1-octene and diethanolamine as model
substrates with strongly opposing polarities. At the end a set of reaction conditions is put
together and then transferred to other amines.
Results and Discussion
To find a suitable catalytic system as well as a solvent system that enables the conversion of
water-soluble amines and olefins, which are insoluble in water, the model substrates
diethanolamine and 1-octene were chosen. Diethanolamine is a functionalized and highly
water-soluble amine.[17] On the other hand, 1-octene is insoluble in water (4.1 mg/L, 25 °C)[18]
and a widely investigated substrate for hydroaminomethylations. The reaction sequence that
follows from these model compounds is depicted in Scheme 2.
Scheme 2: Hydroaminomethylation of a 1-alkene with diethanolamine (only the linear hydroformylation pathway
is shown).
The first step in this particular sequence is the rhodium catalyzed hydroformylation of
1-octene to nonanal 1. The formation of the iso-product 2-methyl-octanal occurs as a side
product (this pathway is omitted in Scheme 2). The resulting n-nonanal (1) can undergo three
different reactions under these conditions: since rhodium is active as a hydrogenation
catalyst, the formation of alcohol 2 is possible. This reaction has been observed, especially in
the presence of amines.[19–21] Another side reaction - irreversible under these reaction
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
conditions - is the aldol condensation, leading to ,-unsaturated branched aldehydes 3. This
reaction is important as it can be catalyzed by secondary amines under hydroformylation
conditions.[22,23]
The condensation of nonanal with diethanolamine yields the corresponding enamine 4. This
reaction is an equilibrium, which means that it is not favored in aqueous systems.
Consequently, the hydrogenation reaction to the desired amine 5 must occur quickly. The
yields (Yx) for the mentioned (intermediate) products are discussed, as well as the selectivity
for the desired amine (S5). In this context, the n- and iso-products are summarized and
branched products that would result from a hydroformylation of internal double bonds were
not observed over the course of our investigations.
Water and 1-butanol proved to be a good choice along with the sulfonated Xantphos ligand
for the hydroformylation in aqueous biphasic systems (Figure 1).[24] The rhodium precursor
[Rh(cod)Cl]2 has shown to be active in many reported hydroaminomethylations[11,14,15,25] and
was employed with the Xantphos ligand in the first evaluations of the organic solvent (Table
1).
Since the functionalized amine diethanolamine is highly water-soluble (log Poct/wat = -1.43[17]
)
and the employed alkene (1-octene) is not, the choice of the organic co-solvent for the
aqueous system is of great importance both in terms of conversion and selectivity. It must
ensure the availability of substrates to the catalyst and provide for sufficient mixing of the
amine and the aldehyde formed. If too little amine is present in the phase containing the
aldehyde, aldol condensation is favored.[26] For this purpose, several short-chain alcohols
were employed (Table 1).
Table 1: Results using different organic co-solvents.
Entry
1.1
Organic solvent
-/-
Y1
0
Y2
0
Y3
0
Y4
0
Y5
0
S5
-/-
1.2
1.3
1.4
1.5
1-butanol
2-butanol
tert-butanol
iso-propanol
ethanol
4
8
9
4
16
5
9
7
2
5
6
4
9
0
0
0
0
30
34
34
30
55
64
61
60
31
1.6
1
17
0
9
Reaction
conditions:
c([Rh(cod)Cl]2)=0.2
mol%,
n(1-octene)=6 mmol,
ligand=Xantphos,
M/P=1/5
n(diethanolamine)=12 mmol, V(Co-solvent)=2.5 mL, V(H2O)=2.5 mL, p(CO/H2)=40 bar, CO/H2=1/1, T=140 °C,
t=6 h, 500 rpm. Yields (Y) are given in % based on 1-octene. S5=Y5/(Y1+Y2+Y3+Y4+Y5). Results determined by
GC-FID using dodecane as internal standard.
Without any additional solvent the reaction did not take place (Entry 1.1), which can be
explained by the insufficiently dissolved 1-octene in the aqueous phase. Ethanol (Entry 1.6)
showed a different influence, with significantly higher amounts of aldol condensation product
formed, which led to decreased selectivity. This could be because it is the only applied
organic solvent that maintained one homogeneous phase after the reaction. When butanols
were used as a solvent (Entry 1.2 – 1.4), phase separation took place. It was reported that
the solvent system consisting of water and 1-butanol is not completely homogenous, even at
elevated temperatures of 140 °C and above, meaning this system constitutes a “narrow
TMS” (TMS= thermomorphic solvent system).[24] This finding was verified for the substrates
present under reaction conditions using a reactor equipped with a sight-glass. Because the
conversion using 1-butanol was the highest, subsequent investigations were carried out
using this solvent system. The aldol condensation, however, was still an important side
reaction, reducing the selectivity to 55% (Entry 1.2). This side reaction was affected by the
amine/aldehyde ratio, which was set at the beginning with a proper alkene/amine ratio (Table
2).
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
Table 2: Results of the investigations regarding the substrate ratio
Entry
2.1
2.2
2.3
2.4
Alkene/amine
Y1
11
3
6
4
Y2
8
3
3
7
Y3
5
31
4
7
3
Y4
0
0
0
0
Y5
21
30
33
35
18
S5
47
45
72
66
72
1/0.5
1/1
1/1.5
1/3
2.5
1/4
1
3
0
Reaction
conditions:
c([Rh(cod)Cl]2)=0.2
mol%,
n(1-octene)=6 mmol,
ligand=Xantphos,
M/P=1/5,
V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, p(CO/H2)=40 bar, CO/H2=1/1, T=140 °C, t=6 h, 500 rpm. Yields (Y) are
given in % based on 1-octene. S5=Y5/(Y1+Y2+Y3+Y4+Y5). Results determined by GC-FID using dodecane as
internal standard.
As expected, the selectivity for 5 employing sub-stoichiometric and stoichiometric amounts of
amine was very low (Entry 2.1 and 2.2). Once an excess of amine is present, selectivity
increases (Entry 2.3 – 2.5). It was shown that the aldol condensation is favored in
hydroaminomethylations if low amounts of amines are present, whereas high amounts of
amine, as compared to aldehydes, tend to form enamines.[26] The highest yield for amines
(35%) was achieved at an alkene to amine ratio of 1/3. Since the amount of diethanolamine
affects phase behavior and, more importantly, the tendency to form aldol condensates,
subsequent investigations were conducted using this substrate ratio.
Since the tandem catalyzed reaction sequence involves two different catalytic processes,
hydroformylation and hydrogenation, the temperature is crucial in terms of adjusting the two
reactions as desired. Temperature also has a strong influence on the miscibility of all
compounds, substrates and solvents. The results achieved at different temperatures are
summarized in Table 3.
Table 3: Results of the investigations regarding the temperature
Entry
3.1
3.2
3.3
3.4
T [°C]
80
100
110
120
130
170
Y1
14
5
5
7
Y2
5
3
1
<1
Y3
5
<1
8
7
6
Y4
0
0
0
0
Y5
76
53
52
28
36
29
S5
76
85
79
67
72
57
3.5
3.6
5
4
3
15
0
0
3
Reaction
conditions:
c([Rh(cod)Cl]2)=0.2
mol%,
n(1-octene)=6 mmol,
n(diethanolamine)=18 mmol,
ligand=Xantphos, M/P=1/5, V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, p(CO/H2)=40 bar, CO/H2=1/1, t=6 h, 500 rpm.
Yields (Y) are given in % based on 1-octene. S5=Y5/(Y1+Y2+Y3+Y4+Y5). Results determined by GC-FID using
dodecane as internal standard.
For instance, a higher temperature increases the miscibility of the liquid phases, whereas the
solubility of gaseous compounds decreases. Temperature also strongly affects the reaction
rates of hydroformylation and hydrogenation. Hydroformylation starts at lower temperatures
than hydrogenation, whereas hydrogenation benefits from higher temperatures more than
hydroformylation does.[6] Surprisingly, the yield of the desired amines was highest (76%) at a
low temperature of 80 °C (Entry 3.1). The highest selectivity observed (85%) was achieved at
100 °C (Entry 3.2), at which point on, selectivity and conversion rates decrease (Entry 3.3 –
3.5). At 170 °C (Entry 3.6) black precipitation indicates that the rhodium catalyst is
irreversibly deactivated. In the present case, lower temperatures seem to favor the desired
reaction pathway. This is most probably due to the temperature-dependent solubility of
hydrogen and carbon monoxide in water.[27]
Comparing the Henry’s law constants of the gaseous compounds in water, solubility
decreases strongly as temperature increases (for details see SI). With temperatures higher
than 50 °C hydrogen is more soluble than CO in the aqueous phase. This is favorable,
because more hydrogen is consumed than carbon monoxide during the reaction. Comparing
the solubility of H2 and CO at 140 °C with their solubility at 100 °C, an increase of 16% for
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
hydrogen and 44% for CO is given at the lower temperature, which explains the better results
at lower temperatures. This effect only seems possible because of the phase-mediating
butanol that makes the olefin substrate available in the aqueous phase.
Taking the considerations from the temperature-dependent solubilities of the syngas
components into account, an excess of hydrogen in the syngas mixture may be
advantageous for the reaction. The selectivity as well as the yield for the desired amines was
the highest using a threefold excess of hydrogen in the syngas mixture (Table 4, Entry 4.2)
though there is no clear trend in selectivity.
Table 4: Results of the investigations on syngas composition
Entry
4.1
4.2
4.3
4.4
CO/H2
1/2
1/3
1/4
1/5
Y1
6
3
3
7
Y2
7
3
8
4
Y3
3
1
1
0
Y4
0
0
0
0
Y5
23
23
20
4
S5
58
77
63
27
Reaction
conditions:
c([Rh(cod)Cl]2)=0.2
mol%,
n(1-octene)=6 mmol,
n(diethanolamine)=18 mmol,
ligand=Xantphos, M/P=1/5, V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, T=140 °C, p(CO/H2)=40 bar, CO/H2=1/1, t=6 h,
500 rpm. Yields (Y) are given in % based on 1-octene. S5=Y5/(Y1+Y2+Y3+Y4+Y5). Results determined by GC-FID
using dodecane as internal standard.
Having tested different pressures, 50 bar was shown to increase conversion rates. Finally,
the amount of rhodium, as well as the metal/ligand ratio was investigated (Table 5). Both are
important factors that affect conversion and selectivity rates within the reaction, as they
determine the structure and composition of the metal complexes in solution.
Table 5: Results of different precursor concentrations.
Entry
5.1
5.2
5.3
5.4
5.5
5.6
5.7
[Rh(cod)Cl]2
0.06 mol%
0.13 mol%
0.2 mol%
0.2 mol%
0.2 mol%
0.2 mol%
0.4 mol%
M/P
1/5
1/5
1/5
1/2
1/7
1/15
1/5
Y1
2
3
4
3
3
<1
2
Y2
2
2
16
1
1
Y3
21
20
5
4
6
Y4
0
0
0
0
0
0
0
Y5
27
33
30
27
21
12
20
S5
52
57
55
77
66
50
61
6
3
6
8
Reaction conditions: n(1-octene)=6 mmol, ligand=Xantphos, n(diethanolamine)=12 mmol, V(1-butanol)=2.5 mL,
V(H2O)=2.5 mL, M/P=1/5, p(CO/H2)=40 bar, CO/H2=1/1, T=140 °C, t=6 h, 500 rpm. Yields (Y) are given in %
based on 1-octene. S5=Y5/(Y1+Y2+Y3+Y4+Y5). Results determined by GC-FID using dodecane as internal
standard.
Low rhodium concentrations (Entry 5.1) achieved only low conversions, whereas high
amounts of catalyst led to increased aldol condensation (Entry 5.3). Going further, the
hydrogenation of the formed aldehyde to alcohol was observed (Entry 5.3). Since both
reactions, hydroformylation and hydrogenation, are catalyzed by the same metal catalyst and
ligand but with different complexations[28], the metal to ligand ratio plays an important role in
selectivity. Typically, with high metal/ligand ratios, the hydroformylation is favored, whereas
low ligand excesses favor hydrogenation.[9,29] Investigations on the metal/ligand ratio showed
no definitive trend in terms of reactivity. In general, the amine yield was higher for lower
amounts of phosphorous (Entry 5.4). Higher amounts of ligand lead to decreasing yields
(Entry 5.5 and 5.6). The selectivity was highest (77%) at a metal/phosphorous ratio of 1/2
(Entry 5.4).
Previous investigations of the influence of different reaction parameters led to new reaction
conditions, which can be considered optimized for hydroaminomethylation in aqueous
solvent systems (Table 6). A precursor concentration of 0.13 mol% ensures high conversion.
Apart from that, transition metal catalyzed side reactions, for example the hydrogenation of
aldehydes to alcohols, are low. This is also maintained by a low ligand excess (M/P=1/2),
which enables both hydroformylation of the olefin and hydrogenation of the enamine. As
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
stated before, high amounts of hydrogen are necessary for the desired reaction pathway.
One way to ensure this is the syngas composition (CO/H2=1/3) and pressure (50 bar).
Furthermore, a temperature of 100 °C enables fast catalytic conversion and high solubility of
the gaseous compounds.
Table 6: Chosen reaction conditions based on previous investigations
Parameter
Value
c([Rh(cod)Cl]2
Metal/phosphorous
Co-Solvent
0.13 mol%
1/2
1-butanol (V/V=1/1)
Syngas pressure
CO/H2
50 bar
1/3
Olefin/amine
Temperature
1/3
100 °C
Taking these reaction conditions into account, the choice of the ligand was re-evaluated. The
sulfonated Xantphos analog Sulfoxantphos as well as the mono-sulfonated
triphenylphosphine was tested (TPPMS, both see Figure 1). TPPMS have been shown to be
beneficial for homogeneously catalyzed reactions in aqueous biphasic reaction mixtures.[30]
In general, sulfonated and thus hydrosoluble ligands may enable recycling of the precious
metal catalyst because the product accumulates in the organic phase.
Figure 1: Structure of employed ligands
As can be seen in Table 7, the product yield was slightly higher using Sulfoxantphos (Entry
7.2) compared to the non-sulfonated ligand (Entry 7.1). Interestingly, there was some
aldehyde left, which could still be converted to the desired product. The TPPMS ligand
showed lower activity than the Xantphos ligands.
Table 7: Results with different ligands under optimized reaction conditions
Entry
7.1
7.2
Ligand
Y1
Y2
Y3
Y4
Y5
S5
Xantphos
Sulfoxantphos
TPPMS
1
13
16
5
2
<1
9
5
3
0
0
0
62
64
51
81
76
73
7.3
Reaction conditions: c([Rh(cod)Cl]2)=0.13 mol%, n(1-octene)=6 mmol, n(diethanolamine)=18 mmol, M/P=1/2,
V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, T=100 °C, p(CO/H2)=50 bar, CO/H2=1/3, t=6 h, 500 rpm. Yields (Y) are
given in % based on 1-octene. S5=Y5/(Y1+Y2+Y3+Y4+Y5). Results determined by GC-FID using dodecane as
internal standard.
A yield-time plot was recorded for these reaction parameters to obtain a deeper
understanding for the course of the reaction. The resulting plot is depicted in Figure 2.
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
100
90
80
70
60
50
40
30
20
10
Y1
Y4
Y2
Y5
Y3
X
0
0
100
200
300
t /min
Figure 2. Yield-time plot under chosen reaction conditions.
400
500
600
Reaction
conditions:
c([Rh(cod)Cl]2)=0.13
mol%,
n(1-octene)=6 mmol,
n(diethanolamine)=18 mmol,
ligand=Sulfoxantphos, M/P=1/2, V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, T=100 °C, p(CO/H2)=50 bar, CO/H2=1/3,
500 rpm. Yields (Y) are given in % based on 1-octene. X=Y1+Y2+Y3+Y4+Y5. Results determined by GC-FID using
dodecane as internal standard.
Interestingly, the first 30 minutes pass without any olefin conversion. This might be explained
by the formation of the catalytically active complex. But after this time conversion takes place
rather slowly over the next two hours. This might be due to insufficient mixture between the
catalyst, which is most probably located in the aqueous phase, and the olefin, which
solubilizes in the organic phase. At this point, a yield of only 7% of the desired amine (Y5)
and an overall conversion of the olefin at 10% was observed. The conversion rate increased
significantly afterwards, leading to an overall yield of 40% after four hours and 50% after
eight hours. It should be stated that at no point was a significant amount of the enamine 4
detected. This is rather exceptional for the time profile of hydroaminomethylations[26,29]
.
Usually, the enamine is formed instantaneously from the aldehyde produced and the
subsequent hydrogenation to the saturated amine takes place successively. On the one
hand, the enamine is not very stable in aqueous media and, consequently, only the
hydrogenated amine, which is no longer in equilibrium, is detected. On the other hand, the
respective aldehyde is not detected in large amounts either. This indicates fast
hydrogenation under these conditions and the hydroformylation step is decisive in
determining the rate. This differentiates the reaction under aqueous biphasic conditions
substantially from hydroaminomethylation, which is typically conducted in one homogenous
phase. In the latter case the hydroformylation is very fast.
It would thus be of interest to use the Sulfoxantphos ligand due to its selectivity and
reactivity, while avoiding the lengthy catalyst formation time. Preformation provided a suitable
workaround, whereby the catalyst was preformed three hours in advance without the olefin in
a syngas atmosphere. This was conducted at room temperature and at reaction temperature
(T=100 °C). After three hours, the olefin was added via a reservoir with syngas pressure. The
results of the following reaction (t=8 h) differ clearly (Figure 3).
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
100
90
80
70
60
50
40
30
20
10
0
Y5
Y2
Y4
Y1
Y3
room temperature
Figure 3. Comparison of the results using catalyst preforming under room temperature and reaction temperature
Reaction conditions: c([Rh(cod)Cl]2)=0.13 mol%, n(1-octene)=6 mmol, n(diethanolamine)=18 mmol,
reaction temperature
ligand=Sulfoxantphos, M/P=1/2, V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, T=100 °C, p(CO/H2)=50 bar, CO/H2=1/3,
500 rpm. Yields (Y) are given in % based on 1-octene. Results determined by GC-FID using dodecane as internal
standard.
Preforming the catalyst at reaction temperature significantly enhances the reactivity of the
catalytic system. After eight hours a yield of 79% of the desired amines was achieved,
constituting the best result of the investigations presented. The active catalyst is not only
formed from the rhodium precursor and the phosphine ligand but also from syngas to form a
rhodium carbonyl hydride complex, as described in the literature.[31] To form this catalytically
active species, higher temperatures and syngas pressure are required, especially with regard
to phase behavior and the availability of carbon monoxide and hydrogen in the liquid phases.
Since the reaction mixture is biphasic after the reaction has taken place, the experiment
procedure was altered slightly to evaluate the potential for catalyst recycling. In particular, the
phase separation time was shortened to a few minutes to avoid catalyst deactivation (see
Experimental section). Since non-converted octene is completely dissolved in the product
phase (1-butanol) it was completely replenished after each reaction. The amine is employed
in a threefold excess and as a consequence, only a third of the original amount is
complemented. The results are shown in Figure 4.
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
6
5
4
3
2
1
0
Rh
P
Y(5)
1st reaction
Figure 4. Reaction yields and leaching of the catalyst in recycling experiments.
Reaction conditions: c([Rh(cod)Cl]2)=0.13 mol%, n(1-octene)=6 mmol,
2nd reaction
3rd reaction
n(diethanolamine)=18 mmol,
ligand=Sulfoxantphos, M/P=1/2, V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, T=100 °C, p(CO/H2)=50 bar, CO/H2=1/3,
500 rpm. Yields (Y) are given in mmol. Results determined by GC-FID using dodecane as internal standard.
Recycling conditions: Recovered aqueous phase, n(1-octene)=6 mmol, n(diethanolamine)=6 mmol,
V(1-butanol)=2.5 mL, T=100 °C, p(CO/H2)=50 bar, CO/H2=1/3, 500 rpm. Leaching determined by ICP-OES and is
given as a percentage of the initial amount. Yields (Y) are given in mmol. Results determined by GC-FID using
dodecane as internal standard.
It can be seen that the yield in the first reaction (23%) is significantly lower than in the
corresponding batch experiment (64%). This is most probably due to the shortened phase
separation period, leading to incomplete product extraction in the organic phase. Catalyst
leaching into the organic phase in this reaction was measured at 7% and 4% for rhodium and
phosphorous, respectively. While this is not sufficient at an industrial scale, the catalyst can
still be reused in the next cycle, as shown. The yield of the second reaction increases (50%),
though at this point this could be product from the first reaction. Unfortunately, catalyst
leaching increases. At the start of the third reaction, a maximum of 76% of the initial amount
of rhodium and 85% of the phosphorous was left. It should be mentioned that neither the
metal nor the ligand were replenished at any time. The third reaction resulted in a yield of
75%, which surely consists of product left from the first two cycles. Nevertheless, the catalyst
was used 2.3 times more efficiently as compared to the batch experiment (Table 7, Entry
7.2). After the third reaction less than half of the catalyst compounds are likely to be present
in the reaction mixture. Furthermore, after the fourth reaction, phase separation did not take
place. A possible explanation for this behavior might be the accumulated 1-butanol in the
aqueous phase. Since the aqueous catalyst phase was not analyzed and the phase
separation was shortened, the amount of leached organic solvent was not determined.
Complete replenishment then leads to an accumulation of 1-butanol, which results in an
outright miscibility of all compounds.
Potential optimizations of this recycling scheme should include a detailed mass balance of all
involved compounds, especially the solvents. Organic solvent that leaches into the aqueous
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
phase interferes with the intended phase behavior and the amount of water formed during
condensation may be an important factor after numerous recycling runs.
Differently functionalized amines were employed to ultimately verify this approach for
conducting hydroaminomethylation in aqueous solvent systems to successfully convert
olefins to functionalized amines (Figure 5).
Reaction conditions: c([Rh(cod)Cl]2)=0.13 mol%, n(1-octene)=6 mmol, n(amine)=18 mmol, ligand=Sulfoxantphos,
M/P=1/2, V(1-butanol)=2.5 mL, V(H2O)=2.5 mL, T=100 °C, p(CO/H2)=50 bar, CO/H2=1/3, 500 rpm. Yields (Y) are
given in % based on 1-octene. Results determined by GC-FID using dodecane as the internal standard.
Figure 5. Substrate scope with different amines
The amino polyol N-methylglucamine is produced by the reductive amination of sucrose and
is
a
promising building block for creating non-ionic surfactants. Enabling the
hydroaminomethylation with this substrate may yield interesting new non-ionic surfactants
that are partially based on a renewable feedstock. Under the conditions mentioned in Table
6, a 50% yield of N,N-methyl-nonyl glucamine was achieved. Employing N-methyl glycine, a
synthetic amino acid, led to yields of 62%. This means that in general, polar amino acids can
be functionalized using the approach presented. In the case of N-methyl glycine the resulting
long chain substituted amino acid could be further functionalized to obtain betaines. To
stress this concept, a highly non-polar amine, N,N-phenyl-naphthyl amine, was employed as
well. This leads to 57% yield, demonstrating that this concept also works with amines that
are not soluble in water.
Conclusion
To produce long chain substituted and highly functionalized amines, it is of interest to employ
water-soluble amines in the hydroaminomethylation. With the model substrates 1-octene and
diethanolamine crucial reaction parameters, i.e. temperature, pressures and concentrations,
were
investigated
systematically.
The
investigations
demonstrate
that
hydroaminomethylation can be conducted efficiently in aqueous solvent systems, though it
incorporates a condensation reaction, forming water. Fast hydrogenation of the intermediate
enamine ensures high yields of the desired amines. For this, a comparably low temperature
is favorable, which is uncommon for hydroaminomethylations. The use of the water-soluble
Sulfoxantphos ligand enables high yields and selectivities, but shows a long catalyst
formation. This can be overcome by preforming the catalyst in the absence of the olefin at
reaction temperature. Finally, this model was applied to amino acids and amino polyols
(N-methyl glycine and N-methyl glucamine). Because N-methylglucamine in particular is
difficult to dissolve in organic solvents, a conversion in aqueous solvent systems is highly
beneficial. Phase separation after the reaction was successfully used for catalyst recycling,
although leaching of solvent and catalyst in the product phase occurs.
The investigations will be extended to different alkenes and terpenes. Furthermore, other
water-soluble amines could be employed.
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
Experimental Section
Chemicals were purchased from Acros Organics (Geel, Belgium), Sigma-Aldrich (Steinheim,
Germany) and TCI (Tokyo, Japan).
The catalyst and ligand were weighed directly into the reaction vessel. The amine was
employed as a stock solution in water, the 1-octene as a stock solution in the organic
solvent. The reactor was closed and pressurized with syngas. The reaction was terminated
by cooling the reactor in an ice-bath. The reaction mixture was transferred to culture tubes
and phase separation took place over night. The product phase was then removed and
analyzed using GC-FID analytics. In addition, the aqueous phase was analyzed by GC-FID,
but no products and/or substrates could be found in any of the samples collected.
Several identical reactions were started in individual reactors and terminated after the given
time to produce the yield-time plot.
In the course of the experiment, all reaction compounds, except the 1-octene, were added
separately to the reaction vessel. As with the previously-mentioned method, the vessel was
closed, treated with syngas and stirred for 3 hours at the desired temperature. After the
reactor was cooled down in an ice-bath, the syngas was devolatilized, and the reaction
solution was stored in an argon atmosphere to insert the 1-octene. Further reaction steps
were performed as described.
The recycling experiments were conducted as mentioned above with slight alterations. In
order to avoid catalyst decomposition, phase separation was conducted in an argon-flushed
schlenk tube and the phase separation time was limited to a few minutes. The product
phases from the recycling experiments were also investigated to evaluate catalyst leaching
(rhodium and phosphorous) using ICP-OES. The replacement of the removed phase was
realized in the form of a stock solution consisting of the alkene, amine and organic solvent.
All products were purified by silica gel column chromatography and calibrated for GC-FID
analysis with the internal standard. Routine gas chromatographic analyses were performed
on an Agilent 7890B instrument (Santa Clara, USA) equipped with a flame ionization detector
(FID) and a HP-5 capillary column (30 m, diameter 0.32 mm, film thickness 0.25 µm)
connected to an auto sampler (7693) and an injector (G4513A). GC-MS analyses of the
products were carried out on an Agilent 5977A MSD (70 eV).
NMR spectra were recorded on Bruker DRX spectrometers. CDCl3 was used as solvent and
standard for chemical shifts, purchased from Deutero.
Spectroscopic data for the main product of these investigations (5):
1H-NMR (400 MHz, d4-MeOH): ppm () = 4.20 – 3.98 (m, 1H), 3.92 – 3.78 (m, 2H), 3.75 –
3.57 (m, 3H), 3.30 – 3.12 (m, 1H), 2.91 – 2.76 (m, 2H), 2.73 – 2.35 (m, 4H), 1.64 – 1.00 (m,
12H), 0.98 - 0.76 (m, 4H). 13C-NMR (100 MHz, d4-MeOH): ppm () = 96.82, 64.27, 61.16,
60.18, 55.19, 51.77, 50.96, 34.29, 32.01, 29.70, 25.17, 22.80, 14.25. EI-MS: m/z = 230 ([M]+,
< 1%), 201 (13), 200 (100), 118 (32), 88 (70), 74 (19), 58 (5), 57 (6), 56 (14), 55 (8).
The spectroscopic data of all other products can be found in the supplementary information.
Acknowledgements
The financial support provided by Clariant is gratefully acknowledged.
We thank Umicore for the donation of the rhodium precursors.
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
References
[1]
P. Roose, K. Eller, E. Henkes, R. Rossbacher, H. Höke, in Ullmann’s Encyclopedia of
Industrial Chemistry, Wiley-VCH, Weinheim, Germany, 2015.
K. Kosswig, in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim,
Germany, 2000.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
P. Ruiz-Castillo, S. L. Buchwald, Chem. Rev. 2016, 116, 12564–12649.
F. Paul, J. Patt, J. F. Hartwig, J. Am. Chem. Soc. 1994, 116, 5969–5970.
C. Chen, X.-Q. Dong, X. Zhang, Org. Chem. Front. 2016, 3, 1359–1370.
D. Crozet, M. Urrutigoïty, P. Kalck, ChemCatChem 2011, 3, 1102–1118.
T. Rische, P. Eilbracht, Synthesis 1997, 1997, 1331–1337.
M. Beller, J. Seayad, A. Tillack, H. Jiao, Angew. Chem. Int. Ed. 2004, 43, 3368–3398.
D. S. Melo, S. S. Pereira-Júnior, E. N. dos Santos, Appl. Catal., A 2012, 411–412, 70–
76.
[10] K. C. B. Oliveira, A. G. Santos, E. N. dos Santos, Appl. Catal., A 2012, 445–446, 204–
208.
[11] A. Behr, T. Seidensticker, A. J. Vorholt, Eur. J. Lipid Sci. Technol. 2014, 116, 477–
485.
[12] J. M. Dreimann, T. A. Faßbach, S. Fuchs, M. R. L. Fürst, T. Gaide, R. Kuhlmann, K. A.
Ostrowski, A. Stadler, T. Seidensticker, D. Vogelsang, H. W. F. Warmeling, A. J.
Vorholt, Chem. Ing. Tech. 2017, 89, 252–262.
[13] B. Zimmermann, J. Herwig, M. Beller, Angew. Chem. Int. Ed. 1999, 38, 2372–2375.
[14] A. Behr, S. Reyer, V. Manz, Chem. Ing. Tech. 2012, 84, 108–113.
[15] A. Behr, A. Wintzer, Chem. Eng. Technol. 2015, 38, 2299–2304.
[16] B. Gall, M. Bortenschlager, O. Nuyken, R. Weberskirch, Macromol. Chem. Phys.
2008, 209, 1152–1159.
[17] M. Wang, L. Fang, C. Ren, T. Li, J. Pharm. Pharmacol. 2008, 60, 429–435.
[18] Y. B. Tewari, M. M. Miller, S. P. Wasik, D. E. Martire, J. Chem. Eng. Data 1982, 27,
451–454.
[19] S. Fuchs, D. Lichte, M. Dittmar, G. Meier, H. Strutz, A. Behr, A. J. Vorholt,
ChemCatChem 2017, 9, 1436–1441.
[20] T. Vanbésien, E. Monflier, F. Hapiot, Green Chem. 2017, 19, 1940–1948.
[21] T. Vanbésien, E. Monflier, F. Hapiot, Green Chem. 2016, 18, 6687–6694.
[22] S. Chercheja, S. K. Nadakudity, P. Eilbracht, Adv. Synth. Catal. 2010, 352, 637–643.
[23] X. Fang, R. Jackstell, R. Franke, M. Beller, Chem. - Eur. J. 2014, 20, 13210–13216.
[24] T. Gaide, J. M. Dreimann, A. Behr, A. J. Vorholt, Angew. Chem. Int. Ed. 2016, 55,
2924–2928.
[25] A. Behr, D. Levikov, E. Nürenberg, RSC Adv. 2015, 5, 60667–60673.
[26] T. Seidensticker, J. M. Vosberg, K. A. Ostrowski, A. J. Vorholt, Adv. Synth. Catal.
2016, 358, 610–621.
[27] R. Fernández-Prini, J. L. Alvarez, A. H. Harvey, J. Phys. Chem. Ref. Data 2003, 32,
903–916.
[28] D. Crozet, A. Gual, D. McKay, C. Dinoi, C. Godard, M. Urrutigoïty, J.-C. Daran, L.
Maron, C. Claver, P. Kalck, Chem. - Eur. J. 2012, 18, 7128–7140.
[29] T. A. Faßbach, T. Gaide, M. Terhorst, A. Behr, A. J. Vorholt, ChemCatChem 2017, 9,
1359–1362.
[30] T. A. Faßbach, F. O. Sommer, A. Behr, S. Romanski, D. Leinweber, A. J. Vorholt,
Catal. Sci. Technol. 2017, 7, 1650–1653.
[31] R. Franke, D. Selent, A. Börner, Chem. Rev. 2012, 112, 5675–5732.
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701463
Advanced Synthesis & Catalysis
Graphic Abstract – TOC File
This article is protected by copyright. All rights reserved.