Palladium- and Nickel-Catalyzed Synthesis of Sodium Acrylate from Ethylene, CO2, and Phenolate Bases: Optimization of the Catalytic System for a Potential Process

Article abstract of DOI:10.1002/ejoc.201501113

The synthesis of sodium acrylate through catalytic carboxylation of ethylene with CO₂ in the presence of a base is a reaction of high interest. To develop a more efficient and sustainable method to access this valuable acrylate monomer, we optimized the system in a one-step homogeneous nickel- or palladium-catalyzed reaction, without the need for stoichiometric amounts of an additional reducing agent. Suitable nontoxic solvents such as anisole instead of the previously reported tetrahydrofuran or chlorobenzene were found to lead to acrylate formation. In combination with appropriate phenolate bases, this could allow a rational process concept for a simple catalyst recycling, product separation, and base regeneration. The carboxylation of ethylene with CO₂ in the presence of phenolate bases to give sodium acrylate was improved to develop the basis for a continuous process concept. The homogeneous nickel- or palladium-catalyzed transformation does not require stoichiometric reductants, and appropriate solvents as well as bases were identified that are suitable for a first recycling protocol.

Full text of DOI:10.1002/ejoc.201501113

FULL PAPER
DOI: 10.1002/ejoc.201501113
Palladium- and Nickel-Catalyzed Synthesis of Sodium Acrylate from Ethylene,
CO2, and Phenolate Bases: Optimization of the Catalytic System for a
Potential Process
Simone Manzini,[a] Núria Huguet,[a,b] Oliver Trapp,[a,c] and Thomas Schaub*[a,b]
Keywords: Synthetic methods / Process development / Homogeneous catalysis / Palladium / C1 building blocks /
High-pressure chemistry
The synthesis of sodium acrylate through catalytic carboxyl-
ation of ethylene with CO2 in the presence of a base is a
reaction of high interest. To develop a more efficient and sus-
tainable method to access this valuable acrylate monomer,
we optimized the system in a one-step homogeneous nickel-
or palladium-catalyzed reaction, without the need for stoi-
chiometric amounts of an additional reducing agent. Suitable
nontoxic solvents such as anisole instead of the previously
reported tetrahydrofuran or chlorobenzene were found to
lead to acrylate formation. In combination with appropriate
phenolate bases, this could allow a rational process concept
for a simple catalyst recycling, product separation, and base
regeneration.
sults of Hoberg et al.[5] The first catalytic carboxylation of
ethylene with CO2 in the presence of NaOtBu using a nickel
catalyst was described by our laboratory as a two-step pro-
cess (Scheme 1, a).[6] The first single-step catalytic system to
obtain lithium acrylate was then reported in 2014 by Vogt et
al.[7] (Scheme 1, b) and towards the desired sodium acrylate
in the same year by our laboratory (Scheme 1, c).[8]
Introduction
The carboxylation of ethylene with CO2 in the presence
of a base can provide an efficient and sustainable route
through which to access acrylate salts, which are of high
industrial importance. Sodium acrylate is produced from
acrylic acid on a large scale as a monomer for use in super-
absorbents.[1] This monomer is nowadays mainly produced
from propylene through a two-stage oxidation process.[2]
The direct synthesis of sodium acrylate from ethylene and
CO2 in the presence of a base could be a very attractive
route that would avoid the use of propylene as feedstock.
CO2 is inexpensive and abundant and, due to the increased
levels of ethane cracking, ethylene is becoming less expens-
ive than propylene. Moreover, ethylene could also be pro-
duced from renewable sources through the dehydration of
bioethanol.
The general strategies to use CO2 as a building block for
the preparation of carboxylic acids involve stoichiometric
or metal-catalyzed reactions activating C–H and C–X
bonds, and through hydrocarboxylation of olefins and alk-
ynes.[3] The carboxylation of olefins with CO2 to access un-
saturated carboxylic acids was investigated over the last
decades in stoichiometric reactions[4] based on the early re-
Scheme 1. Recent methodologies for the synthesis of acrylic acid
salts through carboxylation of ethylene.
[a] Catalysis Research Laboratory (CaRLa),
Im Neuenheimer Feld 584, 69120 Heidelberg, Germany
www.carla-hd.de
In the Vogt system,[7] a nickel-phosphine catalyst is used
and stoichiometric amounts of LiI, an amine base (NEt3),
and Zn are required to obtain the lithium acrylate, which
is a significant drawback for a commercial application. In
addition, lithium acrylate has limited commercial value be-
cause sodium acrylate is the desired product for the synthe-
sis of superabsorbents.
[b] Synthesis and Homogeneous Catalysis, BASF SE,
Carl-Bosch-Str. 38, 67056 Ludwigshafen, Germany
E-mail: thomas.schaub@basf.com
[c] Organisch-Chemisches Institut, Ruprecht-Karls-Universität
Heidelberg,
69120 Heidelberg, Germany
Supporting Information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201501113.
7122
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 7122–7130

Pd- and Ni-Catalyzed Synthesis of Sodium Acrylate
In our system, sodium acrylate can be obtained directly
when a nickel catalyst is used in combination with sodium-
2-fluorophenolate as the base and Zn as an additive to
achieve a TON of up to 107 (Scheme 1, c).[8]
Even though achieving the catalytic synthesis of sodium
acrylate was a significant result, before adapting this reac-
tion to a process, the following points have to be addressed:
(1) Simple separation of sodium acrylate from the reaction
mixture should be integrated into the design. (2) Catalyst
recycling should be possible. (3) The process should allow
high conversion of the base in the carboxylation and simple
base recycling as well as regeneration. (4) The formation
of possible side products from the Kolbe–Schmidt reaction
should be avoided. (5) The use of an additional stoichio-
metric reductant such as Zn should be avoided. (6) The new
process should increase the TON. (7) The use of toxic sol-
vents should be avoided.
Results and Discussion
As shown in previous work on nickel[6,8] and palladium
catalysts[9] reported from our laboratory and by Vogt et
al.,[7] the direct synthesis of acrylic acid from CO2 and eth-
ylene is thermodynamically not feasible.[6] However, the
carboxylation of ethylene is made exergonic by addition of
a base to generate the corresponding sodium acrylate
(Scheme 2).[6]
Scheme 3. Proposed mechanism for the nicke-catalyzed synthesis
of sodium acrylate using sodium phenolates as bases.
sible when water is added after the reaction. By using a
lipophilic and water-stable catalyst, the catalyst could also
be separated via the organic phase (see Figure 1). The so-
dium phenolate in the organic mixture can then be regener-
ated by addition of NaOH under removal of water after
separation of the product.[13] As shown previously in mech-
anistic studies from our laboratory, Zn is not intrinsically
necessary in the catalytic cycle and its use can, in principle,
be avoided.[6,10]
Scheme 2. Free energy for the synthesis of acrylic acid and for so-
dium acrylate.
Unfortunately, it is not possible to perform this reaction
by using NaOH as the base directly because stable carb-
onates are formed immediately by the reaction with
CO2.[6,8,10,11] For this reason it is necessary to use an appro-
priate base that can be used to deprotonate the metallalact-
one intermediate, but will not irreversibly react with CO2
under the reaction conditions (Scheme 3).[6,8,10]
Sodium salts of substituted phenols were found to be the
best compromise between reactivity and basicity (Scheme 1,
c).[8,10] Not all phenolates can be used in this transforma-
tion: di-ortho-substituted or ortho-deactivated phenols are
required to avoid carboxylation in this position (Kolbe–
Schmidt reaction).[12]
Figure 1. Process concept for the synthesis of sodium acrylate from
ethylene, CO2, and NaOH.
To develop an economic process concept, we focused on
adapting the system for use with lipophilic- and ortho-sub-
In this context, it was first necessary to identify whether
stituted phenols. In combination with low toxicity organic any of the previously reported Ni- and Pd-catalysts are
solvents with limited miscibility with water, a simple separa- compatible with di-ortho-substituted sodium phenolates as
tion of the sodium acrylate from the phenol would be pos- the base (Table 1).[14]
Eur. J. Org. Chem. 2015, 7122–7130
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7123

S. Manzini, N. Huguet, O. Trapp, T. Schaub
FULL PAPER
Table 1. Base screening for nickel and palladium catalysts.
cording to previous DFT calculations,[6,10] the two catalysts
were tested with catalytic amounts of Zn and without it.
In case of [PdCl2(COD)]/dcpe, it was possible to reduce
the amount of Zn to a tenth (from 10 to 1 mmol), where-
upon the TON was reduced by only a quarter (38 vs. 29;
see Table 2, entries 1–3). In the absence of Zn the activity
was halved (Table 2, entries 4).
Table 2. Assessing the effect of the amount of zinc and catalyst
precursor.
Entry[a] Catalyst
Additive Amount additive
[mmol]
TON[b]
1
2
3
4
5
6
7
8
PdCl2(COD)
Zn
10
1
0.5
–
10
1
–
–
–
–
–
38[c]
29
22
21
69[c]
56
55
11
14
18
0
–
Zn
Ni(COD)2
–
–
–
–
–
–
[a] Reaction conditions: base (20 mmol), catalyst (0.2 mmol), li-
gand (0.22 mmol), THF (30 mL), ethylene (10 bar), CO2 (20 bar),
20 h. [b] As [a] but 30 mmol of base was added. [c] TON deter-
Pd(OAc)2
9
[Pd(allyl)Cl]2
Pd(Cp)(allyl)
Pd(dba)2
10
11
12
1
mined based on H NMR analysis in D2O as solvent and using 3-
(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt (0.125 mmol)
as internal standard. [d] Result from previous work in our labora-
tory.[8] [e] Result from previous work in our laboratory.[9]
Pd(PPh3)4
–
28
[a] Reaction conditions: base (20 mmol), catalyst (0.2 mmol), dcpe
(0.22 mmol), THF (30 mL), ethylene (10 bar), CO2 (20 bar),
145 °C, 20 h. [b] TON determined based on 1H NMR analysis in
D2O as solvent and using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid
sodium salt (0.125 mmol) as internal standard. [c] The result pre-
According to the preliminary screening, di-alkyl substitu-
tion in the ortho-positions of the phenolate seems to sented in Table 1 is given here for comparison.
strongly effect the activity. In case of 2,6-dimethylphenolate
it was necessary to increase the temperature to 145 °C to
The nickel system without Zn showed a lower decrease
achieve significant conversion (Table 1 entries 7–10). The in the activity when compared with the system with the Pd
[Ni(COD)2]/BenzP* catalyst shows low activity with mono- catalyst (TON of 69 vs. 55; Table 2, entries 5–7), suggesting
and di-alkyl-substituted phenolates at 80 °C. By increasing that the benefit of Zn is to prevent possible side reactions
the temperature to 145 °C it was possible to achieve a TON such as oxidations, which could lead to catalyst deactiva-
of 43 by using 2,6-dimethylphenolate as the base. By chang- tion.
ing the ligand from BenzP* [(R,R)-(+)-1,2-bis(tert-butylm-
It has been found that the addition of ZnII species, in
ethylphosphanyl)benzene] to simple dcpe (bis-1,2-dicyclo- particular by using ZnCl2 and Zn(OTf)2, has no beneficial
hexylphosphinoethane) it is possible to increase the TON effect on the activity (see the Supporting Information). As-
to 69 with the nickel catalyst (Table 1, entries 8 and 9). With suming that the role of Zn is to activate the precatalyst gen-
sodium 2,6-difluorophenolate as the base, negligible cataly- erating the active Pd0 species, we tested further PdII precur-
sis was observed in both cases, probably due to the lower sors. Without Zn, the PdII precursors were less active com-
basicity compared to the other bases (Table 1, entries 3 and pared with [PdCl2(COD)] (Table 2, entries 8–10) as ex-
4). Although the Ni system seems to be superior compared pected according to previous investigations, Ni0 and Pd0
with Pd, several aspects of this reaction still needed to be complexes are active species in the catalytic cycle,[9] which
evaluated, hence both metals were considered for further makes them very sensitive towards traces of oxygen. It
optimization.
seems that the use of Zn can also reduce this sensitivity by
The main drawback in the previous systems was the use regenerating oxidized active species.[15]
of 50 equiv. Zn (relative to the catalyst) as a reductant to
In contrast, a TON of 28 was achieved when [Pd(PPh3)4]/
obtain a significant TON. The use of a solid reductant not dcpe was used as catalyst, which was the highest value ob-
only creates technical challenges in a continuous process, tained when using Pd as catalyst without the addition of
but also makes the system too expensive for an economic Zn (Table 2, entry 12). However, the use of a Pd0 precursor
sodium acrylate production and produces waste. Given that is not a guarantee for an active system, as shown with
Zn is not intrinsically needed for this transformation ac- [Pd(dba)2], which was inactive (Table 2, entry 11).
7124
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 7122–7130

Pd- and Ni-Catalyzed Synthesis of Sodium Acrylate
Given that tetrahydrofuran (THF) is completely miscible
The stability of the catalyst towards traces of water is a
with water and 2,6-dimethylphenol has a significant solubil- very important factor for the development of a continuous
ity in water, it was necessary to identify other solvents and system. After the product separation and base regeneration,
phenols that were suitable for the concept described above. traces of water can remain in the organic phase and could
Based on the catalyst screening, we evaluated the solvent have a negative impact on the catalysis. Therefore, we tested
and base compatibility for [PdCl2(COD)]/dcpe by using a the catalyst activity in solvents saturated with water (see the
catalytic amount of Zn (1 mmol) as well as for [Pd(PPh3)4]/ Supporting Information). Using Pd as catalyst in wet anis-
dcpe and [Ni(COD)2]/dcpe without an additional reduct- ole (water content for water saturated anisole is 2.9 g/L),[17]
ant.
no significant drop in activity was observed, which implies
A convenient solvent for the process should provide: that both catalysts can tolerate a certain amount of water
(1) A phase separation with water; (2) A high boiling point (Table 3, entries 7 and 8). In contrast, with Ni as catalyst,
to remove water in the phenolate regeneration step by distil- the activity dropped significantly, suggesting a low water
lation; (3) Solubility for the sodium phenolates; (4) Low tolerance (Table 3, entry 9).
toxicity. A series of solvents that fulfill these requirements
As shown in Table 4, full conversion of the base was not
were identified and screened with the three catalytic sys- achieved, which could lead to leaching of the base into the
tems. A representative comparison of the results is given in aqueous phase and contamination of the final product. It
Table 3 (for extended solvent screening see the Supporting was therefore necessary to adjust the amount of base to
Information).
achieve higher or even full base conversion, reducing the
phenolate loss into the aqueous phase. Leaching of base
was also due to the solubility of the phenolate in water,
which could be reduced by tuning the lipophilicity of the
base. Base optimization revealed that it was possible to use
half of the base loading with only a small decrease of TON
(Table 4). By using 5 mmol base, TON was greatly reduced,
suggesting an edge of activity at about 10 mmol of base
under the used conditions.
Table 3. Solvent screening.
Entry[a] Solvent
Catalyst
TON[d]
Table 4. Base loading screening for the PdCl2(COD)/dcpe/Zn sys-
tem.
1
2
3
4
5
6
7
8
THF
PdCl2(COD)/Zn[b]
Pd(PPh3)4
29[e]
28[e]
55[e]
45
43
44
39
40
22
6
0
47
1
23
3
Ni(COD)2
anisole
PdCl2(COD)/Zn[b]
Pd(PPh3)4
Ni(COD)2
anisole (wet)[c]
phenyl butyl ether
PdCl2(COD)/Zn[b]
Pd(PPh3)4
Entry[a]
Base [mmol] TON[b] Base conv. [%][c] Base in D2O [%][d]
9
Ni(COD)2
10
11
12
13
14
15
PdCl2(COD)/Zn[b]
Pd(PPh3)4
1
2
3
20
10
5
45[e]
39
14
45
78
56
29
16
10
Ni(COD)2
dibutyl glycol ether PdCl2(COD)/Zn[b]
Pd(PPh3)4
Ni(COD)2
[a] Reaction conditions: Pd (0.2 mmol), ligand (0.22 mmol), Zn
(1 mmol), solvent (30 mL), ethylene (10 bar), CO2 (20 bar), 145 °C,
1
20 h. [b] TON determined based on H NMR analysis in D2O as
[a] Reaction conditions: base (20 mmol), catalyst (0.2 mmol), dcpe
(0.22 mmol), solvent (30 mL), ethylene (10 bar), CO2 (20 bar),
145 °C, 20 h. [b] Zn (1 mmol) added. [c] Solvent saturated with
water (2.9 g/L). [d] TON determined based on 1H NMR analysis
in D2O as solvent and using 3-(trimethylsilyl)propionic-2,2,3,3-d4
acid sodium salt (0.125 mmol) as internal standard. [e] The result
presented in Table 2 is given for comparison.
solvent and using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium
salt (0.125 mmol) as internal standard. [c] Base conversion deter-
mined based on product formation and the amount of base added
at t = 0. [d] Base loss into the water phase determined after fil-
tration and evaporation, the solid residue was dissolved in D2O
and the same internal standard was used for the TON determi-
nation. [e] The result presented in Table 3 is given for comparison.
According to previous results for Pd- and Ni-sys-
Although reduced base loss can be achieved by lowering
tems,[6–10] the highest activity could be achieved with non- the base load (Table 4), the values remained too high, and
protic, oxygen-containing solvents. In case of the Pd cata- more lipophilic substituted sodium phenolates needed to be
lyst, the use of anisole increased the activity, raising the identified to reduce its solubility in water (Table 5).
TON from 28 to over 40 (Table 3, entries 1, 2, 4, and 5). In
The three catalysts showed a different behavior with
the case of the Ni catalyst, the activity seems quite constant, highly substituted phenolates. [PdCl2(COD)]/dcpe/Zn
reaching a TON of around 50 for THF, anhydrous anisole, showed the highest sensitivity towards base modifications.
and for butyl phenyl ether. For the other investigated sol- Increasing the steric hindrance in the ortho-positions re-
vents, only lower activities were observed. To compare the duced the activity greatly. For example, when the steric hin-
three catalysts directly, and to use a low toxicity solvent,[16] drance was increased from methyl, over isopropyl, to tert-
anisole was chosen for further optimization.
butyl substituents in both ortho-position, the TON dropped
Eur. J. Org. Chem. 2015, 7122–7130
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7125

S. Manzini, N. Huguet, O. Trapp, T. Schaub
FULL PAPER
Table 5. Base screening for the palladium system.[a,c]
of the substituents was observed with [Pd(PPh3)4]/dcpe as
the catalyst, and this system exhibited good TON and com-
plete base conversion even with the highly lipophilic
NaBHT base (sodium 2,6-di-tert-butyl-4-methylphenolate)
(Table 5, entry 11).
Regarding the properties of the base required in this pro-
cess concept and the observed activity, we investigated fur-
ther the scope of the [Pd(PPh3)4]/dcpe catalyst system in
the presence of NaBHT as base. To confirm that anisole
remained the optimal solvent, a solvent screening was un-
dertaken with this new combination of base (NaBHT) and
catalyst (Table 6). The results, given in Table 6, confirmed
that anisole was also the solvent of choice for our concept.
Table 6. Solvent screening with NaBHT as base.
Entry[a]
Solvent
TON[b]
1
2
3
4
toluene
anisole
phenyl butyl ether
dibutyl ether
dibutyl glycol ether
5
50[c]
1
3
23
10
[a] Reaction conditions: base (10 mmol), Pd (0.1 mmol), ligand
(0.11 mmol), solvent (30 mL), ethylene (10 bar), CO2 (20 bar),
145 °C, 20 h. [b] TON determined based on 1H NMR analysis in
D2O as solvent and using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid
sodium salt (0.125 mmol) as internal standard. [c] The result pre-
sented in Table 5 is given for comparison.
Another crucial point for the development of an appro-
priate process concept is metal leaching into the aqueous
phase. For simple catalyst separation and suitable product
quality (no heavy metals in the sodium acrylate), the cata-
lyst should be soluble in the organic phase and insoluble in
the product phase after the addition of water to the reaction
mixture. Given that the solubility of the catalyst is influ-
[a] Reaction conditions: base (10 mmol), catalyst (0.2 mmol), dcpe
(0.22 mmol), anisole (30 mL), ethylene (10 bar), CO2 (20 bar),
145 °C, 20 h. [b] As [a] but with 20 mmol base added. [c] TON
determined based on 1H NMR analysis in D2O as solvent and
using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt
(0.125 mmol) as internal standard. [d] Base conversion determined
based on product formation and on the amount of base added at enced by the lipophilicity of the ligand, ligand screening at
t = 0; n.d.: not determined. [e] Base loss into the water phase deter-
mined after filtration and evaporation, the solid residue was dis-
low catalyst loading was carried out with the optimized sys-
tem (Table 7).
solved in D2O and using the same internal standard for the TON
As shown in Table 7, chelating and alkyl-substituted
determination; The lower limit of the analysis is defined by the
phosphine ligands are necessary to obtain catalytically
active Pd species (Table 7, entries 1 and 2), which is consis-
tent with our previous results.[10] The bite angle seems to
play an important role in the activity. Ligands bearing spa-
cers that are longer than C3 or shorter than C2 are not
assumption that the sensitivity of the NMR analysis is ca. 1 mg
phenol in 0.5 mL of solvent; n.d. = not determined. All the base
loss values reported are referred to the reaction with Pd(PPh3)4. [f]
The result presented in Table 4 is given for comparison.
from 78 to a minimum of 4 (Table 5, entries 1, 4 and 7). suitable for this transformation (Table 7, entries 3 and 5–
The presence of para-substituents was also detrimental for 10). σ-Donating ligands such as alkyl-substituted phos-
the activity of this catalyst (Table 5, entries 10 and 15; see phine are also crucial for this transformation because no
also the Supporting Information). Within the same line as catalytic activity could be observed by using aryl-substi-
using [PdCl2(COD)]/dcpe/Zn, [Ni(COD)2]/dcpe showed tuted phosphines such as dppe (Table 7, entries 11 and 12).
only a reduced activity with phenolates bearing tert-butyl In accordance to the results given in Table 6, dcpe is the
substituents in the ortho-positions (Table 5, entries 9, ligand of choice for the proposed process concept. By using
12,15,17). In contrast, no significant effect of the steric bulk [Pd(PPh3)4]/dcpe, the catalyst loading could be reduced to
7126
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 7122–7130

Pd- and Ni-Catalyzed Synthesis of Sodium Acrylate
Table 7. Ligand screening.
0.01 mmol, whereby a TON of 106 was achieved. This is
comparable with the previous reported nickel system[8] for
which 50 equiv. Zn was required. The reason for this in-
crease in activity is probably due to the lower concentration
of sodium acrylate and phenol in the reaction medium be-
cause sodium acrylate has a low solubility in anisole. At
higher product concentrations, the reaction mixture became
heterogeneous, which affects the final TON. Limitations in
our laboratory set up (60 mL steel autoclave equipped with
mechanical stirrer) meant that it was not possible to in-
crease the rate of stirring above 1000 rpm, which led to
problems in mixing the reaction medium when it became
heterogeneous. This effect has also been observed in the
case of highly sterically congested bases (see Table 5, en-
tries 13–18).
Regarding the leaching of palladium at low catalyst load-
ings and using dcpe as ligand, only 1 ppm palladium was
found in the aqueous phase, which is a good compromise
for a useful activity and simple catalyst separation.
An efficient process concept for the synthesis of sodium
acrylate through ethylene carboxylation also requires cata-
lyst recycling, especially at the low TON observed in single
runs. To address this point, we performed the following re-
cycle test: The reaction was set up in an autoclave as de-
scribed in the experimental section. After the reaction time,
the autoclave was cooled to room temperature and de-
pressurized. As the Pd0- and Ni0-phosphine complexes are
presumably highly oxygen-sensitive, we carried out the
recycling procedure under an argon atmosphere. The work-
up and recycle procedure were (Figure 2): (1) The autoclave
was transferred and opened inside an argon atmosphere
glove box. (2) The reaction mixture was transferred to a
Schlenk flask and taken out of the glove box. (3) By using
Schlenk techniques, 30 mL of degassed deuterated water
were added to the reaction mixture. (4) After mixing and
phase separation, the two fractions were separated by sy-
ringe suction of the water phase. (5) The Schlenk flask con-
taining the organic phase was reintroduced into the glove
box and the organic phase was transferred to an autoclave
that was precharged with a fresh amount of base (and Zn
if needed). Recycling of catalyst in the three systems consid-
ered in this publication was assessed, and the results are
given in Table 8.
[PdCl2(COD)]/dcpe in the presence of Zn could be recy-
cled via the organic phase, and a TON of 15 was achieved
for the second run. In the case of [Pd(PPh3)4]/dcpe and
[Ni(COD)2]/dcpe, recycling of the catalyst was not possible
without the addition of Zn in the second cycle. This drop
of activity in the second cycle is mainly due to the limitation
of the recycle procedure applied. As showed in Figure 2, the
recycling is not performed in a sealed system. A series of
different equipment, techniques, materials (autoclave,
Schlenk flasks, syringes, septa) has to be used to perform
the recycling, increasing the probability of contamination
with oxygen, and resulting in rapid deactivation of the cata-
lyst. To overcome this limitation in the laboratory, the ad-
dition of a reductant such as Zn can prevent the formation
[a] Reaction conditions: base (10 mmol), catalyst (0.01 mmol), li-
gand (0.011 mmol), anisole (30 mL), ethylene (10 bar), CO2
(20 bar), 145 °C, 20 h. [b] As [a] but catalyst (0.2 mmol), ligand
(0.22 mmol). [c] TON determined based on 1H NMR analysis in
D2O as solvent and using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid
sodium salt (0.125 mmol) as internal standard. [d] n.d.: not deter-
mined. [e] The result presented in Table 5 is given for comparison. of these deactivated species by reducing the oxidized spe-
Eur. J. Org. Chem. 2015, 7122–7130
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7127

S. Manzini, N. Huguet, O. Trapp, T. Schaub
FULL PAPER
Figure 2. Flow chart of the recycling methodology with additional amount of fresh base.
Table 8. Recycle test with a fresh amount of base.[a], [d]
.
sphere dry box containing an atmosphere of purified argon. Sol-
vents for air- and moisture-sensitive manipulations were dried with
an MBraun SPS 800 solvent purification system, degassed, and
stored over molecular sieves, or dried and deoxygenated by using
reported procedures. [(μ5-Cp)Pd(μ3-allyl)] and the phenolate salts
were prepared according to reported procedures.[8,10,18] All other
reagents were purchased from Sigma–Aldrich or ABCR. Gases
were purchased from Air Liquide. Steel high-pressure reactors
(60 mL capacity) equipped with magnetic overhead stirrer pur-
chased by Premex were used for the carboxylation reactions. 1H,
31P, and 13C NMR spectra were recorded with Bruker Advance
200, 400, 500, or 600 MHz spectrometers. All 1H and 13C NMR
Entry Catalyst
Base
[mmol]
Zn
[mmol]
TON 1 TON 2
1[b]
2[c]
3[c]
4[b]
5[b]
PdCl2(COD)
20 + 20
10 + 10
10 + 10
10 + 10
10 + 10
1 + 1
0 + 0
0 + 1
0 + 0
0 + 0
55
50
50
44
44
15
1
18
0
Pd(PPh3)4
Pd(PPh3)4
Ni(COD)2
Ni(COD)2
10
[a] Reaction conditions: Pd (0.1 mmol), ligand (0.11 mmol), anisole
(30 mL), ethylene (10 bar), CO2 (20 bar), 145 °C, 20 h. [b] 2,6-Di-
methylphenolate was used as base. [c] NaBHT was used as base.
[d] TON determined based on 1H NMR analysis in D2O as solvent
and using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt
(0.125 mmol) as internal standard.
1
chemical shifts are reported in ppm relative to SiMe4 using the H
(residual) and 13C chemical shifts of the solvent as a secondary
standard. 31P NMR signals were referenced to triphenylphosphine.
Elemental analysis and mass spectra were recorded by the analyti-
cal service of the chemistry department of the University of Heidel-
berg. ICP-MS analysis to determine the trace metal content in the
product phase were performed by the analytical department of
BASF SE.
cies. Further investigations on the development of an ap-
propriate recycling protocol are under investigation.
Conclusions
Typical Procedure for the Synthesis of Sodium Acrylate: Inside a
glovebox, a 60 mL steel autoclave was charged with (COD)PdCl2
We were able to further improve the catalytic carb-
oxylation of ethylene with CO2 for the synthesis of sodium
acrylate by considering a sustainable process concept. How-
ever, further improvements are needed. At this stage, the
catalysis is not completely additive free, especially after the
system has been recycled. The reaction should also be
tested in a closed oxygen-free set-up, ideally in a continuous
mode. Despite the obtained TON of 106 with the [Pd(PPh3)4]/
dcpe catalyst system, showing the possibility to access the
(0.22 mmol,
0.057 g),
1,2-bis(dicyclohexylphosphanyl)ethane
(0.20 mmol, 0.093 g), sodium 2,6-dimethylphenoxolate (20 mmol,
2.860 g), and Zn (1 mmol, 0.065 g). The solid mixture was dissolved
in anisole (30 mL). The autoclave was removed from the glovebox
and charged while stirring at 800 rpm with 10 bar of ethylene and
20 bar of CO2 (total pressure 30 bar) for 15 min each at 25 °C. The
autoclave was then heated at 145 °C and stirred for 20 h at 800 rpm.
The autoclave was cooled to 20 °C, the pressure was released and
the reaction mixture was transferred to a 100 mLglass bottle. The
same TON as reported with the nickel system in the pres- autoclave vessel was rinsed with D2O (15 mL) to wash the auto-
clave. To this mixture, 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid so-
dium salt (0.13 mmol, 0.0216 g) was added and additional D2O
(10 mL) was added to the glass bottle. To favor phase separation,
Et2O (40 mL) was added to the mixture. An aliquot of the aqueous
phase was collected, centrifuged and analyzed by 1H NMR spec-
troscopy. The TON was determined by 1H NMR analysis
(200 MHz, 70 scans) according to a reported procedure.[8]
ence of 50 equiv. Zn, this TON is still low. It is necessary
to optimize further the catalyst system to achieve a higher
TON through efficient catalyst recycling. Despite these crit-
ical points, which are under investigation in our laboratory,
we have achieved a big step in the development of a process
concept to apply this interesting route for the production
of valuable sodium acrylate.
Typical Procedure for the Zinc-Free Synthesis of Sodium Acrylate:
Inside a glove box, a 60 mL steel autoclave was charged with
Pd(PPh3)4 (0.22 mmol, 0.234 g), 1,2-bis(dicyclohexylphosphanyl)-
ethane (0.20 mmol, 0.093 g), and sodium 2,6-dimethylphenoxolate
(20 mmol, 2.860 g) in anisole (30 mL). The autoclave was removed
from the glovebox and charged while stirring at 800 rpm with
10 bar of ethylene and 20 bar of CO2 (total pressure 30 bar) for
Experimental Section
General Considerations: All air- and moisture-sensitive manipula-
tions were carried out either by using standard vacuum line,
Schlenk, and cannula techniques or in an MBraun inert atmo-
7128
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 7122–7130

Pd- and Ni-Catalyzed Synthesis of Sodium Acrylate
15 min each at 25 °C. The autoclave was then heated at 145 °C and
stirred for 20 h at 800 rpm. The autoclave was cooled to 20 °C, the
pressure was released and the reaction mixture was transferred into
a 100 mLglass bottle. The autoclave vessel was rinsed with D2O
(15 mL) to wash the autoclave. To this mixture, 3-(trimethylsilyl)-
propionic-2,2,3,3-d4 acid sodium salt (0.13 mmol, 0.022 g) was
added and additional D2O (10 mL) was added to the glass bottle.
To favor phase separation, Et2O (40 mL) was added to the mixture.
An aliquot of the aqueous phase was collected, centrifuged, and
[1]
[2]
M. Frank, Superabsorbents, in: Ullmann’s Encyclopaedia of In-
dustrial Chemistry, vol. 35, Wiley-VCH, Weinheim, Germany,
2003, p. 213–232.
T. Ohara, T. Sato, N. Shimizu, G. Prescher, H. Schwind, O.
Weiberg, K. Marten, H. Greim, Acrylic Acid and Derivates, in:
Ullmann’s Encyclopaedia of Industrial Chemistry, Wiley-VCH,
Weinheim, Germany, 2011, p. 1–18.
[3]
For reviews and highlights on the use of CO2 for carboxylation
reactions, see: a) B. Yu, Z. F. Diao, C. X. Guo, J. CO2 Util.
2013, 1, 60–68; b) M. Aresta, A. DiBenedetto, Dalton Trans.
2007, 2975–2992; c) S. Pulla, C. M. Felton, P. Ramidi, Y. Gar-
tia, N. Ali, U. B. Nasini, A. Ghosh, J. CO2 Util. 2013, 2, 49–
57; J. Takaya, N. Iwasawa, C-1 Building Blocks in Organic Syn-
thesis, in: Science of Synthesis 2014, 46, 263–288; d) A. W.
Kleij, ChemCatChem 2013, 5, 113–115; e) X. Cai, B. Xie, Syn-
thesis 2013, 45, 3305–3324; f) T. Moragas, A. Correa, R. Mar-
tin, Chem. Eur. J. 2014, 20, 8242–8258; g) M. T. Johnson, O. F.
Wendt, J. Organomet. Chem. 2014, 751, 213–220; h) L. Yang,
1
analyzed by H NMR spectroscopy. The TON was determined by
1H NMR (200 MHz, 70 scans) analysis according to a reported
procedure.[8]
Procedure for the Recycling Test with (COD)PdCl2/Zn/dcpe: Inside
a glove box a 60 mL steel autoclave was charged with (COD)PdCl2
(0.22 mmol,
0.057 g),
1,2-bis(dicyclohexylphosphanyl)ethane
(0.20 mmol, 0.093 g), sodium 2,6-dimethylphenoxolate (20 mmol,
2.860 g), and Zn (1 mmol, 0.065 g), and the solid mixture was dis-
solved in anisole (30 mL). The autoclave was removed from the
glove box and charged, while stirring at 800 rpm, with 10 bar of
ethylene and 20 bar of CO2 (total pressure 30 bar) for 15 min each
at 25 °C. The autoclave was heated at 145 °C and stirred for 20 h
at 800 rpm. The autoclave was cooled to 20 °C and, after releasing
the pressure, was introduced into the glove box. The reaction mix-
ture was transferred to a 100 mL Schlenk flask equipped with a
magnetic bar and, outside the glove box, degassed water (30 mL)
was added by using a syringe. The mixture was stirred for 10 min
at room temperature to promote dissolution of sodium acrylate and
the two phase were allowed to settle for 2 min. The water phase
was separated and analyzed as described previously. The organic
phase was reintroduced into the glove box and transferred to a
60 mL steel autoclave previously charged with sodium 2,6-dimeth-
ylphenoxolate (20 mmol, 2.860 g) and Zn (1 mmol, 0.065 g). The
autoclave was then removed from the glove box and charged, while
stirring at 800 rpm, with 10 bar of ethylene and 20 bar of CO2 (to-
tal pressure 30 bar) for 15 min each at 25 °C. The autoclave was
then heated at 145 °C and stirred for another 20 h at 800 rpm. At
time elapsed the work-up and analysis were the same as described
previously.
H. Huang, Chem. Rev. 2015, 115, 3468–3517; i) T. E. Müller
,
W. Leitner, Beilstein J. Org. Chem. 2015, 11, 675–677; j) M.
Peters, T. Mueller, W. Leitner, Tce 2009, 813, 46–47; k) M. Lim-
bach, Advances in Organometallic Chemistry, vol. 63, Burl-
ington, Academic Press, 2015, p. 175–202.
[4]
a) R. Alvarez, E. Carmona, D. J. Cole-Hamilton, A. Galindo,
E. Gutierrez-Puebla, A. Monge, M. L. Poveda, C. Ruiz, J. Am.
Chem. Soc. 1985, 107, 5529–5531; b) A. Galindo, A. Pastor, P.
Pérez, E. Carmona, Organometallics 1993, 12, 4443–4451; c)
R. Fischer, J. Langer, A. Malassa, D. Walther, H. Görls, G.
Vaughan, Chem. Commun. 2006, 2510–2512; d) D. Jin, P. G.
Williard, N. Hazari, W. H. Bernskoetter, Chem. Eur. J. 2014,
20, 3205–3211; e) M. Cokoja, C. Bruckmeier, B. Rieger, W. A.
Herrmann, F. E. Kühn, Angew. Chem. Int. Ed. 2011, 50, 8510–
8537; Angew. Chem. 2011, 123, 8662–8690; f) P. Braunstein, D.
Matt, D. Nobel, Chem. Rev. 1988, 88, 747–764; g) Y. Tsuji, T.
Fujihara, Chem. Commun. 2012, 48, 9956–9964; h) K. Ukai,
M. Aoki, J. Takaya, N. Iwasawa, J. Am. Chem. Soc. 2006, 128,
8706–8707.
a) H. Hoberg, D. Schaefer, J. Organomet. Chem. 1982, 236,
C28–C30; b) H. Hoberg, D. Schaefer, B. W. Oster, J. Or-
ganomet. Chem. 1984, 266, 313–320; c) H. Hoberg, D. Schaefer,
J. Organomet. Chem. 1983, 251, C51–C53; d) H. Hoberg, K.
Jenni, C. Krüger, E. Raabe, Angew. Chem. Int. Ed. Engl. 1986,
25, 810–811; Angew. Chem. 1986, 98, 819–820; e) H. Hoberg,
Y. Peres, A. Milchereit, J. Organomet. Chem. 1986, 307, C41–
C43; f) H. Hoberg, Y. Peres, A. Milchereit, J. Organomet.
Chem. 1986, 307, C38–C40; g) H. Hoberg, K. Jenni, J. Or-
ganomet. Chem. 1987, 322, 193–201; h) H. Hoberg, S. Gross,
A. Milchereit, Angew. Chem. Int. Ed. Engl. 1987, 26, 571–572;
Angew. Chem. 1987, 99, 567–569; i) H. Hoberg, Y. Peres, C.
Krüger, Y. H. Tsay, Angew. Chem. Int. Ed. Engl. 1987, 26, 771–
773; Angew. Chem. 1987, 99, 799–800; j) H. Hoberg, Y. Peres,
A. Milchereit, S. Gross, J. Organomet. Chem. 1988, 345, C17–
C19; k) H. Hoberg, D. Bärhausen, J. Organomet. Chem. 1989,
379, C7–C11; l) H. Hoberg, A. Ballesteros, A. Sigan, J. Or-
ganomet. Chem. 1991, 403, C19–C22; m) H. Hoberg, A. Balles-
teros, A. Sigan, C. Jegat, A. Milchereit, Synthesis 1991, 395–
398; n) H. Hoberg, A. Ballesteros, A. Sigan, C. Jégat, D.
Bärhausen, A. Milchereit, J. Organomet. Chem. 1991, 407,
C23–C29.
[5]
General Procedure for the Determination of Base Residual in the
Water Phase: From the water phase, after the work up, an aliquot
of was collected and filtered through a Millipore 0.5 μm microfilter.
The clear solution was evaporated under vacuum. The residual
1
solid was redissolved in D2O (0.5 mL) and analyzed by H NMR
spectroscopy, integrating the residual base peaks of the base con-
sidered with the 3 CH3 peak of the 3-(trimethylsilyl)propionic-
2,2,3,3-d4 acid sodium salt used as standard in the determination
of the sodium acrylate content. The lower limit of the analysis was
defined by the assumption of the sensitivity of the NMR spectrom-
eter beiong approximately 1 mg phenol in 0.5 mL of solvent.
General Procedure for the Determination of Pd Leaching in the
Water Phase: From the water phase, after the work up, an aliquot
was collected and filtered twice through a Millipore 0.5 μm micro-
filter. The Pd content was determined by ICP-MS analysis (mea-
surements carried out by the certified central analytics laboratories
of BASF SE in Ludwigshafen).
[6]
M. L. Lejkowski, R. Lindner, T. Kageyama, G. É. Bódizs, P. N.
Plessow, I. B. Müller, A. Schäfer, F. Rominger, P. Hofmann, C.
Futter, S. A. Schunk, M. Limbach, Chem. Eur. J. 2012, 18,
14017–14025.
[7]
[8]
C. Hendriksen, E. A. Pidko, G. Yang, B. Schäffner, D. Vogt,
Chem. Eur. J. 2014, 20, 12037–12040.
N. Huguet, I. Jevtovikj, A. Gordillo, M. L. Lejkowski, R.
Lindner, M. Bru, A. Y. Khalimon, F. Rominger, S. A. Schunk,
P. Hofmann, M. Limbach, Chem. Eur. J. 2014, 20, 16858–
16862.
Acknowledgments
CaRLa (Catalysis Research Laboratory) is being co-financed by
the Ruprecht Karls University of Heidelberg and by BASF SE,
Ludwigshafen, Germany.
Eur. J. Org. Chem. 2015, 7122–7130
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7129

S. Manzini, N. Huguet, O. Trapp, T. Schaub
FULL PAPER
[9] S. C. E. Stieber, N. Huguet, T. Kageyama, I. Jevtovikj, P. Ari-
yananda, A. Gordillo, S. A. Schunk, F. Rominger, P. Hofman,
M. Limbach, Chem. Commun. 2015, 51, 10907–10909.
[10] a) P. N. Plessow, A. Schäfer, M. Limbach, P. Hofmann, Orga-
nometallics 2014, 33, 3657–3668; b) , P. N. Plessow, L. Weigel,
R. Lindner, A. Schäfer, F. Rominger, M. Limbach, P. Hof-
mann, Organometallics 2013, 32, 3327–3338.
[11] The calculation of the sodium carbonate formation energy is
based on the following publication: R. Berg, C. E. Vanderzee,
J. Chem. Thermodyn. 1978, 10, 1113–1136.
[12] S. Walendy, G. Francio, M. Hoelscher, W. Leitner, Experiments
in Green and Sustainable Chemistry 2009, 275–277.
[13] a) A. A. Volod’kin, G. E. Zaikov, Russ. Chem. Bull. 2002, 51,
2189–2195; b) A. A. Volod’kin, G. E. Zaikov, Russ. Chem. Bull.
2006, 55, 2220–2225.
[14]
[15]
[16]
The results in Table 1, entries 1, 3, 5, and 7 are reported in
ref.[8] The result presented in Table 1, entry 2 is reported in
ref.[9]
a) A. J. Bard, L. R. Faulkner, Electrochemical Methods. Funda-
mentals and Applications, 2nd ed., 2001; b) W. M. Haynes, CRC
Handbook of Chemistry and Physics, 93rd ed., 2012.
R. K. Henderson, C. Jiménez-González, D. J. C. Constable,
S. R. Alston, G. G. A. Inglis, G. Fisher, J. Sherwood, S. P.
Binks, A. D. Curzons, Green Chem. 2011, 13, 854–862.
[17] The water content was determined in our laboratory by Karl–
Fischer titration.
[18] Y. Tatsuno, T. Yoshida, S. Otsuka, Inorg. Synth. 1979, 19, 220.
Received: August 27, 2015
Published Online: October 8, 2015
7130
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 7122–7130