Rhodium-Catalyzed Hydrocarboxylation: Mechanistic Analysis Reveals Unusual Transition State for Carbon-Carbon Bond Formation

Article abstract of DOI:10.1021/acs.organomet.7b00899

The mechanism of rhodium-COD-catalyzed hydrocarboxylation of styrene derivatives and α,β-unsaturated carbonyl compounds with CO₂ has been investigated using density functional theory (PBE-D2/IEFPCM). The calculations support a catalytic cycle as originally proposed by Mikami and co-workers including β-hydride elimination, insertion of the unsaturated substrate into a rhodium-hydride bond, and subsequent carboxylation with CO₂. The CO₂ insertion step is found to be rate limiting. The calculations reveal two interesting aspects. First, during C-CO₂ bond formation, the CO₂ molecule interacts with neither the rhodium complex nor the organozinc additive. This appears to be in contrast to other CO₂ insertion reactions, where CO₂-metal interactions have been predicted. Second, the substrates show an unusual coordination mode during CO₂ insertion, with the nucleophilic carbon positioned up to 3.6 ? away from rhodium. In order to understand the experimentally observed substrate preferences, we have analyzed a set of five alkenes: an α,β-unsaturated ester, an α,β-unsaturated amide, styrene, and two styrene derivatives. The computational results and additional experiments reported here indicate that the lack of activity with amides is caused by an overly high barrier for CO₂ insertion and is not due to catalyst inactivation. Our experimental studies also reveal two putative side reactions, involving oxidative cleavage or dimerization of the alkene substrate. In the presence of CO₂, these alternative reaction pathways are suppressed. The overall insights may be relevant for the design of future hydrocarboxylation catalysts.

Full text of DOI:10.1021/acs.organomet.7b00899

Article
Cite This: Organometallics XXXX, XXX, XXX−XXX
Rhodium-Catalyzed Hydrocarboxylation: Mechanistic Analysis
Reveals Unusual Transition State for Carbon−Carbon Bond
Formation
Ljiljana Pavlovic,† Janakiram Vaitla,‡ Annette Bayer,‡ and Kathrin H. Hopmann*,†
†Hylleraas Centre for Quantum Molecular Sciences, Department of Chemistry, University of Tromsø-The Arctic University of
Norway, N-9037 Tromsø, Norway
‡Department of Chemistry, University of Tromsø-The Arctic University of Norway, N-9037 Tromsø, Norway
S
* Supporting Information
ABSTRACT: The mechanism of rhodium-COD-catalyzed
hydrocarboxylation of styrene derivatives and α,β-unsaturated
carbonyl compounds with CO2 has been investigated using
density functional theory (PBE-D2/IEFPCM). The calculations
support a catalytic cycle as originally proposed by Mikami and
co-workers including β-hydride elimination, insertion of the
unsaturated substrate into a rhodium−hydride bond, and
subsequent carboxylation with CO2. The CO2 insertion step is
found to be rate limiting. The calculations reveal two interesting
aspects. First, during C−CO2 bond formation, the CO2 molecule
interacts with neither the rhodium complex nor the organozinc
additive. This appears to be in contrast to other CO2 insertion
reactions, where CO2−metal interactions have been predicted.
Second, the substrates show an unusual coordination mode during CO2 insertion, with the nucleophilic carbon positioned up to
3.6 Å away from rhodium. In order to understand the experimentally observed substrate preferences, we have analyzed a set of
five alkenes: an α,β-unsaturated ester, an α,β-unsaturated amide, styrene, and two styrene derivatives. The computational results
and additional experiments reported here indicate that the lack of activity with amides is caused by an overly high barrier for CO2
insertion and is not due to catalyst inactivation. Our experimental studies also reveal two putative side reactions, involving
oxidative cleavage or dimerization of the alkene substrate. In the presence of CO2, these alternative reaction pathways are
suppressed. The overall insights may be relevant for the design of future hydrocarboxylation catalysts.
that the transmetalated ethyl ligand could donate a hydride
directly to the styrene substrate.8 A computational investigation
of a hypothetical reaction involving Rh-pincer complexes
explored the potential of H2 as a hydride donor, but the
study showed that this might lead to a competing hydro-
genation reaction.9 Theoretical studies of a Ni(DBU)2 complex
with ZnEt2 as additive concluded that formation of a Ni hydride
complex might compete with oxidative coupling of the alkene
substrate and CO2 to form a low-lying metallacycle.10
Related reactions involving insertion of CO2 into metal−
alkyl bonds have also been studied computationally, including
copper-catalyzed carboxylation of alkynes13,14 or organo-
boranes,15 nickel-catalyzed carboxylation of aryl or benzyl
halides,16,17 coupling of CO2 with ethylene18,19 or acetylene,20
and insertion of CO2 into rhodium−ethyl or −aryl
bonds.21,22,24 Despite the molecular differences of these
systems, the carbon−CO2 bond formation step is expected to
occur in a similar fashion via a three-membered transition state
INTRODUCTION
■
Hydrocarboxylation of alkenes is a promising strategy for
formation of alkyl carboxylic acids from CO2. In 2008, Rovis
and co-workers reported seminal studies on nickel-catalyzed
hydrocarboxylation of alkenes with CO2.1 Subsequently, related
reactions have been communicated involving different catalysts
based on Cu/Fe,2 Fe,3 Ti,4 Rh,5 or Ru.6
The reaction mechanisms of metal-catalyzed hydrocarbox-
ylation of different types of unsaturated substrates have been
addressed computationally (for a relevant review, see ref 7). A
small number of studies have investigated hydrocarboxylation
of alkene substrates with CO2.8−10 Despite large differences in
catalysts, the different alkene hydrocarboxylation reactions have
been proposed to follow similar patterns, involving formation
of a metal hydride (e.g., through β-hydride elimination from a
transmetalated alkyl ligand), insertion of the alkene substrate
into the metal−hydride bond to form a metal alkyl species, and
subsequent reaction with CO2 to yield the carboxylic acid
(Figure 1A).3,5,8−10 Variations of this mechanism are seen: e.g.,
for an Fe pincer complex with EtMgBr as additive, it has been
proposed that an explicit Fe hydride might not be formed but
Received: December 19, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
the CO2 molecule interacts with neither the rhodium complex
nor the organozinc additive. This is in contrast to most
previous computational studies, which predict a metal−CO2
interaction (Figure 1).8−14,20−22,24 Second, in the C−C bond
forming step, the substrates prefer an η6 coordination mode to
rhodium, with the nucleophilic carbon positioned several
angstroms away from the metal center. Although η6 binding
modes are well-known for coordination of ligands,26−29 to our
knowledge such a binding mode has not previously been
predicted to occur during the conversion of styrene-type
substrates. In order to understand the experimental substrate
preferences, we have analyzed a set of five substrates and show
that the proposed mechanism can explain why esters are the
preferred substrates, whereas amides are unreactive. We further
report experimental studies that reveal two putative side
reactions, which appear to be suppressed in the presence of
CO2.
Figure 1. (A) Generalized mechanism for hydrocarboxylation of
alkenes with CO2 (transmetalation steps not shown, based on refs 3, 5,
and 8−10). (B) Proposed conformation of C−CO2 bond formation
TS for reactions involving Ni,10,11,16,17,20 Cu,12−14 Rh,9,21,22 and Fe/
Mg.8
METHODS
■
(TS), where CO2 is predicted to interact with the metal via the
carbon atom or through an η2-C,O coordination mode (Figure
1B; for a recent review, see ref 23).8−14,20−22,24 An exception is
CO2 insertion into nickel−methyl or −allyl bonds, for which it
has been predicted that CO2 has no interaction with the metal
center during the C−CO2 bond formation step.25
Computational Models. All calculations were performed with the
full catalyst and full substrates (Figure 3), without truncations or
Several of the reported hydrocarboxylation reactions of
alkenes result in the formation of chiral carboxylic acids, but
racemic mixtures are obtained due to the use of symmetric
catalysts.1,3,4 To date, only one asymmetric CO2-based
hydrocarboxylation reaction has been reported, involving a
rhodium complex with a bidentate SEGPHOS ligand, which in
the presence of ZnEt2 carboxylates α,β-unsaturated esters with
up to 66% enantiomeric excess (ee, Figure 2).5 The
Figure 3. Metal complexes and substrates studied here.
symmetry constraints and with a closed-shell spin state. The effect of
the solvent was modeled through inclusion of the polarizable
continuum solvent model IEFPCM (dimethylformamide, which is
the solvent employed in experiments5) in geometry optimizations and
energy evaluations.
Computational Methods. All calculations were performed with
Gaussian 09, Rev. D01.30 The DFT functional PBE,31 including both
the Grimme empirical dispersion correction (D232) and IEFCPM33
(solvent = N,N-dimethylformamide), were employed in all calcu-
lations. We have previously shown that this computational protocol
provides good accuracy.34 In geometry optimizations, the basis set 6-
311G(d,p) was employed on all nonmetal atoms and zinc, whereas
Lanl2DZ (including pseudopotential) was employed on rhodium,
including an extra f polarization function with exponent 1.3535 (this
basis set combination is here referred to as BS1). Big basis set single-
point corrections to the energies were computed employing 6-
311+G(2d,2p) on all nonmetal atoms and zinc, while Lanl2TZ(f) was
used for rhodium (referred to as BS2). Counterpoise corrections (CP)
were computed with BS2 to estimate the magnitude of the basis set
superposition error when several molecules were joined into one
model.36 The CP corrections were found to be small, and they were
not included in the final energies.
Figure 2. Rhodium-catalyzed hydrocarboxylation of alkenes with CO2,
as reported by Mikami and co-workers.5
hydrocarboxylation activity of rhodium catalysts with bidentate
ligands appears to be highly dependent on both the nature of
the ligand and the properties of the substrate.5 With Rh−COD,
α,β-unsaturated esters and styrene derivatives with electron-
withdrawing para substituents are converted with reasonable
yields; however, α,β-unsaturated amides and styrene itself are
completely unreactive.5
Detailed insights into the reaction mechanism may be
valuable for improving the activity, substrate scope, and
selectivity of rhodium-catalyzed hydrocarboxylation reactions.
Here we have studied the mechanism of rhodium−COD/ZnEt2
mediated alkene hydrocarboxylation using DFT methods on
the full molecular system. In addition to the mechanistic details,
our aim was to elucidate the substrate preferences of this
reaction and to understand how CO2 and the substrate interact
with the rhodium complex. Our theoretical analysis revealed
two interesting aspects. First, during C−CO2 bond formation,
The thermodynamic parameters were obtained from frequency
calculations and were corrected to the experimental temperature of
273 K.
In order to convert computed free energies (ΔG°1atm,BS1) from a 1
atm to a 1 M standard state, a standard state conversion factor (SS)
was included, which can be calculated as R1T ln(R2TΔn), where R1 =
B
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
8.31447 J K−1 mol−1, R2 = 0.08206 L atm K−1 mol−1, T = temperature
in K, and Δn = change in number of moles.34,37 For an association
reaction A + B → C, the change in moles is Δn = −1 and the above
expression becomes −1.89 kcal/mol at 298 K (−1.69 kcal/mol at 273
K). Unless explicitly stated otherwise, reported energies correspond to
solution standard state Gibbs free energies at 273 K (ΔG°1M,273K),
including big basis set corrections to the electronic energy (E) and
standard state corrections (SS): ΔG°1M,273K = ΔG°1atm,BS1,273 K − ΔEBS1
complex. Finally, transmetalation with diethylzinc forms a zinc
carboxylate, with simultaneous regeneration of the Rh−Et
species. Upon acidic workup, zinc carboxylate is protonated to
provide the α-aryl carboxylic acid.
The computed energy profile is shown in Figure 5, with the
Rh(DMF)Et species used as an energetic reference. We have
used the α,β-unsaturated ester methyl 2-phenylacrylate (Figure
3) in our calculations of the full reaction cycle. The first step, β-
H elimination, occurs via a four-membered TS (TSβ‑Hyd) with
an activation barrier of 6.8 kcal/mol. The subsequent low-
barrier hydride transfer to the terminal carbon of the substrate
agrees with the experimentally observed product, α-aryl
carboxylic acid.5 The alternative hydride transfer pathway
leading to the β-aryl carboxylic acid is higher in energy (Figure
S1 in the Supporting Information).
+ ΔEBS2 + SS273 K
.
Experimental Details. Diethylzinc (1.0 M in hexane solution) and
[RhCl(cod)]2 were purchased from Aldrich. Methyl 2-phenylacrylate
and N,N-dimethyl-2-phenylacrylamide were prepared according to the
literature.5 DMF was dried over 4 Å molecular sieves. Conversions
were monitored by thin-layer chromatography (TLC) with Merck
precoated silica gel plates (60 F254). Visualization was accomplished
either with UV light or by immersion in potassium permanganate or
5% alcoholic phosphomolybdic acid (PMA) followed by light heating
with a heating gun. High-resolution mass spectra (HRMS) were
recorded from MeOH solutions on an LTQ Orbitrap XL instrument
(Thermo Scientific) in positive electrospray ionization (ESI) mode.
The mechanistic cycle was studied with methyl 2-phenylacrylate as
substrate using stoichiometric catalyst/substrate mixtures, monitored
by NMR (for further details, see the Supporting Information).
Our calculations indicate that the Rh−alkyl intermediate
formed after substrate insertion is energetically very stable, 12.4
kcal/mol below the Rh(DMF)Et species (Figure 5). The
optimized geometry of Rh−alkyl reveals an η2 coordination
mode of the phenyl ring to the metal (Figure 6A). This type of
interaction is well-known and has been described in the
literature.39 The Rh−Cipso and Rh−Cortho distances are 2.26 and
2.27 Å, respectively. We propose that several resonance
structures contribute to this binding mode (Figure S2 in the
Supporting Information). An alternative binding mode
involving the ester moiety was found to be higher in energy
by 3.1 kcal/mol (see Figure S3 in the Supporting Information).
In the next step, insertion of CO2 occurs to give a
carboxylated rhodium complex, with a barrier of 14.4 kcal/
mol relative to the Rh−alkyl species (TS-CO2-I, Figure 5). The
carbon−carbon bond formation TS reveals an unusual binding
mode of the substrate, with the phenyl ring coordinated to
rhodium in an η6 fashion, whereas the nucleophilic carbon is
positioned 3.6 Å away from the metal center (Figures 6B and
7A). Five of the Rh−Cphenyl distances are between 2.24 and 2.37
Å. We predict that the negative charge on the substrate is
delocalized over the nucleophilic carbon and the ester oxygen,
which is stabilized by two CH···O interactions (distances 2.14
and 2.45 Å, respectively, Figure 6B). Note that, for substrates
lacking an adjacent carbonyl, different resonance structures are
predicted, with charge delocalization over seven carbon atoms
(Figure S4 in the Supporting Information). We optimized an
alternative TS geometry, where the carbonyl group of the ester
interacts with the metal (TS-CO2-II, CO···RH distance of
2.16 Å, Figure 7B), but this is 1.8 kcal/mol higher in energy in
comparison to TS-CO2-I. Interestingly, both of these TSs show
no interaction of the CO2 molecule with rhodium. If CO2 is
placed closer to rhodium (TS-CO2-III and TS-CO2-IV, Figure
S5 in the Supporting Information), the barrier increases by 5−6
kcal/mol, whereas if CO2 is bound to rhodium in an η2 fashion
(TS-CO2-V, Figure S6 in the Supporting Information), the
barrier is 23.5 kcal/mol above TS-CO2-I.
RESULTS AND DISCUSSION
■
Mechanistic Investigation. The mechanism for rhodium-
catalyzed hydrocarboxylation as proposed by Mikami and co-
workers5 is shown in Figure 4, with minor modification on the
Figure 4. Mechanism for Rh−COD catalyzed hydrocarboxylation,
based on the previous proposal5 and computations performed here.
We also tested if CO2 might prefer to interact with the
organozinc additive (in analogy to the results by Yuan and co-
workers, who proposed that interactions between zinc and CO2
may take place10). However, the optimized geometry of the C−
C bond formation TS in the presence of ZnEt2 does not
indicate that a CO2−Zn interaction occurs (see Figure S7 in the
Supporting Information). This is in line with results by Lv et
al.,24 who concluded that an interaction of CO2 with zinc is
energetically disfavored.
basis of our results. The active rhodium hydride species is
generated in two steps. The first step involves transmetalation
of an ethyl group from diethylzinc to the precatalyst. This leads
to formation of a Rh−Et species, which we predict may loosely
coordinate a DMF solvent molecule to form Rh(DMF)Et. The
complex then loses DMF and undergoes β-hydride elimination
to give a Rh−H species. Next, insertion of the alkene substrate
into the Rh−H bond takes place to produce a Rh−alkyl
intermediate.38 Two possible pathways for alkene insertion
exist due to the asymmetry of the double bond. Insertion of
CO2 into the Rh−alkyl species gives the carboxylated rhodium
The computed energies (Figure 5) indicate that insertion of
CO2 may be rate determining. The carboxylic acid is
C
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 5. Computed free energy profile (kcal/mol, 273 K, PBE-D2/BS2[IEFPCM]//PBE-D2/BS1[IEFPCM] level of theory) for rhodium−COD
catalyzed hydrocarboxylation of methyl 2-phenylacrylate.
Figure 6. Schematic representation of the energetically preferred
conformations of (A) the Rh−alkyl intermediate, displaying an η2
coordination mode of the ester, and (B) the TS for CO2 insertion
(TS-CO2-I), displaying an η6 coordination of the substrate, with
additional CH···O interactions.
subsequently transmetalated to zinc, and the active rhodium
species is regenerated. A TS for the transmetalation step could
not be located, but we do not expect this step to be rate
limiting, as shown by Lv et al. for a related Rh-catalyzed
carboxylation reaction.24 We optimized an Rh−CO2−Zn
intermediate, where both metals, Rh and Zn, interact with
the carboxyl group. The computed energy is −20.1 kcal/mol
below the Rh−DMF−ethyl species (Figure 5; for the optimized
geometry, see Figure S8 in the Supporting Information).
We also performed computations on an alternative reaction
pathway, involving formation of a metallacycle from CO2,
Figure 7. Optimized geometries with methyl 2-phenylacrylate of (A)
alkene, and the rhodium complex. This was tested in analogy to
the preferred CO2 insertion TS with η6-coordination (TS-CO-I) and
the results reported for nickel-catalyzed hydrocarboxylation,
(B) an energetically higher- lying CO2 insertion TS displaying a Rh−
OCester interaction (TS-CO2-II). Distances are given in angstroms.
where a low−lying metallacycle was observed in calculations.10
However, a rhodacycle intermediate was here found to be
higher in energy and was excluded (see Figure S9 in the
Supporting Information).
D
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
We extended our analysis by additionally performing
experimental studies (Figures S10−S20 in the Supporting
Information) to shed further light on the catalytic steps of the
hydrocarboxylation reaction. When [RhCl(cod)]2 (1 equiv)
was treated with Et2Zn (1 equiv) in DMF-D7 in an NMR tube
at 0 °C, an immediate color change from pale yellow to dark
red was observed. The 1H NMR spectrum indicated formation
of Rh−H species with a diagnostic resonance at −11.74 ppm
(Figure 8).40 Then, 2.5 equiv of methyl 2-phenylacrylate was
Figure S15 in the Supporting Information). Formation of
methyl benzoylformate can be explained with a side reaction of
the rhodium complex with O2 (from air) to generate a
Rh(cod)(peroxo) complex, which can further react with alkene
to give methyl benzoylformate.42 To support our hypothesis
and confirm the formation of methyl benzoylformate, we
carried out the same reaction using 2 mol % of [RhCl(cod)]2
and 1.2 equiv of ZnMe2 in the presence of air, which gave
methyl benzoylformate in 52% yield (for a mechanistic
hypothesis, see Figure S15 in the Supporting Information).
Under ordinary reaction conditions, where a CO2 atmosphere
is present from the initiation of the reaction, the side products
discussed here are not formed, and instead methyl 2-
phenylacrylate provides a 99% yield of the expected
carboxylated product.5 Therefore, we speculate that, under
the reaction conditions, CO2 both inhibits the formation of the
rhodium−peroxo complex (due to the reduced presence of O2)
and prevents dimerization of the alkene by promoting
carboxylation of the rhodium−alkyl intermediate instead. Our
analysis thus points to a serendipitous role of CO2 in
suppressing putative side reactions.
Investigation of the Substrate Selectivity. In order to
understand the experimental substrate preferences, we have
performed computational and experimental studies of five
substrates: an α,β-unsaturated ester and amide, styrene, and
two para-substituted styrene derivatives (Figure 3). For two of
these substrates, the ester and the 1-(4-vinylphenyl) ethan-1-
one, Mikami and co-workers reported yields of 89 to >99% in
the Rh−COD catalyzed hydrocarboxylations, whereas styrene
and 1-methoxy-4-vinylbenzene gave 0% yield.5 Our exper-
imental results (see Figure S10 in the Supporting Information)
are in agreement with earlier results, showing good conversion
for methyl 2-phenyl acrylate and 1-(4-vinylphenyl) ethan-1-
one. No reaction was detected for an N,N-dimethyl-substituted
amide, in agreement with Mikami’s studies on a related amide.5
We evaluated two hypotheses for the lack of activity of
certain substrates: (a) lack of reactivity is due to substrate-
mediated catalyst inactivation and (b) lack of reactivity is due to
an overly high barrier for CO2 insertion. To evaluate hypothesis
a, we computed the energies (ΔGbond strength) for binding the five
substrates or a solvent molecule to rhodium (Table 1 and
Figure 10). The ester has the strongest binding energy (−20.9
1
Figure 8. H NMR of [Rh]-H peak.
placed in the NMR tube and the resulting complex was
characterized. The absence of the olefinic protons of methyl 2-
phenylacrylate with chemical shifts at 6.37 and 5.90 ppm and
the Rh−H chemical shift at −11.74 ppm indicates that insertion
of the alkene into the rhodium hydride took place. However,
subsequent bubbling of CO2 into the NMR tube (for 10 min)
did not give the carboxylated product. To investigate the failure
of the expected carboxylation step, we analyzed the reaction
mixture by HRMS and GCMS prior to CO2 bubbling. The
results indicate formation of a complex reaction mixture
including dimerized methyl 2-phenylacrylate and methyl
benzoylformate (Figure 9).
The rhodium-catalyzed alkene dimerization is not a
surprising observation41 and supports the formation of a
rhodium−alkyl intermediate (for a mechanistic hypothesis, see
Table 1. Computed Binding Energies and CO2 Insertion
Barriers for Different Substrates
computed barrier for
computed energy for CO2 insertion (relative
binding to rhodium
to Rh−alkyl
exptl
(ΔGbond strength
,
intermediate),
yield
a
b
b
substrate/ligand
DMF
kcal/mol, 298 K)
kcal/mol
(%)
−7.1
methyl 2-
−20.9
14.4
16.4
>99
0
phenylacrylate
1-methoxy-4-
vinylbenzene
−15.9
styrene
−17.3
−17.2
16.7
18.7
0
1-(4-vinylphenyl)
ethan-1-one
89
N,N-dimethyl-2-
phenylacrylamide
−17.3
19.8
0
a
b
See Figure 10. Energies at the PBE/BS1(IEFPCM) level. From ref 5
Figure 9. Stoichiometric NMR experiments providing evidence of
except for the amide, which was studied here (see Figure S10 in the
reaction intermediates as well as side reactions.
Supporting Information).
E
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
kcal/mol). The other substrates bind more weakly, but still
considerably more strongly in comparison to the solvent DMF.
A correlation between binding strengths and experimental
conversions is not evident.
observed with ester and amide as substrates but are not possible
for styrene-type substrates that lack a carbonyl group (for the
latter, possible resonance structures are depicted in Figure S6 in
the Supporting Information). It can also be seen that the CO2
molecule does not interact with the metal center in any of the
TSs. The conformational freedom of the CO2 molecule implies
that there potentially exist a large number of energetically close-
lying TSs.
An analysis of the free energies shows that the ester has the
lowest barrier (14.4 kcal/mol), whereas the highest activation
energy was found for the amide (19.8 kcal/mol). This indicates
that the barrier for CO2 insertion could explain why esters are
the preferred substrates and amides are unreactive.5 Styrene
and 1-methoxy-4-vinylbenzene have similar barriers of 16.7 and
16.4 kcal/mol, respectively, which are ∼2 kcal/mol above that
of the ester, in line with the lack of reactivity in experiment.5
However, 1-(4-vinylphenyl)ethan-1-one has a relatively high
CO2 insertion barrier of 18.7 kcal/mol, in disagreement with
the experimentally observed conversion. We cannot exclude
that a lower CO2 insertion TS exists for this particular
substrate. Overall, the correlation for the four other substrate
indicates that the CO2 insertion barrier could explain the
substrate preferences observed in rhodium-catalyzed hydro-
carboxylation reactions.
We then performed an experimental test, where we
hydrocarboxylated a 1:1 mixture of amide and ester (Figure
11). The ester showed full conversion, whereas the amide was
unreactive. This confirms that the amide is not inactivating the
catalyst. On basis of these results, hypothesis (a) was excluded.
We then proceeded to evaluate hypothesis (b) by computing
the alkyl intermediates and TSs for CO2 insertion for the five
substrates (Table 1). Interestingly, the energetically lowest
lying TS geometries all show an η6 coordination mode of the
substrate (Figures 7A and 12). CH···OC interactions are
Figure 10. Computation of Rh−substrate bond strength.
CONCLUSIONS
■
The detailed mechanism of the highly regioselective rhodium-
catalyzed hydrocarboxylation of styrene derivatives and α,β-
unsaturated carbonyl compounds with CO2 was investigated
using DFT theory, combined with experimental studies. The
computational investigations reproduce the experimentally
observed regioselectivity and reveal that, during the C−CO2
bond formation step, the CO2 molecule interacts with neither
the rhodium complex nor the organozinc additive. This is in
contrast to predictions on related reactions, which propose a
Figure 11. Rh-catalyzed hydrocarboxylation of a 1:1 mixture of amide
and ester.
Figure 12. Optimized geometries for the CO2 insertion TS for the substrates (A) 1-methoxy-4-vinylbenzene, (B) styrene, (C) 1-(4-
vinylphenyl)ethan-1-one, and (D) N,N-dimethyl-2-phenylacrylamide. Distances are given in angstroms.
F
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
CO2−metal interaction.8−14,20−24 In addition, we find that, for
the five studied substrates, the energetically lowest lying TS
geometries all show an unusual coordination mode of the
substrate through the phenyl ring instead of through the
nucleophilic carbon. Our calculations indicate that the rate-
limiting CO2 insertion barrier could provide an explanation of
why esters are reactive and amides are unreactive. A possible
poisoning effect of the amide was ruled out. Our experimental
results are in agreement with the computationally computed
mechanism. In addition, the experimental studies revealed two
possible side reactions involving substrate dimerization or
oxidative cleavage of the alkene, which were suppressed when
the reaction was performed under a CO2 atmosphere. The
insights obtained here may be relevant for designing novel
rhodium-catalyzed reactions for carbon−carbon bond forma-
tion with CO2.
(4) Shao, P.; Wang, S.; Chen, C.; Xi, C. Cp2TiCl2-Catalyzed
Regioselective Hydrocarboxylation of Alkenes with CO2. Org. Lett.
2016, 18, 2050−2053.
(5) Kawashima, S.; Aikawa, K.; Mikami, K. Rhodium-Catalyzed
Hydrocarboxylation of Olefins with Carbon Dioxide. Eur. J. Org. Chem.
2016, 2016, 3166−3170.
(6) Murata, K.; Numasawa, N.; Shimomaki, K.; Takaya, J.; Iwasawa,
N. Construction of a visible light-driven hydrocarboxylation cycle of
alkenes by the combined use of Rh(I) and photoredox catalysts. Chem.
Commun. 2017, 53, 3098−3101.
(7) Fan, T.; Chen, X.; Lin, Z. Theoretical studies of reactions of
carbon dioxide mediated and catalysed by transition metal complexes.
Chem. Commun. 2012, 48, 10808−10828.
(8) Ren, Q.; Wu, N.; Ying, C.; Fang, J. DFT Study of the
Mechanisms of Iron-Catalyzed Regioselective Synthesis of α-Aryl
Carboxylic Acids from Styrene Derivatives and CO2. Organometallics
2016, 35, 3932−3938.
(9) Ostapowicz, T. G.; Holscher, M.; Leitner, W. Catalytic
̈
Hydrocarboxylation of Olefins with CO2 and H2 -a DFT Computa-
tional Analysis. Eur. J. Inorg. Chem. 2012, 2012, 5632−5641.
(10) Yuan, R.; Lin, Z. Computational Insight into the Mechanism of
Nickel-Catalyzed Reductive Carboxylation of Styrenes using CO2.
Organometallics 2014, 33, 7147−7156.
(11) Vummaleti, S. V. C.; Talarico, G.; Nolan, S. P.; Cavallo, L.;
Poater, A. How easy is CO2 fixation by M−C bond containing
complexes (M = Cu, Ni, Co, Rh, Ir)? Org. Chem. Front. 2016, 3, 19−
23.
(12) Lv, X.; Wu, Y.-B.; Lu, G. Computational exploration of ligand
effects in copper-catalyzed boracarboxylation of styrene with CO2.
Catal. Sci. Technol. 2017, 7, 5049−5054.
(13) Fan, T.; Sheong, F. K.; Lin, Z. DFT Studies on Copper-
Catalyzed Hydrocarboxylation of Alkynes Using CO2 and Hydro-
silanes. Organometallics 2013, 32, 5224−5230.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.organo-
met.7b00899.
Additional computational results as described in the text
and experimental details (PDF)
Optimized coordinates (XYZ)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail for K.H.H.: kathrin.hopmann@uit.no.
(14) Wang, Q.; Jia, J.-F.; Guo, C.-H.; Wu, H.-S. Mechanistic
investigation of Cu(I)-mediated three-component domino reaction of
asymmetrical alkynes with carbon dioxide: Theoretical rationale for the
regioselectivity. J. Organomet. Chem. 2013, 748, 84−88.
ORCID
Annette Bayer: 0000-0003-3481-200X
Kathrin H. Hopmann: 0000-0003-2798-716X
(15) Dang, L.; Lin, Z.; Marder, T. B. DFT Studies on the
Carboxylation of Arylboronate Esters with CO2 Catalyzed by
Copper(I) Complexes. Organometallics 2010, 29, 917−927.
(16) Sayyed, F. B.; Tsuji, Y.; Sakaki, S. The crucial role of a Ni(I)
intermediate in Ni-catalyzed carboxylation of aryl chloride with CO2: a
theoretical study. Chem. Commun. 2013, 49, 10715−10717.
(17) Sayyed, F. B.; Sakaki, S. The crucial roles of MgCl2 as a non-
innocent additive in the Ni-catalyzed carboxylation of benzyl halide
with CO2. Chem. Commun. 2014, 50, 13026−13029.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work has been supported by the Research Council of
Norway through a FRINATEK grant (No. 231706) and a
Centre of Excellence Grant (No. 262695), by the Tromsø
Research Foundation (No. TFS2016KHH), by Notur-The
Norwegian Metacenter for Computational Science through
grants of computer time (Nos. nn9330k and nn4654k), and by
NordForsk (No. 85378) and the members of the Nordic
Consortium for CO2 Conversion (UiT-The Arctic University of
Norway, Uppsala University, Stockholm University, KTH
Royal Institute of Technology, Aarhus University, University
of Oslo, University of Bergen, Helsinki University, and
University of Iceland).
(18) Graham, D. C.; Mitchell, C.; Bruce, M. I.; Metha, G. F.; Bowie,
J. H.; Buntine, M. A. Production of Acrylic Acid through Nickel-
Mediated Coupling of Ethylene and Carbon DioxideA DFT Study.
Organometallics 2007, 26, 6784−6792.
(19) Plessow, P. N.; Schafer, A.; Limbach, M.; Hofmann, P. Acrylate
̈
Formation from CO2 and Ethylene Mediated by Nickel Complexes: A
Theoretical Study. Organometallics 2014, 33, 3657−3668.
(20) Graham, D. C.; Bruce, M. I.; Metha, G. F.; Bowie, J. H.; Buntine,
M. A. Regioselective control of the nickel-mediated coupling of
acetylene and carbon dioxide − A DFT study. J. Organomet. Chem.
2008, 693, 2703−2710.
REFERENCES
■
(21) Ostapowicz, T. G.; Holscher, M.; Leitner, W. CO Insertion into
̈
2
(1) Williams, C. M.; Johnson, J. B.; Rovis, T. Nickel-Catalyzed
Reductive Carboxylation of Styrenes using CO2. J. Am. Chem. Soc.
2008, 130, 14936−14937.
Metal-Carbon Bonds: A Computational Study of Rh(I) Pincer
Complexes. Chem. - Eur. J. 2011, 17, 10329−10338.
(22) Qin, H.; Han, J.; Hao, J.; Kantchev, E. B. Computational and
experimental comparison of diphosphane and diene ligands in the Rh-
catalysed carboxylation of organoboron compounds with CO2. Green
Chem. 2014, 16, 3224−3229.
(2) Shirakawa, E.; Ikeda, D.; Masui, S.; Yoshida, M.; Hayashi, T.
Iron−Copper Cooperative Catalysis in the Reactions of Alkyl
Grignard Reagents: Exchange Reaction with Alkenes and Carbome-
talation of Alkynes. J. Am. Chem. Soc. 2012, 134, 272−279.
(3) Greenhalgh, M. D.; Thomas, S. P. Iron-Catalyzed, Highly
Regioselective Synthesis of α-Aryl Carboxylic Acids from Styrene
Derivatives and CO2. J. Am. Chem. Soc. 2012, 134, 11900−11903.
(23) Obst, M.; Pavlovic, Lj.; Hopmann, K. H. Carbon-Carbon Bonds
with CO2: Insights from Computational Studies. J. Organomet. Chem.
2018, DOI: 10.1016/j.jorganchem.2018.02.020.
G
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(24) Lv, X.; Zhang, L.; Sun, B.; Li, Z.; Wu, Y.-B.; Lu, G.
Computational studies on the Rh-catalyzed carboxylation of a
C(sp2)−H bond using CO2. Catal. Sci. Technol. 2017, 7, 3539−3545.
(25) Schmeier, T. J.; Hazari, N.; Incarvito, C. D.; Raskatov, J. A.
Exploring the reactions of CO2 with PCP supported nickel complexes.
Chem. Commun. 2011, 47, 1824−1826.
(37) Cramer, C. J. Essentials of Compuational Chemistry: Theories and
models; Wiley: Chichester, U.K., 2004; p 378.
(38) Throughout the text, Rh−alkyl refers to the alkyl formed from
the alkene substrate. The rhodium−ethyl complex is referred to as
Rh−ethyl.
(39) Catellani, M.; Mealli, C.; Motti, E.; Paoli, P.; Perez-Carreno, E.;
Pregosin, P. S. Palladium-Arene Interactions in Catalytic Intermedi-
ates: An Experimental and Theoretical Investigation of the Soft
Rearrangement between η1 and η2 Coordination Modes. J. Am. Chem.
Soc. 2002, 124, 4336−4346.
(26) Canepa, G.; Brandt, C. D.; Iig, K.; Wolf, J.; Werner, H.
Exploring the reactions of CO2 with PCP supported nickel complexes.
Chem. - Eur. J. 2003, 9, 2502−2515.
(27) Nolte, M. J.; Gafner, G.; Haines, L. M. An Example of the Co-
ordination of the Tetraphenylboron Anion to a Transition Metal
through an Arene Ring. J. Chem. Soc. D 1969, 1406−1407.
(28) Shao, L.; Geib, S. J.; Badger, P. D.; Cooper, J. N. [2 + 2] Cross
Coupling of Benzene and Tropylium Ligands in Reductively Activated
Piano Stool Complexes of Mn, Cr, and W. J. Am. Chem. Soc. 2002,
124, 14812−14813.
(29) Doux, M.; Ricard, L.; Mathey, F.; Le Floch, P.; Mezailles, N.
Synthesis and Reactivity of the First η6-Rhodium(I) and η6-
Iridium(I)Complexes of 2,6-Bis(trimethylsilyl)phosphinines. Eur. J.
Inorg. Chem. 2003, 2003, 687−698.
(40) Del Rosal, I.; Maron, L.; Poteau, R.; Jolibois, F. DFT
calculations of 1H and 13C NMR chemical shifts in transition metal
hydrides. Dalton Trans. 2008, 30, 3959−3970.
(41) Kumaraswamy, G.; Rambabu, D. Hydroxy-rhodium(I) catalyzed
regioselective Michael addition of cyclic enones. Tetrahedron Lett.
2012, 53, 1042−1044.
(42) Bonnemann, H. Rhodium-Catalyzed C,C-Double Bond
̈
Cleavage by Molecular Oxygen. Helv. Chim. Acta 1983, 66, 177−184.
(30) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision D.01; Gaussian, Inc., Wallingford, CT, 2009.
(31) (a) Becke, A. D. A Density-fnnctional exchange-energy
approximation with correct asymptotic behavior. Phys. Rev. A: At.,
Mol., Opt. Phys. 1988, 38, 3098−3100. (b) Lee, C.; Yang, W.; Parr, R.
G. Development of the Colle-Salvetti correlation-energy formula into a
functional of the electron density. Phys. Rev. B: Condens. Matter Mater.
Phys. 1988, 37, 785−789.
(32) Grimme, S. Semiempirical GGA-Type Density Functional
Constructedwith a Long-Range Dispersion Correction. J. Comput.
Chem. 2006, 27, 1787−1799.
(33) (a) Tomasi, J.; Mennucci, B.; Cammi, R. Quantum Mechanical
Continuum Solvation Models. Chem. Rev. 2005, 105, 2999−3093.
̀
(b) Tomasi, J.; Mennucci, B.; Cances, E. The IEF version of the PCM
solvation method: an overview of a new method addressed to study
molecular solutes at the QM ab initio level. J. Mol. Struct.:
THEOCHEM 1999, 464, 211−216. (c) Cances, E.; Mennucci, B.;
̀
Tomasi, J. A new integral equation formalism for the polarizable
continuum model: Theoretical background and applications to
isotropic and anisotropic dielectrics. J. Chem. Phys. 1997, 107,
3032−3041.
(34) Hopmann, K. H. How Accurate is DFT for Iridium-Mediated
Chemistry? Organometallics 2016, 35, 3795−3807.
(35) Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.; Hollwarth,
̈
̈
A.; Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
̈
A set of f-polarization functions for pseudo-potential basis sets of the
transition metals SC-Cu, Y-Ag and La-Au. Chem. Phys. Lett. 1993, 208,
111−114.
(36) (a) Boys, S. F.; Bernardi, F. The calculation of small molecular
interactions by the differences of separate total energies. Some
procedures with reduced errors. Mol. Phys. 1970, 19, 553−566.
(b) Simon, S.; Duran, M.; Dannenberg, J. J. How does basis set
superposition error change the potential surfaces for hydrogen-bonded
dimers? J. Chem. Phys. 1996, 105, 11024−11031.
H
DOI: 10.1021/acs.organomet.7b00899
Organometallics XXXX, XXX, XXX−XXX