Selective Iron-Catalyzed N-Formylation of Amines using Dihydrogen and Carbon Dioxide

Article abstract of DOI:10.1021/acscatal.7b03834

A family of iron(II) carbonyl hydride species supported by PNP pincer ligands was identified as highly productive catalysts for the N-formylation of amines via CO₂ hydrogenation. Specifically, iron complexes supported by two different types of PNP ligands were examined for formamide production. Complexes containing a PNP ligand with a tertiary amine afforded superior turnover numbers in comparison to complexes containing a bifunctional PNP ligand with a secondary amine, indicating that bifunctional motifs are not required for catalysis. Systems incorporating a tertiary amine containing a PNP ligand were active for the N-formylation of a variety of amine substrates, achieving TONs up to 8900 and conversions as high as 92%. Mechanistic experiments suggest that N-formylation occurs via an initial, reversible reduction of CO₂ to ammonium formate followed by dehydration to produce formamide. Several intermediates relevant to this reaction pathway, as well as iron-containing deactivation species, were isolated and characterized.

Full text of DOI:10.1021/acscatal.7b03834

Research Article
pubs.acs.org/acscatalysis
Cite This: ACS Catal. 2018, 8, 1338−1345
Selective Iron-Catalyzed N‑Formylation of Amines using Dihydrogen
and Carbon Dioxide
Upul Jayarathne,† Nilay Hazari,‡ and Wesley H. Bernskoetter*,†
†The Department of Chemistry, University of Missouri, Columbia, Missouri 65211, United States
‡The Department of Chemistry, Yale University, New Haven, Connecticut 06520, United States
S
* Supporting Information
ABSTRACT: A family of iron(II) carbonyl hydride species
supported by PNP pincer ligands was identified as highly
productive catalysts for the N-formylation of amines via CO2
hydrogenation. Specifically, iron complexes supported by two
different types of PNP ligands were examined for formamide
production. Complexes containing a PNP ligand with a tertiary
amine afforded superior turnover numbers in comparison to complexes containing a bifunctional PNP ligand with a secondary
amine, indicating that bifunctional motifs are not required for catalysis. Systems incorporating a tertiary amine containing a PNP
ligand were active for the N-formylation of a variety of amine substrates, achieving TONs up to 8900 and conversions as high as
92%. Mechanistic experiments suggest that N-formylation occurs via an initial, reversible reduction of CO2 to ammonium
formate followed by dehydration to produce formamide. Several intermediates relevant to this reaction pathway, as well as iron-
containing deactivation species, were isolated and characterized.
KEYWORDS: carbon dioxide, iron, formamide, amine, carbamic acid
formic anhydride, and formic acid.6 In comparison to these
synthetic routes, the reductive functionalization of CO2 offers a
potential environmentally benign alternative for the synthesis of
formamides. Likewise, the use of H2 in lieu of more toxic silane
INTRODUCTION
■
The widespread chemical utilization of carbon dioxide (CO2)
would be a significant step toward realizing the potential of our
renewable carbon resources.1 The conversion of CO2 into
or borane reducing agents is likely to minimize the environ-
value-added chemicals would also offer an economically
mental impact of this approach to N-formylation.7
beneficial outlet for carbon capture technologies.2 Unfortu-
Transition-metal-catalyzed N-formylation using CO2 and H2
was investigated as early as 1970, when Haynes et al. reported
that Vaska’s complex, [(Ph3P)2Ir(CO)Cl], could catalyze the
nately, despite the low toxicity and widespread availability of
CO2, it is rarely utilized as a C1 feedstock for synthesizing fine
or bulk chemicals.3 Two prominent exceptions are in the
formation of dimethylforamide (DMF) with a maximum
production of salicylic acid and urea.3e,f The relative paucity of
turnover number (TON) of 1200.8 In subsequent years other
industrial processes which utilize CO2 as a renewable synthon is
due, in part, to its inherent thermodynamic and kinetic stability.
These factors often necessitate harsh reaction conditions and
highly reactive catalysts to facilitate chemical transformations
with high yields and selectivities, resulting in impractical
systems. In recent years, in order to accelerate the discovery of
sporadic syntheses of DMF via CO2/H2-derived N-formylation
appeared in the literature, each using precious metals to
facilitate the reaction.9,10 In 1996, Noyori and co-workers
reported a ruthenium catalyst, [RuCl2(PMe3)4], that N-
formylates dimethylamine into DMF with an impressive
TON of 370000. A year later, Baiker and co-workers found
systems for the utilization of CO2, our laboratories have
that switching the ancillary phosphine ligand from PMe3 to
reported detailed investigations into the hydrogenation of CO2
to formate by a series of pincer-supported iron catalysts.4
Ph2PCH2CH2PPh2 (dppe) nearly doubles the production of
DMF, and they obtained a TON of 740000 using
Herein we describe an extension of these studies and report a
[RuCl2(dppe)2]. In 2003, the same laboratory expanded these
system catalyzed by an earth-abundant metal for the production
studies to a wider range of amine substrates, most notably
of formamides from CO2, H2, and amines, as well as
including the formylation of aniline (TON = 930), which has
mechanistic experiments which elucidate key features of the
proven to be one of the most challenging substrates. Jessop also
reaction pathway.
reported the synthesis of formanilide from the formylation of
The N-formylation of amines to formamides is performed
aniline using the catalyst [RuCl2(PMe3)4] in the presence of
the strong base 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU). In
industrially on a large scale for a number of different
applications.5a For example, formamides are used as solvents
for ionic compounds as well as feedstocks for the synthesis of
pharmaceuticals, pesticides, and herbicides.5 Traditionally, the
carbonyl moiety is introduced into a formamide using relatively
hazardous reagents such as chlorals, carbon monoxide, acetic
Received: November 9, 2017
Revised: January 1, 2018
© XXXX American Chemical Society
1338
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
Figure 1. First base metal catalysts for N-formylation of amines with CO2.
a
Table 1. Solvent Screening for N-Formylation of Morpholine Using Precatalysts 1 and 2
c
THF (TON)
4 h 16 h
toluene (TON)
dioxane (TON)
ethyl acetate (TON)
catalyst
4 h
1935
16 h
4 h
16 h
4 h
16 h
b
[(iPrPNP)FeH(CO)] (1)
[(iPrPMeNP)FeH(CO)(BH4)] (2)
1930
2440
2660
3420
2340
3020
1700
1720
a
b
Reaction conditions: PH = 500 psi, PCO = 500 psi, 5 mL of reaction solution, 2.5 μmol of [Fe], 12.5 mmol of morpholine. TONs based on the
2
2
c
amount of 4-formylmorpholine production as determined by GC-FID. Reported values are the average of two or more trials. 1-
Morpholinoethanone and ethanol were observed as additional products.
2015, Ding and co-workers reported the most productive N-
formylation catalyst reported thus far, [(PhPMeNP)RuClH-
(CO)], obtaining a TON of 1850000 for the synthesis of 4-
formylmorpholine over 3 days under 70 atm of total CO2/H2
(1/1) pressure at 120 °C.10
There is currently considerable interest in the replacement of
toxic precious-metal catalysts with systems containing earth-
abundant base metals.11 In general, the potential for economic
and environmental advantages motivates most of these efforts,
but the transition to base-metal catalysts also offers the
opportunity for new catalytic selectivity, substrate preference,
and even improved performance in comparison with precious-
metal systems. To date there have been few reports of highly
active well-defined base-metal catalysts for the N-formylation of
amines via reductive CO2 hydrogenation.12,13 Two leading
examples were independently reported by the laboratories of
Beller and Milstein using iron and cobalt catalysts, respectively
(Figure 1).13 Beller and co-workers demonstrated that a
tetraphosphine-supported iron(II) complex is an excellent
precatalyst for N-formylation of dimethyl and diethylamine,
achieving TONs up to 5100 over 20 h under 100 bar total
pressure of H2/CO2 (7/3) at 100 °C. However, the substrate
scope was not examined further. Milstein and co-workers
reported a more broadly applied cobalt pincer catalyst which
transformed a range of primary and secondary amines to the
corresponding formamides. Although the cobalt system could
produce high conversion for several formamides, the observed
productivity was quite modest with a maximum TON of 130
Figure 2. Pincer iron precatalysts screened for N-formylation activity.
Reaction conditions are identical with those applied in Table 1.
over 36 h under 60 bar total pressure of H2/CO2 (1/1) at 150
°C. Our own laboratories have reported similar pincer iron
complexes, [(iPrPNP)FeH(CO)] (1) and [(iPrPNMeP)FeH-
(CO)(BH4)] (2) (iPrPNP = N{CH2CH2(PiPr2)}2; iPrPNMeP =
MeN{CH2CH2(PiPr2)}2), which are highly active for the
hydrogenation of CO2 in the presence of Lewis acids and
poorly nucleophilic amines, such as DBU.14b This activity
motivated further examination of these catalysts for N-
formylation of a diverse group of more nucleophilic amine
substrates in search of more productive base-metal catalysts.
RESULTS AND DISCUSSION
■
Catalytic N-Formylation. Our investigations of N-
formylation catalysis with CO2 and H2 began by screening 1
for activity with three amine substrates, aniline, diphenylamine
and morpholine, under conditions similar to those used in
1339
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
catalysis with precious-metal systems.10 A 0.02 mol % loading
of 1 in THF under 1000 psi (∼69 bar) total pressure of H2/
CO2 (4/1) at 120 °C for 16 h only produced formamide in the
case of morpholine, although the unoptimized TON of 2700
was quite promising. The absence of conversion using the
aromatic amines is perhaps unsurprising, given that these
substrates are also problematic for precious-metal catalysts,
owing to their relatively weak basicity and nucleophilicity.9b,c
The catalytic conversion of morpholine to 4-formylmorpho-
line was used as a benchmark N-formylation reaction to
optimize reaction conditions prior to examining substrate scope
and exploring mechanistic considerations. Complete details of
the reaction optimization are provided in the Supporting
Information. A total pressure of 1000 psi (∼69 bar) of a 1/1
H2/CO2 mixture was the preferred combination of gases, as
enriching the mixture in either H2 or CO2 decreased
conversion. Decreasing the total pressure of the reaction even
to 750 psi also produced a substantial drop in TON, suggesting
that pressures greater than 1000 psi may enhance the reaction
further. However, a limit of 1000 psi was used in this study due
to safety considerations. The reaction temperature of 120 °C
used in the initial trials was found to be optimal following
variation between 100 and 140 °C. Altering the reaction solvent
demonstrated that similar catalyst performance is observed in
THF and toluene, while TONs diminish slightly in dioxane and
ethyl acetate during 4 h reaction trials (Table 1). It is notable
that N-formylation works well in both ethereal and hydro-
carbon solvents.
In recent years, our laboratories and others have developed a
number of closely related pincer-supported iron (pre)catalysts
for reversible hydrogenation reactions.14 In many cases, more
than one such iron complex proved to be catalytically active,
which motivated a comparative catalytic study of a range of
well-defined (pre)catalysts (Figure 2). The highest TONs were
obtained for the iron complexes bearing an N-methyl version of
the PNP ligand, [(iPrPNMeP)FeH(CO)(BH4)] (2) and
[(iPrPNMeP)Fe(H)2(CO)] (3). These species show very similar
activities, which is in agreement with prior experiments
demonstrating that 2 rapidly eliminates BH3 in the presence
of excess amine.14b The absence of observable catalytic activity
using (iPrPNMeP)Fe(H)(BH4) (4) suggests the presence of the
ancillary carbonyl ligand is crucial to reductive functionalization
of CO2 in this system. Complex 2 also performed comparably
well in toluene (Table 1). These screening results mirror many
of those in our prior investigations into the hydrogenation of
CO2 to formate salts using iron catalysts with DBU as an
external base.14b There too, the best productivity was obtained
using 2 and 3, which contain the tertiary PNMeP ligand.
One notable difference between the two catalytic processes
to produce formate salts and formamide is the influence of
Lewis acidic lithium salts. In CO2 hydrogenation to formate,
TONs for all catalysts were dramatically enhanced (up to 20-
fold) by the addition of lithium triflate (LiOTf); however, the
addition of 20 equiv of LiOTf (relative to iron) noticeably
reduced formamide production using catalysts 1 and 2 (Table
2; entries 1, 2 and 5, 6). Further consideration of this effect is
discussed in the Supporting Information. Hypothesizing that a
stronger base may accelerate any potential dihydrogen
activation or deprotonation steps in catalysis, we also tested
DBU as an additive. Though the presence of this hindered base
did marginally improve formamide production, it was deemed
not sufficiently effective to warrant its inclusion in the standard
reaction conditions.
Table 2. Catalytic Effect of Additives on N-Formylation
Reaction
a
amt, equiv
b
c
entry catalyst LiOTf DBU 2-fluorophenol TON
yield, %
1
2
3
4
5
6
7
8
1
1
1
1
2
2
2
2
1930
1030
1910
1680
2440
615
39
21
38
34
49
12
51
39
20
20
20
20
20
20
2530
1940
a
Reaction conditions: PH = 500 psi, PCO = 500 psi, 0.14 mmol of
2
2
[Fe], 7 mmol of morpholine with appropriate additive in 5 mL of THF
b
heated at 120 °C for 4 h. TONs based on the amount of 4-
formylmorpholine production as determined by GC-FID. Reported
c
values are the average of two or more trials. Yield reported relative to
amine loading.
As a counterpoint to DBU addition, an acidic alcohol, 2-
fluorophenol, was tested as an additive. Several other studies of
CO2 hydrogenation have noted an accelerating effect from
acidic alcohols,15 but in this case the phenol had a slight
inhibitory influence on TON. The results of the additive
experiments indicate that there is no advantage to including
these reagents to our standard N-formylation; however, the
data raise some mechanistic questions which are discussed
further in the Supporting Information.
Having empirically devised a set of optimized conditions for
N-formylation, we examined the amine substrate scope using 2
as the precatalyst (Table 3). For experimental convenience
these comparative catalytic trials were conducted over 4 h
intervals, though a notable increase in conversion was obtained
with several substrates by extending reactions to 16 h (entries 3,
13, 17, and 19). A collection of primary and secondary aliphatic
and benzylic amines was successfully converted to formamides,
all with essentially complete selectivity (as observed by GC-
FID) and with TONs up to ∼4600 over 4 h. In one case (entry
15), decreasing the catalyst loading and extending reaction
times afforded a remarkable TON of 8900, surpassing the
production of formamide from previously reported base-metal
catalysts. Several −OR and −OH functionalized amines also
exhibited moderate activity (entries 10−12). Overall, the
productivity of catalyst 2 for N-formylation with CO2 and H2
was remarkable in comparison to the handful of other reported
first-row-metal catalysts. In many cases, our iron system
achieves TONs which are 10−100-fold higher than those
reported for well-defined manganese, cobalt, iron, and copper
catalysts.12c,13a,c,12d It is also notable that, in cases where
comparable TONs have been achieved, other existing first-row-
metal catalysts require ≥20 h (under similar temperature and
pressure) to reach this production, in comparison to only 4 h
using catalyst 2. While the state of the art precious-metal
catalysts are still superior to the productivity of 2, this iron-
based system provides a significant advance in the development
of base-metal systems.
1340
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
Table 3. Iron-Catalyzed N-Formylation via Reductive CO2
Hydrogenation
Table 3. continued
a
based on the amount of formamide production as determined by GC-
c
FID. Reported values are the average of two or more trials. Yield
d
e
reported relative to amine loading. TON (yield) after 16 h. TON
(yield) after 16 h using 2.5 μmol of [Fe] and 37.5 mmol of amine.
Closer examination of the results in Table 3 indicates several
trends in substrate preference. In general, the acyclic secondary
amines show higher conversion than the corresponding primary
amines as can be observed from comparing entries 3 and 4, 8
and 9, and 10 and 12. However, steric considerations can alter
this trend, as seen in entries 6 and 20. The iron-catalyzed N-
formylation favors amine substrates bearing electron-donating
substituents (entries 5−7), which is analogous to the trend
observed with more widely studied precious-metal catalysts.9b,10
This preference for electron-rich amines could arise from either
the enhanced basicity of the substrate or improved nucleophil-
icity. It is often difficult to completely separate these two
influences, but certainly enhanced catalytic performance is only
loosely correlated with the pKa of the amine (Table S5). Clearly
steric constraints also play a strong role in catalytic efficiency, as
cyclohexylamine (entry 13) has a TON of 900 while
dicyclohexylamine (entry 14) affords little to no conversion.
Similarly, conversion of 1-heptylamine (entry 1) was
significantly greater than that of 2-heptylamine (entry 2),
where the −NH2 unit is attached to a secondary carbon (entry
2). Steric influences may also contribute to the excellent
performance of the cyclic amine substrates (entries 15−19),
which exhibit considerable activity over the 4 h reaction period.
Notably, when isoquinoline (entry 19) is allowed to react for
16 h, only a modest increase in TON is obtained, suggesting a
greater catalyst deactivation with this substrate.
Mechanistic Considerations. Several prior studies of
transition-metal-catalyzed N-formylation from CO2 and H2
have proposed the common mechanistic pathway depicted in
Scheme 1.13a,9g The pathway begins with a metal hydride
promoted reduction of CO2 to formate. Amine-assisted
extrusion of the formate, along with activation of dihydrogen,
would then regenerate the metal hydride species and liberate an
ammonium formate salt. Finally, dehydration of ammonium
formate would yield the observed formamide product followed
by extrusion of an ammonium formate salt. Though this oft-
proposed route appears quite reasonable, minimal direct
evidence in support of the pathway has been gathered to
date.13a,9g
During the course of catalytic N-formylation experiments our
laboratories made several observations relevant to elucidating
the reaction pathway. Most notably, at the completion of
several catalytic trials a colorless precipitate was observed upon
cooling, depressurizing, and opening the stainless steel
autoclave reactor. This species was initially thought to be the
carbamate salt resulting from residual amine substrate reacting
with the high pressure of CO2 (eq 1).
[R2NCO2][R2NH2] + H2 ⇄ CO2 + 2R2NH + H2
⇄ [R2NH2][O2CH] + R2NH
(1)
Analysis by NMR spectroscopy confirmed that this was
indeed true for most substrates where a precipitate formed.
However, in the cases of cyclohexylamine and phenethylamine
NMR analysis revealed the solid also contained the
a
Reaction conditions: PH = 500 psi, PCO = 500 psi, 2.5 μmol of [Fe],
2
2
b
12.5 mmol of amine in 5 mL of THF heated at 120 °C for 4 h. TONs
1341
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
Scheme 1. Generic Pathway for N-Formylation of Amines via CO2 Reduction
a
Scheme 2. Dehydration and Dehydrogenation of Formate Salt in the Presence and Absence of 2
a
Reaction conditions: PH = 500 psi, PCO = 500 psi, 2.5 μmol of [Fe], 6.25 mmol of amine in 5 mL of THF heated at 120 °C for 4 h.
2
2
Figure 3. (a) Molecular structure of 7 with ellipsoids at 30% probability. Hydrogen atoms have been removed for clarity. Selected bond lengths (Å)
and angles (deg): Fe(1)−P(1) 2.2130(7), Fe(1)−P(2) 2.2081(7), Fe(1)−N(1) 2.1449(2), Fe(1)−C(18) 1.7034(2); P(1)−Fe(1)−P(2) 164.64(3),
Fe(1)−O(3)−C(19) 125.9 (2). O(3)−C(19)−O(2) 128.4(2). (b) Molecular structure of 8 with ellipsoids at 30% probability. Most hydrogen atoms
have been removed for clarity. Selected bond lengths (Å) and angles (deg): Fe(1)−P(1) 2.205(1), Fe(1)−P(2) 2.209(1), Fe(1)−N(1) 2.18(1),
Fe(1)−C(18) 1.726(5), Fe(1)−C(19) 1.736(5); P(1)−Fe(1)−P(2) 166.65(5), C(18)−Fe(1)−C(19) 117.1(2).
corresponding ammonium formate salt, supporting the
hypothesis that formate salts are intermediates in the formation
of formamides. This possibility was further tested by taking
samples of cyclohexylamonnium and phenylethylammonium
formate16 and placing them under the catalytic conditions.
Surprisingly, the conversion of the formate salt to the
corresponding formamide was quite poor in both cases, with
the majority of ammonium formate reverting to the amine and
CO2 (Scheme 2). When the same experiments were conducted
in the absence of 2, dehydration to the formamide product was
nearly quantitative under the same conditions. This suggests
that, while ammonium formate salts are viable and observable
intermediates to formamide production, the dehydration step is
not metal catalyzed and in fact competes with an iron-mediated
dehydrogenation back to the amine and CO2.
2 in benzene-d6 were treated with ≥6 equiv of pyrrolidine,
placed under 2 atm of H2/CO2 (1/1), heated at 50 °C, and
1
monitored by H and 31P NMR spectroscopy. Sample spectra
from these experiments may be found in the Supporting
Information. NMR spectra taken immediately after sample
preparation, and prior to gas addition, exhibit resonances
identical with those previously reported for the iron(II)
dihydride compound 3.14b This confirms that 2 functions as a
precatalyst for 3 and is consistent with their near-identical
catalytic performances. Pressurizing the sample with H2 and
CO2 resulted in rapid CO2 insertion into the Fe−H bond and
formation of the previously reported iron(II) formate complex
[(iPrPNMeP)FeH(CO)(HCO2)] (6).14b In addition to complex
6, a second iron-containing product was detected with a
1
signature Fe−H resonance at −24.83 ppm in the H NMR
In order to obtain mechanistic details regarding the
speciation of the iron complexes, catalytic N-formylation
experiments were carried out in J. Young NMR tubes under
modified reaction conditions. In typical experiments, samples of
spectrum and a peak at 89.7 ppm in the 31P NMR spectrum.
This new species was identified as the pyrrolidine carbamate
complex [(iPrPNMeP)FeH(CO)(C4H8NCO2)] (7). Complex 7
was independently prepared by exposing isolated samples of 6
1342
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
Scheme 3. Proposed Reaction Pathway for Catalytic N-Formylation
to pyrolidine under an H2 atmosphere, and its molecular
structure is depicted in Figure 3. Similar carbamato structures
have been previously reported on biorelevant metals.17
Continued monitoring of the modified catalytic reaction for
12 h at 50 °C resulted in peaks indicative of 1-formylpyrrolidine
formation in the 1H NMR spectrum along with partial
decomposition of the iron species to free ligand and an
iron(0) dicarbonyl species, [(iPrPNMeP)Fe(CO)2] (8) (Figure
3).18 Despite the partial decomposition, complexes 6 and 7
remained the major organometallic species in solution after 12
h.
The dramatic difference between the pressure and temper-
ature conditions of the standard catalytic reaction and the NMR
model reaction require some caution in using the NMR data to
draw conclusions in regard to the mechanism of catalysis.
However, it seems reasonable to use the observations in
outlining the general reaction landscape (Scheme 3). The initial
activation of 2 to the iron(II) dihydride species 3 by pyrrolidine
was self-evident in the NMR spectra and will likely be nearly
instantaneous in the catalytic reaction when 5000 equiv of base
is present. As has been described in prior studies,14b the
reaction of pyrrolidine with 6. However, under our standard
catalytic conditions, we hypothesize that 7 could also originate
from the reaction of 3 with the ammonium carbamate salt,
which is likely present in significant concentrations and was
observed as a precipitate following several catalytic trials. We
were initially concerned that formation of complex 7 could be a
deactivation pathway, but employing isolated samples of 7 in
our standard catalytic reaction afforded a TON of 4200 for
pyrrolidine N-formylation. This productivity is comparable to
that using 2 or 3 as the (pre)catalyst and suggests that 7 is
likely a resting state, not a deactivation species. In contrast,
formation of the iron(0) dicarbonyl complex 8 does appear to
be a deactivating process, as it produced a TON of only 520
under analogous catalytic conditions. The pathway for the
formation of 8 is not clear, though the secondary amine ligated
congener, [(CyPNHP)Fe(CO)2], is proposed to form via H2
reductive elimination from the corresponding iron(II) dihy-
dride species [(iPrPNHP)Fe(H)2(CO)].14f A similar route for
H2 elimination from 3 (which is present as cis and trans
isomers) could account from the formation of 8.
CONCLUDING REMARKS
insertion of CO2 into 3 to produce the iron(II) formate
■
complex 6 appears to be rapid even at low pressure. The
subsequent reaction of 6 with amine to generate an ammonium
formate (and later formamide following dehydration) is
consistent with isolation of these salts following several catalytic
reactions (vide supra). An accompanying activation of H2
would regenerate 3. The more surprising observation was the
formation of the iron(II) carbamate complex 7. In the NMR
tube and synthetic experiments, 7 was obtained from the
The catalytic activity of 1 and 2 for the production of
formamides from amines, CO2, and H2 represents an advance
toward the goal of using earth-abundant metals to harness
renewable carbon sources. Perhaps more significantly, the
superior performance of the tertiary amine ligated 2 over the
secondary amine supported 1 demonstrates that metal−ligand
bifunctional reactivity is not always required to obtain good
activity with these base-metal systems. This is a feature which
1343
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
might be easily overlooked, given the significant recent
successes in base-metal−ligand bifunctional catalyst devel-
opment.4,13a,14e,19 Precatalyst 2 is remarkably productive for
N-formylation within the context of base-metal catalysis,
achieving TONs in excess of 4500 and high conversions over
relatively short reaction times. This level of productivity is
essentially state of the art for earth-abundant metals, though
further catalyst development will be required to match the
performance obtained with precious-metal systems. On the
basis of the mechanistic insights outlined here, obviating
catalyst deactivation via formation of iron dicarbonyl species is
one avenue which may lead to significant TON enhancement.
Although the pathway for iron dicarbonyl complex formation
has not been fully elucidated, it appears likely to involve a
bimolecular event between catalytic species. Approaches such
as catalyst site isolation and intermolecular steric repulsion to
prevent these events are among the methods currently under
consideration in our laboratories. Additionally, the CO2
reduction technology described here could potentially be
coupled with other transformations of formamides to broaden
the scope of CO2 utilization or provide net pathways to CO2-
based chemical hydrogen storage materials.
above flask using a syringe. The mixture was stirred for 1 h and
washed with THF (5 mL × 3) and filtered to obtain a white
solid. This solid was dried under vacuum overnight.
Synthesis of [(iPrPMeNP)FeH(CO)(C4H8NCO2)] (7). Inside
a glovebox the iron formate complex 6 (0.035 g, 0.08 mmol)
was measured into a vial and transferred to a 50 mL Schlenk
bomb using THF (5 mL). Pyrrolidine (0.127 mL, 1.6 mmol)
was added to the bomb, and H2 (1 atm) was introduced using a
high-vacuum line. This mixture was stirred overnight at 50 °C.
Upon evaporation of the solvent a yellow solid was obtained.
This solid was dissolved in 5 mL of pentane and concentrated
to ∼3 mL. After the solution was left overnight at −35 °C, a
yellow precipitate formed, which was isolated by filtration.
Yield: 0.018 g (44%). This solid contained a small amount of
[(iPrPMeNP)Fe(CO)2)]. Crystals of 7 suitable for X-ray
diffraction were obtained by leaving a concentrated solution
1
of this solid in pentane at −35 °C overnight. H NMR (THF-
d8): δ 3.16 (s, 4H, CH2(py)), 2.91−1.91 (br, 11H, ligand CH2,
overlap with N-Me)), 1.64 (br, 4H, CH2(py), overlap with
i
solvent peak), 1.51−0.74 (br, 28H, Pr), −24.66 (t, Fe-H).
31P{1H} NMR (THF-d8): δ 89.2 (major), 90.2 (minor).
13C{1H} NMR (THF-d8): δ 223.48 ((Fe-CO)), 162.23 (Fe−
OCOPy), 65.83 (t, CH2), 47.23 (s, CH2), 45.26 (CH2), 27.06
(t, PiPr2), 26.58 (t, PiPr2), 20.38 (s, PiPr2), 19.74 (s, PiPr2),
19.18 (s, PiPr2), 18.08 (s, PiPr2).
EXPERIMENTAL SECTION
■
General Considerations. All manipulations were carried
out using standard vacuum, Schlenk, cannula, or glovebox
techniques. Hydrogen was purchased from Airgas and used as
received. Catalysts 1−4 and 6 were prepared as previously
described.14b All amines were purchased from Aldrich or Fisher
Scientific and purified by sublimation or vacuum distillation
after drying over CaH2 or NaH. Inorganic materials were
purchased from Strem and isotopically labeled materials from
Cambridge Isotope Laboratories. All nonvolatile solids were
dried under vacuum at 50 °C overnight. Solvents were dried
and deoxygenated using literature procedures.20 1H, 13C, and
31P NMR spectra were recorded on a Bruker 300 MHz DRX,
500 MHz DRX, or 600 MHz spectrometer at ambient
Synthesis of [(iPrPMeNP)Fe(CO)2]. Inside a glovebox a
sample of KC8 (0.039 g, 0.282 mmol, 2.1 equiv) was measured
into a vial and transferred to a 100 mL Schlenk bomb using
THF (∼5 mL). This mixture was frozen using a cold well, and a
crystalline sample of [(iPrPMeNP)FeCl2], (0.069 g, 0.134 mmol,
1 equiv) was added. The bomb was immediately removed from
the box and kept frozen until CO (1 atm) was added. The
mixture was then stirred at room temperature for 3 h. The
volatiles were removed under reduced pressure, and the residue
was extracted with THF (3 mL × 4) and filtered through a bed
of Celite. The filtrate was evaporated under reduced pressure,
and an orange solid was obtained. This solid was washed with
pentane (2 mL × 3) and dried for 4 h to obtain a
spectroscopically pure product. Yield: 0.032 g (0.074 mmol,
51%). Crystals suitable for X-ray diffraction were obtained by
leaving a concentrated solution of [(iPrPMeNP)Fe(CO)2)] in
temperature, unless otherwise noted. H and 13C chemical
1
shifts are referenced using the solvent resonances; 31P chemical
shifts are referenced to an external standard of H3PO4. Probe
temperatures were calibrated using ethylene glycol and
methanol as previously described.21 High-pressure catalytic
hydrogenation reactions were performed using a Parr 5500
series compact reactor with glass insert.
1
THF at −35 °C overnight. H NMR (C6D6): δ 2.30 (m, 2H,
CH2), 2.18−2.02 (m, 4H, CH2), 2.01−1.89 (m, 2H, CH2,
overlap with N-Me), 1.94 (s, 3H, N-Me, overlap with CH2),
1.55 (sep, 2H, PiPr2), 1.06−1.43 (m, 26H, PiPr2). 31P{1H}
NMR (C6D6): δ 104.6. 13C{1H} NMR: δ 223.13 (Fe-CO),
220.48 (Fe-CO), 64.64 (t, CH2), 52.32(s, CH2), 26.98(t, PiPr2),
26.24 (t, PiPr2), 23.42 (t, PiPr2), 17.96 (s, PiPr2), 17.67 (s,
PiPr2).
General Method for Catalytic N-Formylation of
Amines. Inside a glovebox, a glass reactor liner (50 mL) was
charged with the amine (12.5 mmol) and THF (5 mL). Then a
solution of 1−6 (2.5 μmol) in THF was added to this mixture
via microsyringe. The Parr reactor was sealed and removed
from the glovebox. The reactor was pressurized with
commercial-grade CO2 at ambient temperature (500 psi),
followed by commercial-grade H2 (500 psi). After the allotted
time (4−16 h), the reactor was cooled by submersion in an ice
bath and the remaining gas was slowly vented. The product
ASSOCIATED CONTENT
■
S
* Supporting Information
1
solution was then analyzed by H NMR spectroscopy and/or
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.7b03834.
GC-FID using mesitylene as a standard. Control experiments
using FeCl2 as a catalyst using this procedure gave no
detectable conversion.
Additional experimental details, discussion, and selected
General Method for the Synthesis of Formate Salts.
Inside the glovebox the required amine was measured into a
round-bottom flask sealed with a rubber septum. This flask was
removed from the glovebox and cooled to 0 °C using an ice
bath. Degassed formic acid (1.2 equiv) was transferred to the
NMR spectra (PDF)
Crystallographic data (CIF)
Crystallographic data (CIF)
Crystallographic data (CIF)
1344
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345

ACS Catalysis
Research Article
(9) (a) Kayaki, Y.; Shimokawatoko, Y.; Ikariya, T. Adv. Synth. Catal.
2003, 345, 175−179. (b) Schmid, L.; Canonica, A.; Baiker, A. Appl.
Catal., A 2003, 255, 23−33. (c) Munshi, P.; Heldebrant, D. J.;
McKoon, E. P.; Kelly, P. A.; Tai, C.-C.; Jessop, P. G. Tetrahedron Lett.
2003, 44, 2725−2727. (d) Kayaki, O.; Suzuki, T.; Ikariya. Chem. Lett.
AUTHOR INFORMATION
Corresponding Author
*E-mail for W.H.B.: bernskoetterwh@missouri.edu.
ORCID
Nilay Hazari: 0000-0001-8337-198X
Wesley H. Bernskoetter: 0000-0003-0738-5946
Notes
■
2001, 30, 1016−1017. (e) Krocher, O.; Koppel, R. A.; Baiker, A. Chem.
̈
̈
Commun. 1997, 453−454. (f) Jessop, P. G.; Hsiao, Y.; Ikariya, T.;
Noyori, R. J. Am. Chem. Soc. 1994, 116, 8851−8852. (g) Jessop, P. G.;
Hsiao, Y.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1996, 118, 344−
355. (h) Schreiner, S.; Yu, J. Y.; Vaska, L. J. Chem. Soc., Chem.
Commun. 1988, 602. (i) Schreiner, S.; Yu, J. Y.; Vaska, L. Inorg. Chim.
Acta 1988, 147, 139−141.
(10) Zhang, L.; Han, Z.; Zhao, X.; Wang, Z.; Ding, K. Angew. Chem.,
Int. Ed. 2015, 54, 6186−6189.
(11) (a) Chirik, P. J.; Gunnoe, T. B. ACS Catal. 2015, 5, 5584.
(b) Chirik, P.; Morris, R. Acc. Chem. Res. 2015, 48, 2495. (c) Fuerstner,
A. ACS Cent. Sci. 2016, 2, 778−789.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the U.S. Department of Energy,
Office of Science, Basic Energy Sciences, Catalysis Science
Program, under Award DE-SC0018222 and the Curators of the
University of Missouri.
(12) (a) Liu, H.; Mei, Q.; Xu, Q.; Song, J.; Liu, H.; Han, B. Green
Chem. 2017, 19, 196−201. (b) Affan, M. A.; Jessop, P. G. Inorg. Chem.
2017, 56, 7301−7805. (c) Dubey, A.; Nencini, L.; Fayzullin, R. R.;
Nervi, C.; Khusnutdinova, J. R. ACS Catal. 2017, 7, 3864−3868.
(d) Kumar, S.; Jian, S. L. RSC Adv. 2014, 4, 64277−64279. (e) Minato,
M.; Zhou, D.-Y.; Sumiura, K.; Oshima, Y.; Mine, S.; Ito, T.; Kakeya,
M.; Hoshino, K.; Asaeda, T.; Nakada, T.; Osakada, K. Organometallics
2012, 31, 4941−4949.
(13) (a) Daw, P.; Chakraborty, S.; Leitus, G.; Diskin-Posner, Y.; Ben-
David, Y.; Milstein, D. ACS Catal. 2017, 7, 2500−2504. (b) Federsel,
C.; Ziebart, C.; Jacstell, R.; Jennerjahn; Baumann, W.; Beller, M. Chem.
- Eur. J. 2012, 18, 72−75. (c) Federsel, C.; Boddien, A.; Jacstell, R.;
Jennerjahn, R.; Dyson, P. J.; Scopelliti, R.; Laurenczy, G.; Beller, M.
Angew. Chem., Int. Ed. 2010, 49, 9777−9780.
REFERENCES
■
(1) (a) Dincer, I. Energy Policy 1999, 27, 845−854. (b) Lewis, N. S.;
Nocera, D. G. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15729−15735.
(b1) Dimitratos, N.; Lopez-Sanchez, J.; Hutchings, G. Top. Catal.
2009, 52, 258−268. (c) Darensbourg, D. J. Inorg. Chem. 2010, 49,
10765−10780. (d) Riduan, S. N.; Zhang, Y. Dalton Trans. 2010, 39,
3347−3357. (e) Cokoja, M.; Bruckmeier, C.; Rieger, B.; Herrmann,
W. A.; Kuhn, F. E. Angew. Chem., Int. Ed. 2011, 50, 8510−8537.
̈
(f) Appel, A. M.; Bercaw, J. E.; Bocarsly, A. B.; Dobbek, H.; DuBois, D.
L.; Dupuis, M.; Ferry, J. G.; Fujita, E.; Hille, R.; Kenis, P. J. A.; Kerfeld,
C. A.; Morris, R. H.; Peden, C. H. F.; Portis, A. R.; Ragsdale, S. W.;
Rauchfuss, T. B.; Reek, J. N. H.; Seefeldt, L. C.; Thauer, R. K.;
Waldrop, G. L. Chem. Rev. 2013, 113, 6621−6658.
(14) (a) Werkmeister, S.; Junge, K.; Wendt, B.; Alberico, E.; Jiao, H.;
Baumann, W.; Junge, H.; Gallou, F.; Beller, M. Angew. Chem., Int. Ed.
2014, 53, 8722−8726. (b) Zhang, Y.; MacIntosh, A.; Wong, J. L.;
Bielinski, E. A.; Williard, P. G.; Mercado, B. Q.; Hazari, N.;
Bernskoetter, W. H. Chem. Sci. 2015, 6, 4291−4299. (c) Chakraborty,
(2) (a) Federsel, C.; Jackstell, R.; Beller, M. Angew. Chem., Int. Ed.
2010, 49, 6254−6257. (b) Peters, M.; Kohler, B.; Kuckshinrichs, W.;
Leitner, W.; Markewitz, P.; Muller, T. E. ChemSusChem 2011, 4,
1216−1240. (c) Wang, W.-H.; Himeda, Y.; Muckerman, J. T.;
Manbeck, G. F.; Fujita, E. Chem. Rev. 2015, 115, 12936−12937.
(d) Klankermayer, J.; Wesselbaum, S.; Beydoun, K.; Leitner, W.
Angew. Chem., Int. Ed. 2016, 55, 7296−7343.
̈
̈
S.; Lagaditis, P. O.; Forster, M.; Bielinski, E. A.; Hazari, N.;
̈
Holthausen, M. C.; Jones, W. D.; Schneider, S. ACS Catal. 2014, 4,
3994−4003. (d) Schneck, F.; Assmann, M.; Balmer, M.; Harms, K.;
Langer, R. Organometallics 2016, 35, 1931−1943. (e) Rezayee, N. M.;
Samblanet, D. C.; Sanford, M. S. ACS Catal. 2016, 6, 6377−6383.
(f) Bielinski, E. A.; Lagaditis, P. O.; Zhang, Y.; Mercado, B. Q.;
(3) (a) Feedstocks for the Future; Bozell, J., Patel, M. K., Eds.;
American Chemical Society: Washington, DC, 2006; ACS Symposium
Series 921. (b) Renewable Raw Materials: New Feedstocks for the
Chemical Industry; Ulber, R., Sell, D., Hirth, T., Eds.; Wiley-VCH:
Weinheim, Germany, 2011. (c) Carbon Dioxide as Chemical Feedstock;
Aresta, M., Ed.; Wiley-VCH: Weinheim, Germany, 2011. (d) Trans-
formation and Utilization of Carbon Dioxide; Bhanage, B. M., Arai, M.,
Eds.; Springer: Berlin, 2014. (e) Lindsey, A. S.; Jeskey, H. Chem. Rev.
1957, 57, 583−620. (f) Krase, N. W.; Gaddy, V. L. J. Ind. Eng. Chem.
1922, 14, 611−615.
Wurtele, C.; Bernskoetter, W. H.; Hazari, N.; Schneider, S. J. Am.
̈
Chem. Soc. 2014, 136, 10234−10237.
(15) Munshi, P.; Main, D. A.; Linehan, J. C.; Tai, C.-C.; Jessop, P. G.
J. Am. Chem. Soc. 2002, 124, 7963−7971.
(16) Dhake, K. P.; Tambade, P. J.; Singhal, R. S.; Bhanage, B. M.
Green Chem. Lett. Rev. 2011, 4, 151−157.
(17) Walther, D.; Ruben, M.; Rau, S. Coord. Chem. Rev. 1999, 182,
67−100.
(4) Bernskoetter, W. H.; Hazari, N. Acc. Chem. Res. 2017, 50, 1049−
1058.
(18) For full details on the independent preparation, characterization,
and molecular structure of 8, see the Supporting Information.
(19) (a) Fillman, K. L.; Bielinski, E. A.; Schmeier, T. J.; Nesvet, J. C.;
Woodruff, T. M.; Pan, C. J.; Takase, M. K.; Hazari, N.; Neidig, M. L.
Inorg. Chem. 2014, 53, 6066−6072.
(5) (a) Bipp, H.; Kieczka, H. Formamides. In Ullmann’s Encyclopedia
of Industrial Chemistry; Wiley-VCH: Weinheim, Germany, 2003; Vol.
15, pp 36−47. (b) Dunetz, J. R.; Magano, J.; Weisenburger, G. A. Org.
Process Res. Dev. 2016, 20, 140−177. (c) Reddy, N. V.; Kumar, P. S.;
Reddy, P. S.; Kantam, M. L.; Reddy, K. R. New J. Chem. 2015, 39,
805−809.
(20) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518−1520.
(6) Gerack, C. J.; McElwee-White, L. Molecules 2014, 19, 7689−
7713.
(21) Sandstrom, J. Dynamic NMR Spectroscopy; Academic Press: New
̈
York, 1982.
(7) (a) Fang, C.; Lu, C.; Liu, M.; Zhu, Y.; Fu, Y.; Lin, B.-L. ACS
Catal. 2016, 6, 7876−7881. (b) Nale, D. B.; Bhanage, B. M. Synlett
2016, 27, 1413−1417. (c) Lv, H.; Xing, Q.; Yue, C.; Lei, Z.; Li, F.
́ ́
Chem. Commun. 2016, 52, 6545−6548. (d) Gonzalez-Sebastian, L.;
Flores-Almo, M.; García, J. J. Organometallics 2015, 34, 763−769.
(e) Nguyen, T. V. Q.; Yoo, W.-J.; Kobayashi, S. Angew. Chem., Int. Ed.
2015, 54, 9209−9212. (f) Jacquet, O.; Frogneux, X.; Gomes, C. D. N.;
Contat, T. Chem. Sci. 2013, 4, 2127−2131. (g) Jacquet, O.; Gomes, C.
D. N.; Ephritikhine, M.; Contat, T. ChemCatChem 2013, 5, 117−120.
(8) Haynes, P.; Slaugh, L. H.; Kohnle, J. F. Tetrahedron Lett. 1970,
11, 365−368.
1345
DOI: 10.1021/acscatal.7b03834
ACS Catal. 2018, 8, 1338−1345