Key words carbonylation - molybdenum - multicomponent reactions - palladium - catalysis
1
Introduction
There has been considerable development in the carbonylation chemistry field since
two of us authored the first Account on Mo(CO)6 -mediated CO gas-free carbonylative reactions.[1 ] The use of nongaseous CO sources has achieved general acceptance within the synthetic
organic community and this update covers results from more than 50 new articles. As
a consequence, this new Account, which covers the literature from October 2011 to
May 2018, is substantially different from the first review. It contains more complex
examples of carbonylative processes and new technologies such as the use of two-chamber
systems for lab-scale synthesis and multicomponent reactions (MCRs). Highlighted methodologies
were to a large extent selected from the authors’ own laboratories.
In the late 1930s, hydroformylation with syngas (the Roelen reaction)[2 ] and hydrocarboxylation with carbon monoxide and water (the Reppe reaction)[3 ] were discovered. However, the finding by Heck and co-workers in 1974 that organohalides
could be carbonylatively coupled with aliphatic alcohols and amines by employing catalytic
amounts of Pd(0) represented a major step forward.[4 ]
[5 ]
[6 ]
The use of CO as a one-carbon building block has many advantages. The catalytic 1,1-insertion
of the carbonyl moiety is highly atom-efficient and provides a valuable synthetic
handle for further structural elaboration of the resulting carbonyl compound. Furthermore,
carbonylations are in essence three-component reactions, and by varying the organohalide
and nucleophile component, considerable product diversity can easily be achieved.
Thus, Pd-catalyzed carbonylation reactions such as aminocarbonylation, alkoxycarbonylations,
hydroformylations, and carbonylative cross-coupling reactions are now essential tools
for radiochemists[7 ]
[8 ] as well as for synthetic and medicinal chemists.[6,9 ]
Despite the huge potential, the acute toxicity, flammable nature, and requirement
for specialized lab equipment, such as metal reactors, in combination with the difficulty
to detect leakages of the colorless and odorless gas have deterred synthetic chemists
from fully applying the useful carbonylation methods despite their synthetic advantages.
As a result, much recent effort has been invested in developing more convenient and
safer methods for handling the toxic carbon monoxide gas.[6 ]
[10 ]
[11 ]
[12 ]
2
Recent Developments
2.1
New CO Sources
In order to avoid handling of gaseous CO, several methods employing a variety of CO
precursors/sources have been developed (Figure [1 ]). One approach has been to utilize molecules with carbonyl motifs, which by the
exposure to transition metals, additives, base, or heat will release CO. Examples,
include alkyl- and arylformates,[11 ]
[13 ]
[14 ]
[15 ]
[16 ] aldehydes,[17 ] formic acid,[18 ] formamides and N -formylsaccharin,[19 ]
[20 ]
[21 ] carbon dioxide,[22 ]
[23 ] and metal carbonyls,[24 ]
[25 ] such as the highly versatile Mo(CO)6.
[26 ]
[27 ]
[28 ] However, several of the mentioned CO sources require an additional transition metal,
strong base, or high temperatures to release CO gas. Alternatively, the use of metal
carbonyls will generate stoichiometric amounts of another transition metal as waste.
Indeed, Mo(CO)6 has been reported to possess catalytic activities[29 ]
[30 ] in addition to reducing aromatic nitro functionalities at elevated temperatures.[31 ]
Figure 1 Representative selection of various CO sources reported in the literature
2.2
Two-Chamber System for ex Situ CO Generation
The issues with compatibility of the CO-generating reaction with the CO-consuming
reaction may impose severe limitations on the scope of nongaseous carbonylation reactions.
An elegant approach that circumvents these problems was developed by Skrydstrup et
al., in which CO was liberated ex situ following Pd-catalyzed decomposition of 9-methylfluorene-9-carbonyl
chloride (COgen). A special two-chamber glassware system was developed to keep the
carbonylation and the decarbonylation reaction mixtures separate, to avoid problems
with incompatibility (Figure [2 ]).[32 ] A similar approach was later described, in which a reaction of silacarboxylic acid
with a fluoride source liberated CO.[33 ] Both these methods allow the use of stoichiometric or substoichiometric amounts
of CO as well as a possibility to introduce an isotopically labeled carbonyl group.[32 ]
[33 ]
Figure 2 Two-Chamber vial after radical carbonylation reaction. Left-hand chamber (CCO ) contains DBU/Mo(CO)6 , right-hand chamber (Crxn ) contains reactants.[34 ]
Ex situ generation of carbon monoxide from solid CO sources, by using two-chamber
glassware, has made it possible to use various carbonylation reactions for small-scale
applications in a standard laboratory since lower pressures of CO can be used, which
in turn eliminates the need for pressurized vessels. There are now several nongaseous
CO sources reported, intended both for in situ and ex situ use, including the base-mediated
decomposition of oxalyl chloride[35 ] and chloroform,[36 ]
[37 ] which have been reported as effective CO-generating strategies for carbonylation
chemistry. Notably, the latter allow the preparation of 13 C- and 14 C-labeled carbonyl derivatives.
Finally, metal carbonyls such as Mo(CO)6 offer a convenient solid CO source suitable for both in situ and ex situ gas release.[1 ]
[38 ] CO is readily released from Mo(CO)6 either by ligand exchange with e.g. DBU[27,39 ] or MeCN,[40 ]
[41 ] or at elevated reaction temperatures.[26 ] However, because of the potential reduction of nitro groups[31 ] and precipitation of molybdenum complexes after the release of CO (complicating
product purification)[42 ] ex situ protocols have been developed for the use of Mo(CO)6 -mediated carbonylations.[38 ]
[43 ] The carbonylative work presented in this review will fully focus on Mo(CO)6 as the nongaseous CO source.
2.3
Multicomponent Carbonylations
A multicomponent reaction is defined as a reaction with three or more reaction components
that react to form a single product that contains essentially all of the atoms of
the starting materials.[44 ]
[45 ] The components may be separate molecular entities or they may be different functional
groups in bifunctional reagents.[44,46 ] As such, carbonylative coupling reactions, comprising the coupling of an electrophile,
CO, and a nucleophile, constitute a three-component reaction. However, carbonylation
reactions with less than four components are not usually categorized as MCRs, because
the CO component is generally fixed, unless different carbon or oxygen isotopes are
employed.[47 ]
Many well-known noncarbonylative MCRs, such as the Mannich,[48 ] Strecker,[49 ] Biginelli,[50 ]
[51 ] Passerini,[52,53 ] and Ugi[54 ] reactions utilize carbonyl derivatives, for example in the form of aldehydes or
ketones, to install additional carbons. The ability to incorporate one-carbon fragments
by Pd(0)-catalyzed carbonylations from an additional source of organo(pseudo)halide
starting materials is one of the reasons why carbonylation chemistry is such a powerful
complement to the field of MCRs. The advance of carbonylation chemistry and the development
of numerous new methods[6 ]
[10 ]
[55 ] has spurred an increased research interest in carbonylative MCRs.[47 ]
[56 ]
There are several advantages to carbonylative MCRs: (i) They are highly atom economical
because nearly all atoms of the starting materials are incorporated into the product.
(ii) The rapid assembly of simple starting materials to generate cyclic and acyclic
scaffolds with increased molecular complexity is readily achieved. Furthermore, by
secondary transformations, for example by using bifunctional reagents or secondary
reactions, a wider chemical space can be reached. This strategy has been successful
in the synthesis of various heterocycles.[57 ]
[58 ]
[59 ] (iii) Limiting the number of steps of a reaction and ideally the number of isolated
intermediates, is both time- and cost-effective. (iv) The waste generated from a reaction,
e.g. from unreacted starting materials and solvents used in purification processes,
is kept to a minimum.
Considering the many benefits of carbonylative MCRs it is not surprising that the
methodology has increased in popularity. However, the majority of carbonylative MCRs
are performed by using gaseous CO[47 ] and only two examples of Mo(CO)6 -mediated MCRs were reported before October 2011.[60 ]
[61 ] In order to meet the demands of convenient and safe methods in the future it will
be of importance to develop carbonylative MCRs that are compatible with nongaseous
CO sources.
Carbonylations with N and O Nucleophiles
3
Carbonylations with N and O Nucleophiles
Most of the Pd(0)-catalyzed carbonylation reactions involve the coupling of a suitable
carbon halide, or pseudo halide, (RX) starting material (e.g. aryl or vinyl) and
a nucleophile (e.g. amine, amide, alcohol, water) in the presence of CO, although
alternative organopalladium precursors can be used through C–H activation or transmetalation
(Scheme [1 ]).
Scheme 1 General depiction of a palladium-catalyzed carbonylation reaction generating the
essential R–Pd intermediate through (a) oxidative addition of RX (b) C–H activation
of RH, or (c) transmetalation through an organometallic reactant RM
The three-component reaction between an organic (pseudo)halide, CO, and an amine to
yield amides is known as an aminocarbonylation reaction. The corresponding reaction with an amide nucleophile is an amidocarbonylation . The reaction with an alcohol is an alkoxycarbonylation , and the process using water is a hydroxycarbonylation . A base is typically required to abstract a proton and usually a ligand is added
to modulate the reactivity of the palladium complexes in the catalytic cycle. Most
commonly, a phosphine ligand is used and several ligand properties may be considered
when designing a catalytic system, such as electronic and steric properties. Moreover,
various steps of the catalytic cycle will be facilitated by different ligand characteristics.
For example, oxidative addition will be promoted by electron-rich phosphine ligands,
whereas the CO 1,1-insertion will be favored by electron-deficient phosphine ligands.[10 ]
[62 ] Typically, Pd(II) salts are used as precatalysts, mainly because of their enhanced
stability compared to available Pd(0) complexes. Therefore, as an initial step before
entering the catalytic cycle, the Pd(II) precatalyst will be reduced by a solvent
molecule, ligand, or CO to a 14-electron Pd(0) complex.[62 ] Once the active catalyst is formed, the first step in the catalytic cycle is the
insertion of Pd(0) into the R–X bond, resulting in oxidation of the Pd(0) species
to a square-planar organopalladium(II) complex (oxidative addition). The rate of oxidative
addition is strongly dependent on the nature of the C–X bond, where strong C–X bonds
will be less reactive. As a result, iodides react more readily than other halides
or such as chlorides (I > OTf ≥ Br > Cl ~ Ts).[6 ] With this argument also follows that electron-deficient C–X substrates are more
susceptible to oxidative addition, and electron-donating ligands that increase the
electron density on Pd will promote oxidative addition.[10 ]
[63 ]
Next, coordination of CO to the Pd center accompanied by ligand displacement and 1,1-insertion
generates the acylpalladium species. The introduction of the nucleophile can occur
either directly on the acyl carbon and thereby releasing the carbonyl compound or
the nucleophile can coordinate to the vacant site on palladium (nucleophilic attack).
Abstraction of a proton from the nucleophile with base and subsequent reductive elimination
then yields the carbonylated product and regenerates the catalytically active Pd(0)
species.[62 ]
[64 ]
During our early work on Mo(CO)6 -mediated carbonylation reactions,[1 ] the use of aryl nitro-group-containing substrates was precluded because of their
facile reduction by Mo species present in the reaction mixture.[31 ] To overcome this problem, we started using a variant of Skrydstrup’s bridged two-chamber
system, in which the carbon monoxide releasing Mo(CO)6 was physically separated from the catalytic reaction mixture (Scheme [2 ]).[12 ]
[38 ] The two-vial system was constructed by fusing two standard pyrex vials through
a borosilicate cylinder and designed to fit DRYSYN™ system, making it both cheap
and extremely convenient to use (Figure [2 ]). The benefits of separating the catalytic and CO-releasing components were clearly
demonstrated through the efficient and high-yielding transformation of various nitro-group-containing
aryl and heteroaryl iodides and bromides into the corresponding benzamides. Notably,
the same reactions conducted in a single-vial setting resulted in substantial competing
nitro group reduction. This ex situ approach is today our preferred method for conducting
Mo(CO)6 -mediated carbonylations and has been employed in the vast majority of reactions performed
in our laboratory since 2012, the only exceptions being when high temperatures are
required to enable the carbonylation of particularly unreactive substrates.
Scheme 2 Carbonylation of aryl halides with use of Mo(CO)6 as an ex situ CO source
The aminocarbonylation of aryl halides with cyanamide by using CO generated ex situ
from Mo(CO)6 to produce N -cyanobenzamides has previously been described (Scheme [3 ]).[65 ]
[66 ]
[67 ] The method was compatible with both aryl iodides at 65 °C and bromides at 85 °C
in moderate to good yields. The mechanism is believed to follow a general aminocarbonylation
reaction with cyanamide acting as a nucleophile through the terminal amine group.
Scheme 3 Carbonylation of aryl halides with cyanamide as nucleophile to yield N -cyanobenzamides
Amino acids are often considered to be challenging nucleophiles due to the inductive
withdrawing nature of the carboxyl group and the additional steric bulk imparted by
the α substituent. In 2013, the aminocarbonylation of 5-aryl-4-iodo-1H -indazoles with a phenylalanine amide nucleophile was reported (Scheme [4 ]).[43 ] The ex situ generation of CO was again leveraged to prepare a number of constrained
H–Phe–Phe–NH2 analogues as part of a medicinal chemistry campaign. The aminocarbonylation reaction
was particularly efficient (72–85%) given that the reaction took place at a hindered
ortho position.
Scheme 4 Example of an amincarbonylation with use of a challenging phenylalanine amide nucleophile
The reductive properties of Mo(CO)6 can also be harnessed to enable the use of nitroarenes as nitrogen sources in aminocarbonylation
reactions. Driver and co-workers have disclosed a dual C–H functionalization/aminocarbonylation
process using 2-pyridyl substituted arenes as R–Pd precursors and nitroarenes as nitrogen
donors (Scheme [5 ]).[68 ] The reaction scope was broad with respect to both reaction components; however,
the requirement for a 2-pyridyl substituent represents a practical limitation. In
a conceptually related study, Wang et al. utilized the Mo-mediated reductive ring-opening
of anthranils as an efficient strategy to generate 2-aminobenzaldehyde derivatives
in situ (Scheme [5 ]).[69 ] In the presence of an aryl iodide and Pd/C, the reactive intermediates were conveniently
transformed into N -(2-carbonylaryl)benzamides in moderate to good yields. Notably, the reaction was
conducted by using water as a green solvent, although the requirement for organic
solvents in the aqueous work-up and silica gel chromatography limit the environmental
benefits of the overall process.
Scheme 5 Top: Directed aminocarbonylation of C(sp2 )–H bonds with use of nitroarenes as amine precursors. Bottom: Aminocarbonylation
of aryl iodides with use of anthranils as amine precursors
The Wu group has recently reported on two related nitrogen-directed C–H functionalization/aminocarbonylation
strategies for the synthesis of 3-methyleneisoindolin-1-ones and 2-arylindazolones
from acyl hydrazones[70 ] and azoarenes,[71 ] respectively (Scheme [6 ]). In both cases, coordination through a Lewis basic nitrogen atom followed by C–H
insertion generated the key R–Pd precursors, which underwent subsequent CO insertion
and intramolecular nucleophilic attack to afford the desired compounds. Interestingly,
no Mo-mediated reductive cleavage of the potentially sensitive substrates was detected.
Scheme 6 Directed carbonylative synthesis of 3-methyleneisoindolin-1-ones (top) and 2-arylindazolones
(bottom) by a C–H annulation strategy; BQ = benzoquinone
Sulfonyl isocyanates are versatile and valuable building blocks in synthetic chemistry;
however, their utility is hampered by a lack of commercial availability, stability,
and methods for their preparation. Recently, sulfonyl azides have been exploited as
convenient precursors for the in situ generation and functionalization of sulfonyl
isocyanates under carbonylative conditions.[72 ] The reaction was found to proceed under ligand-free conditions by using simple PdCl2 and the use of aryl amine or alcohol nucleophiles afforded sulfonyl ureas or carbamates,
respectively (Scheme [7 ]). Mechanistically, the reaction was believed to occur in an analogous fashion to
the general carbonylation mechanism, with oxidative addition on the sulfonyl azide
group leading to a nitrene–palladium complex. Subsequent CO insertion and reductive
elimination furnishes the key sulfonyl isocyanate intermediate, which can then be
trapped by an appropriate nucleophile to afford the desired products. In the case
of aliphatic amines, the carbonylative reaction pathway is disfavored and a competing
direct SN 2 process leads to the exclusive formation of substituted sulfonamides, rather the
expected sulfonyl ureas.
Scheme 7 Substrate-controlled carbonylative synthesis of sulfonyl carbamates or acyl sulfonyl
ureas
The indole scaffold is one of the most important and pervasive structures in organic
and medicinal chemistry[73 ] and new methods to access functionalized indoles are continually in demand. In 2015,
the groups of Wu and Langer disclosed the carbonylative synthesis of N- benzoylindoles from indole and aryl iodides (Scheme [8 ]).[74 ] The reaction scope was explored with a variety of aryl iodides and a significant
preference for electron-rich substrates was noted.
Scheme 8 Carbonylative synthesis of N -benzoylindoles with use of indole as a nucleophile
The use of aromatic amine nucleophiles in aminocarbonylation chemistry can often be
problematic because of their inherent low nucleophilicity. Very recently, Piguel and
co-workers reported an efficient procedure for the carbonylative coupling of various
heteroaryl bromides with aryl and heteroaryl amine nucleophiles (Scheme [9 ]).[75 ] The precatalyst PdCl2 (dppf) was found to be particularly effective in promoting the reaction, producing
a diverse array of products in moderate to excellent yields. The reaction could even
be extended to include the double functionalization of 2,6-diaminopyridine with 3-bromopyridine
to produce an interesting 2,6-diamidopyridine product. The Wu and Langer groups have
also utilized Mo(CO)6 as a CO source in the aminocarbonylation of challenging 2-aminobenzonitrile nucleophiles
using aryl bromides.[76 ] In this case, the use of Pd(OAc)2 and cataCXium A [di(1-adamantyl)-n -butylphosphine, BuPAd2 ] at elevated temperatures was optimal for the reaction (Scheme [9 ]). The yields in many cases were only moderate; however, given the highly challenging
nature of the nucleophile, these results are still rather impressive. In addition,
the one-pot oxidative cyclization of a selected number of the N -(2-cyanoaryl)-benzamide products to give 2-aryl quinazolinone derivatives was also
demonstrated.
Scheme 9 Top: Aminocarbonylation of N-heterocycles with arylamine nucleophiles. Bottom: Carbonylative
synthesis of N -(2-cyanoaryl)-benzamides from aryl halides and 2-aminobenzonitriles
Scheme 10 Carbonylative synthesis of acyl sulfonimidamides and acyl sulfinimides from aryl
or vinyl halides
Similarly, the use of sulfonamide and related nucleophiles is often associated with
lower reactivity because of the inductive effects from the neighboring oxygen atoms.
The groups of Sandström and Arvidsson have described the carbonylative synthesis of
interesting acyl sulfur-containing carboxylic acid bioisoteres using sulfonimidamide[77 ]
[78 ] and sulfonamide[79 ] nucleophiles, respectively. In the former case, (hetero)aryl and vinyl halides or
triflates were suitable reaction partners and were successfully coupled with a selected
number of aryl sulfonimidamide nucleophiles (Scheme [10 ]). In general, the use of (hetero)aryl substrates led to higher yields of the target
compounds and this was attributed to the lower stability of the vinyl acyl sulfonimidamide
products. Interestingly, complete thermolytic removal of the Boc-protecting group
was shown to occur at the same temperature (80 °C) as the carbonylation reaction suggesting
that background Boc deprotection may be a contributing factor to the low reaction
yields obtained with these substrates. The corresponding carbonylative synthesis of
acyl sulfinimides was limited to (hetero)aryl iodide substrates and some erosion in
ee was noted over the course of the reaction (Scheme [10 ]).
In 2016, the groups of Rodríguez, Arraýas, and Carretero published an impressive study
detailing the carbonylative cyclization of (N -SO2 Py)-protected amines using a γ-C(sp3 )–H activation strategy.[80 ] The choice of protecting group and the use of substoichiometric amounts of Mo(CO)6 were essential for obtaining high yields of the γ-lactam products (Scheme [11 ]). The substrate compatibility was demonstrated on a wide variety of different amino
acid and aliphatic amine derivatives containing suitably disposed γ-methyl or γ-methylene
groups. The reaction could also be extended to C(sp2 )–H carbonylative cyclization, and in the case of substrates containing two potentially
reactive C–H groups, carbonylation took place at the more acidic C(sp2 )–H bond. Importantly, the reaction was equally efficient on a gram scale and the
sulfonamide directing group was readily removed by using Mg turnings in MeOH under
sonication. Mechanistically, the reaction was suggested to occur through a Pd(II)-mediated
C–H activation followed by CO insertion and reductive elimination with subsequent
reoxidation of the Pd(0) species by benzoquinone and AgOAc closing the catalytic cycle.
Scheme 11 Directed carbonylative cyclization of amines and amino acids through C(sp3 )–H functionalization; HFIP = 1,1,1,3,3,3-hexafluoro-2-propanol, BQ = benzoquinone
The in situ formation of sulfonyl isocyanates has also been exploited in the carbonylation
of sulfonyl azides with amide nucleophiles.[81 ] The reaction pathway was shown to be dependent on the amide nucleophile with tertiary
amides reacting through a [2+2] cycloaddition/decarboxylation cascade to afford substituted
sulfonyl amidine products (Scheme [12 ]). In contrast, primary and secondary amides proceeded through a more conventional
pathway and nucleophilic attack at the amine nitrogen led to the formation of various
acyl sulfonyl ureas (Scheme [12 ]).
Scheme 12 Substrate- controlled carbonylative synthesis of sulfonyl amidines or acyl sulfonyl ureas
The alkoxylcarbonylation reaction, which is typically defined by the use of oxygen-centered
nucleophiles, is one of the most useful methods for transforming R–Pd precursors into
carboxylic acid and ester derivatives. In 2016, the Wu group reported the Pd(II)-catalyzed
alkoxycarbonylation of C(sp2 )–H bonds using an array of aliphatic alcohol nucleophiles (Scheme [13 ]).[82 ] The use of pendant nitrogen-containing heterocycles (2-pyridine, pyrazole, and pyrimidine)
as directing groups and benzoquinone as a co-oxidant and a Pd ligand was crucial for
reactivity. Under the optimized conditions a range of ester products were obtained
in moderate yields, and good selectivity towards Pd(0)-labile functional groups (bromide
and chloride) was observed. Taszarek and Ressig have recently demonstrated that alkenyl
triflates and nonaflates are also competent substrates for alkoxycarbonylations using
water or methanol as the nucleophile.[83 ] However, only a limited number of substrates were examined and the scope and limitations
of the process are yet to be determined.
Scheme 13 Directed alkoxycarbonylation of C(sp2 )–H bonds; BQ = benzoquinone
One of the most important factors in determining the practical utility of synthetic
methodology is its applicability to real world substrates that fall outside the limited
chemical space usually covered in screening tables. Pleasingly, the Mo(CO)6 -mediated carbonylation reaction has been applied in a wide range of target-molecule-based
studies, primarily medicinal chemistry campaigns aimed at the discovery of new compounds
against multiple different indications. Although a detailed description of the design,
synthesis, and evaluation of these compounds is outside the scope of this account,
some representative examples of target compounds synthesized by using this approach
are given in Scheme [14 ].[84 ]
[85 ] It is clearly evident from the chemotypes represented in Scheme [14 ] that this reaction is not just an academic curiosity and can be effectively utilized
to access a wide range of structurally complex and biologically relevant molecules.
Scheme 14 Selected examples of biologically relevant target compounds synthesized by Mo(CO)6 -mediated alkoxy- or aminocarbonylation. References: VACht,[86 ] kinase,[87 ] AT2 R,[88 ] ghrelin,[89 ] (±)-ampelopsin B,[90 ] CDK8,[91 ] neurotensin,[92 ] α3β4 nAChR ligand[93 ]
All of the examples described in the preceding pages have been conducted under a homogenous
catalysis regime. Despite its immense popularity and utility, homogeneous catalysis
suffers from two major drawbacks. Firstly, the expensive transition-metal catalyst
is often discarded following the reaction because of problems associated with catalyst
recovery. Secondly, contamination of the final product with traces of transition metals
can be troublesome, especially in a good manufacturing practice (GMP) production setting.
These issues have led to the development of numerous immobilized-palladium catalysts
that can be used to catalyze a wide range of cross-coupling reactions, under heterogeneous
conditions.[94 ] In 2015, Hajipour and co-workers reported the synthesis and application of immobilized
palladium containing magnetic nanoparticles [ImmPd(0)-MNPs] in the amino- and alkoxycarbonylation
of aryl iodides (Scheme [15 ]).[95 ] In both cases the substrate scope was wide and good to excellent yields of the carbonylated
products were obtained by using a catalyst loading of 0.14 mol%. Catalyst reusability
was also assessed and a slight decrease in conversion was noted after four consecutive
runs. A control experiment by using a hot-filtration test was conducted to determine
whether the catalytic activity was due to Pd leaching from the solid support, and
no catalytic activity was detected in the filtrate, consistent with the operation
of a heterogeneous catalytic reaction.
Scheme 15 Amino- and alkoxylcarbonylation of aryl iodides by using immobilized palladium-containing
magnetic nanoparticles [ImmPd(0)-MNPs]
Carbonylative Cross-Coupling Reactions with Organometallics
4
Carbonylative Cross-Coupling Reactions with Organometallics
The typical cross-coupling reactions with an organometallic reactant (or more correctly
a transmetalation substrate) are available in a carbonylative version, with one exception,
namely the Kumada cross-coupling. The first reports of the Pd-catalyzed carbonylative
Stille coupling of aryl diazonium salts with organotin reagents appeared in 1982 and
1987.[96 ] Shortly thereafter, Echavarren and Stille presented a similar carbonylative coupling
of aryl triflates with organostannanes.[97 ] Further developments expanded the scope of this reaction and reactions of various
electrophiles with different stannanes have been performed in the presence of CO gas.[45 ] More recently, Nilsson and coworkers presented in situ Mo(CO)6 -assisted cross-coupling of arylstannanes with aryl triflates and aryl bromides using
Pd catalysis.[98 ]
[99 ] In 2015, the CO-free methodology was expanded by the research group of Iranpoor,
using nickel catalysis and predominantly chromium hexacarbonyl as the CO source but
also Mo(CO)6 (Scheme [16 ]).[100 ]
Scheme 16 A nickel-catalyzed carbonylative Stille cross-coupling reaction
Horváth and Rábai introduced the term “fluorous” as an analog to aqueous because of
the special properties of highly fluorinated compounds.[101 ] The poor solubility of fluorous systems in different organic solvents and water
is due to low surface tension, low intermolecular interactions, high density, and
low dielectric constants. The main field of application for fluorous chemistry is
in the fluorous biphasic catalysis method, since the often complex and expensive
catalyst can be recycled. Originally the fluorous-tagged catalyst was dissolved in
the fluorous solvent and the substrate and reagents were added to the organic phase,
which is immiscible with the perfluorocarbons at room temperature. On heating, the
reaction medium becomes homogeneous and the reaction occurs. By utilizing fluorous
extractions or fluorous chromatography the perfluoro-tagged catalyst can be separated
and reused.[102 ] In 2014, Lo and Lam published an article in which they presented expedient Mo(CO)6 -mediated carbonylative Suzuki cross-couplings using a fluorous oxime-based palladacycle
as catalyst (A ) under aqueous or neat conditions.[103 ] By employing microwave heating (MW) with in situ release of carbon monoxide followed
by fluorous silica gel column chromatography, unsymmetric aryl ketones were obtained
in high yields (Scheme [17 ]). The fluorous Pd catalyst was recycled five times and a number of biologically
relevant molecules were synthesized.
Scheme 17 Aqueous carbonylative Suzuki–Miyaura cross-coupling reactions
Alkyl halides have been elusive substrates in transition-metal-mediated cross-coupling
reactions because of their slow oxidative addition and the risk of beta elimination.[104 ] During the last decade, the use of alkyl halides as coupling agents has increased
with the application of radical chemistry to the traditional cross-coupling protocols.[105 ] By creating a single-electron transfer (SET) event, in which an alkyl radical is
generated, the challenges related to the oxidative addition step can be circumvented.
By employing visible-light photocatalysis, unactivated alkyl halides have been used
as substrates in cross-coupling reactions and functionalized under mild conditions,
and displayed great functional group tolerance. The alkyl radical can thus be generated
in catalytic amounts with the aid of an organometallic- or organic dye-based photocatalyst.
Upon irradiation with visible light, the photocatalyst forms an excited-state species
capable of transferring an electron to generate the alkyl radical.
For carbonylative cross-couplings with alkyl halides as substrates, SET methodology
has been employed but the generation of the alkyl radical has varied. Notably, whereas
visible-light photocatalysis has been utilized in an aminocarbonylation,[34 ] the methods have generally applied a combination of intense light or UV irradiation,
elevated carbon monoxide pressures, or elevated temperatures.[106 ]
[107 ]
[108 ]
[109 ]
[110 ]
[111 ]
[112 ]
In 2017, Odell et al reported that the use of visible-light irradiation together with
Pd(0) catalysis enabled the carbonylative Suzuki cross-coupling of unactivated alkyl
iodides and alkyl bromides (Scheme [18 ]).[113 ] The reaction was performed under ambient temperature and pressure whilst utilizing
Mo(CO)6 as an ex situ solid source of carbon monoxide. The methodology represents a very
convenient and accessible reaction procedure, which allowed the preparation of a range
of functionalized aryl alkyl ketones including the antipsychotic drug, melperone.
For a recent example of biaryl ketone synthesis by metal-carbonyl-mediated Suzuki
cross-coupling methodology, see also the published work by Jung et al.[114 ]
Scheme 18 Visible-light-mediated carbonylative Suzuki–Miyaura cross-coupling reactions
Examples of palladium-catalyzed carbonylative Negishi couplings reported in the literature
use CO gas to carry out the carbonylation, and the reaction time is several hours
(ca. 20–30 h).[115 ]
[116 ]
[117 ]
[118 ]
[119 ] Two new Mo(CO)6 -promoted in situ protocols for carbonylative Negishi cross-couplings were developed
for aryl iodides and aryl bromides (Scheme [19 ]).[120 ] The carbonylative cross-coupling reactions were carried out by using commercially
available benzylzinc bromide in closed vials at 90–120 °C for 0.5–1 hours, providing
a set of diarylated ethanones, a common pharmacophore among several pharmaceuticals,
in moderate to high isolated yields (47–84%).
Scheme 19 Carbonylative Negishi cross-coupling reactions
Benzoylacetonitriles are highly useful building blocks in pharmaceutical and material
chemistry fields.[121 ] A straightforward in situ method for the synthesis of benzoylacetonitriles through
CO-free palladium(0)-catalyzed Hiyama-type carbonylative cross-coupling employing
Mo(CO)6 was published in 2012 (Scheme [20 ]).[122 ] The key reactant, trimethylsilyl acetonitrile, was activated by CuF2 and reacted smoothly at 80 °C. The reaction showed good tolerance toward functional
groups such as alkoxy, bromo, chloro, ester, ketone, and nitrile moieties.
Scheme 20 Carbonylative Hiyama-type cross-coupling reaction
Carbonylative Cascade Reactions
5
Carbonylative Cascade Reactions
In 2013, Wu and co-workers presented a synthesis of 2-aminobenzoxazinones from 2-bromoanilines
and isocyanates employing Mo(CO)6 as the CO source (Scheme [21 ]).[123 ] The authors proposed that the corresponding urea is formed in an initial step from
2-bromoaniline and phenylisocyanate. Following oxidative addition of the C–Br bond,
CO insertion, and reductive elimination furnished 2-aminobenzoxazinones in good yields.
Remarkably, the alternative 3-phenylquinazoline-2,4-(1H ,3H )dione was not detected and it was suggested that Mo(CO)6 might act as a Lewis acid and aid in chemoselectivity. With the developed method,
22 examples were prepared in 40–90% isolated yield for a diverse set of reagents.
The scope of the reaction was also expanded to 2-bromophenylisocyanate and aniline,
potentially increasing the scope of the reaction.
Scheme 21 Synthesis of 2-aminobenzoxazinones from 2-bromoanilines and isocyanates by carbonylation/cyclization
Scheme 22 Palladium-catalyzed carbonylative Sonogashira/cyclization sequence for the synthesis
of 4-quinolones
The quinolone scaffold is one of the most frequently occurring heterocyclic fragments
in small-molecule drugs.[124 ] The interest in the quinolone scaffold largely explains the constant development
of new synthetic strategies for the synthesis of that heterocyclic core. In 2015,
we disclosed a nongaseous synthesis of 4-quinolones from ortho -iodoanilines and terminal acetylenes, in which CO is released in situ from Mo(CO)6 .[41 ] Two methods were developed allowing for either rapid assembly of the quinolone scaffold
or introduction of potentially labile substituents such as nitro or bromide groups
(Scheme [22 ]). In method A, Pd2 (dba)3 and dppf were used to efficiently catalyze the reaction under MW irradiation at 120
°C for 20 minutes. With the presence of a secondary amine in the reaction mixture,
the cyclization was completed in situ, providing 13 examples of 4-quinolones in one
step in 29–85% isolated yield. As expected, the high reaction temperature in combination
with the presence of Mo(CO)6 in the reaction mixture significantly reduced the yield of the nitro-substituted
quinolone. To circumvent this problem a second method was developed to allow the introduction
of chemically labile groups. Method B employed acetonitrile as solvent which has been
used to generate CO from Mo(CO)6 at room temperature (Scheme [22 ]).[40 ]
[60 ] In addition, electron-rich precatalyst [(t -Bu)3 PH]BF4 was used as a ligand. As a result, the reaction could be performed at room temperature
providing 20 examples of quinolones in 32–84% isolated yield. Notably, both nitro
and bromide groups were successfully introduced in good yields.
Jafarpour developed a versatile Mo(CO)6 -promoted route to 3,4-disubstituted 2(1H )-quinolones using monoprotected 2-iodoanilines, which involved palladium(0)-catalyzed
carbonylative annulation of internal alkynes (Scheme [23 ]).[125 ] In addition to the successful application with disubstituted alkynes, the use of
norbornene furnished the corresponding (dihydro)quinolin-2(1H )-ones in good yields. With both annulation substrates free 2-iodoaniline afforded
lower yields than N -protected 2-iodoanilines.
Scheme 23 Palladium-catalyzed carbonylative annulation reaction for the synthesis of 2(1H )-quinolones
Adding to the available strategies for the synthesis of quinolones and chromones (flavones),
Ghosh and co-workers recently reported a carbonylative Sonogashira annulation sequence
(Scheme [24 ]). The heterocycles were prepared from 2-iodoanilines or 2-iodophenols in the presence
of a benzimidazole-based Pd–N-heterocyclic carbene catalyst (Pd–NHC) in moderate to
excellent yields and good functional group tolerance.[126 ]
Scheme 24 Palladium-catalyzed carbonylative Sonogashira/cyclization sequence for the synthesis
of 4-quinolones and 4H -chromen-4-ones
Further adding to the carbonylation/cyclization strategy, a method was developed for
the synthesis of functionalized 2-aminoquinazolinones (Scheme [25 ]).[127 ] It was envisioned that by changing the nucleophilic coupling partner different heterocyclic
structures could readily be assembled. Accordingly, 2-iodoanilines were carbonylatively
coupled with cyanamide to yield an N -cyanobenzamide intermediate,[65 ] which following a heating step could undergo cyclization. Both unsubstituted and
N 1-substituted quinazolinones were readily obtained by precipitation in moderate to
excellent yields with broad substrate scope. Recently, this approach has been extended
to the synthesis of 4H -benzo[e ][1,3]oxazin-4-ones from 2-iodophenols or 2-bromophenols by using either Mo(CO)6 or a range of nongaseous CO-sources.[128 ]
Scheme 25 Synthesis of 2-aminoquinazolinones by carbonylation/cyclization from ortho -iodoanilines and cyanamide
In 2017, we developed a four-component carbonylation/amination two-step one-pot protocol
for the synthesis of N -acylguanidines (Scheme [26 ]).[129 ] The reaction was initiated by the formation of an N -cyanobenzamide intermediate from the carbonylative coupling of aryl iodides and bromides
with cyanamide. A sequential amination step provided access to a large variety of
N -acylguanidines in moderate to excellent yields. During the optimization studies an
impurity was detected which gave rise to 31 P–13 C couplings in 13 C NMR spectra when Pd(PPh3 )4 was used as the catalyst. It was suspected that monodentate phosphine ligands might
attack the N -cyanobenzamide intermediate and therefore a ligand screening was performed, which
revealed DPEphos as the optimal ligand for this reaction. In addition, the acylguanidine
moiety was utilized as a precursor in the preparation of three different heterocycles.
Scheme 26 Carbonylation/amination two-step one-pot synthesis of N -acylguanidines
Nitrocompounds are inexpensive and readily available precursors to many N -containing compounds. However, the limitations of the reduction of the nitro group
impose considerable constraints for the use of such precursors in e.g. Pd-catalyzed
aminocarbonylations. In 2014, Wu and co-workers reported the synthesis of 4(3H )-quinazolinones from 2-bromoformanilides and organonitro compounds in which Mo(CO)6 served as a CO source as well as a reducing agent of the nitro group and a cyclization
promoter (Scheme [27 ]).[130 ] Under the given conditions aromatic and aliphatic nitro compounds as well as electron-rich
and electron-deficient substituents were successfully used to produce 4(3H )-quinazolinones in 26 examples in moderate to excellent yields.
Scheme 27 Carbonylative synthesis of 4(3H )-quinazolinones from 2-bromoformanilides and organonitro substrates
N -Substituted phthalimide derivatives have been explored for their biological activities.
In 2013, Langer and co-workers described a Pd-catalyzed double carbonylation of 1,2-dibromoarenes
with amines (Scheme [28 ]). The reaction was used to prepare N -substituted phthalimides from a wide range of aliphatic and aromatic amines and 1,2-dibromoarenes
in moderate to good yields.[131 ] A somewhat related Mo(CO)6 -mediated but Pd-free method to generate benzimidazoles and benzoxazoles was published
by Vidavalur in 2015.[132 ]
Scheme 28 Carbonylative synthesis of phthalimides from 1,2-dibromoarenes
The combination of C–H activation and carbonylation in heterocyclic synthesis is a
particularly desirable approach in the development of new and sustainable synthetic
strategies. Wu and co-workers have developed a carbonylative cyclization of N -aryl-pyridine-2-amines and internal alkynes by C–H activation with which 2-quinolinone
derivatives were prepared in moderate to good yields (Scheme [29 ]).[133 ] The developed strategy was used with electron-rich and electron-poor substituents
with good yields and nonsymmetric alkynes were incorporated with good regioselectivity.
Partly on the basis of a kinetic isotope effect experiment the reaction was suggested
to proceed through C–H activation with the pyridyl acting as the directing group.
Next, alkyne insertion followed by CO insertion produces an acyl palladium intermediate.
The desired product was obtained after reductive elimination and finally reoxidation
of Pd(0) by benzoquinone and AgOAc led to regeneration of the active Pd(II) catalytic
species.
Scheme 29 Carbonylative annulation of N -aryl-pyridine-2-amines with internal alkynes by C–H activation providing 2-quinolinones
In 2016, the same strategy was applied to norbornene by Wu and co-workers. The Pd-catalyzed
carbonylative C–H bond annulation of arenes with norbornene as the coupling partner
was reported for the synthesis of 5-(pyridine-2-yl)hexahydro-7,10-methanophenanthridin-6(5H )-one scaffold (Scheme [30 ]). With this more challenging alkene coupling partner, the desired heterocycle was
obtained in low to moderate yields.[134 ]
Scheme 30 Carbonylative annulation of N -aryl-pyridine-2-amines with norbornene by C–H activation
Carbonylative Cascade, Multistep Reactions
6
Carbonylative Cascade, Multistep Reactions
In 2016, Lee reported a one-pot synthesis of benzoylacetonitriles through sequential
carbonylation and decarboxylation.[135 ] A reaction of methyl cyanoacetate, an acetonitrile equivalent, with an aryl iodide
in the presence of a palladium(0) catalyst and Mo(CO)6 provided a beta-keto cyanoester, which was treated with LiI in water at 130 °C to
afford the benzoylacetonitrile in good yields (Scheme [31 ]).
Scheme 31 One-pot two-step carbonylation–decarboxylation process to provide benzoylacetonitriles
An impressive two-step palladium-catalyzed and iodine(III)-mediated β-fluorocarboxylation
of alkenes was presented by Liu et al. in 2017 (Scheme [32 ]).[136 ] The cooperative electrophilic alkene activation–carbonylation process smoothly gave
the β-fluoro ester with high regioselectivity by using Mo(CO)6 as the solid CO source.
Scheme 32 Two-step fluoriniation alkoycarbonylation synthesis of β-fluoro esters
7
Summary and Outlook
A broad array of new, convenient, and efficient Pd-catalyzed carbonylative Mo(CO)6 -mediated reactions have been developed and reported in the last six years. The use
of nongaseous CO sources has gained general acceptance resulting in a larger overall
usage in modern organic synthesis, and especially in natural product synthesis, bioorganic
chemistry, and medicinal chemistry. Furthermore, the increased use of two-chamber
systems for ex situ CO generation and the development of carbonylative cascade reactions
have further increased the interest in CO-free protocols. We anticipate that the methods
reported in this account will further stimulate the development within the field.
