Palladium-catalyzed aromatic azidocarbonylation

Article abstract of DOI:10.1002/anie.201200078

Aryl iodides smoothly react with NaN₃ and CO in the presence of a Pd/Xantphos catalyst to give aroyl azides (ArCON₃) in 75-92 % yield. The reaction occurs under mild reaction conditions (1 atm, 20-50 °C) and exhibits high functional-group tolerance. (Xantphos=9,9-dimethyl-4,5- bis(diphenylphosphino)xanthene)

Full text of DOI:10.1002/anie.201200078

.
Angewandte
Communications
DOI: 10.1002/anie.201200078
Azidocarbonylation
Palladium-Catalyzed Aromatic Azidocarbonylation**
Fedor M. Miloserdov and Vladimir V. Grushin*
Aroyl azides, ArCON3, are valuable intermediates and
building blocks in the synthesis of various useful compounds,
including isocyanates (the Curtius rearrangement), aromatic
amides, iminophosphoranes,[1] and oxazoles.[2] Conventionally
used general methods to synthesize aroyl azides are limited to
diazotization of hydrazides and reactions of NaN3 with acid
chlorides, mixed anhydrides, and N-acyl benzotriazoles.[1,3]
All of these procedures involve highly reactive chemicals
which put significant limitations on functionalities that can be
present on the aromatic ring of the substrate. The develop-
ment of methodologically new, highly functional-group toler-
ant, catalytic routes to aroyl azides is particularly desirable
and has been sought after.
bond under identical reaction conditions.[7] Moreover,
upon treatment with CO (1 atm), the N3 ligand in these
complexes is converted into NCO (with concomitant
release of N2) at as low as room temperature [Eq. (2);
THF = tetrahydrofuran].
*
Carboxylic acid salts, esters, amides, and all other products
that are formed in the reaction shown in Equation (1) are
thermally stable and hence easily survive the conditions
that conventionally employ elevated temperatures and CO
pressure. In contrast, aroyl azides are thermally unstable,
undergoing the Curtius rearrangement at temperatures
(808C and above) typically used for the carbonylation
[Eq. (1)].
Since its discovery by Heck and co-workers[4] in 1974, the
palladium-catalyzed reaction of aryl halides ArX with CO
and nucleophiles [Eq. (1); Tf = trifluoromethanesulfonyl] has
*
The product, ArCON3, is not only thermally unstable but
also highly reactive toward tertiary phosphines that
conventionally comprise palladium catalysts for the car-
bonylation and other coupling reactions. The exceedingly
facile Staudinger reaction[8] of the aroyl azide product with
the reversibly dissociated PR3 ligand would lead irrever-
sibly to the corresponding phosphinimine ArC(O)N = PR3
and, as a result, to rapid catalyst deactivation.
been widely used for the synthesis of various carboxylic acids
and their derivatives such as esters and amides.[5] Although it
would be natural to consider this type of carbonylation for the
À
synthesis of ArCON3 [Nu = N3 in Eq. (1)], this transforma-
tion has never been reported in spite of the substantial
amount of work done in the area by numerous research
groups exploring various nucleophiles in this process.[5]
The total lack of reports on palladium-catalyzed azidocarbo-
nylation, however, is not surprising considering the following:
As follows from the above, the azidocarbonylation differs
principally, in a number of respects, from the already known
palladium-catalyzed aromatic carbonylation reactions
[Eq. (1)]. Each of the three aforementioned reactivity
patterns alone seriously challenges the feasibility of the
reaction [Eq. (1)] for Nu = N3À, if not buries altogether the
very idea of palladium-catalyzed aromatic azidocarbonyla-
tion. Nevertheless, in this report we demonstrate that, against
all expectations, this transformation is not only possible but
can even be remarkably clean and high yielding for a carefully
selected catalytic system.
We have carried out a large number of experiments to
explore the possibility of transforming PhX (X = I, Br) into
PhCON3 using CO and NaN3 in the presence of various
palladium catalysts. We were delighted to find, as a result of
these exploratory runs, that Pd/Xantphos is an excellent
catalyst for azidocarbonylation of iodobenzene. Herein we
present only a very succinct summary of our scouting and
optimization experiments that are described in full detail in
the Supporting Information.
À
*
Migratory insertion of CO into the Pd C bond of
[(R3P)2Pd(Ph)(X)] (X = I, Br, Cl), a key step in the
catalytic loop governing the reaction [Eq. (1)], has long
been known[6] to readily and cleanly occur at room
temperature and atmospheric pressure. In sharp contrast,
complexes of the type [(R3P)2Pd(Ph)(N3)] have been
reported to not undergo CO insertion into the Pd–Ph
[*] F. M. Miloserdov, Prof. V. V. Grushin
The Institute of Chemical Research of Catalonia
Avgda. Paꢀsos Catalans 16, 43007 Tarragona (Spain)
E-mail: vgrushin@iciq.es
[**] We thank Drs. Jordi Benet-Buchholz, Eduardo C. Escudero-Adꢁn,
Marta Martꢂnez Belmonte, and Eddy Martin for single-crystal X-ray
diffraction studies and Prof. Piet W. N. M. van Leeuwen and
Maxim A. Novikov for helpful discussions. The ICIQ Foundation
and Consolider Ingenio 2010 (Grant CSD2006-0003) are thankfully
acknowledged for support of this research.
1. Pd/Xantphos was the only system that competently
catalyzed the azidocarbonylation reaction. Other catalysts
based on Pd(OAc)2 or [Pd2dba5][9] precursors and various
ligands, including PPh3, o-Tol3P, Cy3P, tBu3P, dppe, dppp,
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201200078.
3668
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 3668 –3672

Angewandte
Chemie
dppb, dppf, rac-BINAP, iPr-Xantphos, tBu-Xantphos, and
DPEphos were inefficient. The yield of PhCON3 from PhI
using these ligands was only 0–6% under a variety of
reaction conditions.
2. The highest catalytic activity was observed for a 1:1 Pd0/
Xantphos molar ratio, in accord with the literature data.[10]
It is noteworthy that Xantphos[11] has been found[12] to be
a superior ligand for a series of palladium-catalyzed
carbonylation reactions.
3. In many instances, the product (ArCON3) and the starting
material (ArI) exhibited nearly identical Rf parameters
and hence could not be separated by column chromatog-
raphy. Therefore, to isolate the product in pure form the
reaction should be driven to full conversion. Toward this
goal, we found that nearly quantitative conversions of PhI
could be obtained by running the reaction in the presence
of water.[13] The promoting effect of water was particularly
powerful when the reaction was performed under biphasic
conditions[14] with THF/hexanes (2:1 v/v) as the organic
phase of choice.[15] Although high conversions and yields
of up to 87–88% could be obtained under such biphasic
conditions with only 1% Pd, we settled on 2 mol% of the
catalyst in order to reach full conversion that is critical for
isolation of the product in pure form (see above).
With optimized reaction conditions in hand, a series of
aryl iodides (1) were successfully converted to the corre-
sponding aroyl azides (2; Scheme 1). All conversions were
quantitative, thus allowing isolation of the pure products in
high yield. The reaction exhibited high functional-group
tolerance and smoothly and cleanly occurred with both
electron-withdrawing (2e–l, 2p) and electron-donating (2b–
d, 2n) substituents on the aromatic ring. The yields of the
isolated products were usually in the range of 80–90% with
a few exceptions, including the lower yield of the doubly
azidocarbonylated benzene (2m, 52%) obtained from p-
diiodobenzene and the higher, 92% yield of m-toluoyl azide
2n.
Both 2-iodothiophene and 3-iodopyridine were cleanly
converted to the corresponding azidocarbonyl derivatives (2q
and 2r) in 87 and 75% isolated yield, respectively. In contrast,
2-iodopyridine produced a mixture of products under similar
reaction conditions, including 2-PyNCO, 2-PyCONH2, and 2-
PyNH2. This outcome is likely due to the lower thermal
Scheme 1. Palladium-catalyzed azidocarbonylation of iodoarenes. Reac-
tion conditions: ArI (1 mmol), NaN3 (1.2 mmol), [Pd2dba5][9]
(0.01 mmol; 2 mol% Pd), Xantphos (2 mol%) in THF (2 mL), hexanes
(1 mL), and water (3 mL) at 238C. All yields are isolated yields. [a] At
508C (oil bath). [b] Isolated product contained 2% of PhCON3 as
a result of P-Ar/Pd-Ar’ exchange. [c] Isolated product contained dba.
[d] Obtained from 1,4-diiodobenzene (1 mmol) and double amounts of
all reagents and solvents. dba=dibenzylideneacetone.
[16]
stability of 2-PyCON3 and, possibly, to the well-known[17]
dimerization of 2-pyridyl palladium derivatives through the N
atoms of the pyridine ring.
Under the same reactions conditions, ortho-substituted
iodoarenes gave mixtures of the corresponding amide, amine,
and diarylurea, as was the case with 2-iodotoluene. The
reaction was more selective toward the formation of ureas for
2-iodoanisole and 1-iodonaphthalene: N,N’-bis(1-naphthyl)-
urea (3a; see Scheme 2) was obtained pure in 76% yield and
2-iodoanisole gave N,N’-bis(2-methoxyphenyl)urea (3b) that
was isolated in 74% yield (ca. 80% purity by 1H NMR
spectroscopy; see the Supporting Information). The different
reaction outcome for 2-substituted iodoarene substrates is
hardly surprising because ArCON3 bearing an ortho substitu-
ent are 50–200 times more reactive toward the Curtius
rearrangement than their m and p isomers.[16,18] As a result,
an ortho-substituted aroyl azide originally produced in the
reaction quickly rearranges to the corresponding isocyanate
that may be partially converted into ArNH2 under the
reaction conditions. The aniline then adds to the as yet
unreacted ArNCO to form the urea (Scheme 2).
Importantly, aroyl azides produced in the reaction can be
used for further transformations without isolation. This
concept was successfully demonstrated with 4-fluoroiodoben-
zene, as shown in Scheme 3. Carrying out the azidocarbony-
Angew. Chem. Int. Ed. 2012, 51, 3668 –3672
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 3669

.
Angewandte
Communications
Scheme 4. Azidocarbonylation mechanisms.
Scheme 2. In situ formation of ureas from the originally produced
ortho-substituted ArCON3.
resultant s-aryl PdII iodide to the key intermediate that
reductively eliminates ArCON3: 1) anionic ligand exchange
to produce 7 with subsequent carbonylation and 2) migratory
À
insertion of CO leading to 8, followed by IÀ/N3 ligand
exchange.
The proposed reaction pathways (Scheme 4) were con-
firmed by independent mechanistic studies. Both intermedi-
ates, [(Xantphos)Pd(Ph)(N3)] (7) and [(Xantphos)Pd-
(COPh)(I)] (8), were synthesized and fully characterized,
including by single-crystal X-ray diffraction (Figure 1).[23]
Bubbling CO through a [D6]benzene solution of 7 or adding
Scheme 3. Reaction scope experiments. Yields were determined by
19F NMR spectroscopy. PMHS=polymethylhydrosiloxane.
lation in the presence of PMHS (2 equiv) afforded 4-
fluorobenzamide (4) in nearly quantitative yield. It is worth
noting that one-step aminocarbonylation of aryl halides to
primary benzamides is a challenging transformation.[19] In
another experiment, after 4-fluorobenzoyl azide (2j) was
produced (Scheme 1) the entire reaction mixture was treated
with PPh3 (1.1 equiv) to bring about the transformation of 2j
to iminophosphorane 5 that was formed in 85% overall yield
(Scheme 3). Running the azidocarbonylation in a toluene/
water (1:1) solvent mixture allowed for conveniently carrying
out the Curtius rearrangement by heating the toluene phase
containing the ArCON3 product. In this way 4-fluorophenyl-
isocyanate (6) was obtained in 86% yield, as calculated on the
basis of the amount of 4-fluoroiodobenzene used. Aryl
isocyanates are industrially important intermediates[20] that
are widely used in the synthesis of various valuable chemicals,
including carbamates, anilines, and ureas. Tkatchenko et al.[21]
have patented the nickel-catalyzed reaction of haloarenes
with NaOCN to produce aryl isocyanates and their deriva-
tives in up to 50% yield. Direct palladium-catalyzed trans-
formation of aryl halides into aryl isocyanates is unknown.
Attempts to azidocarbonylate bromoarenes were unsuccess-
ful.[22]
Figure 1. ORTEP drawings of 7 (left) and 8 (right) with H atoms
omitted and thermal ellipsoids drawn to the 50% probability level.
Bu4N+ N3 to a solution of 8 in [D6]benzene under CO at
À
room temperature resulted in immediate reductive elimina-
tion leading to a single 31P NMR-observable compound (s, d =
10.5 ppm). The same species was formed, as detected by
31P NMR spectroscopy, upon treatment of [Pd2dba5] with
Xantphos and CO in benzene or toluene, thus indicating that
the signal at d = 10.5 ppm was likely from a zero-valent
palladium carbonyl phosphine complex. Layering the ben-
zene solution with ether produced X-ray quality crystals of
[(Xantphos)Pd(CO)2] co-crystallized with [(Xantphos)Pd-
(dba)] and benzene in a 1:1:1.5 molar ratio. The X-ray
determination of the structure of [(Xantphos)Pd(CO)2][23]
resonating at d = 10.5 ppm in the 31P NMR spectra (see
above) provided unambiguous evidence for the existence of
the two reaction pathways, via 7 and via 8, leading to the final
product that is apparently produced from [(Xantphos)Pd-
(COPh)(N3)] (Scheme 4). The competition between the two
pathways involving 7 and 8 is expected to strongly depend on
A plausible mechanism for the azidocarbonylation reac-
À
tion is presented in Scheme 4. After the Ar I bond oxida-
tively adds to Pd0, two reaction pathways could lead from the
3670 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 3668 –3672

Angewandte
Chemie
À
Chem. Rev. 2011, 111, 2177; c) R. Grigg, S. P. Mutton, Tetrahe-
the CO pressure, N3 concentration, and the nature of the
dron 2010, 66, 5515; d) A. Brennfꢁhrer, H. Neumann, M. Beller,
Angew. Chem. 2009, 121, 4176; Angew. Chem. Int. Ed. 2009, 48,
4114; e) C. F. J. Barnard, Organometallics 2008, 27, 5402.
[6] P. E. Garrou, R. F. Heck, J. Am. Chem. Soc. 1976, 98, 4115.
[7] Y.-J. Kim, Y.-S. Kwak, S.-W. Lee, J. Organomet. Chem. 2000, 603,
152.
[8] For a recent review of the Staudinger reaction, see: D. M. Iula in
Name Reactions for Functional Group Transformations (Ed.: J. J.
Li, E. J. Corey), Wiley, Hoboken, 2007, p. 129.
[9] The empirical composition [Pd2dba5] was derived from the
elemental analysis data. For details, see: A. V. Ushkov, V. V.
Grushin, J. Am. Chem. Soc. 2011, 133, 10999.
[10] L. M. Klingensmith, E. R. Strieter, T. E. Barder, S. L. Buchwald,
Organometallics 2006, 25, 82.
medium, among numerous other factors. Regardless of which
of the two internal pathways prevails, the process invariably
leads to the aroyl azide product. What is crucial to the overall
catalytic process,
however,
is
that
unlike the
À
[(R3P)2Pd(Ph)(N3)] [Eq. (2)], 7 does not form a Pd NCO
species by N2 loss upon treatment with CO.
The uniqueness of the catalytic system based on Xantphos
is not fully understood. The remarkable ability of Xantphos to
promote various transformations at the metal center is well-
documented,[11] including particularly challenging ones, such
II [24]
À
as Ar CF3 reductive elimination from Pd . Although this
behavior of Xantphos has often been linked to its wide bite
angle,[11] the latter might not always be the key contributing
factor.[25] Studies of the critical role of Xantphos in the
azidocarbonylation reaction are currently underway in our
laboratories.
[11] For selected reviews on Xantphos and related ligands, see:
a) P. C. J. Kamer, P. W. N. M. van Leeuwen, J. N. H. Reek, Acc.
Chem. Res. 2001, 34, 895; b) Z. Freixa, P. W. N. M. Van Leeuwen,
Dalton 2003, 1890; c) Z. Freixa, P. W. N. M. Van Leeuwen,
Coord. Chem. Rev. 2008, 252, 1755.
In conclusion, we have demonstrated, for the first time,
the possibility of palladium-catalyzed aromatic azidocarbo-
nylation. Careful catalyst selection and mechanistic insights
have led to the identification of reaction conditions that are
sufficiently mild to circumvent the multiple dangers of failure
stemming from the Curtius rearrangement of the aroyl azide
product, its Staudinger reaction with the phosphine ligand,
and, as a result, immediate catalyst degradation. The catalytic
reaction smoothly occurs at temperatures as low as 25–508C
and 1 atm to cleanly give aroyl azides from the corresponding
aryl iodides, CO, and NaN3. Unlike the traditional routes to
ArCON3, our method does not require stoichiometric quan-
tities of acid activators (Cl-containing reagents or oxidants),
and therefore can be used to prepare otherwise hardly
accessible aroyl azides such as 2p that bears a formyl group on
the ring. The reaction exhibits high functional-group toler-
ance and can also be conveniently used for one-pot, two-step
procedures furnishing primary benzamides, iminophosphor-
anes, isocyanates, and ureas in high yield without isolation of
the primary benzoyl azide product.
[12] J. R. Martinelli, D. A. Watson, D. M. M. Freckmann, T. E.
Barder, S. L. Buchwald, J. Org. Chem. 2008, 73, 7102.
[13] a) It has also been found that the reaction can be promoted by
a reductant such as Zn dust or PMHS. This effect might be due to
the reduction of the PdII catalyst precursor to Pd0 rather than by
the coordinated phosphine.[13b] As in palladium-catalyzed cyan-
ation of haloarenes,[13c] the reductant also recuperates deacti-
vated catalyst. For example, in one experiment, after the
reaction stopped at 64% conversion, the addition of Zn
revitalized the process and 90% conversion was quickly
achieved. However, both Zn and PMHS could reduce, at least
partially, the aroyl azide product to ArCONH2 (see the
Supporting Information; b) V. V. Grushin, Chem. Rev. 2004,
104, 1629; c) S. Erhardt, V. V. Grushin, A. H. Kilpatrick, S. A.
Macgregor, W. J. Marshall, D. C. Roe, J. Am. Chem. Soc. 2008,
130, 4828.
[14] a) In THF/H2O (1:1), the reaction stalled after quickly reaching
80–90% conversion because of catalyst deactivation with the
iodide released (see the Supporting Information). The inhibiting
effect of iodide[14b,c] could be eliminated by running the reaction
in an H2O/organic solvent biphasic system, probably because in
this way no excess of inorganic salts could be accumulated in the
water-immiscible organic phase. Azide in excess does not
deactivate the catalyst: b) Q. Shen, T. Ogata, J. F. Hartwig, J.
Am. Chem. Soc. 2008, 130, 6586; c) B. P. Fors, N. R. Davis, S. L.
Buchwald, J. Am. Chem. Soc. 2009, 131, 5766.
[15] Although the reaction afforded ArCON3 in 97% yield at > 99%
conversion in H2O/benzene (1:1) or H2O/toluene (1:1), benzene
is toxic and the boiling point of toluene (1108C) would prevent
isolation of the volatile and thermally unstable ArCON3
product. After a series of tests it was found that the aromatic
organic phase can be successfully substituted with nontoxic, low-
boiling THF/hexanes (2:1 v/v).
[16] H. Saikachi, T. Kitagawa, A. Nasu, H. Sasaki, Chem. Pharm.
Bull. 1981, 29, 237.
[17] a) R. Bertani, A. Berton, F. Di Bianca, B. Crociani, J. Organo-
met. Chem. 1986, 303, 283; b) K. Isobe, K. Nanjo, Y. Nakamura,
S. Kawaguchi, Bull. Chem. Soc. Jpn. 1986, 59, 2141.
Received: January 4, 2012
Published online: February 28, 2012
Keywords: aroyl azides · aryl iodides · azides · carbonylation ·
.
palladium
[1] For general reviews, see: a) E. F. V. Scriven, K. Turnbull, Chem.
Rev. 1988, 88, 297; b) S. Brꢀse, C. Gil, K. Knepper, V.
Zimmermann, Angew. Chem. 2005, 117, 5320; Angew. Chem.
Int. Ed. 2005, 44, 5188; c) S. Brꢀse, K. Banert, Organic Azides:
Syntheses and Applications, Wiley, Chichester, 2010.
[2] I. Cano, E. Alvarez, M. C. Nicasio, P. J. Perez, J. Am. Chem. Soc.
2011, 133, 191.
[3] A. R. Katritzky, K. Widyan, K. Kirichenko, J. Org. Chem. 2007,
72, 5802.
[4] a) A. Schoenberg, I. Bartoletti, R. F. Heck, J. Org. Chem. 1974,
39, 3318; b) A. Schoenberg, R. F. Heck, J. Org. Chem. 1974, 39,
3327; c) A. Schoenberg, R. F. Heck, J. Am. Chem. Soc. 1974, 96,
7761.
[18] The steric bulk of the ortho substituent weakens the conjugation
between the aromatic ring and CON3, thus destabilizing the
latter and making the molecule more prone to the Curtius
rearrangement. See: Y. Yukawa, Y. Tsuno, J. Am. Chem. Soc.
1958, 80, 6346.
[19] X.-F. Wu, H. Neumann, M. Beller, Chem. Eur. J. 2010, 16, 9750.
[20] H. Ulrich, Chemistry and Technology of Isocyanates, Wiley,
Chichester, 1997.
[5] For recent reviews on palladium-catalyzed carbonylation reac-
tions of aryl halides, see: a) X.-F. Wu, H. Neumann, M. Beller,
Chem. Soc. Rev. 2011, 40, 4986; b) J. Magano, J. R. Dunetz,
Angew. Chem. Int. Ed. 2012, 51, 3668 –3672
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3671

.
Angewandte
Communications
[21] I. Tkatchenko, R. Jaouhari, M. Bonnet, G. Dawkins, S. Lecolier,
U.S. Patent 4749806, 1987.
[22] Although ligands, solvents, and temperature were varied in
a broad range, the conversion never exceeded 5% even for more
reactive, electron-deficient aryl bromides. It is likely that at
the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[24] V. V. Grushin, W. J. Marshall, J. Am. Chem. Soc. 2006, 128,
12644.
[25] V. I. Bakhmutov, F. Bozoglian, K. Gꢂmez, G. Gonzꢃlez, V. V.
Grushin, S. A. Macgregor, E. Martin, F. M. Miloserdov, M. A.
Novikov, J. A. Panetier, L. V. Romashov, Organometallics 2012,
DOI: 10.1021/om200985g.
À
elevated temperatures required for Ar Br oxidative addition to
Pd0, the catalyst is quickly deactivated by the Staudinger
reaction of the ArCON3 product and the stabilizing phosphine
ligand (see above).
[23] CCDC 860745
(7),
860746
(8),
and
860747
([(Xantphos)Pd(CO)2]·[(Xantphos)Pd(dba)]·1.5C6H6) contain
3672
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 3668 –3672