Copper(I)/N-Heterocyclic Carbene (NHC)-catalyzed addition of terminal alkynes to trifluoromethyl ketones for use in continuous reactors

Article abstract of DOI:10.1002/adsc.201300802

A copper(I)/N-heterocyclic carbene complex-catalyzed addition of terminal alkynes to trifluoromethyl ketones at low loading is described. The developed process functions well using a range of terminal alkynes but functions best when an aryl trifluoromethyl ketone is used. This substrate scope is well-suited for the production of active pharmaceutical ingredients (APIs) such as efavirenz. In this vein, we demonstrate that the described method can be translated into a flow process laying the framework for a completely continuous synthesis of efavirenz in the future. Copyright

Full text of DOI:10.1002/adsc.201300802

COMMUNICATIONS
DOI: 10.1002/adsc.201300802
Copper(I)/N-Heterocyclic Carbene (NHC)-Catalyzed Addition
of Terminal Alkynes to Trifluoromethyl Ketones for Use in
Continuous Reactors
Camille A. Correia,a D. Tyler McQuade,a,c,* and Peter H. Seebergera,b
a
Department of Biomolecular Systems, Max Planck Institute of Colloids and Interfaces, Am Mꢀhlenberg 1, 14476
Potsdam, Germany
E-mail: tyler.mcquade@mpikg.mpg.de
Institute for Chemistry and Biochemistry, Freie Universitꢁt Berlin, Arnimallee 22, 14195 Berlin, Germany
Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Florida 32306, USA
b
c
E-mail: mcquade@chem.fsu.edu
Received: September 3, 2013; Revised: October 21, 2013; Published online: December 11, 2013
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201300802.
Abstract: A copper(I)/N-heterocyclic carbene com-
plex-catalyzed addition of terminal alkynes to tri-
fluoromethyl ketones at low loading is described.
The developed process functions well using a range
of terminal alkynes but functions best when an aryl
trifluoromethyl ketone is used. This substrate scope
is well-suited for the production of active pharma-
ceutical ingredients (APIs) such as efavirenz. In this
vein, we demonstrate that the described method
can be translated into a flow process laying the
framework for a completely continuous synthesis of
efavirenz in the future.
phenylacetylene to 2,2,2-trifluoroacetophenone over
1–2 days. More recently, Carreira[5] described an auto-
catalytic improvement for the diethylzinc, alkyllithi-
um mediated alkynylation first developed by Merck.[6]
With an interest in establishing continuous routes into
critical medicines for developing world applications,
we sought a 1,2-addition strategy that could be incor-
porated into a multi-step continuous production pro-
cess of the anti-HIV drug, efavirenz.[7] Efavirenz is
a critical anti-HIV drug used in combination therapy
in the developing world and while its price has de-
creased due to competition from generics companies,
the cost still remains too high for millions of potential
patients.[8] The published production process by
Merck involves a six-step synthesis from p-chloroani-
line.[6] Similarly, Lonza recently detailed a shorter
four-step synthesis of efavirenz from 1,4-dichloroben-
Keywords: alkynes; continuous synthesis; copper;
microreactor; trifluoromethyl ketones
zene.[9] One of the C C bond forming steps common
À
to both synthetic procedures involves the 1,2-addition
Earth-abundant transition metal complexes to cata- of cyclopropylacetylene to aryl trifluoromethyl ke-
lyze carbon-carbon bond formation are becoming in- tones, via stoichiometric zinc. While the reported
creasingly attractive as the cost of rare transition routes by Merck and Lonza have many attractive at-
metals rise. The in-situ generation of metal acetylides tributes, both are batch processes and as years of liter-
and production of the highly functionalized propar- ature have demonstrated the commoditization of
gylic alcohols or amines opens the door to a variety a large-scale process is almost always more efficient
of subsequent transformations. Consequently, it is no when made continuous.[10]
surprise that the metal-catalyzed addition of terminal
alkynes to aldehydes, ketones and aldimines has been
Of the existing catalytic approaches for adding alk-
AHCTUNGTREGyNNNU nes to trifluoromethyl ketones, Cu(I) is attractive en-
extensively studied.[1] Despite the prevalence of the vironmentally and has the potential for easy conver-
trifluoromethyl group in many active pharmaceutical sion to a continuous process. We examined catalytic
agents and agrochemicals,[2] the alkynylation of tri- systems that could perform well at low loadings and
fluoromethyl ketones is much less developed. Shiba- under mild conditions, and are tolerant of higher tem-
saki et al.[3] first demonstrated that copper-diphos- peratures without loss of activity. The major consider-
phine and diamine complexes catalyzed in-situ activa- ation in identifying an active copper catalyst was its
tion and alkyne addition. Li[4] and co-workers used an ability to increase the nucleophilicity of the acetylide.
“on-water” silver complex to catalyze the addition of Typically electron-rich, bulky ligands can simultane-
Adv. Synth. Catal. 2013, 355, 3517 – 3521
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3517

Camille A. Correia et al.
COMMUNICATIONS
Table 1. Optimization of the copper-catalyzed alkyne addi-
tion to trifluoromethyl ketones.
Figure 1. Copper(I)/NHC catalysts.
Entrya
[Cu]/NaO-t-Bu (mol%)
T [oC]
Yield [%]
1
2
3
–
ACHTUNGTRENNUNG
T
40
40
40
40
40
50
50
50
50
ND
86
77
ND
ND
99
97
78
48
the Cu(I)–NHC complex is necessary as both removal
of copper or use of simple CuCl did not yield any
product (entries 1 and 4). As expected, sodium tert-
butoxide is crucial for the reaction (entry 5), both to
form the active Cu alkoxide and then, once protonat-
ed (t-BuOH), it facilitates proton transfer from the
terminal alkyne to the product.[14] Almost quantitative
conversion was achieved at increased temperatures
(entry 6) and this temperature increase enabled the
use of lower catalyst loadings (entry 7). Two equiva-
lents of alkyne were necessary to achieve the highest
yields, likely due to loss of material via evaporation
as no evidence of Glaser–Hay products was identified
(entry 8).
4
5[b]
6
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
N
N
7
8[c]
9[d]
E
[a]
Conditions: 0.2 mmol 1a, 2 equiv. 2a, copper catalyst [Cu]
and base in 0.2 mL tetrahydrofuran (THF) for 16 h
under argon (Ar).
NaO-t-Bu removed.
1.5 equiv. alkynes.
[b]
[c]
[d]
8 h. ND=none detected by NMR.
We chose the conditions as laid out in Table 1,
entry 7 as our optimum system, and proceeded to ex-
ously activate the alkyne while preventing the forma- amine the scope of the reaction (Table 2). Addition
tion of inactive polymeric copper–alkyne species. The products were obtained in good to moderate yields
strong electron-donating ability of N-heterocyclic car- for the alkynylation of aromatic trifluoromethyl ke-
benes (NHCs) as well as the Brønsted basic proper- tones with various terminal alkynes. The electronics
ties of copper(I) alkoxides prompted us to examine of both partners played an important role in this reac-
copper(I)–NHC alkoxides for the production of tri- tion. In some cases, higher temperatures were neces-
fluoromethylpropargylic alcohols. We predicted that sary to obtain full conversion (3d–3i). Even at these
these Cu(I)–NHC alkoxides would catalyze 1,2-addi- elevated temperatures the Cu(I)–NHC complexes re-
tions under milder conditions allowing the use of mained active. It should be noted that the increased
lower boiling point alkynes, for example. cyclopropy- boiling point of phenylacetylene enabled us to reduce
lacetylene, a critical feature in the synthesis of APIs the number of alkyne equivalents used.[15]
such as efavirenz.
Reactions involving the electron-rich electrophile,
From initial reports of their use in the conjugate 4-(trifluoroacetyl)toluene, and electron-poor nucleo-
ethylation of cyclic enones in 2001 by Woodward,[11] phile, 4-fluorophenylacetylene, were sluggish and
copper/N-heterocyclic carbenes have proven to be re- therefore needed longer reaction times to obtain full
markably versatile catalysts.[12] Recently the Cu(I)– conversion to 3f and 3h, respectively. The alkynyla-
NHC-catalyzed carboxylations of terminal alkynes tion of aliphatic 3-phenyl-1,1,1-trifluoroacetone was
have been described.[13] These examples clearly dem- also accomplished, albeit in lower yield. Additionally,
onstrate that Cu(I)–NHC complexes can form reac- the method supports a larger scale reaction where
tive Cu acetylides. Herein, we report that copper(I)– a 72% isolated yield of 3a was obtained on the 1-
NHC complexes can catalyze the addition of terminal mmol scale in batch. This lower yield may be due to
alkynes to trifluoromethyl ketones at low loadings less efficient heating and mixing typical of larger scale
(3 mol%). We demonstrate that this system has the heterogeneous batch reactions.
necessary attributes to enable a continuous alkynyla-
tion process.
We have a strong interest in developing low cost,
high efficiency synthetic procedures for APIs to ad-
Initially, the alkynylation of trifluoromethyl ketones dress developing world needs.[16] One potential strat-
was optimized in batch, using 2,2,2-trifluoroaceto- egy to achieve these goals is to create fully or semi-
ACHTUNGTRENNUNGphenone and cyclopropylacetylene as substrates automated continuous synthetic routes by reducing
(Table 1). Although both (IPr)CuCl and (IMes)CuCl labor and waste costs while at the same time provid-
(Figure 1) catalyzed the desired 1,2-addition (entries 2 ing APIs of higher quality (via continuous monitor-
and 3), the more hindered (IPr)CuCl complex provid- ing). The first step to creating a continuous process is
ed higher yields. Control experiments confirm that to establish that each synthetic step can be achieved
3518
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 3517 – 3521

Copper(I)/N-Heterocyclic Carbene (NHC)-Catalyzed Addition of Terminal Alkynes
Table 2. Scope of the copper-catalyzed alkynylation of trifluoromethyl ketones in batch.
[a]
508C for 16 h. Isolated yields based on 1 with NMR yields reported in parenthesis.
408C for 24 h.
708C for 16 h.
708C for 24 h.
[b]
[c]
[d]
[e]
908C for 16 h.
in flow with good throughput. The fact that the
To overcome these problems, we tested various
Cu(I)–NHC-based alkynylation reported here can cat- mixed solvent systems[17] and discovered that a near
alyze the addition of cyclopropylacetylene to aryl tri- neutral density slurry was formed by using a dimethyl-
fluoromethyl ketones suggests it as a key step in the formamide (DMF), benzene and THF mixture in
synthesis of the anti-viral API efavirenz. Adaptation
of this reaction to flow conditions could serve as prog-
ress towards more attractive syntheses of efavirenz in-
termediates A and B, the trifluoromethyl-substituted
quaternary propargylic alcohols (Scheme 1). Figure 2
depicts our flow set-up where two solutions (solution
1=substrates; solution 2=in situ formed catalyst) are
delivered to a T-mixer. The resulting reaction mixture
would then proceed into a heated reactor loop.
Translating our method to flow required us to ad-
dress solubility obstacles. The optimized solvent
system defined in Table 1 resulted in reduced yields
due to the lower solubility of the base, the active
(IPr)CuO-t-Bu catalyst and sodium chloride in THF.
Thus solids settling in the syringes and tubing caused
inconsistent catalyst concentrations throughout the
system. With cost in mind, we also needed to deliver
3 mol% of the catalyst without leaving catalyst
behind or adhering to the delivery syringe.
Scheme 1. Merckꢃs and Lonzaꢃs retrosynthetic analyses of
efavirenz.
Adv. Synth. Catal. 2013, 355, 3517 – 3521
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3519

Camille A. Correia et al.
COMMUNICATIONS
Figure 2. Flow set-up and application to biologically relevant intermediates.
a 2:3:1 ratio.[18] This slurry settled slower than when creasing throughput and providing a potential key
THF was used alone. To further circumvent adher- step towards an overall continuous process. Finally,
ence and catalyst settling, we decided to make use of we predict that this method opens the door to realiz-
an injection loop submerged horizontally in a sonica- ing Cu(I)–NHC-promoted asymmetric alkynylations
tion bath.[19] An in-line check valve (CV) and a 7-bar based on the simple manipulation of the backbone
back pressure regulator (BPR) were added to allow and ready availability of NHC ligands.
constant flow through the 10-mL reactor loop at tem-
peratures above the solventꢃs boiling point. Likewise,
these conditions enabled the use of cyclopropylacety- Experimental Section
lene well above its boiling point and underscore the
value of a flow chemical system. To achieve these General Procedure for the Cyclopropylacetylene
conditions in batch would require a sealed, pressur- Addition to 2,2,2-Trifluoromethylacetophenone (in
ized vessel limiting the potential scale-up and Batch)
throughput.[20] Using the set-up depicted in Figure 2,
NaO-t-Bu
(3 mol%,
0.006 mmol)
and
(IPr)CuCl
the reaction was performed on a 1-mmol scale (0.5M)
of the trifluoromethyl ketone, good yields of 3a and
intermediates A and B were obtained in less than
3.5 h, almost five times faster than the batch process.
The shorter reaction time can be attributed to more
efficient heating and mixing of the heterogeneous re-
action in the flow reactor, as well as the increase in
temperature and pressure. It should be noted that al-
though we were able to decrease the amount of
alkyne required in the flow device, it was still used in
(0.006 mmol) were placed in a sealable tube with magnetic
stir bar. The tube was evacuated and backfilled with argon,
then 0.1 mL anhydrous, degassed THF was added. 2,2,2-Tri-
fluoroacetophenone (0.2 mmol) and cyclopropylacetylene
(0.4 mmol, 34 mL) were then added, followed by 0.1 mL
THF. The tube was flushed with argon, sealed and placed in
an oil bath at 508C for 16 h. The heterogeneous mixture was
cooled then flushed with ethyl acetate through a short
column of silica gel. NMR yields were obtained using mesi-
tylene as an internal standard. Compound 3a was isolated
by flash column chromatography on silica gel using a 5:1
mixture of hexanes and ethyl acetate. Full details and char-
acterization data can be found in the Supporting Informa-
tion.
À
excess. It has been established that the C C bond
forming step in copper(I) alkynylation of aldehydes is
a reversible process.[21]As such, it is possible that the
excess alkyne is necessary for good yields due to a re-
versible addition to trifluoromethyl ketones in the
presence of copper(I) alkoxides.
In summary, we have identified a copper(I)/N-het-
erocyclic carbene catalyst that can be used at low
loadings for the addition of both aliphatic and aro-
matic alkynes to trifluoromethyl ketones in good
yields in batch. The method also enabled the transla-
General Procedure for the Cyclopropylacetylene
Addition to 2,2,2-Trifluoromethylacetophenone (in
Flow)
NaO-t-Bu (3 mol%, 0.03 mmol) and (IPr)CuCl (0.03 mmol)
were mixed with 2 mL DMF/benzene/THF (2:3:1 ratio) to
form a homogenous slurry and loaded into an injection loop
tion of this reaction to a flow chemistry modality, in- under argon. 2,2,2-Trifluoroacetophenone (1 mmol) and cy-
3520
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 3517 – 3521

Copper(I)/N-Heterocyclic Carbene (NHC)-Catalyzed Addition of Terminal Alkynes
clopropylacetylene (1.8 mmol) were dissolved in 2 mL THF
and loaded into a second injection loop under argon. The
loops were connected with a switchable IDEX T-mixer and
fed into a 10-mL [1.6 mm (1/16“) O.D. (outer diameter)ꢄ
0.3 mm (0.012”) I.D. (inner diameter)] PTFE tubing at 908C
and at a rate of 25 mLminÀ1 each under 7 bar back pressure.
The resulting product mixture was collected and compound
3a isolated by flash column chromatography on silica gel
using a 5:1 mixture of hexanes and ethyl acetate. Full details
and characterization data can be found in the Supporting In-
formation.
Logue, G. Trainor, P. S. Anderson, S. K. Erickson-Vitta-
nen, Med. Chem. 2000, 43, 2019–2030.
[8] Access Campaign, Medicins Sans Frontieres, “Untan-
gling the web of antiretroviral price reductions, 16th
edn., July 2013”. Retrieved from http://msfaccess.org/
site/default/files/AIDS_Report_UT-
W16_ENG_2013.pdf, accessed in August 2013.
[9] D. Dai, X. Long, B. Luo, A. Kulesza, J. Reichwagen, Y.
Guo, (Lonza Ltd), PCT Int. Appl. WO2012097510,
2012.
[10] a) R. L. Hartman, J. P. McMullen, K. F. Jensen, Angew.
Chem. 2011, 123, 7642–7661; Angew. Chem. Int. Ed.
2011, 50, 7502–7519; b) L. Malet-Sanz, F. Susanne, J.
Med. Chem. 2012, 55, 9, 4062.
[11] P. K. Fraser, S. Woodward, Tetrahedron Lett. 2001, 42,
2747.
[12] For reviews see: a) J. D. Egbert, C. S. Cazin, S. P.
Nolan, Catal. Sci. Technol. 2013, 3, 912–926; b) S. Diez-
Gonzalez, N. Marion, S. P. Nolan, Chem. Rev. 2009,
109, 3612–3676; c) S. Diez-Gonzalez, S. P. Nolan, Syn-
lett 2007, 2158–2167.
[13] a) D. Yu, Y. Zhang, Proc. Natl. Acad. Sci. USA, 2010,
107, 20184; b) W.-Z. Zhang, W.-J. Li, X. Zhang, H.
Zhou, X.-B. Lu, Org. Lett. 2010, 12, 4748.
[14] Soluble bases typically used in copper-catalyzed alkyny-
lations were also tested. Triethylamine and cesium car-
bonate furnished the product in only 11% and 58%
yields, respectively.
[15] Please refer to the Supporting Information for full re-
action conditions of each substrate.
Acknowledgements
The authors gratefully acknowledge the Max-Plank Society
for generous financial support.
References
[1] For selected reviews, see: a) B. M. Trost, A. H. Weiss,
Adv. Synth. Catal. 2009, 351, 963–983; b) E. M. Car-
reira, D. E. Frantz, in: Science of Synthesis – Stereose-
lective Synthesis, Vol. 2, (Ed.: G. A. Molander), Georg
Thieme Verlag, Stuttgart, 2011, pp 497–515; c) C.-J. Li,
Acc. Chem. Res. 2010, 43, 581–590; d) W.-J. Yoo, L.
Zhao, C.-J. Li, Aldrichimica Acta 2011, 44, 43–51; e) L.
Zani, C. Bolm, Chem. Commun. 2006, 4263–4275.
[2] a) K. Muller, C. Faeh, F. Diederich, Science 2007, 317,
1881; b) S. Purser, P. R. Moore, S. Swallow, V. Gouver-
neur, Chem. Soc. Rev. 2008, 37, 320; c) W. K. Hagmann,
J. Med. Chem. 2008, 51, 4359; d) P. Kirsch, Modern Flu-
oroorganic Chemistry: Synthesis Reactivity, Applica-
tions. Wiley-VCH, Weinheim, 2004; e) C. Isanbor, D. J.
OꢃHagan, J. Fluorine Chem. 2006, 127, 303.
[16] a) F. Lꢅvesque, P. H. Seeberger, Angew. Chem. 2012,
124, 1738–1741; Angew. Chem. Int. Ed. 2012, 51, 1706–
1709; b) D. Kopetzki, F. Lꢅvesque, P. H. Seeberger,
Chem. Eur. J. 2013, 19, 5450–5456.
[17] The active catalyst (IPr)CuO-t-Bu is soluble in ben-
zene. (IPr)CuCl and NaO-t-Bu could be pre-mixed in
benzene or a mixture of benzene and THF for easier
delivery into the microreactor, however lower yields
were obtained with this solvent system.
[3] R. Motoki, M. Kanai, M. Shibasaki, Org. Lett. 2007, 9,
2997–3000.
[4] G.-J. Deng, C.-J. Li, Synlett 2008, 1571–1573.
[5] N. Chinkov, A. Warm, E. M. Carreira, Angew. Chem.
2011, 123, 3014–3018; Angew. Chem. Int. Ed. 2011, 50,
2957–2961.
[6] a) L. A. Radesca, Y. S. Lo, J. R. Moore, M. E. Pierce,
Synth. Commun. 1997, 27, 4373–4384; b) M. E. Pierce,
R. L. Parsons, L. A. Radesca, Y. S. Lo, S. Silverman,
J. R. Moore, Q. Islam, A. Choudhury, J. M. Fortunak,
D. Nguyen, et al., J. Org. Chem. 1998, 63, 8536–8543;
c) A. Choudhury, J. R. Moore, M. E. Pierce, J. M. For-
tunak, I. Valvis, P. N. Confalone, Org. Process Res. Dev.
2003, 7, 324–328.
[7] a) R. C. Rizzo, M. Udier-Blagovic, D.-P. Wang, E. K.
Watkins, M. B. Kroeger Smith, R. H. Smith, J. Triado-
Rives, W. L. Jorgensen, J. Med. Chem. 2002, 45, 2970–
2987; b) J. W. Corbett, S. S. Ko, J. D. Rodgers, L. A.
Gearhart, N. A. Magnus, L. T. Bacheler, S. Diamond, S.
Jeffrey, R. M. Klabe, B. C. Cordova, S. Garber, K.
[18] Please see the Supporting Information for optimization
of the alkynylation reaction in flow.
[19] For examples of handling slurries in flow, see: a) T.
Horie, M. Sumino, T. Tanaka, Y. Matsushita, T. Ichi-
mura, J.-I. Yoshida, Org. Process Res. Dev. 2010, 14,
405; b) J. Sedelmeier, S. V. Ley, I. R. Baxendale, M.
Baumann, Org. Lett. 2010, 12, 3618; c) T. Noel, J. R.
Naber, R. L. Hartman, J. P. McMullen, K. Jensen, S. L.
Buchwald, Chem. Sci. 2011, 2, 287–290.
[20] The comparable batch reaction of 3a was performed on
the 1-mmol scale with 1.8 equivalents of cyclopropyla-
cetylene in THF (1M) at 908C for 200 min in a tightly
sealed reaction tube. An NMR yield of 84% was ob-
tained.
[21] a) Y. Asano, H. Ito, K. Hara, M. Sawamura, Organo-
metallics 2008, 27, 5984–5996; b) T. Ishii, R. Watanabe,
T. Moriya, H. Ohmiya, S. Mori, M. Sawamura, Chem.
Eur. J. 2013, 19, 13547–13553.
Adv. Synth. Catal. 2013, 355, 3517 – 3521
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3521