Dehydrogenation of perfluoroalkyl ketones by using a recyclable oxoammonium salt

Article abstract of DOI:10.1002/ejoc.201300392

A novel dehydrogenation reaction of perfluoroalkyl ketones by the oxoammonium salt 4-acetylamino-2,2,6,6-tetramethylpiperidine-1-oxoammonium tetrafluoroborate (4-NHAc-TEMPO⁺BF₄^-, Bobbitt's salt, 1) is described. The reaction proceeds under mildly basic conditions and appears to be unique to perfluoroalkyl ketones. A proposed mechanism for this unusual transformation is given. The byproduct of the reaction, 4-acetylamino-2,2,6,6-tetramethyl-1-piperidinyloxy (1a), can easily be recovered and used to regenerate the oxoammonium salt. The dehydrogenation of perfluoroalkyl ketones by using an oxoammonium salt is reported. The reaction proceeds under mildly basic conditions and affords α,β-unsaturated products in fair to excellent yields. The reaction likely proceeds through a two-step sequence. The spent oxidant can easily be recovered and used to regenerate the oxoammonium salt. Copyright

Full text of DOI:10.1002/ejoc.201300392

SHORT COMMUNICATION
DOI: 10.1002/ejoc.201300392
Dehydrogenation of Perfluoroalkyl Ketones by Using a Recyclable
Oxoammonium Salt
Trevor A. Hamlin,[a] Christopher B. Kelly,[a] and Nicholas E. Leadbeater*[a,b]
Keywords: Dehydrogenation / Oxoammonium salts / Fluorine / Alkenes / Ketones
A novel dehydrogenation reaction of perfluoroalkyl ketones
by the oxoammonium salt 4-acetylamino-2,2,6,6-tetrameth-
ylpiperidine-1-oxoammonium tetrafluoroborate (4-NHAc-
unique to perfluoroalkyl ketones. A proposed mechanism for
this unusual transformation is given. The byproduct of the
reaction, 4-acetylamino-2,2,6,6-tetramethyl-1-piperidinyloxy
TEMPO+BF4–, Bobbitt’s salt, 1) is described. The reaction pro- (1a), can easily be recovered and used to regenerate the
ceeds under mildly basic conditions and appears to be
oxoammonium salt.
Introduction
Oxoammonium salts are environmentally benign, re-
cyclable, metal-free species that can facilitate oxidation un-
der extremely mild conditions.[1] As such, they have gained
recent popularity as alternatives to more traditional oxidiz-
ing reagents.[2–9] Much of the literature involving oxoam-
monium salts centers around the conversion of alcohols
into the corresponding carbonyl species. A range of func-
tional groups can be accessed by using this strategy includ-
ing aldehydes,[2a] carboxylic acids,[2b] ketones,[2a] and even
esters.[2c,2d] As part of a broad program to enhance the pro-
file of oxoammonium salt based transformations, we be-
came interested in the possibility of utilizing 4-acetylamino-
2,2,6,6-tetramethylpiperidine-1-oxoammonium tetrafluoro-
Scheme 1. Previously observed result when attempting oxidation of
1,1,1-trifluoro-4-phenylbutan-2-ol.[10]
Results and Discussion
Initially, we sought to convert a model aliphatic α-CF3
alcohol, 1,1,1-trifluoro-4-phenylbutan-2-ol, directly into its
corresponding α,β-unsaturated TFMK 3a. However, several
complications arose when attempting this endeavor. First, a
large excess amount of salt was required to ensure complete
conversion into the desired material. In addition, 1,5-diaza-
bicyclo(4.3.0)non-5-ene (DBN) was required for initial oxi-
dation of the alcohol and, as a result of competitive decom-
position of this base, an extensive purification procedure
was required to isolate the product.
Given these complications, we chose instead to start from
aliphatic TFMK 2a, which could readily be prepared by an
alternative method we recently developed.[12] Treatment
with DBN and the oxoammonium salt did facilitate rapid
dehydrogenation of this TFMK, which indicates that the
dehydrogenation event is independent of alcohol oxidation.
However, an excess amount of salt was still required, and
we again encountered significant difficulty when purifying
the product at the end of the reaction. Seeking to circum-
vent these roadblocks, we wondered if a pyridyl base could
be used in place of DBN. We found that this was indeed
the case. Upon treatment of 2a with the oxoammonium salt
(2.6 equiv.) and pyridine (2.2 equiv.) in dichloromethane as
–
borate (4-NHAc-TEMPO+BF4 , Bobbitt’s salt, 1) for the
oxidation of α-trifluoromethyl alcohols to trifluoromethyl
ketones (TFMKs).[10] Although successful in that aim, we
encountered an unusual side product when attempting to
oxidize aliphatic α-CF3 alcohols, namely, the over-oxidized
α,β-unsaturated TFMKs (Scheme 1). With essentially no
precedence for this type of direct dehydrogenation transfor-
mation by an oxoammonium species,[11] we sought to ex-
ploit this serendipitous finding. We report our results
herein.
[a] Department of Chemistry, University of Connecticut,
55 North Eagleville Road, Storrs, Connecticut 06269-3060,
U.S.A.
Fax: +1-860-486-2981
E-mail: nicholas.leadbeater@uconn.edu
Homepage: www. http://chemistry.uconn.edu/
[b] Department of Community Medicine & Health Care,
University of Connecticut Health Center, The Exchange,
263 Farmington Avenue, Farmington, Connecticut 06030,
U.S.A.
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201300392.
3658
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2013, 3658–3661

Dehydrogenation of Perfluoroalkyl Ketones
the solvent, we observed slow but steady formation of de- Table 1. Scope of the dehydrogenation of various trifluoromethyl
ketones.[a]
hydrogenated product 3a at room temperature. Seeking to
accelerate the rate of dehydrogenation, we opted for the
more basic 2,6-lutidine. Although a marked increase in the
reaction rate was noted, the requisite reaction time was still
quite long (Ͼ48 h). We therefore performed the reaction at
reflux, which allowed us to shorten the reaction time to 24–
48 h. Using these conditions (Scheme 2), we were not only
able to obtain the pure dehydrogenation product in good
yield, but we were also able to recover spent oxidant 1a.
Scheme 2. Dehydrogenation of 2a under the optimized conditions.
Seeking to explore the scope of this transformation, we
prepared a number of aliphatic TFMKs and subjected them
to our optimized dehydrogenation conditions. The results
are shown in Table 1. Initially we chose to screen a variety
of hydrocinnamyl derivatives. We found that both electron-
rich (Table 1, Entries 2–4) and electron-poor (Table 1, En-
tries 5–7) substrates readily underwent dehydrogenation to
give similar yields of the product, albeit at different rates.
A significant rate enhancement in the systems bearing elec-
tron-deficient arenes was observed, and these reactions
reached completion typically in 12–16 h. The sterically en-
cumbering o-isopropyl ether (Table 1, Entry 4) did undergo
facile dehydrogenation but required a slightly longer reac-
tion time to ensure complete conversion. A representative
heterocyclic system (Table 1, Entry 10) and a polycyclic sys-
tem (Table 1, Entry 11) were next screened, and they af-
forded their corresponding dehydrogenation products in
good yield. Suspecting the driving force for dehydrogena-
tion to be extended conjugation (and an activated benzyl
methylene moiety), we shifted our attention to non-aryl-
substituted substrates. As expected, we obtained substan-
tially diminished yields and far longer reaction times were
required (Table 1, Entries 12 and 13). In the case of 2l
(Table 1, Entry 12), although full conversion was achieved,
desired product 3l could not be obtained in acceptable pu-
rity or yield. To probe further the importance of an acti-
vated β-methylene, we prepared a diketo substrate (Table 1,
Entry 14). Such a system would possess far more acidic β-
hydrogen atoms. Dehydrogenation of this substrate pro-
ceeded extraordinarily fast and gave the unsaturated prod-
uct in excellent yield, which corroborated the necessity for
an activated β-methylene moiety. It should be stressed that
the spent oxidant could easily be recovered and reclaimed,
which make the dehydrogenation process inherently re-
[a] Conditions: TFMK (1 equiv.),
1 (2.6 equiv.), 2,6-lutidine
cyclable.
To gain further mechanistic insight, we prepared α- and
β-methyl hydrocinnamyl substrates (Table 1, Entries 15 and
(2.2 equiv.), CH2Cl2 (0.1 m in starting TFMK). [b] Isolated yield
after purification. [c] Substrate was completely consumed by GC–
MS analysis. [d] Contained inseparable impurities following
workup.
16). Interestingly, both failed to undergo dehydrogenation,
Eur. J. Org. Chem. 2013, 3658–3661
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3659

T. A. Hamlin, C. B. Kelly, N. E. Leadbeater
SHORT COMMUNICATION
Figure 1. Plausible mechanistic pathway of the dehydrogenation reaction.
but in the case of the β-methyl system, by GC analysis the pentafluoroethyl group readily dehydrogenated to yield the
starting material was completely consumed. Intrigued by desired alkene in superior yield to its trifluoromethyl ana-
this finding, we conducted a brief 19F NMR spectroscopy logue (Table 2, Entry 1). Interestingly, the less-destabilizing
study by using model system 3a. We found that three pentafluorophenyl-substituted system (Table 2, Entry 2)
unique fluorinated species were present within 30 min of gave a very low yield of the chalcone-like product. On the
starting the reaction. Whereas two of these signals corre- basis of the posited mechanism, we believe that the interme-
sponded to starting material and product respectively, the diate of the α-alkylation was much less prone to hydration,
third signal was initially puzzling, as it appeared at a rather and hence, either competitive α-oxidation or lack of further
unique chemical shift (–86.2 ppm). We postulated that reactivity contributed to the marked low yield. Aliphatic
rather than proceeding in a concerted pathway, the reaction ketones (Table 2, Entries 3 and 4) failed to undergo dehy-
likely was a two-step process. The first step involved pre- drogenation, which is in line with the proposed mechanism.
cedented α-alkylation of the intermediate enolate (or enol),
whose formation is likely facilitated by lutidine.[3,14] Once
alkylated, the carbonyl undergoes rapid hydration by any
Table 2. Scope of dehydrogenation of various perfluoroalkyl and
trace amount of water in the system (i.e., from the salt or
solvent).[13] On the basis of the 19F chemical shift of similar
α-alkoxy trifluoromethyl hydrates,[15] we believe this ex-
plains the third peak observed in the NMR spectroscopy
studies. This deters the known α-oxidation pathway, by
decreasing the acidity of the α-proton, opening up an E2-
like elimination pathway. Elimination of the alkoxyamine
solely affords the E isomer. This likely results from steric
clash between the β-substituent and the bulky tetrameth-
ylpiperidine moiety, which restricts the conformational free-
dom of the molecule. The resultant gem-diol undergoes ra-
pid dehydration, as the equilibrium substantially favors the
keto form of α,β-unsaturated TFMKs over the gem-diol.[15]
The proposed mechanism for the reaction is shown in Fig-
ure 1.
aliphatic ketones.[a]
With a plausible mechanism in hand, we sought to test
whether other perfluoroalkyl substrates and even alkyl sys-
tems would undergo dehydrogenation under our optimized
[a] Conditions: TFMK (1 equiv.),
(2.2 equiv.), CH2Cl2 (0.1 m in starting TFMK). [b] Isolated yield
1 (2.6 equiv.), 2,6-lutidine
conditions (Table 2). As expected, a substrate bearing a after purification.
3660
www.eurjoc.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2013, 3658–3661

Dehydrogenation of Perfluoroalkyl Ketones
Conclusions
Acknowledgments
This work was supported by the National Science Foundation
CAREER Award (CHE-0847262). The authors especially acknow-
ledge Dr. James Bobbitt for helpful discussions, troubleshooting
assistance, and aid in salt preparation. Dr. You-Jun Fu is thanked
for his assistance in obtaining the HRMS of novel compounds. The
authors also acknowledge Owen Insel, Mike Mercadante, Timothy
Monos, and Rebecca Wiles of the University of Connecticut for
We have described the first direct dehydrogenation of
simple ketones using an oxoammonium salt. The reaction
can be conducted under mild conditions and gives good
yields of the corresponding α,β-unsaturated products. While
specific to perfluoroalkyl ketones, an array of β-substituents
is tolerated. The proposed mechanism involves a two-step
process invoking α-alkylation followed by an E2-like elimi- useful discussion/technical assistance.
nation. Although this study was focused on perfluoroalkyl
[1] a) M. A. Mercadante, C. B. Kelly, J. M. Bobbitt, L. J. Tilley,
ketones, extension of this reaction to other substrates bear-
ing electron-withdrawing groups may be possible. Further
studies of this and other methods for oxoammonium salt-
mediated dehydrogenation are currently underway.
N. E. Leadbeater, Nat. Protoc. 2013, 8, 666–676; b) C. B. Kelly,
Synlett 2013, 24, 527–528; c) J. M. Bobbitt, C. Bruckner, N.
Merbouh, Org. React. 2010, 74, 103–424; d) J. M. Bobbitt, N.
Merbouh, Org. Synth. 2005, 82, 80–86.
[2] a) J. M. Bobbitt, J. Org. Chem. 1998, 63, 9367–9374; b) J. Qui,
P. P. Pradhan, N. B. Blanck, J. M. Bobbitt, W. F. Bailey, Org.
Lett. 2012, 14, 350–353; c) N. Merbouh, J. M. Bobbitt, C.
Bruckner, J. Org. Chem. 2004, 69, 5116–5119.
[3] N. A. Eddy, C. B. Kelly, M. A. Mercadante, N. E. Leadbeater,
G. Fenteany, Org. Lett. 2012, 14, 498–501.
Experimental Section
[4] H. Richter, O. Mancheño García, Org. Lett. 2011, 13, 6066–
6069.
[5] H. Richter, R. Rohlmann, O. Mancheño García, Chem. Eur. J.
2011, 17, 11622–11627.
Synthesis of (E)-1,1,1-Trifluoro-4-phenylbut-3-en-2-one (3a) as a
Representative Procedure for Dehydrogenation: To a 250 mL round-
bottomed flask equipped with stir bar was added trifluoromethyl
ketone 2a (1.42 g, 7 mmol, 1 equiv.), 2,6-lutidine (1.65 g,
15.4 mmol, 2.2 equiv.), and CH2Cl2 (70.0 mL). The reaction mix-
ture was stirred for 10 min, and salt 1 (5.46 g, 18.2 mmol,
2.6 equiv.) was then added to the flask. The flask was equipped
with a reflux condenser, and the contents were stirred for 24 h at
reflux. Once the reaction was judged complete (GC–MS analysis),
the solvent was removed in vacuo to yield a thick red residue. An-
hydrous diethyl ether (ca. 30 mL) was added to the flask (causing
precipitation of the spent oxidant 1a), and the mixture was allowed
to stir for 10 min. Any large pieces of 1a were crushed into fine
particles to release any trapped product. The solution was filtered
through a fritted funnel to isolate spent oxidant 1a. The solid was
washed thoroughly with anhydrous diethyl ether. The resulting fil-
trate was dry-packed on silica (using 1.5–2 weight equivalents rela-
tive to the ketone). The dry-packed material was loaded on to a
pad of silica gel (3.5 weight equivalents relative to the ketone) and
rinsed with approximately 80–100 mL of hexane/EtOAc (95:5). The
eluting solvent was removed in vacuo by rotary evaporation in a
room-temperature water bath to afford the pure α,β-unsaturated
TFMK product (0.9086 g, 65%) as a clear yellow oil. 1H NMR
(400 MHz, CDCl3): δ = 7.02 (dd, J = 15.89, 0.73 Hz, 1 H) 7.41–
7.57 (m, 3 H) 7.60–7.70 (m, 2 H) 7.97 (d, J = 15.89 Hz, 1 H) ppm.
13C NMR (100 MHz, CDCl3): δ = 116.68 (q, JC,F = 291.20 Hz,
CF3) 116.90 (CH) 129.50 (C) 129.51 (CH) 132.59 (CH) 133.60
(CH) 150.41 (CH) 180.26 (q, JC,C-F = 35.90 Hz, 1 C) ppm. 19F
NMR (377 MHz, CDCl3): δ = –80.67 ppm. GC–MS (EI): m/z (%)
= 200 (56) [M]+, 199 (12), 132 (10), 131 (100), 103 (81), 102 (13),
77 (42), 69 (6), 51 (16). HRMS (ESI+): calcd. for C10H7F3O [M +
H]+ 201.0527; found 201.0548.
[6] a) P. P. Pradhan, J. M. Bobbitt, W. F. Bailey, Org. Lett. 2006,
8, 5485–5487; b) T. Takata, Y. Tsujino, S. Nakanishi, K. Naka-
mura, E. Yoshida, T. Endo, Chem. Lett. 1999, 937–938.
[7] a) M. Shibuya, M. Tomizawa, Y. J. Iwabuchi, J. Org. Chem.
2008, 73, 4750–4752; b) M. Shibuya, M. Tomizawa, Y. Iwabu-
chi, Org. Lett. 2008, 10, 4715–4718; c) J.-M. Vatèle, Synlett
2008, 1785–1788; d) J.-M. Vatèle, Synlett 2009, 2143–2145; e)
J.-M. Vatèle, Tetrahedron 2010, 66, 904–912.
[8] M. Hayashi, M. Shibuya, Y. Iwabuchi, Org. Lett. 2012, 14,
154–157.
[9] H. Richter, O. Mancheño García, Eur. J. Org. Chem. 2010,
4460–4467.
[10] C. B. Kelly, M. A. Mercadante, T. A. Hamlin, M. H. Fletcher,
N. E. Leadbeater, J. Org. Chem. 2012, 77, 8131–8141.
[11] Mancheño García and co-workers recently reported an oxoam-
monium salt mediated C–C bond-forming reaction from benz-
ylic C(sp3)–H bonds adjacent to an oxygen or nitrogen atom
by dehydrogenative coupling with enolizable carbonyls, see
ref.[9] To the best of our knowledge, the only known dehydro-
genative oxidation by an oxoammonium species was reported
by us, see ref.[3]
[12] D. M. Rudzinski, C. B. Kelly, N. E. Leadbeater, Chem. Com-
mun. 2012, 48, 9610–9612.
[13] During our studies we noted that the addition of water to the
reaction mixture (0.2 mL) greatly accelerated the rate of the
reaction, but complicated the isolation. Additionally, rigor-
ously dried crystalline oxoammonium salt (see ref.[3] for a dry-
ing protocol) caused the reaction to proceed at a much slower
rate, which indicated that the low levels of water found in the
powdered form did assist in the reaction.
[14] a) V. A. Golubev, R. V. Miklyush, Zh. Org. Khim. 1972, 8,
1376–1377; b) Y. Liu, T. Ren, Q. Guo, Chin. J. Chem. 1996, 14,
252–258.
Supporting Information (see footnote on the first page of this arti-
[15] J. Iskra, D. Bonnet-Delpon, J.-P. Bégué, J. Fluorine Chem. 2005,
cle): Experimental details, characterization data for the products,
126, 549–554.
1
and copies of the H NMR, 13C NMR, and 19F NMR spectra for
Received: March 14, 2013
all intermediates and products.
Published Online: May 7, 2013
Eur. J. Org. Chem. 2013, 3658–3661
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3661