Acid-Catalyzed Activation of Peroxyketals: Tunable Radical Initiation at Ambient Temperature and below

Article abstract of DOI:10.1021/acs.orglett.6b02419

This study details how peroxyketals, commercially available thermal initiators, and structurally related peroxides are activated in the presence of an acid catalyst to generate radicals at room temperature and below. This simple combination of two substrates was shown to efficiently initiate a variety of radical processes. This phenomenon is rationalized by the acid-catalyzed in situ formation of highly unstable alkenyl peroxides which readily decompose into initiating radical species.

Full text of DOI:10.1021/acs.orglett.6b02419

This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Letter
pubs.acs.org/OrgLett
Acid-Catalyzed Activation of Peroxyketals: Tunable Radical Initiation
at Ambient Temperature and Below
Bertrand Schweitzer-Chaput,*,† Esther Boess, and Martin Klussmann*
Max-Planck-Institut fur Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 Mulheim an der Ruhr, Germany
̈
̈
S
* Supporting Information
ABSTRACT: This study details how peroxyketals, commer-
cially available thermal initiators, and structurally related
peroxides are activated in the presence of an acid catalyst to
generate radicals at room temperature and below. This simple
combination of two substrates was shown to efficiently initiate
a variety of radical processes. This phenomenon is rationalized by the acid-catalyzed in situ formation of highly unstable alkenyl
peroxides which readily decompose into initiating radical species.
adical chemistry offers unique reaction pathways, often
commercial low temperature initiators are cumyl peroxyneo-
decanoate and the diazo compound V-70, which have 10 h half-
life temperatures of only 38 and 30 °C, respectively.2c,3 While
these are commercial and useful products, it is evident that a
cold chain must be strictly followed throughout transport and
storage. To alleviate the associated safety hazards, other
strategies have been developed to generate radicals at low
temperature.2b,e Activation of peroxides by UV light, metal salts,
or amines is a common method.4 Other strategies rely on
photoinitiation5 or on the instability of organoborane and zinc
compounds toward molecular oxygen.6 However, storage of the
latter reagents requires strictly inert conditions, and control of
the initiation rate can be difficult.6d
R
giving products that are inaccessible by ionic mecha-
nisms.1 To initiate radical-chain reactions, compounds that
generate radical species by the homolytic scission of a weak
bond are commonly used, for example, peroxides like tert-
butylperoxide and 1a (Scheme 1a) and diazo compounds like
Scheme 1. Selected Radical Initiators, Including 10 h Half-
Life Temperatures2c,d,3
Here, we show that the combination of geminal bisperoxides
and derivatives with an acid catalyst initiates radical reactions at
ambient temperature and below (Scheme 1c).
We recently reported that ketones and hydroperoxides
generate oxyl and α-keto alkyl radicals in the presence of acid.7
These radicals can be exploited synthetically for addition
reactions to olefins to give low molecular weight products7a,8 or
polymers.9 We postulated alkenyl peroxides 2 as intermediates
that rapidly decompose due to their very weak O−O bonds.7 In
polymerization reactions, we observed that geminal bisperoxide
1a can be used instead of the combination of ketone and
hydroperoxide and leads to increased rates.9 Interestingly,
peroxyketals like 1a are common thermal initiators2e,10 but also
potentially explosive (Figure 1).11
It has been shown that thermal runaway hazards of
peroxyketals are dramatically increased in the presence of
strong acids12 and that acid addition shortens curing times
when they are used.2e We hypothesized that this phenomenon
is due to the in situ formation of unstable alkenyl peroxides,
and that this behavior could be exploited in a controlled
manner for radical initiation purposes (Scheme 1c). We
therefore evaluated the potential of geminal bisperoxides as
AIBN (Scheme 1b).2 Thermal bond homolysis is usually
achieved at elevated temperatures, and many initiators have
been developed to vary the rate of initiation at a given
temperature (Scheme 1).
Designing low-temperature radical initiators requires com-
promises. The desired compound has to be sufficiently unstable
to decompose at low temperature but needs to be stable
enough to be produced, transported, and handled. Examples of
Received: August 12, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b02419
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(22%, entry 5) and trichloroacetic acid (18%, entry 6) gave
reasonable conversions, while acetic acid was inactive (entry 7).
Nitric acid falls out of this trend and is more efficient than its
pKa value would suggest, possibly due to alternative pathways
involving its oxidizing properties (96%; entry 4).13 Eventually,
all reactions gave high yields when allowed to reach full
conversion (80−95% yield of 8 after 24 to 72 h), showing that
the acid catalyst only influences the initiation rate, formation of
the alkenyl peroxide intermediate presumably being the rate-
controlling step. Scandium(III) triflate, a Lewis acid, was also
found to be competent (69%; entry 8).
Figure 1. Geminal bisperoxides and related compounds used in this
study.
Different commercial peroxyketal solutions were then
evaluated using methanesulfonic acid as a standard catalyst of
medium reactivity. Compound 3a (Trigonox D; 50% weight)
proved to be less efficient than 1a (45%, entry 3), giving only
10% of product 8 after 1 h and 76% after 48 h (entry 11).
Compound 4 (Trigonox 301; 41% weight) only showed low
conversion after 2 days of reaction (8%, entry 15). Compound
5 (Luperox DHD-9, 32% weight) was found to be slightly more
reactive than 4, giving a trace amount of 8 after 1 h and 12%
after 48 h (entry 16).
radical initiators at ambient temperature in reactions that do
not incorporate fragments of the initiator. For this purpose, we
selected the radical bromination of fluorene 7 by N-
bromosuccinimide (NBS) at room temperature as a benchmark
reaction (Table 1).
Table 1. Wohl−Ziegler Bromination of Fluorene at Ambient
a
Temperature.
Based on this strong influence of the peroxyketal structure on
its reactivity, we decided to synthesize structural analogues,
hoping to design more reactive ones toward acid activation.
The synthesis of these compounds is easily accomplished by
reaction of ketones, ketals, and aldehydes, respectively, with
hydroperoxides (see the Supporting Information). We first
looked at the effect of the leaving group. Compound 1b proved
to be less reactive than 1a (21%, entry 9). Interestingly, the
azide containing 1c was as efficient as 1a, giving 47% of 8 after
1 h (entry 10). We then evaluated aromatic substituents around
the peroxyketal moiety to enhance the driving force of
elimination toward alkenyl peroxides. Compound 3b was
indeed more effective than 3a (20%, entry 12 compared to
entry 11), while 3c was much less efficient (33% after 48 h,
entry 13), showing that a benzyl substituent is beneficial while
an aromatic substituent is highly detrimental. Peroxyacetal 3d
was very unreactive, only giving small amounts of product after
48 h (9%, entry 14), showing the requirement of a ketone-
derived bisperoxide for efficient acid activation. In light of these
results and the increased reactivity of cyclic 1a compared to the
acyclic 3a, we reasoned that a cyclic and benzylic peroxyketal
would be very reactive. Indeed, 6a was found to be slightly
more reactive than 1a (50%, entry 17), while 6b was the most
efficient of the structures evaluated, giving 74% of 8 after 1 h of
reaction (entry 18).
b
c
entry
peroxide
acid (pKa)
yield (%)
1
1a
1a
1a
1a
1a
1a
1a
1a
1b
1c
3a
3b
3c
3d
4
H2SO4 (−8)
72
67
45
96
22
18
0
2
p-TsOH (−2.8)
CH3SO3H (−1.9)
HNO3 (−1.4)
CF3CO2H (0.23)
CCl3CO2H (0.66)
AcOH (4.76)
3
4
5
6
7
d
8
Sc(OTf)3
69
9
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
CH3SO3H (−1.9)
21
47
10
11
12
13
14
15
16
17
18
e
10; 76
20
e
0; 33
e
0; 9
e
0; 8
e
5
1; 12
6a
6b
50
74
a
All reactions were performed on a 0.5 mmol scale in 5 mL of
dichloromethane under an argon atmosphere at room temperature
b
(22−23 °C) and with exclusion of light for 1 h. pKa value in water
c
given in parentheses. Yields determined by 1H NMR using CH2Br2 as
We were then interested in finding the limits of reactivity
with regard to temperature (Scheme 2).
Lowering the temperature had a detrimental effect on the
efficiency of the reaction (see the Supporting Information for
d
e
an internal standard. Reaction performed in acetonitrile. After 48 h
of reaction.
The bromination proceeded efficiently using a commercial
solution of peroxyketal 1a (Trigonox 22, 50% weight in mineral
oil) in combination with different Brønsted acids. Control
experiments confirmed the requirement for both acid and
peroxyketal, with no conversion being observed after 24 h if
either one of these components was omitted. A clear trend
following the pKa value of the acid catalyst can be seen:
stronger acids give faster conversion. Sulfuric and p-
toluenesulfonic acid display similar behavior with 72% and
67% of 8 after 1 h, respectively (entries 1 and 2).
Methanesulfonic acid gave a slightly lower yield (45%; entry
3), and only the activated carboxylic acids trifluoroacetic acid
Scheme 2. Wohl−Ziegler Bromination at Cryogenic
Temperatures
B
DOI: 10.1021/acs.orglett.6b02419
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
additional details). However, when the reaction was performed
at −10 °C with stoichiometric amounts of 1a and
methanesulfonic acid, a satisfying 60% yield of 8 was obtained
after 24 h. We compared this result with our original
combination of ketone and hydroperoxide7a and found that,
while it is still a competent system, its efficiency is considerably
lower than using the preformed peroxyketal 1a (60% vs 17%).
Lowering the temperature to −20 °C showed only trace
conversion with this combination, while the more reactive 6b
still gave 25% of 2 after 24 h.
Scheme 4. Radical Decarboxylation with an Acid-Sensitive
Compound
We evaluated a selection of classical radical reactions using
the commercial solution of peroxyketal 1a and methanesulfonic
acid for initiation (Scheme 3).
The supposed in situ generation of unstable alkenyl
peroxides and the corresponding radicals (Scheme 1C) was
supported by a trapping experiment in the presence of styrene,
which led to a γ-peroxyketone, indicative of the formation of α-
keto-alkyl and oxyl radicals (see the Supporting Information).
No attempt was made yet to observe alkenyl peroxide
intermediates spectroscopically, which might be difficult since
these compounds are believed to be highly reactive. In fact,
only a single characterization study is known for a special type
of heterocyclic alkenylperoxides.16
Scheme 3. Application of the 3a/Acid Combination for
Selected Radical Reactions at Room Temperature
These different reactions demonstrate the generality of this
initiation system. Abstraction of a hydrogen atom from suitable
substrates, presumably by a tert-butoxyl radical, is possible, as
shown by the success of the Wohl−Ziegler and thiol−ene
reactions. Direct abstraction of halogen atoms, presumably by
the complementary carbon centered radical formed, is also
possible, as in the ATRA of CCl4. If the radicals generated are
not reactive enough to initiate chains themselves, they are
competent in initiating reactions relying on the use of a hydride
mediator like TMS3SiH. The exact mechanism of initiation and
fate of the initiating radicals will not be obvious in many
reactions, since hydrogen atom transfer, for example, will
generate cyclohexanone and tert-butyl alcohol, which would
also form by hydrolysis and O−O bond homolysis via any other
mechanism, respectively. Detailed studies to clarify this issue
are planned for a later stage.
In conclusion, we describe how peroxyketals and related
compounds, normally used as thermal initiators, can be
activated in the presence of an acid catalyst to efficiently
initiate a wide variety of radical processes at room temperature
and down to −20 °C. The rate of radical formation was shown
to depend on the strength and concentration of the activating
acid as well as the structure of the peroxide. Both parameters
offer the opportunity to fine-tune the reaction rate. This
method of generating radical initiators in situ, believed to be
alkenyl peroxides, has the advantage that the substrates are
commercially available and easily synthesized, respectively,
tolerant to air and moisture, and thermally quite stable. It can
be envisioned that even more reactive peroxyketals or
combinations with special catalysts can be developed to further
lower the initiation temperatures.
Addition of a thiol to a terminal olefin proceeded smoothly
in the presence of 20 mol % of 1a and 10 mol %
methanesulfonic acid to give product 9 in 88% yield. Similarly,
an atom-transfer radical addition (ATRA) of carbon tetra-
chloride was successfully initiated. Although it required a larger
amount of 1a, 60% of product 10 was obtained. The TMS3SiH-
mediated reduction of 9-iodophenanthrene was performed,
giving phenanthrene 11 in an excellent 94% yield, requiring
only low amounts of peroxyketal and acid. The reductive
cyclization of 12 was similarly successful, giving benzofurane 13
in 79% yield. Disappointingly, n-Bu3SnH seemed incompatible
with this initiation system, and no conversion was observed in
these two reactions when it replaced TMS3SiH. However,
TMS3SiH has proven to be a suitable replacement for n-
Bu3SnH-mediated reactions and avoids most toxicity and
purification problems associated with stannane reagents.14
Obviously, the reaction conditions involving a strong acid
will be incompatible for some compounds. We investigated this
in more detail by performing a Barton−McCombie deoxyge-
nation on an acid-sensitive ketal, 14.15 While longer reaction
times generally led to decomposition, the reaction could be
performed successfully with 72% yield of product 15 and 28%
of leftover starting material after 1 day, as determined by NMR
spectroscopy using an internal standard (Scheme 4, also see the
Supporting Information).
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b02419.
This outcome illustrates that acid-sensitive substrates or
products can be compatible with this method, especially given
that reaction time, acid strength, and concentration can be
further fine-tuned.
Experimental details including synthesis and character-
ization of starting materials and products (PDF)
C
DOI: 10.1021/acs.orglett.6b02419
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
in mineral oil which were used as such and are safe to be shipped and
stored. See the Supporting Information for further details..
(12) Hsueh, K.-H.; Chen, W.-T.; Chu, Y.-C.; Tsai, L.-C.; Shu, C.-M.
J. Therm. Anal. Calorim. 2012, 109, 1253.
(13) Control experiments revealed that nitric acid alone does not
promote any reaction but that, in the presence of tBuOOH, significant
conversion is achieved after 24 h (80%).
(14) (a) Chatgilialoglu, C.; Timokhin, V. I. Silanes as Reducing
Reagents in Radical Chemistry. In Encyclopedia of Radicals in
Chemistry, Biology and Materials; John Wiley & Sons, Ltd, 2012.
(b) Studer, A.; Amrein, S. Synthesis 2002, 2002, 835.
(15) Lopez, R. M.; Hays, D. S.; Fu, G. C. J. Am. Chem. Soc. 1997, 119,
6949.
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: bschweitzer@iciq.es.
*E-mail: klusi@mpi-muelheim.mpg.de.
Present Address
†Institut Catala
43007 Tarragona, Spain.
̀
d’Investigacio
́
Quimica, Av. Paisos Catalans 16,
́
̈
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
(16) Kropf, H.; Ball, M. Justus Liebigs Ann. Chem. 1976, 1976, 2331.
■
We thank the DFG (KL 2221/3-1 and Heisenberg scholarship
to M.K., KL 2221/4-1) and the MPI fuer Kohlenforschung
(postdoctoral scholarship to B.S.-C.) for funding.
REFERENCES
■
(1) (a) Encyclopedia of Radicals in Chemistry, Biology and Materials;
Chatgilialoglu, C., Studer, A., Eds.; John Wiley & Sons, 2012. (b)
Radicals in Synthesis I & II; Gansauer, A., Ed.; Springer: Berlin, 2006;
̈
Vols. 263−264. (c) Zard, S. Z. Radical Reactions in Organic Synthesis;
Oxford University Press: Oxford, 2003. (d) Radicals in Organic
Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-VCH, Weinheim, 2001.
(e) Jasperse, C. P.; Curran, D. P.; Fevig, T. L. Chem. Rev. 1991, 91,
1237. (f) Giese, B. Radicals in Organic Synthesis: Formation of Carbon−
Carbon Bonds; Pergamon Press: Oxford, 1986; Vol. 5.
(2) (a) Androvic,
2016, 42, 5133. (b) Lalevee
̌
L.; Bartac
́ ̌ ́
ek, J.; Sedlak, M. Res. Chem. Intermed.
́
, J.; Fouassier, J. P. Overview of Radical
Initiation. In Encyclopedia of Radicals in Chemistry, Biology and
Materials; John Wiley & Sons, Ltd., 2012. (c) Sheppard, C. S. Peroxy
Compounds. In Encyclopedia of Polymer Science and Engineering; John
Wiley & Sons, Inc., 1985; Vol. 11, p 1. (d) Sheppard, C. S. Azo
Compounds. In Encyclopedia of Polymer Science and Engineering; John
Wiley & Sons, Inc., 1985; Vol. 2, p 143. (e) Sheppard, C. S.; Kamath,
V. R. Polym. Eng. Sci. 1979, 19, 597.
(3) Kita, Y.; Gotanda, K.; Murata, K.; Suemura, M.; Sano, A.;
Yamaguchi, T.; Oka, M.; Matsugi, M. Org. Process Res. Dev. 1998, 2,
250.
(4) Kwon, T.-Y.; Bagheri, R.; Kim, Y. K.; Kim, K.-H.; Burrow, M. F. J.
Investig. Clin. Dent. 2012, 3, 3.
(5) (a) Angnes, R. A.; Li, Z.; Correia, C. R. D.; Hammond, G. B. Org.
Biomol. Chem. 2015, 13, 9152. (b) Fouassier, J.-P.; Morlet-Savary, F.;
́
Lalevee, J.; Allonas, X.; Ley, C. Materials 2010, 3, 5130.
(6) (a) Zaremski, M. Y.; Garina, E. S.; Gurskii, M. E.; Bubnov, Y. N.
Polym. Sci., Ser. B 2013, 55, 304. (b) Akindele, T.; Yamada, K.-i.;
Tomioka, K. Acc. Chem. Res. 2009, 42, 345. (c) Renaud, P.;
Beauseigneur, A.; Brecht-Forster, A.; Becattini, B.; Darmency, V.;
Kandhasamy, S.; Montermini, F.; Ollivier, C.; Panchaud, P.; Pozzi, D.;
Scanlan, E. M.; Schaffner, A.-P.; Weber, V. Pure Appl. Chem. 2007, 79,
223. (d) Ollivier, C.; Renaud, P. Chem. Rev. 2001, 101, 3415.
(7) (a) Schweitzer-Chaput, B.; Demaerel, J.; Engler, H.; Klussmann,
M. Angew. Chem., Int. Ed. 2014, 53, 8737. (b) Schweitzer-Chaput, B.;
́
Kurten, T.; Klussmann, M. Angew. Chem., Int. Ed. 2015, 54, 11848.
(8) (a) Boess, E.; Karanestora, S.; Bosnidou, A.-E.; Schweitzer-
Chaput, B.; Hasenbeck, M.; Klussmann, M. Synlett 2015, 26, 1973.
(b) Xia, X.-F.; Zhu, S.-L.; Zeng, M.; Gu, Z.; Wang, H.; Li, W.
Tetrahedron 2015, 71, 6099.
(9) Vandenbergh, J.; Schweitzer-Chaput, B.; Klussmann, M.; Junkers,
T. Macromolecules 2016, 49, 4124.
(10) (a) Watanabe, Y.; Ishigaki, H.; Suyama, S. Polym. J. 1992, 24,
257. (b) Drumright, R. E.; Kastl, P. E.; Priddy, D. B. Macromolecules
1993, 26, 2246.
(11) Although we never experienced any problems in the
experiments as described in this work, precautions should be taken
when working with neat peroxides. Neat 4 and 5 are described as
explosive; see: Milas, N. A.; Golubovic,
5824 The commercial bisperoxides 1a, 3a, 4, and 5 come as solutions
́
A. J. Am. Chem. Soc. 1959, 81,
D
DOI: 10.1021/acs.orglett.6b02419
Org. Lett. XXXX, XXX, XXX−XXX