Silylated cyclohexadienes as new radical chain reducing reagents: Preparative and mechanistic aspects

Article abstract of DOI:10.1021/ja0341743

Various silylated 1,4-cyclohexadienes are presented as superior tin hydride substitutes for the conduction of various radical chain reductions. Debrominations, deiodinations, and deselenations can be performed using these environmentally benign reagents. Furthermore, Barton - McCombie-type deoxygenations using silylated cyclohexadienes are described. Radical cyclizations, ring expansions, and Giesetype addition reactions with the new tin hydride substitutes are presented. The polymerization of styrene can be regulated using silylated cyclohexadienes. Rate constants for hydrogen atom abstraction from two 1-silyl-cyclohexadienes by primary C-radicals were determined. The effects of the cyclohexadiene substituents on the reaction outcomes are discussed. Finally, qualitative EPR experiments on silyl radical expulsion from silylated cyclohexadienyl radicals are presented.

Full text of DOI:10.1021/ja0341743

Published on Web 04/17/2003
Silylated Cyclohexadienes as New Radical Chain Reducing
Reagents: Preparative and Mechanistic Aspects
Armido Studer,*,† Stephan Amrein,† Florian Schleth,† Tobias Schulte,† and
John C. Walton*,‡
Contribution from the Fachbereich Chemie der UniVersita¨t Marburg, Hans-Meerwein-Strasse,
D-35032 Marburg, UniVersity of St. Andrews, School of Chemistry,
St. Andrews, Fife, United Kingdom KY16 9ST
Received January 15, 2003; E-mail: studer@mailer.uni-marburg.de; jcw@st-andrews.ac.uk
Abstract: Various silylated 1,4-cyclohexadienes are presented as superior tin hydride substitutes for the
conduction of various radical chain reductions. Debrominations, deiodinations, and deselenations can be
performed using these environmentally benign reagents. Furthermore, Barton-McCombie-type deoxygen-
ations using silylated cyclohexadienes are described. Radical cyclizations, ring expansions, and Giese-
type addition reactions with the new tin hydride substitutes are presented. The polymerization of styrene
can be regulated using silylated cyclohexadienes. Rate constants for hydrogen atom abstraction from two
1-silyl-cyclohexadienes by primary C-radicals were determined. The effects of the cyclohexadiene
substituents on the reaction outcomes are discussed. Finally, qualitative EPR experiments on silyl radical
expulsion from silylated cyclohexadienyl radicals are presented.
Scheme 1. Silylated Cyclohexadienes as Radical Chain Reducing
Reagents - the Concept
Introduction
Organotin hydrides such as Bu3SnH, Ph3SnH, and Me3SnH
have successfully been used to mediate preparative radical chain
processes in chemistry over the last 40 years.1 However, there
are several drawbacks associated with these tin-based reagents.
First of all, organostannanes are toxic.2 In addition, special
handling during disposal of tin residues is necessary, and
problems with product purification are often encountered. It is
therefore not surprising that many approaches have been
presented to solve the so-called tin-problem in radical chemistry.3a
For example, tris(trimethylsilyl)silane3b and tetraphenyldisilane3c
have successfully been used to conduct reductive tin free radical
reactions. For a more thorough compilation of tin hydride
substitutes, we refer to a recently published review article on
this topic.3d
1-carbamoylcyclohexa-2,5-dienes are effective sources of
aminoacyl radicals that ring close to yield â- and γ-lactams.4d,f
On the basis of these results, the German group has introduced
silylated 1,4-cyclohexadienes of type 1 as new, readily prepared
radical chain reducing reagents.5 The cyclohexadiene bis-allylic
CH2-moiety acts as the H-donor in these chain reactions.
Reduction of a radical R with a reagent of type 1 affords
cyclohexadienyl radical 2. Rearomatization of 2 then provides
the corresponding silyl radical, which is able to propagate the
chain by reaction with a starting halide, xanthate or phenyl-
selenide R-X (Scheme 1). Arene 3 is formed as the byproduct.
We have also found that these silylated cyclohexadienes can
be used in radical hydrosilylation reactions.6
One of us (JCW) has successfully used functionalized
cyclohexadienes in radical chain reactions.4 For example,
† Fachbereich Chemie der Universita¨t Marburg.
‡ University of St. Andrews.
(1) For reviews, see: (a) Neumann, W. P. Synthesis 1987, 665. (b) Pereyre,
M.; Quintard, J.-P.; Rahm, A. Tin in Organic Synthesis; Butterworth:
London, 1987. (c) RajanBabu, T. V. In Encyclopedia of Reagents for
Organic Synthesis; Paquette, L., Ed.; Wiley: New York, 1995; Vol. 7, p
5016.
(2) (a) Ingham, R. K.; Rosenberg, S. D.; Gilman, H. Chem. ReV. 1960, 60,
459. (b) Boyer, I. J. Toxicology 1989, 55, 253.
(3) (a) Baguley, P. A.; Walton, J. C. Angew. Chem. 1998, 110, 3272; Baguley,
P. A.; Walton, J. C. Angew. Chem., Int. Ed. Engl. 1998, 37, 3072. (b)
Chatgilialoglu, C. Acc. Chem. Res. 1992, 25, 188. (c) Sugi, M.; Togo, H.
Tetrahedron 2002, 58, 3171, and references therein. (d) Studer, A.; Amrein,
S. Synthesis 2002, 835.
(4) (a) Binmore, G.; Walton, J. C.; Cardellini, L. J. Chem. Soc., Chem. Commun.
1995, 27. (b) Baguley, P. A.; Binmore, G.; Milne, A.; Walton, J. C. Chem.
Commun. 1996, 2199. (c) Baguley, P. A.; Walton, J. C. J. Chem. Soc.,
Perkin Trans. 1 1998, 2073. (d) Jackson, L. V.; Walton, J. C. Chem.
Commun. 2000, 2327. (e) Jackson, L. V.; Walton, J. C. J. Chem. Soc.,
Perkin Trans. 2 2001, 1758. (f) Bella, A. F.; Jackson, L. V.; Walton, J. C.
J. Chem. Soc., Perkin Trans. 2 2002, 1839.
Herein, we report in full detail the use of silylated cyclo-
hexadienes as superior tin hydride substitutes. In addition, the
(5) Studer, A.; Amrein, S. Angew. Chem. 2000, 112, 3196; Angew. Chem.,
Int. Ed. Engl. 2000, 39, 3080.
9
5726
J. AM. CHEM. SOC. 2003, 125, 5726-5733
10.1021/ja0341743 CCC: $25.00 © 2003 American Chemical Society

New Radical Chain Reducing Reagents
A R T I C L E S
Table 1. Reduction of Bromoadamantane with Reagent 4 under
Different Conditions
entry
initiator
solvent
time (h)
yield (%)a
1
2
3
4
5
6
7
8
9
AIBN
AIBN
AIBN
AIBN
AIBN
AIBN
Et3B (0.1 equiv)/O2
Et3B (0.2 equiv)/O2
Et3B (0.2 equiv)/O2
air
benzene
hexane
toluene
THF
t-BuOH
MTBE
hexaneb
hexaneb
hexanec
hexane
14
5.5
14
14
14
14
5
4
14
9
88
99
28
78
42
60
82
99
-
10
99
a mol % adamantane determined by GC-analysis using tetradecane as
internal standard. b Reaction was conducted at 80 °C. c Reaction was
conducted at room temperature.
reaction in hexane provided adamantane in quantitative yield
(entry 2). Reductions in toluene, THF, tert-butylmethyl ether
(MTBE) and t-BuOH were not that efficient (entries 3-6).
Therefore, most of the following experiments were conducted
in hexane.
In addition to AIBN, other radical initiators can be used for
the reduction of bromoadamantane with reagent 4. Initiation
8
with Et3B (0.1 equiv)/O2 afforded adamantane in 82% yield
(entry 7). A quantitative conversion was observed upon using
0.2 equiv of Et3B (entry 8). However, the same reaction at room
temperature afforded only traces of adamantane (entry 9). As
will be presented below, the silyl radical expulsion from the
cyclohexadienyl radical deriving from 4 is a very fast process
at ambient temperature. Therefore, the slow H-transfer step,
namely the reduction of the adamantyl radical, is the reason
for the failure of the room temperature reduction. Pleasingly,
the reduction can be conducted perfectly well under an
atmosphere of air, without using an initiator, to give adamantane
in quantitative yield (entry 10).
Figure 1. Various silylated cyclohexadienes studied.
use of silylated cyclohexadienes as regulators for the radical
polymerization of styrene is discussed. Moreover, we present
EPR spectroscopic studies of silyl radical expulsion from the
intermediate silylated cyclohexadienyl radicals.
Results and Discussion
Reductive Defunctionalizations using Silylated Cyclohexa-
dienes. As recently described,6c the silylated cyclohexadienes
can readily be prepared from the corresponding dienes in a one-
pot silylation alkylation procedure (see also the Supporting
Information). The silylated dienes 4-15 used for the present
study are depicted in Figure 1.
Most of the initial experiments were conducted with reagent
4, which was readily prepared on a 30-40 g-preparatiVe scale.
As a test reaction, the reduction of bromoadamantane was
investigated first (equation 1).
We then examined the reduction of bromoadamantane under
the optimized conditions using the Si-reagents 5-15 (Table 2).
The thexyldimethylsilyl reagent 6 and the tri-isopropylsilylated
cyclohexadiene 7 could be used as very efficient tin hydride
substitutes. In both cases, quantitative reduction of bromo-
adamantane was obtained (entries 1,2). However, a disappoint-
ingly low conversion was obtained with the TMS-derivative 5
(35%, entry 3). We soon realized that 5 was not stable under
the applied conditions. We further found that this diene slowly
decomposed even at 4 °C and was therefore not a suitable tin
hydride substitute. With the monomethoxy-reagent 8, conversion
was not complete under the optimized conditions (entry 4). Upon
increasing the amount of AIBN (1 equiv) an 85% yield of
adamantane was obtained (entry 5). The reduction with 8 could
also be performed with di-tert-butylperoxide (DTBP) (hexane,
sealed tube 140 °C, 77%, entry 6). Debromination with
Si-reagent 10 (lacking the methoxy groups) under the optimized
conditions provided adamantane in only 6% yield (entry 7). A
50% conversion was obtained upon increasing the amount of
initiator and reagent (entry 8). We were pleased to find that a
quantitative reduction was obtained with DTBP as an initiator
at 140 °C (entry 9). We suspected that the diminished yields in
the AIBN-experiments at 70 °C were due to inefficient initiation.
Indeed, a quantitative reduction was obtained at 70 °C if di-
The yields were determined using GC-analysis with tetra-
decane as internal standard. The reduction did not go to
completion in refluxing benzene (0.2 M) using R,R′-azoiso-
butyronitrile (AIBN) as initiator (88%, Table 1, entry 1). It is
well-known that silyl radicals can add to benzene and are thus
prevented from propagating the radical chain.7 Indeed, the same
(6) (a) Amrein, S.; Timmermann, A.; Studer, A. Org. Lett. 2001, 3, 2357. (b)
Amrein, S.; Studer, A. Chem. Commun. 2002, 1592. (c) Amrein, S.; Studer,
A. HelV. Chim. Acta 2002, 85, 3559.
(7) Cole, S. J.; Kirwan, J. N.; Roberts, B. P.; Willis, C. R. J. Chem. Soc.,
Perkin Trans. 1 1991, 103. See also Chatgilialoglu, C.; Ingold, K. U.;
Scaiano, J. C. J. Am. Chem. Soc. 1983, 105, 3292.
(8) Miura, K.; Ichinose, Y.; Nozaki, K.; Fugami, K.; Oshima, K.; Utimoto, K.
Bull. Chem. Soc. Jpn. 1989, 62, 143.
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5727

A R T I C L E S
Studer et al.
Table 2. Reduction of 1-Bromoadamantane with Reagents 5-15
in Hexane under Different Conditions
Table 3. Radical Reductions of Various Substrates Using 4, 7 and
10 (1.3 equiv) and AIBN (0.3 equiv) in Refluxing Hexane (0.2 M)
Si-reagent
(equiv)
entry
initiator (equiv)
temp, time
yield (%)a
1
2
3
4
5
6
7
8
6 (1.3)
7 (1.3)
5 (1.3)
8 (1.3)
8 (1.3)
AIBN (0.3)
AIBN (0.3)
AIBN (0.3)
AIBN (0.3)
AIBN (1.0)
DTBP (0.3)
AIBN (0.3)
AIBN (1.0)
DTBP (0.3)
TBHN (0.3)
DTBP (1.0)
TBHN (1.0)
AIBN (0.3)
AIBN (0.3)
AIBN (0.3)
AIBN (0.3)
AIBN (1.0)
DTBP (0.3)
80 °C, 5 h
80 °C, 4.5 h
80 °C, 7 h
70 °C, 7 h
70 °C, 7 h
99
99
35
35
85
77
6
50
99
99
90
86
25
c
8 (1.3)
140 °C, 2.5 h
70 °C, 5 h
10 (1.3)
10 (2.0)
10 (2.0)
10 (1.3)
11 (2.0)
11 (2.0)
9 (1.3)
12 (1.3)
13 (1.3)
14 (1.3)
15 (1.3)
15 (1.3)
70 °C, 7.5 h
140 °C, 2.5 h
70 °C, 0.8 h
140 °C, 1.5 h
70 °C, 4.0 h
80 °C, 14 h
80 °C, 7 h
80 °C, 7 h
80 °C, 8 h
70 °C, 7 h
140 °C, 2.5 h
9b
10
11b
12
13
14
15
16
17
18
c
27
9
14
a Determined by GC-analysis using tetradecane as internal standard.
b Reaction was conducted in decane. c Zero yields also obtained with DTBP
and TBHP.
tert-butylhyponitrite (TBHN)9 was used as an initiator (entry
10). Obviously, the activating methoxy groups are necessary if
AIBN is used as initiator.
In contrast to the bismethoxylated TMS-cyclohexadiene 5,
described above, the TMS-diene 11 was a stable compound and
could readily be stored. As expected, reduction of bromo-
adamantane with 11, lacking the activating methoxy groups,
could not be accomplished using AIBN as an initiator. However,
satisfying yields were obtained with DTBP (entry 11) or TBHN
(entry 12). Reduction of bromoadamantane under the optimized
conditions with the benzannelated cyclohexadiene 9 provided
adamantane in only 25% yield (entry 13). No further experi-
ments were conducted with 9. Our model reduction was
intensively studied with the silylated dihydroanthracene deriva-
tives 12 and 13. The reduction was performed using various
initiators under different conditions (entries 14, 15). However,
the formation of adamantane was never observed. Finally, the
silylated cyclohexadienes 14 and 15 lacking the geminal methyl
group were tested as tin hydride substitutes. It turned out that
the methyl group was essential for the success of the radical
chain reduction. Only very low conversions were obtained with
these reagents under different conditions (entries 16-18).
From these initial experiments we concluded that the bis-
methoxylated cyclohexadienes 4, 6, and 7 are very efficient
reagents for radical debromination. For cyclohexadienes lacking
the methoxy groups, such as 10 and 11, reduction worked well
only if alkoxyl radicals were used to initiate the chain reactions.
As will be discussed in detail below, the methoxy groups
strongly affected the rate of the rearomatization (silyl radical
expulsion). Furthermore, the stability of the cyclohexadienyl
radicals, and probably also the rate constant for the H-transfer
step, correlated with the cyclohexadiene substituents (failed
AIBN-initiation!). The cyclohexadienyl radicals generated from
the silylated anthracene derivatives 12 and 13 are rather stable.
Due to the low resonance energy of anthracenes, silyl radical
expulsion from those radicals did not occur under the reaction
conditions (see EPR experiments below) which eventually led
a Yield determined by GC analysis using tetradecane as internal standard.
b Isolated yield. c TBHN (0.3-1.5 equiv). d With 2.0 equiv of 4 and 0.6 equiv
of AIBN. e With 3 equiv of 7 and 1 equiv of AIBN.
to chain termination. In addition, we have shown that the
geminal methyl group is essential for our reagents. The
regioselectively (ortho-substituents in 14 and 15 efficiently block
the R-position, see below) generated silylated cyclohexadienyl
radicals derived from 14 and 15 underwent partial oxidation to
the corresponding silylated arenes which is a chain termination
process.10
Cyclizations and Intermolecular Additions. We then
decided to study the scope and limitations of our new tin hydride
substitutes. Some typical radical chain defunctionalizations were
studied using Si-reagents 4, 7, and 10 under the conditions
specified in Table 3. Primary, secondary, and aromatic halo-
genides were efficiently reduced (entries 1-11). Moreover,
Barton-McCombie type deoxygenations11 using xanthates and
thionocarbonates12 could efficiently be conducted with our new
reagents (entries 12-17). Deselenation of a tertiary phenyl-
selenide worked equally well as shown in entry 18. Very
recently, Carreira successfully used reagent 4 for a radical
deselenation on the way to his synthesis of Leucascandrolide
(10) Kira, S.; Sugiyama, H.; Sakurai, H. J. Am. Chem. Soc. 1983, 105, 6436.
(11) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans. 1 1975,
1574.
(9) Mendenhall, G. D. Tetrahedron Lett. 1983, 24, 451.
(12) Robins, M. J.; Wilson, J. S. J. Am. Chem. Soc. 1981, 103, 932.
9
5728 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

New Radical Chain Reducing Reagents
A R T I C L E S
Table 5. Comparison of the Rate Constants for H-Atom
Table 4. Giese-type Additions with Si-Reagents 4 (Ad )
1-Adamantyl)
Abstraction by Primary Alkyl Radicals from Silylated
Cyclohexadienes with Those of Various Alternative Donors at 70
1
2
entry
R −X
R
product
time (h)
yield (%)
°Ca
1
2
3
4
5
Ad-Br
Ad-Br
Ad-I
Ad-Br
Ad-Br
Ad-Br
t-BuI
t-BuI
t-BuI
i-PrI
EtI
CN
16
17
17
18
19
20
21
21
21
22
23
7
7
7
3
8
14
14
14
14
14
14
59
42
48
40
16
58
95
85
80
89
88
H-donor
n-Bu3SnH
n-Bu3GeH
(Me3Si)3SiH
Et3SiH
4
10-5kH/M-1 s-1
ref
CO2Me
CO2Me
COMe
SOPh
SO2Ph
SO2Ph
SO2Ph
SO2Ph
SO2Ph
SO2Ph
55
2.75
11
0.04
1.0
0.7
18a
18b
19
20
tw
6
7
8a
9b
10
11
10
tw
4e
1-Et-1-CO2H-CHDb
0.14
a tw signifies this work. b H-atom abstraction from 1-ethylcyclohexa-
2,5-diene-1-carboxylic acid by the ethyl radical.
a 0.15 equiv of AIBN were used. b 0.10 equiv of AIBN were used.
(85%, entry 8). With 0.1 equiv of AIBN a further decrease of
the yield was observed (80%, entry 9).
A.13 This application clearly proves that our method can be
applied in complex natural product syntheses.
Other common sets of synthetic reactions that utilize trialkyl-
tin hydrides are radical cyclizations and intermolecular addition
reactions. Cyclization of o-iodophenyl allyl ether (eq 2, 82%)
and the Beckwith-Dowd ring enlargement reaction14 could be
conducted with Si-reagent 4 (equation 3, 49%). The o-
iodophenyl allyl ether cyclization could also be performed with
reagent 10 (1.5 equiv, TBHN (0.6 equiv), hexane (0.2 M), reflux,
65% yield). Giese-type additions work very well with the Si-
reagent 4 (eq 4, f 16-23, Table 4). The radical precursor, the
Si-reagent (1.3 equiv), AIBN (0.3 equiv), and the olefin (1.3
equiv) were dissolved in hexane (1.2 M) and heated to reflux.
Removal of the solvent and purification yielded the inter-
molecular addition products. Importantly, we did not have any
problems with product purification. Bromoadamantane was
successfully added to acrylonitrile (f 16, 59%, entry 1). We
found that increasing the amount of acrylonitrile did not improve
the yield. Thus, upon using 5 equiv of acrylonitrile under
otherwise identical conditions a decrease of the yield was
observed (30%), and hydrosilylation of acrylonitrile occurred
in significant amounts (35%). Hydrosilylation was also a
problem in the reaction with methyl acrylate (f 17, 42%, entry
2). To suppress the hydrosilylation, we repeated the reaction
with methyl acrylate using iodoadamantane as radical precursor.
It is known that I-abstraction is about 1 order of magnitude
faster than Br-abstraction.15 Indeed, a slightly increased yield
was obtained (48%, entry 3), however, hydrosilylation could
not be completely suppressed. Reductive addition of bromo-
adamantane to methyl vinyl ketone (f 18, 40%) and phenyl
vinyl sulfoxide (f 19, 16%) occurred in moderate yields (entries
4 and 5). Phenyl vinyl sulfone turned out to be an exceptionally
good acceptor. Good to excellent yields were obtained for the
intermolecular reductive addition under standard conditions (0.3
equiv AIBN, 1.3 equiv olefin) using Ad-Br (f 20, 58%), t-BuI
(f 21, 95%), i-PrI (f 22, 89%), and EtI (f 23, 88%, entries
6, 7, 10, 11). Finally, we also tested whether the amount of
AIBN could be decreased. To this end, the reaction of t-BuI
with phenyl vinyl sulfone was conducted with 0.15 equiv of
AIBN. A slightly lower but still satisfactory yield was obtained
It is important to note that the syringe pump technique is not
necessary to conduct these intermolecular additions. Further-
more, the olefin is used in only a small excess (1.3 equiv).
Kinetics of H-Atom Transfer from Silylated Cyclohexa-
dienes. Hydrogen-transfer from our reagents to C-radicals is
expected to be much slower than for the analogous process using
the toxic Bu3SnH. We determined the rates of hydrogen transfer
from reagents 4 and 10 to a primary radical using the 5-exo-
cyclization of the 5-hexenyl radical as the radical clock.16,20
From measurements of the amounts of methylcyclopentane and
1-hexene produced in reactions at 70 °C, rate constants for
abstraction of the bisallylic H-atoms by the 5-hexen-1-yl radical
(kH) were estimated (Table 5).
Rate constants for the reductions of C-radicals with cyclo-
hexadienes have previously been determined and are in broad
agreement with the values obtained in the present study.17 It
follows that the silyl substituents in our cyclohexadienes have
no profound effect on the rate constant for the H-transfer
reaction. Table 5 shows that 4 and 10 transfer H-atoms at least
(16) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317. Newcomb, M. In
Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-VCH:
Weinheim: 2001; Vol. 1, p 317.
(13) Fettes, A.; Carreira, E. M. Angew. Chem. 2002, 114, 4272; Fettes, A.;
Carreira, E. M. Angew. Chem., Int. Ed. Engl. 2002, 41, 4098.
(14) Beckwith, A. L. J.; O′Shea, D. M.; Gerba, S.; Westwood, S. W. J. Chem.
Soc., Chem. Commun. 1987, 666. Dowd, P.; Choi, S. C. J. Am. Chem.
Soc. 1987, 109, 3493. For a review, see: Dowd, P.; Zhang, W. Chem.
ReV. 1993, 93, 2091.
(17) Hawari, J. A.; Engel, P. S.; Griller, D. Int. J. Chem. Kinet. 1985, 17, 1215.
Newcomb, M.; Park, S. U. J. Am. Chem. Soc. 1986, 108, 4132. See also
4(e) and references therein.
(18) (a) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc.
1981, 103, 7739. (b) Lusztyk, J.; Maillard, B.; Lindsay, D. A.; Ingold, K.
U. J. Am. Chem. Soc. 1983, 105, 3578.
(19) Chatgilialoglu, C.; Dickhaut, J.; Giese, B. J. Org. Chem. 1991, 56, 6399.
(20) Newcomb, M. Tetrahedron 1993, 49, 1151.
(15) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc. 1982,
104, 5123.
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5729

A R T I C L E S
Studer et al.
Table 6. Si-Reagent 4 as a Regulator for the Polymerization of
Styrene
without addition of the regulator provided PS with an Mn of
12 900 and a broader PDI (5.0, entry 2). Thus, reagent 4 did
indeed beneficially regulate the radical polymerization of styrene
leading to decreases in both Mn and PDI. The same trend was
observed for all the other styrene polymerizations (compare
entries 3 and 4, 6 and 7, 8 and 9, 10, and 11). In all the
4-mediated polymerizations a comparatively low PDI (2.0-2.5)
was obtained. The molecular weight of the desired PS could be
adjusted depending on the AIBN/4-ratio. It is important to note
that with Et3SiH as an additive, the Mn-value could also be
decreased, however, the PDI remained very high (4.7, entry 5).
We also determined the chain transfer constant Cx for regulator
4 according to Mayo at 100 °C.21d
entry
4 (mol %)
AIBN (mol %)
conversion (%)
M
n
PDI
1
2
3
1
1
1
0.5
0.5
86
>95
86
>95
91
93
93
77
87
59
87
46
39
8000
12 900
19 600
38 600
19 900
24 800
45 600
24 400
40 500
31 900
61 700
46 300
40 300
2.5
5.0
2.1
4.2
4.7
2.1
4.2
2.0
4.9
2.0
3.4
2.1
3.4
1
4
5a
6
0.5
1
1
1
1
0.388
0.388
0.275
0.275
0.163
0.163
0.05
0.05
7
8
9
10
11
12
13
a Et3SiH (0.5 mol %) was added.
1/Mn ) 1/Mo + Cx [4]/[styrene]
(2)
55 times more slowly than Bu3SnH,18a about two times more
slowly than Bu3GeH,18b and about an order of magnitude more
slowly than tris(trimethylsilyl)silane ((Me3Si)3SiH).19 However,
the kH values of 4 and 10 are significantly greater than kH for
Et3SiH20 and this explains why the silylated cyclohexadienes
can sustain efficient chains whereas Et3SiH cannot. The 2- and
6-methoxy substituents of 4 should extend the resonance
stabilization in the cyclohexadienyl radical (2) and hence would
be expected to augment the rate of H-donation. Thus, the
somewhat smaller kH value for 10 (which lacks the methoxy
substituents) in comparison with 4, is in accord with expectation.
Comparison with analogous data for the 1-ethylcyclohexa-2,5-
diene-1-carboxylic acid,4e (Table 5) shows an even smaller kH
value for this material, again suggesting that the 2,6-methoxy
substituents significantly enhance the H-donation ability of 4.
This factor may contribute considerably to the efficiency of 4
as mediator of the chain reactions.
Silylated Cyclohexadienes as Regulators for the Polym-
erization of Styrene. The transfer reaction in radical polym-
erization describes a process in which further growth of the
individual polymer molecule is prevented but which does not
interfere with the kinetic chain. For example, the polymer radical
reacts with a regulator forming a dead polymer and a new radical
which can reinitiate a new chain. If the activity of the radical
derived from the regulator is similar to the polymer radical,
then the transfer reaction will have no influence on the overall
polymerization kinetics, but the molecular weight of the polymer
produced will be decreased.21a Thiols are often used as regulators
to control the molecular weight in radical polymerizations.21a-c
We conceived that our reagents would also be suited to this
purpose because reduction/polymer chain termination with our
Si-reagents will generate a silyl radical which in turn can
reinitiate another polymer chain. We therefore carried out a
small program along this line. The polymerizations were
conducted in sealed tubes in neat styrene using varying amounts
of AIBN as initiator in the presence of varying amounts of
reagent 4 at 80 °C and were stopped after 6 h. The conversion
was determined gravimetrically. The polydispersity indices
(PDI) and the molecular weights of the polymers were analyzed
using size exclusion chromatography (SEC). As control experi-
ments the polymerizations were also conducted without Si-
reagent 4. The results are summarized in Table 6.
Cx is a dimensionless number and is defined by eq 2 (Mn )
molecular weight obtained in the presence of regulator; M0 )
molecular weight obtained in the absence of regulator). For
regulator 4 a chain transfer constant of 0.45 × 10-3 was
determined (see the Supporting Information). The Cx for 4 is
smaller than chain transfer rate constants obtained for thiol
regulators.21a-c This is not surprising because H-transfer to
C-centered radicals is much faster from thiols than from
cyclohexadienes.20 Importantly, along with the regulation of the
molecular weight, cyclohexadiene 4 is able to significantly
decrease the PDI.
Spectroscopic Study of Silyl Radical Expulsion from
Silylated Cyclohexadienes. Solutions of individual silylated
compounds in neat DTBP, or with DTBP diluted in cyclo-
propane, were photolyzed (500W Hg lamp) and radical produc-
tion was monitored by EPR spectroscopy. For the 1-TMS
compound 11 a spectrum with a dtt line structure was observed
in the temperature range 150-310 K. The hyperfine splittings
(hfs) [a(1H) ) 12.5, a(2H) ) 8.4, a(2H) ) 2.4 G at 250 K, g
) 2.0027] were very similar to those of structurally related
cyclohexadienyl radicals;4f,22 hence, the spectrum was easily be
recognized as that of a cyclohexadienyl radical of type 2. This
spectrum weakened at higher temperatures and was replaced
above ca. 350 K by a new 7-line spectrum [g ) 2.0029, a(6H)
) 20.2 G]. These EPR parameters were very similar to those
reported for the Me2C‚OSiR3 radical [R ) Me; a(6H) ) 21.0,
R ) Et; a(6H) ) 20.3, R ) Si(TMS)3; a(6H) ) 20.0 G],23,24
and therefore, we attribute the spectrum to Me2C•OSiMe3
radicals formed by addition of released Me3Si• radicals to
acetone. Acetone would be expected to build up at higher
temperatures from thermal dissociation of the initial t-BuO•
radicals.
As a further check on radical production, samples containing
11, DTBP and Me3CBr (1 equiv) were photolyzed in the
resonant cavity. The same cyclohexadienyl radical (type 2) was
(21) (a) Logemann, H. In Methoden der Org. Chemie, (Houben-Weyl) 4th ed.;
1987; vol. E 20/1, p 66. (b) Gregg, R. A.; Alderman, D. M.; Mayo, F. R.
J. Am. Chem. Soc. 1948, 70, 3740. (c) Scott, G. P.; Wang, J. C. J. Org.
Chem. 1963, 28, 1314. (d) Mayo, F. R. J. Am. Chem. Soc. 1943, 65, 2324.
(22) Berndt A. In Lando¨lt-Bornstein; Magnetic Properties of Free Radicals;
Fischer, H., Hellwege, K., Eds.; Springer: Berlin, 1977; Vol. 9b, p 452;
1987, vol. 17c, p 88.
(23) Neumann, W. P.; Schroeder, B.; Ziebarth, M. Liebigs Ann. Chem. 1975,
2279.
Using 1% of AIBN together with 1% of 4 under the
conditions depicted above afforded polystyrene (PS) with an
Mn of 8000 with a PDI of 2.5 (entry 1). The control experiment
(24) Bower, H.; McRae, J.; Symons, M. C. R. J. Chem. Soc. A 1971, 2400.
Cooper, J.; Hudson, A.; Jackson, R. A. J. Chem. Soc., Perkin Trans. 2
1973, 1933. Krusic, P. J.; Chen, K. S.; Meakin, P.; Kochi, J. K. J. Phys.
Chem. 1974, 78, 2036.
9
5730 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

New Radical Chain Reducing Reagents
A R T I C L E S
Scheme 2. Reactions of Si-centered Radicals under EPR
Conditions
2D) when n-PrBr was present in solution. Similar results were
obtained with the mono-methoxy derivative 8. No cyclohexa-
dienyl radical was detected, even at 145 K, however, the •SiMe2-
Bu-t radical was not observed in this case. Low-temperature
spectra in cyclopropane showed a broad doublet spectrum (hfs
Figure 2. EPR spectra obtained on photolysis of solutions containing 4
and DTBP. A: compound 4 and DTBP at 145 K in cyclopropane. B:
reagent 4 in neat DTBP at 315 K showing radical 26. C: reagent 4, DTBP
and t-BuBr at 320 K showing the central 6 lines of the t-Bu• radical
(including second order structure). D: reagent 4, DTBP and n-PrBr at 315
K showing the n-Pr• radical.
•
ca. 12 G) that might be due to addition of SiMe2Bu-t to 8 to
generate radical 25 (Scheme 2). That •SiMe2Bu-t radicals were
produced efficiently was shown by the observation of 26 at T
> ca. 340 K, and by the observation of the t-Bu• radical at T >
270 K.
observed up to temperatures of ca. 325 K. Above this the t-Bu•
radical started to appear and its spectrum dominated at 350 K
and above (see the Supporting Information). These observations
strongly support the mechanism of Scheme 1 and indicate that
dissociation of radical 2 (R′′ ) H) becomes rapid at 325 K and
above. A very similar sequence of spectra was obtained from
compound 10, the only difference being that the spectra at 350
K (in the absence of t-BuBr) showed Me2C•OSiMe2Bu-t, plus
additional complex spectra that were probably due to silyl radical
addition to the substrate.
It is evident from these results that 4 and 8 are very efficient
sources and that Si-radicals are released at temperatures below
270 K and probably even as low as 150 K, as judged by the
absence of cyclohexadienyl type species, and the direct observa-
•
tion of SiMe2Bu-t from 4.
EPR experiments with the benzannelated cyclohexadiene 9
showed a poorly resolved spectrum at lower temperatures that
was probably due to the benzannelated cyclohexadienyl radical.
Dissociation of this species required higher temperatures, as
judged by the appearance of the acetone adduct 26 only above
360 K. For the silylated dihydroanthracene derivatives 12 and
13, ill-resolved spectra that were probably the delocalized
cyclohexadienyl analogues were observed at lower temperatures,
but dissociation was evidently very difficult because the acetone
adduct was not detected from either compound even at 360 K.
The low-temperature spectrum from the silylated cyclohexa-
diene 15, lacking the geminal methyl group, was as expected
for the cyclohexadienyl radical 2 (R1dR2dR3d Me) [a(2H) )
2.87 G, a(6H) ) 7.93 G, a(1H) ) 12.49 G, a(1H) ) 23.13 G
at 250 K]. At higher temperatures the spectrum evolved through
several stages to a complex spectrum at 345 K that probably
consisted of bis-silylated radical 27 together with isomeric
species (see the Supporting Information). Radical 27 could have
been formed by coupling of 2 with a silyl radical and subsequent
abstraction of an H-atom, or by partial oxidation to the
corresponding silylated arene followed by addition of a silyl
radical. By either mechanism a mixture of isomers should result.
None of the acetone adduct radical 26 was detected. Radical
28, which could easily be generated from 15 by abstraction of
the bis-allylic H-atom at C(1), was not detected. However, minor
amounts could easily have been hidden under the complex
spectra of 2 and 27.
Somewhat different behavior was encountered with the mono-
8 and di-methoxy 4 derivatives. Photolysis of a solution of 4
and DTBP at 145 K gave rise to the spectrum shown in Figure
2 A. This consisted of 7 broad lines with a spacing of a(6H) )
6.4 G. The known spectrum of the Me3Si• radical has a(9H) )
6.3 G25 and therefore spectrum A can most likely be attributed
to the t-BuMe2Si• radical. The central line was somewhat
enlarged, possibly because of overlap with adduct radicals
derived from the Si-radical and substrate 4 (see below).
Even when photolysis was carried out at 100 K, with solutions
of 4 in n-propane, no sign of the corresponding cyclohexadienyl
radical 2 was obtained. Above ca. 310 K, where acetone
formation became appreciable,26 the Me2C•OSiMe2Bu-t radical
(26) (Figure 2B) was observed, confirming the release of the
silyl radical. Solutions containing 4, DTBP and alkyl bromides
were also examined. Figure 2C shows the high quality spectrum
of the t-Bu• radical (central 6 lines only) obtained in the presence
of t-BuBr (1 equiv). This spectrum started to appear at ca. 270
•
K. That the released SiMe2Bu-t radical could also abstract
bromine from primary bromides under EPR conditions was
demonstrated by the observation of the n-Pr• radical (Figure
(25) Cook, M. D.; Roberts, B. P. J. Chem. Soc. Chem. Commun. 1983, 264.
Carmichael, I. J. Phys. Chem. 1985, 89, 4727.
(26) The concentration of the Me2C•OSiMe2Bu-t radical in Figure 2B was ca.
10-7 M and hence acetone concentrations of this order of magnitude, or
greater, are required for the Si-radical adducts to be detectable.
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5731

A R T I C L E S
Studer et al.
Table 7. Temperature Ranges (K) over Which Cyclohexadienyl
Type Radicals (CHD), Acetone Adduct Radicals 26, and t-Bu•
Radicals Were Spectroscopically Observed for Various
Si-Reagents
effective because it led to rapid silyl radical release and
simultaneously was beneficial to the H-atom transfer step.
H-transfer from reagent 4 to primary C-radicals was about 10
times slower than the reduction of primary C-radicals with (Me3-
Si)3SiH. This offers some advantages, especially for Giese-type
addition reactions, where the direct reduction often competes
with the desired intermolecular addition reaction. Giese-type
additions with reagent 4 can be conducted without using a large
excess of the olefin and without using syringe pump techniques.
Similarly, reagent 4 should be effective for radical cascade
(domino) sequences where a series of rearrangement steps is
required to occur without reduction of the individual intermedi-
ates. We have also shown that Si-reagent 4 can be used as
regulator for the polymerization of styrene. Molecular weight
can be adjusted and depends on the initiator/Si-reagent-ratio.
Polymerizations mediated by reagent 4 afford polystyrenes with
rather narrow PDIs (∼2.0).
T range
•
Si-reagent
T range CHD 2
acetone adduct 26
T range t-Bu
4
8
11
10
9
13
12
15
a
>310
>340
>350
>350
>360
n.o.b
>270
>270
>325
150-310
150-310
240-340
240f360
240f360
240-340
n.o.
n.o.
a SiMe2Bu-t observed at lower temps. b n.o. signifies not observed.
Data relevant to silyl radical release from the series of
silylated reagents is collected in Table 7. These spectroscopic
observations agree unequivocally with the deductions about
reagent efficiency from the preparative experiments with 1-bro-
moadamantane. For the methoxy-substituted reagents 4 and 8,
cyclohexadienyl radicals were not observed because dissociation
Our exploratory EPR experiments showed that silylated
cyclohexadienes are good sources of Si-centered radicals for
spectroscopic work. Direct observation of silyl radicals will not
in general be possible because they react with the starting
cyclohexadienes. Triethylsilane is the classic choice for use in
conjunction with an alkyl bromide for EPR observation of
C-centered radicals at low temperatures, whereas Me3SnSnMe3
has been preferred for higher temperature work.27 Reagent 4
offers some advantages at higher temperatures (>270 K) where
Et3SiH initiated spectra often fade and when avoidance of
stannanes is desirable. In addition, some C-centered radicals
(particularly σ-radicals) abstract hydrogen from the ethyl groups
of Et3SiH, as a side reaction giving Et2Si(H)CH•CH3 radicals,
that interfere with the desired spectrum.28 This will not be a
problem with reagent 4 because the t-BuMe2Si group does not
possess secondary H-atoms adjacent to silicon. Good quality
spectra of various transient radicals can be obtained by use of
4 in conjunction with a reaction partner having a high affinity
for Si such as an alkyl bromide or a carbonyl compound. Thus,
4 and analogues will be useful in a general sense for spectro-
scopic detection, characterization and monitoring of radical
intermediates in various processes. EPR experiments with
reagents 4-15 provided good support for the proposed chain
mechanism and showed that Si-radical formation was most
efficient for the methoxy-substituted cyclohexadienes, was
difficult for the benzannelated derivative 9 and was too slow
to detect for the dihydroanthracene reagents in the accessible
temperature range. Reagents 4, 7, and 10 provide versatile and
environmentally friendly additions to the toolkit of the synthetic
chemist and are superior to current favorites in many applica-
tions.
•
and release of SiMe2Bu-t was very rapid even at T < 270 K.
This was supported by the observation of 26 and t-Bu• at
comparatively low temperatures. This supports the classification
of these two reagents as very efficient Si-radical sources. For
compounds 10 and 11 cyclohexadienyl radical formation
proceeded well but dissociation and silyl radical release became
appreciable at somewhat higher temperatures (325 K). For the
dihydroanthracene derivatives 12, 13 silyl radical release was
not observed by EPR in the accessible temperature range and
this explains the zero adamantane yields of Table 2. Although
H-abstraction from reagent 15 gave mainly the desired cyclo-
hexadienyl radical 2 the reaction was diverted into other
channels at higher temperatures.
Conclusions
We reported the syntheses of various silylated cylohexadienes.
These dienes are readily prepared on a large scale and can be
used as efficient tin hydride substitutes. Product purification is
straightforward. In contrast to tris(trimethylsilyl)silane,3b which
is probably the most successful tin hydride substitute to date,
our most promising Si-reagent 4 is a solid and can be readily
stored and handled. Most of the common reductive radical
reactions can be conducted perfectly well with the new
Si-reagents. Excellent yields are achievable with primary,
secondary, and tertiary organo-iodides and -bromides. Yields
approaching 90% can also be obtained with organo-chlorides
and thionocarbonates. Reductions of xanthates and phenyl
selenides were also realized in useful yields. The radical chain
processes were tolerant of hydroxy and carbonyl functional
groups. With this reagent, the alternative dissociation of
intermediate 2, releasing a methyl radical, was not observed in
any case and hence product contamination with silylated-arene
byproducts was nonexistent.
Experimental Section
General. Solvents were purified by standard methods. Compounds
sensitive to air and moisture were handled under argon using Schlenk
techniques. FC: Merck or Fluka silica gel 60 (40-63 µm); at ca. 0.4
bar. GC: Hewlett-Packard 5890 chromatograph using Hewlett-Packard
HP-5, Macherey-Nagel Optima δ-3, or Supelco γ-DEX 120 columns.
Melting points: Bu¨chi 510 apparatus; uncorrected. I. R. spectra:
recorded on a Perkin-Elmer 782 or Bruker IFS-200 spectrophotometer.
An advantage of silylated cyclohexadienes is that their
properties can be tuned to suit particular synthetic needs. Silyl
radicals of differing reactivity may be incorporated. The two
key chain propagation steps, i.e., silyl radical release from radical
2 and H-atom transfer, can both be controlled to some extent
by introduction of suitable substituents into the cyclohexadiene
ring. The 2,6-dimethoxy substitution of 4 was particularly
(27) Hudson, A.; Jackson, R. A. J. Chem. Soc. Chem. Commun. 1969, 1323.
(28) (a) Della, E. W.; Head, N. J.; Mallon, P.; Walton, J. C. J. Am. Chem. Soc.
1992, 114, 10730. (b) Binmore, G. T.; Walton, J. C.; Adcock, W.; Clark,
C. I.; Krstic, A. R. Magn. Reson. Chem. 1995, 33, S53.
9
5732 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

New Radical Chain Reducing Reagents
A R T I C L E S
MS: Recorded on a VG Tribid, Varian CH7 (EI); IonSpec Ultima,
Finnigan MAT TSQ 700 or a Finnigan MAT 95S (ESI) in m/z (% of
basis peak). SEC: Pump: Merck Hitachi L-6200 A. Columns: Polymer
Laboratories PL gel 5 µm GUARD and MIXED-C. Mobile phase: THF
(1 mL/min). Detector: Shodex RI-101. Molar weights and PDI were
determined using polystyrene samples of known molecular weight
distribution. For data analysis Polymer Laboratories Cirrus GPC Online
software (Version 1.0) was used. EPR spectra were obtained with a
Bruker EMX 10/12 spectrometer operating at 9.5 GHz with 100 kHz
modulation. Samples of the substrate (0.3 to 40 mg) and di-tert-butyl
peroxide (0.01 to 0.1 mL), in 4 mm od quartz tubes were photolyzed
in the resonant cavity by light from a 500 W super pressure mercury
lamp. For reactions performed in cyclopropane (up to 0.5 mL), the
solutions were degassed on a vacuum line using the freeze-pump-
thaw technique, and tubes were flame sealed. In all cases where spectra
were obtained, hfs were assigned with the aid of computer simulations
using the Bruker Simfonia software package. For quantitative measure-
ments, signals were double integrated using the Bruker WinEPR
software.
Removal of the solvent in vacuo and purification by FC (Et2O/pentane
1:10) afforded 3-deoxy-1,2;5,6-di-O-isopropylidene-R-d-glucofurano-
side (111 mg, 91%) as a colorless solid.
Giese-Type Addition using Reagent 4 (Representative Example).
1-Bromoadamantane (213 mg, 1 mmol), Si-reagent 4 (349 mg, 1.3
mmol), AIBN (49 mg, 0.3 mmol), and acrylonitrile (85 µL, 1.3 mmol)
were dissolved in hexane (3 mL) under Ar. The reaction mixture was
heated to reflux for 4 h. Removal of the solvent in vacuo and
purification by FC (Et2O/pentane 1:30) afforded 3-adamantylpropioni-
trile 16 (112 mg, 0.59 mmol, 59%) as a colorless oil.
Polymerizations (Representative Example). In a Schlenk tube,
AIBN (33 mg, 0.203 mmol) and Si-reagent 4 (141 mg, 0.524 mmol)
were dissolved in freshly distilled styrene (6 mL, 52.4 mmol). The
mixture was degassed with three freeze-thaw cycles. The Schlenk tube
was flushed with argon, sealed and heated to 80 °C for 6 h. After
cooling to room temperature, the polymer was dissolved in CH2Cl2
and dried at 60 °C (high vacuum, 12 h). Conversion was determined
gravimetrically: 93%. The polymer was characterized by SEC (average
values of three runs): Mn ) 24 800; Mw ) 51 000; PDI ) 2.1.
Reduction of Bromoadamantane with Reagent 4 (Representative
Example). 1-Bromoadamantane (107.5 mg, 0.50 mmol), Si-reagent 4
(174.5 mg, 0.65 mmol) and AIBN (24.5 mg, 0.15 mmol) were dissolved
in hexane (2.5 mL) under Ar and heated to reflux. The course of the
reaction was monitored by GC. The yield was determined by GC-
analysis using tetradecane (130 µL, 0.50 mmol) as an internal standard.
After 5.5 h, the conversion was found to be complete and the yield of
adamantane was determined to be 99%.
Barton-McCombie Deoxygenation with Reagent 4 (Representa-
tive Example). 1,2;5,6-Di-O-isopropyliden-3-O-phenoxythiocarbonyl-
R-D-glucofuranoside (197 mg, 0.50 mmol), Si-reagent 4 (174 mg, 0.65
mmol) and AIBN (25 mg, 0.15 mmol) were dissolved in hexane (2.5
mL) under Ar. The reaction mixture was heated to reflux for 18 h.
Acknowledgment. Dedicated to Prof. Reinhard W. Hoffmann
on the occasion of his 70th birthday. We thank the Swiss Science
National Foundation (2100-055280.98/1), the Fonds der Chem-
ischen Industrie and the EPSRC for funding.
Supporting Information Available: Full experimental details
(including analytical data) for the preparation of 4-15. Rep-
resentative EPR spectra obtained from compounds 8, 9, 11, 12,
and 15. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA0341743
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5733