вход по аккаунту


On the Trail of Xanthates Some New Chemistry from an Old Functional Group.

код для вставкиСкачать
On the Trail of Xanthates:
Some New Chemistry from an Old Functional Group
Samir Z. Zard
Dedicated Professor Sir Derek Barton
The use of free radical reactions in organic synthesis has witnessed an extraordinary development in recent
years. When properly conceived, radical
processes often exhibit many of the
properties desired by synthetic organic
chemists, such as flexibility, mildness,
and selectivity. Unfortunately, the number of synthetically useful radical-generating systems is still limited, and most
applications have relied on tin hydride
chemistry. Secondary U-alkyl-S-methyl
xanthates, for example, react with tributyltin hydride to give the corresponding alkane (the Barton-McCombie reaction). It was, however, found that the
isomeric 0-methyl-S-alkyl xanthates
undergo cleavage of the weaker carbonsulfur bond and that a chain reaction
can be sustained without tin or other
heavy metals. A variety of synthetically
interesting free radicals can thus be produced and captured, the last propagating step being a reversible transfer of the
xanthate group. S-Propargyl xanthates
represent a special class. Their radical
chemistry can be easily overshadowed
by hitherto unknown but equally interesting nonradical behavior. Upon heating, a thermal [3,3] sigmatropic rearrangement occurs to give the allenyl
isomer, which is in equilibrium with a
1. Introduction
Xanthates, now classed under the more systematic but less
evocative name of dithiocarbonates (or carbonodithioates) ,
have been known for almost two centuries. Xanthate salts for
example were reported as early as 1822 by Zeise;['] yet, in comparison with other functional groups, their chemistry has remained largely unexplored. A chemist's first encounter with
xanthates is usually through the Chugaev elimination'21 and
perhaps as intermediates in the manufacture of viscose from
cellulose.[31Later, budding organic chemists eventually come
across the more recent Barton- McCombie reaction[41for deoxygenating secondary alcohols. Unlike the widely known but
little used Chugaev elimination, the Barton-McCombie radical
deoxygenation has had a major impact in organic synthesis,
especially for the modification of carbohydrates, and as a convenient source of radicals from alcohols in general.[51A chance
[*] Dr. S. Z. Zard
Laboratoire de Synthese Organique associe au CNRS
Ecole Polytechnique, F-91128 Palaiseau Cedex (France)
Fax: Int. code + ( I ) 6933-3010
Institut de Chimie des Substances Naturelles
F-91198 Gif-Sur-Yvette (France)
Angew. Chem. Int. Ed. Engl. 1997, 36, 612-685
novel betaine structure. This species is at
the heart of a number of new transformations, including formal [3 + 21 cycloadditions, formation of esters with
inversion in the case of secondary alcohols, conversion into 1,3-dithiol-2-ones,
generation of cisoid dienes, carbon-carbon bond formation through reaction
with acid chlorides etc. This account
provides a brief description of this original radical and nonradical chemistry of
xanthates, an old family of compounds
that still harbors many mysteries.
Keywords: alkynes * C-C coupling
cycloadditions radicals * xanthates
observation made in the course of a mechanistic study of this
important reaction started our exploration in this area. Our
peregrinations were hardly planned from the outset; we simply
went where the chemistry took us. The main purpose of this
review article is to relate our findings in a brief, candid manner.
2. Radical Reactions of Xanthates
2.1 The Story of a Discovery
The Barton-McCombie reaction was conceived on the basis
of the mechanistic reasoning shown in Scheme
The wellknown affinity of tin for sulfur was used to advantage by treating stannyl radicals with the thiocarbonyl group of a xanthate
such as 1, with the expectation that the intermediate adduct 2
would undergo a preferential /3-scission of the carbon-oxygen
bond to give radical R . Hydrogen abstraction from tributyltin
hydride provides the desired alkane and a tributylstannyl rddical to propagate the chain. The logical co-product 3 is unstable
with respect to elimination of carbon oxysulfide. In practice the
process works best with xanthates derived from secondary alcohols. Deoxygenation of primary alcohols usually requires higher
reaction temperatures and yields are not always satisfactory,
0 VCH Veriugsgeseiischuft mhH. 0-69451 Weinhrim,1YY7
$17.50+ SOi0
S. Z. Zard
S-methyl and S-2-propyl xanthates derived from cholestanol(5
and 6) to compete for one equivalent of stannane (Scheme 3): If
attack occurs on the sulfide sulfur atom, steric hindrance should
slow the reaction of the isopropyl derivative. If the thiocarbonyl
group was the reacting center, little difference in the rates would
be observed. Unexpectedly, the outcome of the actual experiment was neither of these two possibilities. As it turned out, the
isopropyl analogue 6 reacted much faster than the methyl
derivative 5, which was recovered largely unchanged. And the
product was not cholestane, as would be anticipated from a
Barton-McCombie reaction, but propane!
Scheme 1. The original Barton- McCombie mechanism.
whereas xanthates of tertiary alcohols are especially prone to
Chugaev elimination and are thus difficult to handle in general.[41
Almost a decade later, the mechanistic picture put forward
by Barton and McCombie was questioned by Beckwith and
Barker,[61who observed an ESR signal corresponding to the
hitherto unknown alkoxythiocarbonyl radical 4 when a xanthate was irradiated in the presence of hexabutyldistannane.
Although these conditions are not those normally applied in the
deoxygenation, this finding raised the possibility of an alternative route (Scheme 2) whereby the stannyl radical, in fact, reacts
- J.
Scheme 3. The key competition experiment.
Two conclusions, summarized in Scheme 4, could be drawn
from this experiment : a) the addition of the tin radical on the
Scheme 2. The Beckwith-Barker mechanism.
with the sulfide sulfur atom. The alkoxythiocarbonyl radical 4
created by this S,2 process was assumed to extrude carbon
oxysulfide by analogy with its oxygen counterpart, which was
known to lose carbon dioxide.
The paper of Beckwith and Barker triggered a more systematic study in the Barton group, and the results strongly supported
a slightly modified version of original me~hanisrn.~~]
on the thiocarbonyl group rather than on the sulfide sulfur atom
was further confirmed by an elegant trapping experiment by
Bachi and Bosch.[*]In retrospect, one of the key experiments
in the Barton studyf7]was to allow an equimolar mixture of
I slow
I fast
Scheme 4. Mechanistic rationale for the competition experiment
Samir Z. Zard was born in 1955 in r f , Nigeria. His training as a chemist started at the
American University of Beirut, then at Imperial College, London, and finally at the Universitk
Paris-Sud, France, where he received his doctorate under the supervision of Professor Sir Derek
Barton. In 1981 he began working at the Centre National de la Recherche Scientifique
(C.N.R.S.), but at the same time remained member of the group of Professor Barton at the
Institut de Chimie des Substances Naturelles in Gif-sur- Yvette. When Professor Barton moved
to the Texas A & M University in far distant College Station, Zard took up an additional
part-time post as lecturer at the Ecole Polytechnique in Palaiseau-only 6 k m from Gif. His
main research interests concern the study and development of new reactions and processes.
Angew,. Chem. Int. Ed. Engl. 1997,36, 672-685
thiocarbonyl group is fast and reversible, and b) the rate-determining step is p-scission of the carbon-oxygen or the carbonsulfur bond, the latter being the preferred mode (at least kinetically) when radicals of similar stability are produced (in this
case, both the isopropyl and 3-cholestanyl radicals are secondary). Such a reversible addition is not limited to stannyl
radicals: a similar reversible addition of carbon radicals on the
thiocarbonyl group of thiohydroxamate esters had to be invoked a little later in connection with the Barton decarboxylation reaction.[g'
2.2. A New Radical-Generating System
Considering these facts, we pursued the possibility of using
the relatively easy rupture of a carbon-sulfur bond in xanthates
in a much more general radical-chain process. The manifold of
mechanisms underlying our work is set out in Scheme 5. It basically represents the behavior of a xanthate such as 7 that cannot
- IR'I-dE
Scheme 5 General manifold of reactions for the generation and capture of radicals
from xanthates
undergo cleavage of the carbon-oxygen bond in the BartonMcCombie mode. Following a chemical or photochemical initiation step, a radical R thus generated has the choice of either
reacting with the starting xanthate (path A) or with a given trap,
say an olefin, placed in the medium (path B). The former possibility leads to an adduct 8 for which 8-scission of the carbonoxygen is very unfavorable since it would produce a methyl
radical, which is thermodynamically less stable than R'. Rupture of either of the carbon-sulfur bonds, on the other hand,
leads back to R' and the starting xanthate 7. The same will apply
if the methyl group on the oxygen is replaced by another group,
as long as the stability of its corresponding radical is comparable
to, but preferably lower than that of R' (in most applications a
primary substituent is sufficient).
Angeu. Chem. Int. Ed. Engl. 1997, 36, 672-685
In other words, the reaction of radical R' with xanthate 7 is
degenerate. This path does not consume R' and therefore is not
in competition with the olefinic trap (that is, paths A and B
effectively do not compete). This point is at the heart of the
whole system. Because R is not irreversibly quenched by its
precursor, its effective lifetime in the medium becomes longer,
allowing it to be captured by unreactive traps that cannot normally be employed with other radical-generating systems (e.g.,
tin or mercury hydrides). This property, also shared by the
so-called atom-transfer or Kharasch reactions (e.g., bromine or
iodine atom transfer)
opens many synthetic possibilities for
constructing frameworks hitherto accessible only with difficulty
or not at all. Thus, capture of R leads to an adduct radical 9,
which in turn can react with the starting xanthate 7 to provide,
after two reversible steps, another xanthate 10; in the process a
new carbon-carbon bond and a new carbon-sulfur bond have
been created.
In fact, and to our delight, xanthates turned out to be a
convenient source of a variety of radicals that can be handled
according to the general Scheme 5-and a source not only of
radicals, as will become apparent later.
At this stage several comments on the radical aspect can be
made in anticipation of the results that will be shortly described.
No heavy or toxic metals are involved in such a process, and the
starting materials are cheap and readily available (for example,
0-ethyl-S-alkyl xanthates can be made by nucleophilic displacement of an alkyl halide, tosylate, methylsulfonate etc. with
potassium U-ethyl xanthate, which in turn is made by mixing
potassium hydroxide and carbon disulfide in ethanol) .[31
Furthermore, the end product is also a xanthate and can be
employed as a starting point for another radical sequence or
modified further by using the immensely rich chemistry of sulfur. However, looked at from a different angle, this fact at the
same time constitutes a limitation, since the reversibility of the
xanthate group transfer means that the last two propagating
steps (9+ 10) represent an equilibrium that is normally easier to
bias in the forward direction if R' is more stable than the adduct
radical 9. This point has to be kept constantly in mind when
designing a synthetic sequence.
The feasibility of this new system for generating and capturing radicals was easily demonstrated with a variety of combinations of xanthates and inter- and intramolecular olefinic
traps." The experimental procedure is usually exceedingly
simple: mere heating of the xanthate and trap in an inert solvent
(benzene, cyclohexane, dichloroethane, toluene, chlorobenzene,
etc.) under an inert atmosphere and in the presence of a catalytic
amount of an appropriate initiator (dibenzoyl or dilauroyl peroxide, di-tert-butyl peroxide etc., depending on the reaction
temperature, which i s usually the boiling point of the solvent;
azobisobutyrylnitrile (AIBN) is not generally suitable since it
gives rise to isobutyryl radicals that are too stable to trigger the
process) .[' A selection of examples illustrating the intermolecular case is pictured in Scheme 6. The array of functionality that
can be obtained is quite amazing and has interesting synthetic
implications. For instance, the derivative arising from addition
to phenyl vinyl sulfone can act as a precursor of a thioaldehyde
upon hydrolytic cleavage of the xanthate group,[131whereas the
chemistry of compounds formed from reaction with allyl- and
vinyltrimethylsilane remains to be explored. Most of the olefinic
S. Z. Zard
example given, cyclization is followed by opening of one
cyclopropane ring) and in do lone^['^] (by cyclization onto
an aromatic ring, which takes place slowly if at all) can be
produced. For the last example, taken from a recent application of the reaction by the group of Speckamp and
Hiemstra in Amsterdam,[’51 none of the usual radical
methods was successful in accomplishing the desired cyclization.
E = S i M e s 51%
E = S O z P h 52%
2.3. A Study in Yellow
As mentioned above, radicals of various types can be
generated with the appropriate xanthates. Thus acyl radicals can be made from S-acyl xanthates, such as 0-ethyl
derivative 11 (Scheme 8). These substances are yellow in
color (the word xanthate comes from the Greek xanthos
meaning yellow, since the xanthate salts made in the early
days were yellow)[161and have been known for almost a
~entury.“’~In fact, their sensitivity to visible light
Scheme 6. Examples of intermolecular radical additions with xanthate transfer. (The reactions were run in cyclohexane with dilauroyl peroxide as initiator; neoPn = neopentyl.)
traps shown are not “activated” and therefore cannot normally
be used with many other radical methods such as those based on
tin. Another practical advantage, of some importance when
operating on a large scale, is that the reactions can be run in a
fairly concentrated medium, typically 0.5 M (and sometimes even
without solvent), which is between tenfold and more than a
hundredfold more concentrated than the medium for most other radical processes. Last (but not least), some radicals, such as
the trifluoroacetonyl radical as in the second example
(Scheme 6), are not readily available by other methods.
The intramolecular variant is even easier to perform and rings
of various sizes and substitution patterns can be obtained
(Scheme 7).[”. 14,
It is noteworthy that p-lactams (in the
cat. dilauroyl
cat. tBuOOtBu
cat. tBuOOtBu
Scheme 7. Examples of radical cyclizations with xanthate transfer
Scheme 8. Formation of an S-acyl xanthate and its decomposition by an ionic chain
(leading to acyl radicals) was established by Barton, George,
and Tomoeda[’*] nearly thirty years before our own study, but
at the time a radical recombination mechanism was invoked
instead of a chain process (Scheme 5, with R replaced by RCO).
The preparation of S-acyl xanthates is, in principle, quite simple
and merely involves treating an acid chloride with a xanthate
salt such as potassium ethyl xanthate. However, there was a
hidden twist: like the chemists before us, we found that the
results were not very reproducible.[”~ 18] Variable amounts of
the thioanhydride 12 and 0,O-diethyl xanthic anhydride 13
were formed in addition to the desired S-acyl xanthate 11. The
explanation is that the S-acyl xanthate can decompose completely by an ionic chain mechanism under the influence of a
small amount of the xanthate salt (Scheme 8). Indeed, addition
of a small quantity of potassium ethyl xanthate to a pure sample
of S-benzoyl 0-ethyl xanthate caused its complete conversion
into thiobenzoic anhydride (12, R = Ph) and 0,O-diethylxanthic anhydride 13.[”]We therefore modified the experimental procedure so as to avoid an excess of the xanthate salt at any
given time (the salt is added to a slight excess of the acid chloride
in the dark).
Acyl radicals can be generated by irradiating the corresponding S-acyl xanthater”] with visible light (or by the usual chemical initiation) and captured in an inter- or intramolecular
fashion by using appropriate olefinic traps (Scheme 9). The
Angew. Chem. Int Ed. Engl. 1997, 36, 672-685
hv (vis)
X = 0,70%
x = s. 93%
(X =O)
Scheme 9. Examples of radical reactions with S-acyl xanthates involving radical
capture and xanthate transfer in the final step. Cu = copper powder, vis = visible
xanthate function in the adduct is in the /$position with respect
to the carbonyl group and may be easily eliminated with base o r
by heating with copper powder (the xanthate group may also be
reductively removed with tributylstannane) . In the case of tertiary (or benzylic) derivatives, loss of carbon monoxide occurs
from the acyl radical, and it is the tertiary (or benzylic) radical
that is finally trapped, as shown by the last example in Scheme 9.
One of the key assumptions in the Beckwith- Barker mechanism (Scheme 2) for the Barton - McCombie deoxygenation
concerns the ability of an alkoxythiocarbonyl radical, ROk=S
(4), to extrude carbon oxysulfide. Loss of carbon dioxide from
the oxygen analogue ROC=O was known to be comparatively
slow ( k at 25 ' C is of the order of lo5 s- for R = tert-butyl and
much lower when R is secondary or primary),[201and it was not
clear at the outset what effect the sulfur would have on the rate
of fragmentation. The main problem in answering this question
was how to generate an alkoxythiocarbonyl radical cleanly and
unambiguously in order to study its behavior. For example, a
selenocarbonate precursor can be quite suitable for creating a n
alkoxycarbonyl radical,[211but the corresponding selenothiocarbonate would not constitute an unambiguous source of an
alkoxythiocarbonyl radical, because the thiocarbonyl group is
at least as reactive, and probably more so, than the selenide
group (Scheme 10).
A simple solution to this problem turned up in the guise of a
xanthic anhydride, also a long-known member of the xanthate
family.[221Our first encounter with such a compound was with
the 0,O-diethyl derivative 13, which appeared as an undesirable
side product in the preparation of S-acyl xanthates (Scheme 8).
The reaction of xanthic anhydride 14 with stannyl radicals can
take place by two possible pathways: a) attack at the thiono
sulfur atom or b) at the sulfide sulfur. But, because of the symAngun.. Chmii. In1 Ed. EngI. 1997. 36. 612-685
Scheme 10. Comparison of the conventional preparation of alkoxythiocarbonyl
with two possible reaction paths and the route alkoxycarbonyl radicals.
metry inherent in the structure of the xanthic anhydride, both
processes lead to a n identical outcome, namely an alkoxythiocarbonyl radical and a tin xanthate (Scheme 11); it does not
therefore matter which sulfur atom the tin radicals attack. Using
this approach, we could show that in the case of the cholestanyl
derivative extrusion of carbon oxysulfide proceeds readily.[231
The assumption of Beckwith and Barker16] was thus justified,
even though their mechanistic picture for the Barton- McCombie deoxygenation is not correct.
Scheme 11. Unambiguous generation of alkoxythiocarbonyl radicals from xanthic
But over and beyond revealing the propensity of alkoxythiocarbonyl radicals to expel carbon oxysulfide (competition experiments indicated this to be a t least an order of magnitude
faster than loss of carbon dioxide from alkoxycarbonyl radicals), xanthic anhydrides 14 have an interesting chemistry of
their own. In the absence of stannane, they undergo a clean
radical chain transformation into a symmetrically substituted
xanthate upon irradiation with visible light (like S-acyl xanthates, these substances are yellow) or on heating in the presence
of a chemical initiator (Scheme 12). Since such anhydrides are
easily made from alcohols via the xanthate salt, the process
constitutes a method for generating radicals from primary or
secondary alcohols (xanthic anhydrides derived from tertiary
alcohols are generally not stable with respect to a Chugaev-type
elimination) and a way to replace homolytically a carbon- oxygen bond with a carbon-sulfur bond, as exemplified by the
conversion of the xanthic anhydride derived from glucose
(Scheme 12).[231
But this same symmetry which allowed us to answer a mechanistic question introduced a limitation from the synthetic view
point. Only half of the R - 0 bonds in the xanthic anhydride are
S. Z. Zard
hv R(vis)
O S A O E t
- cos
hv (vis)
Scheme 12. Radical chain reactions of xanthic anhydrides.
~ O i S A O E t
exchanged for R-S bonds. Attempts to use unsymmetrical xanthic anhydrides R-0-CS-S-CS-OR,
where the O R unit is
derived from a cheap alcohol, resulted only in frustration. The
two thiocarbonyl groups had more or less the same reactivity
and no selectivity could be achieved. The way out of this difficulty was in fact quite simple. All we had to do was to replace
the thiocarbonyl bound directly to the OR group of the unsymmetrical xanthic anhydride with the much less radicophilic carbonyl group. This gives what may be called an S-alkoxycarbonyl xanthate, R-0-CO-S-CS-OR,
such as 15. The
simplest members of this family have been known for a long
time,[241but again almost nothing about their radical chemistry
had been reported. By analogy with the previous xanthates,
irradiation with visible light of these (also yellow) substances
should lead to alkoxycarbonyl radicals. This concept is outlined
in Scheme 13. We chose R = Et as example for clarity and also
because essentially all the derivatives we examined were prepared by reaction of a chloroformate with potassium ethyl xanthate.Iz5]As stated above, alkoxycarbonyl radicals were known
to extrude carbon dioxide relatively slowly, especially in the case
of primary and secondary derivatives. However, unlike earlier
methods (for example by the reaction of a selenocarbonate with
tributyltin hydride;'''] see scheme lo), our system cannot react
as intended unless the alkoxycarbonyl radical loses carbon dioxide (Scheme 13, path B) or is trapped in some other way, since
its reaction with its xanthate precursor 15 is reversible and degenerate (path A). Thus, in the absence of a trap, the S-alkoxycarbonyl xanthate is converted by a chain reaction into an ordinary xanthate, in which the R - 0 bond has been replaced by an
R-S bond. The applicability of this process to tertiary, secondary, and even primary derivatives is demonstrated by the
examples in Scheme 13. Alteratively, the initial alkoxycarbonyl
radical may be captured by a suitably located, internal double
bond in what constitutes a simple approach to lactones. The
cyclization step must of course be faster than loss of carbon
dioxide. One application of this variation (Scheme 14) represents the final steps in a synthesis of cinnamolide; base-induced
p-elimination of the xanthate group introduces the required
Scheme 13. Radical chain reactions of S-alkoxycarbonyl xanthates.
Scheme 14 Capture of an alkoxycarbonyl radical by an internal double bond as
exemplified in the synthesis of cinnamolide.
2.4. A Brief Return to Tin Chemistry
In addition to the various carbon-centered radicals available
through the appropriate xanthate precursors, we found that
even stannyl radicals can be generated in a similar fashion by
Angew. Chem. Int. Ed. Engl. 1997, 36, 612-685
using a stannyl xanthate reagent.12’] This variation allowed us
to combine the highly useful reactivity of tin radicals with the
advantages that accrue from the introduction of a xanthate
group and the reversibility of its transfer. After some experimentation, we selected 0-neopentyl-S-triphenylstannylxanthate 16
as the reagent of choice. This nicely crystalline compound is
easily made and stored. In its simplest form, the process allows
the replacement of an X group with a xanthate (Scheme 15),
where X represents a group that can be abstracted by tin radicals
(bromide, iodide, another xanthate etc.). The process is best
triggered by addition of a few mol percent of hexabutylditin and
irradiation with visible light.
hv ( i s ) reflux
cat. Bu3Sn-SnBu3
a :66%
Scheme 15. Generation of triphenylstannyl radicals from S-triphenylstannyl
xanthate (16).
The first two examplesf2*]displayed in Scheme 16 illustrate
the transformation of even hindered bromides into the corresponding xanthates. Not only would a traditional nucleophilic
substitution be difficult to perform in these cases, but since
radicals are involved, the reaction can also take place with inversion or retention depending on the steric environment in the
substrate. The third example represents a modification of the
Barton-McCombie reaction where, instead of the usual reduction, the C - 0 bond in the xanthate group of the substrate is
exchanged for a C-S bond in a new xanthate derived from
the reagent.[271This is yet another homolytic route for modifying alcohols that may be compared to those presented in
Schemes 12 and 14.
As with traditional tin chemistry, a cyclization step can be
incorporated, so that a C-C bond is formed in addition to the
xanthate transfer, as illustrated by the last example in
Scheme 16.[’*’ Thus one obtains products of xanthate transfer
without actually starting with a xanthate as we have done so far;
it is perhaps worth pointing out that in these instances the
corresponding initial xanthates would have been quite
difficult to prepare, whereas access to the starting halides is
fairly trivial.
3. The Case of 5’-Propargyl Xanthates
3.1. Catching Visible Light
Ordinary alkyl xanthates are not normally sensitive to visible
light, in contrast to the yellow S-acyl xanthates. However, we
found that if a small amount of S-benzoyl xanthate is added to
the reaction mixture, xanthate transfer from alkyl xanthates can
indeed be effected with visible light, a process which otherwise
Ango?. Chem. Inl. Ed. Engl. 1997. 36. 672-685
/3 :4%
Scheme 16. Examples of transformations with S-triphenylstannyl xanthate (yield in
parenthesis is based on recovered starting material; identical conditions for all
reactions: Tr = triphenylmethyl)
would have required UV irradiation (or initiation by a peroxide). The principle of this contrivance is exposed in Scheme 17,
along with a couple of examples;[’O~
291 the S-benzoyl xanthate
is regenerated in the process and is indeed catalytic. This could
also be applied to a multiple radical addition sequence with two
olefinic traps of different polarities, as shown in Scheme 18. A
p-fluorobenzyl radical is produced from the corresponding xanthate and captured with N-benzyl maleimide to give adduct 17,
itself a xanthate and thus a starting point for another addition
to ally1 acetate to finally yield 18.f291Since all the xanthate
transfer steps are reversible, it is not actually necessary to isolate
intermediate 17. All the ingredients can be mixed and irradiated
with visible light to give the end product 18 (a rnulticomponent
reaction in the now popular jargon of combinatorial chemistry!). The reaction stops at 18 (thus avoiding polymerization)
because it is more difficult to form an unstabilized secondary
radical from 18 than it is to generate either a benzyl radical from
the benzyl xanthate or a carbonyl-stabilized radical from the
first adduct 17.
S. Z. Zard
capture and
xanthate transfer
Scheme 19 Unexpected reaction of an S-propargyl xanthate with an electrophilic
The simplest and most natural explanation for the
serendipitous formation of 21 in the context of our work
seemed to be a hitherto unprecedented 5-endo-digonal
cyclization of the intermediate radical 22 (Scheme 19).
This rationale was quite satisfying in that it seemed to
demonstrate in a spectacular way the long life of radicals
PhCOSCSOEt (cat.) EtO
‘ T O M e
generated by the xanthate method, allowing them to
hv (vis)
indulge in the difficult acrobatics required for a 5-endoS
digonal cyclization. Alas, in the event our wishful specuScheme 17. Use of S-benzovl xdnthate in catalvtic amounts to generate radicals from
lations were not borne out by experiment, and we had to
S-alkyl xanthates with visible light.
suffer what Thomas Huxley described succintly a century ago as “the tragedy of science-the slaying of a beautiful
hypothesis by an ugly fact”.
To adduce evidence for such a pathway, we attempted to
PhCOSCSOEt (cat )
hv (VIS)
generate radical 22 from the “normal” product 20 to show that
MeOCSS6 N C H z P b
it can indeed cyclize. But try as we might, we could not convert
xanthate 20 into its cyclopentene isomer 21. Our hypothesis that
* F O . .
cyclopentene 21 is the result of a 5-endo-digonal radical cyclization was finally completely shattered when, among other experJ OAC
iments, we found that neither light nor an S-benzoyl xanthate
PhCOSCSOEt (cat.)
PhCOSCSOEt (cat )
were really necessary for the reaction to occur. Mere heating of
hv (VIS)
hv (ns)
the propargyl xanthate and the electrophilic olefin in toluene or
chlorobenzene gave the observed cyclopentene derivative 21.
Under these conditions, of course, the minor radical adduct 20
Scheme 18 Multiple addition of a xanthate to two oletinic traps of different polaris not formed.
PhCOSCSOEt (cat.)
hv ( i s )
3.3. A New Betaine Rears Its Head
3.2. Serendipity Yet Again
Because the addition of the benzoyl radical to a xanthate
group is reversible, the radicals that can be generated by the
above “trick” must be of comparable or greater stability than
the benzoyl radical itself. The procedure is hence best suited for
producing resonance-stabilized radicals (benzyl, allyl, or those
located next to a carbonyl group, etc.). With these limitations in
mind, we set out to apply this approach to create propargylic
radicals. These stabilized radicals had hardly been studied in the
past and, as far as we could tell, had never been employed in
synthesis. In the event, and under the same experimental conditions, the simplest S-propargyl xanthate 19 did produce the
expected adduct 20 from N-benzyl maleimide but only in a dismal yield (Scheme 19). To our utter surprise, the major product
turned out to be cyclopentene 21, in yields of up to 50%.~301
Eventually, another mechanistic hypothesis presented itself
(Scheme 20). A thermal sigmatropic rearrangement of the
Scheme 20. Mechanism for the formal (3 +2] cycloaddition process
Angen. Chem. Int. Ed. EngI. 1997. 36, 672-685
the counteranion. It is thus ideally placed for performing a substitution to give an ester with inversion of configuration
(Scheme 22).[341
S-propargyl xanthate affords an S-allenyl isomer, itself in equilibrium with a betaine 23 of a novel type. The negatively charged
end of this species appears to be sufficiently nucleophilic to
promote a Michael type addition to an electron-poor olefin
followed by cy~lization.[~’~
Evidence for a stepwise rather than
a concerted mechanism will be presented in Section 3.4.
Unfortunately, from a synthetic perspective this approach to
cyclopentenes was limited to highly reactive olefins. But the
proposed mechanism raised intriguing questions about the existence and reactivity of such a betaine intermediate, to which we
then turned our attention. One approach involved capturing the
betaine with the simplest of electrophiles, the proton. It seemed
to us that if the S-propargyl xanthate was heated in the presence
of a sufficiently weak proton source (perhaps a carboxylic acid)
that was incapable of reacting either with the starting propargyl
xanthate or with the allenyl xanthate but could protonate the
more basic betaine, then a simple examination of the end products would provide evidence for such a species. Indeed, when
xanthate 24 was heated in toluene in the presence of benzoic
acid, a good yield of 1,3-dithiol-2-one 25 was produced
along with methyl benzoate (Scheme 21).r311The CH, group
I ?-
I s ’
gs7 \
A O - R
Scheme 22. Mechanism of ester formation from an S-propargyl xanthate. The carboxylic acid may in fact be replaced by any acid H-X of comparable acidity to give
a compound R-X, again with inversion at R.
In practice, this turns out to be a highly effective and simple
method for making esters by mere heating of a carboxylic acid
with the appropriate S-propargyl xanthate. The examples displayed in Scheme 23 show the formation of various esters, some
of which are quite complex. Moreover, the process is tolerant of
many of the functional groups commonly encountered in natural products. The last example involves inversion of the 3-position in cholestanol as well as providing a protected galacturonate ester. We thus have in hand a reasonably efficient, cheap,
and convenient method for the inversion of secondary alcohols
/ - J
Scheme 21. Capture of the betaine of type 23 with a weak acid.
that initially carried the xanthate group in the substrate 24 was
converted into a methyl group in the product 25. Direct protonation of the starting propargylic xanthate (which can be
achieved by the much stronger trifluoroacetic acid)[321would
have given the isomeric 1,3-dithiol-2-one 26, whereas reaction at
the allene level should have led to a six-membered ring derivative (27). Neither of these two possible compounds was formed
to any significant extent, so the product obtained can only be the
result of protonation of the betaine as shown. It may be mentioned in this context that another synthesis of 1,3-dithiol-2ones by a radical process based on bis(xanthate)s has recently
been devised by Y. G a r e a ~ . [ ~ ~ ’
The transformation just described was initially designed as a
probe for the existence of the betaine, but as a method for
making esters (methyl benzoate in the example just discussed)
the whole process acquires a completely different dimension.
Upon protonation the heterocyclic unit, in fact, becomes a leaving group, and the desired nucleophile (the carboxylate) is
Angew Chem. Int. Ed. EnRI. 1997,36, 672-685
R = Me (98%)
R = Me (95%)
PhCH,, reflux
Scheme 23. Ester formation on use of S-propargyl xanthates (methyl esters were
obtained by heating the carboxylic acid with 0-methyl-S-propargyl xanthate or
with O-methyl-S-(l-methyl-2-propynyl)
xanthate in refluxing toluene or chlorobenzene.
S. Z. Zard
that complements the popular Mitsunobu reaction,[351a process
that necessitates the use of a hazardous and expensive azodicarboxylate ester in conjunction with a phosphane, usually
triphenylphosphane. Both of these reagents are undesirable for
large scale work.
The reaction is, however, not limited to ester formation, since
the protonation step may, in principle, be effected by any weak
proton donor. Halides can therefore be prepared if the corresponding strong haloacid is incorporated in a “subdued” form
such as an ammonium or pyridinium salt. For instance, heating
0-8-cholestanyl-S-propargyl xanthate with triethylammonium
hydrogenfluoride gives 3a-fluorocholestane in 59 ‘/a yield, the
remainder being 2-cholestene (Scheme 24) .[341 Incidentally, and
PhCH,, reflux
PhCH,, reflux
PhCH,, reflux
Scheme 25. Reaction of S-propargyl xanthates with active methylene compounds.
3.4. More Nonradical Chemistry from
S-Propargyl Xanthates
PhCH,. reflux
The intermediate betaine is sufficiently nucleophilic to be captured by an acid chloride. This creates a new carbon-carbon
bond in a very simple manner and leads to ketones of unusual
structures (Scheme 26)
Aliphatic or aromatic acid chlorides
may be used but it is important, at least on a small scale, to add
an acid scavenger such as cc-pinene to remove any adventitious
hydrogen chloride, which otherwise would prematurely protonate the betaine.
Scheme 24. Preparation of halides (e.g. fluorides) and alkylation of nitrogen heterocycles containing an acidic hydrogen.
simultaneously with our work, the group of W. B. Motherchlorobyene
well found that fluorides could also be made from S-methyl
xanthates by oxidation with 4-methyl(difl~oroiodo)benzene.[~~~
(- MeCI)
Nitrogen heterocycles containing sufficiently acidic protons can be alkylated in the same way by a propargylic
xanthate as illustrated by the last two examples in
Scheme 24 involving the methylation of 5-phenyltetrazole and the nearly quantitative conversion of theobromine
into caffeine upon heating with 0-methyl-S-propargyl xanScheme 26. Reaction of S-propargyl xanthates with acid chlorides
We observed another interesting transformation when we attempted to protonate the betaine with a carbon acid of the active
But perhaps the most interesting variation in this nonradical
methylene type. Under the same conditions, a condensation
chemistry of S-propargyl xanthates is to have a leaving group
reaction took place to give an alkylidene dithiole through attack
on the ring carbon and loss of an alcohol molecule (Scheme 25).
within the xanthate itself. Starting from an alkynediol such as
Symmetrical carbon acids were used to avoid structural compli2-butyne-I ,4-diol, it is very easy to prepare propargyl derivative
cations from geometrical isomers. Yields were quite variable
28 containing a xanthate and a benzoyloxy group on either side
and depended strongly on the nature of the xanthate and active
of the triple bond. The betaine produced from such a compound
methylene ~ o m p o n e n t . ~ ~ ~ ]
can expel a benzoate and collapse into the highly reactive, rigid682
Angew Chem. hi.Ed. Engl. 1997,36,672-685
ly cisoid, and at the time novel diene 29 (and methyl benzoate as
the co-product) .[401 This diene (and its more substituted congeners) can be intercepted by Diels-Alder cycloaddition with a
variety of dienophiles including acetylenes and quinones, as
shown by the examples in Scheme 27. The adducts may be aro-
Scheme 28. Formation and intramolecular capture of a diene
2 , 6 - d i m e t h f l \ ~ s ~ o
gives an open chain product 31 (Scheme 29). The
Michael addition is therefore followed by elimination of the
acetate group and not by cyclization to give cyclopentene
derivative 32 (cf. Scheme 19) .[431 This observation provides
compelling evidence against concertedness during the formation
of the cyclopentene ring and exposes yet another aspect of the
reactivity of the intermediate betaine.
Scheme 27. Formation of dienes from S-propargyl xanthates.
Scheme 29. Addition-elimination at a betaine derived from an S-propargyl xanthate.
matized through oxidation, sometimes simply by heating under
air in 2-propanol containing triethylamine. Highly substituted
aromatic derivatives with two sulfur atoms (in protected form)
in an ortho disposition are thus made readily accessible. Such
compounds are usually difficult to construct by conventional
An intramolecular variant of this reaction, in which only one
diastereomer is produced, is illustrated in Scheme 28.c4l1 In a
sense, the propargylic xanthate can be viewed as a loaded spring
that, upon heating, reveals a reactive diene, ready to pounce on
the nearest olefinic trap. It is worth underlining the simplicity
with which the precursor is assembled and the complexity of the
resulting product, thanks to the powerful Diels-Alder cycloaddition step.
It may be recalled (Schemes 19 and 20) that heating an Spropargyl xanthate with an electrophilic olefin gives rise to a
cyclopentene derivative. If a leaving group is now incorporated
into the olefin, the course of the reaction is completely modified.
For instance, heating propargylic xanthate 19 with unsaturated
nitrile 30, a compound easily accessible by the Baylis-Hillman
4. Back to Radicals
But what about propargyl radicals? While trying to generate
these species we unwittingly stumbled across this unusual, nonradical chemistry of propargylic xanthates. The manifold of
radical reactions was in fact overshadowed by the thermal sigmatropic rearrangement leading to the allene and then to the
betaine. However, during our study (and in hindsight not surprisingly), we noticed that the sigmatropic rearrangement was
significantly slower, and even quite sluggish, at or below 80 "C,
when the alkyne terminus bore substituents. In the light of these
considerations, it seemed reasonable that with substituted
propargylic xanthates and under appropriate reaction conditions, namely initiation at around 80 "C, the formation and
trapping of propargyl radicals could perhaps overtake the
rearrangement to the isomeric allene. Indeed, heating the
trimethylsilyl derivative in the presence of N-benzyl maleimide
and a small amount of dilauroyl peroxide under reflux in cyclohexane gave a high yield of the "normal" radical
adduct 34
Angew. Chem. Inr Ed Engl. 1997, 36,612-685
S. Z. Zard
(Scheme 30).[441For comparison, in Scheme 19 with the same
trap, the unsubstituted analogue 19 afforded mainly cyclopentene derivative 22 and only little of the radical addition product
20. Work in progress indicates that even a methyl group is suflicient to allow useful radical chemistry, and that the intramolecular variant of this process is also quite effective. As far as we
can tell, the addition of propargyl radicals has never been accomplished previously with such efficiency.
di-lauroyl peroxide (cat.)
cyclohexane I A
I ' "
Me3Si rsyOneoPn
Cyclohexane I A
Scheme 30. Generation and capture of propargyl radicals.
Our xanthate-based method appears to be well-suited to accommodate the inherent sluggishness of these species. The very
existence of the degenerate background reaction between the
radical produced and its xanthate precursor (path A in
Scheme 5, Section 2.2) provides the radical with a longer lifetime to survive a relatively slow trapping step caused by the
unreactive nature of either the radical or the trap. All these
observations open the way to many synthetic opportunities, for
it is now possible by a judicious choice of substrates and experimental parameters to exploit both the radical and nonradical
aspects of the chemistry of propargyl xanthates.
5. Summary and Outlook
The trail of xanthates has taken us forward; but it would be
foolish to assume that this old functional group has relinquished
all its secrets. Too many questions still remain to be answered,
even if what we have learned so far allows us to conjecture on
some future developments.
For instance, what other useful radicals will become accessible through xanthates? Phosphorus-centered radicals look
quite appealing, for the recent explosive popularity in the application of radical-based methods in synthesis does not seem to
have caught up with these species. We also have a keen interest
in nitrogen radicals, especially iminyls, and xanthates could
prove useful in this area too. More fundamentally, the kinetics
of the xanthate transfer need to be determined in order to gain
a greater mastery of the system. In principle, the xanthate functionality can be replaced by other related groups. Preliminary
studies thus seem to indicate that trithiocarbonates are equally
capable of sustaining the same type of radical chain reactions,
but further work is necessary to ascertain whether they offer any
special advantages.
The nonradical
Of S-propargyl xanthates also offers many opportunities for speculation. What other elec684
trophiIes are capable of reacting with the betaine? How important is aromaticity in stabilizing this species? To what extent can
its reactivity be enhanced? How would extra conjugation affect
its behavior? And so on and so forth. Our efforts so far have
been of an exploratory nature, but since most of the new reactions form carbon -carbon bonds, new synthetic strategies can
now be conceived and perhaps implemented. Some work directed towards such an aim is currently under way.
I wish to record my gratitude and appreciation to my collaborators, whose names appear in the references. The work described in
this review is largely a testimony to their enthusiasm, skill, and
diligence. I am also indebted to the C N R S , Ecole Polytechnique,
the Royal Society, RhBne-Poulenc Chimie, and Roussel-Uclaf for
generous$nancial and material support. It is an amusing twist of
fate that potassium ethyl xanthate was one of the first organic
compounds I prepared as a teenager in a small room next to the
kitchen, to my mother's utter disgust. It is perhaps only fair that
I should also thank her and the rest of my family for putting up
with such smelly activity for a good many years.
Received: May 13, 1996 [A167IE]
German version: Angew Chem. 1997,109,724-737
[I] W C. Zeise, J Chem. Phys. 1822,35. 173, ibid, 1822,36, 1 ; see also ref. [3].
pp. 1-6.
[2]L. Chugaev, Chem. Ber. 1899,32, 3332.H. R. Nace, Org React. 1962, 12,
57 - 1 00.
[3]S.Ramachandra Rao, Xunthates and Related Compounds,Marcel Dekker h e . ,
New York, 1971,chap. 9.
141 D. H.R. Barton, S . W. McCombie. J Chem. Soc. Perkin Trans. 1. 1975,15741585;D. H. R. Barton, Halfa Century ofFree Radical Chemistry, Cambridge
University Press. Cambridge, 1993;D. H. R. Barton, S . I. Parekh, C:L. Tse,
Tetrahedron Lett. 1993.34,2733-2736.
[5]W. Hartwig, Tetruhedron 1983,39,2609-2645;D.Crich, L. Quintero, Chem.
Rev. 1989.89. 1413- 1432.
[6]A. L. J. Beckwith, P. J. Barker, J Chem. SOC.Chem. Commun. 1984,683-684.
[7]D. H. R. Barton, D. Crich. A. Lobberding, S . Z Zard J. Chem. Soc.. Chem.
Commun. 1985,646-647;Tetrahedran 1986,42.2329-2338.
[S]M. D.Bachi, E. Bosch. J Chem. SOC.Perkin Trans. 1. 1988.1517-1519;M. D.
Bachi, E. Bosch, D. Denenmark, D. Girsh, J Org. Chem. 1992,57,6803-6810;
D . Crich, Tetrahedron Leit. 1988,29, 5805-5806
[9] D. H.R Barton, D. Bridon, I. Fernandez-Picot, S . 2. Zard. Tetrahedron 1987,
[lo] D. P. Curran in Comprehensive Organic Synthesis, Vo1.4 (Eds.: B. M. Trost.
I. Fleming), Pergamon Press, Oxford, 1991,pp. 715-831;D.P. Curran, Synthesis 1988,417-439and 489-513.
[ll] P. Delduc, C. Tailhan, S . Z. Zard, J. Chem. SOC.Chem. Cammun. 1988,308310.
[12]1. E. Forbes. C. Tailhan, S . Z. Zard. Tetrahedron Lett. 1990, 31, 2565-2568;
M.-P. Denied, B. Quiclet-Sire, S. Z. Zard, unpublished observations.
1131 W. M. McGregor, D. C. Sherrington, Chem. SOC.Rev. 1993,22,199-204.
[14]J. Axon, L. Boiteau, J. Boivin, J. E. Forbes. S . Z. Zard. Tetrahedron Lett. 1994,
35. 1719- 1712;L. Boiteau, PhD Thesis. 1994,Universite Paris-Sud, Orsay:
B. Quiclet-Sire, J.-B. Saunier, S . 2. Zard, unpublished observations.
[15] J. H. Udding, H. Hiemstra, W. N. Speckamp, J Org. Chem. 1994,59,37213725;J H. Udding, J. P. M. Giesselink, H. Hiemstra, W N. Speckamp. ibid.
[16]A. Nickon, E. F. Silversmith, Organic Chemistry: The Name Game. Pergamon
Press, Oxford, 1987,p.318.
[17]G. Bulmer, F.G. Mann, J Chem. SOC.1945,677-680 and references cited
[I81 D. H. R. Barton, M. V. George, M. Tomoeda, J. Chem. Sor. 1962,1967-1974.
1191 C. Tailhan, PhD Thesis, 1990,Universite Paris-Sud, Orsay.
[20]D. Griller. B. P. Roberts, J. Chem. Sac. Perkin Trans. 2.1972,747-751;K . U.
Ingold, B. Maillard, J. C. Walton, ibid. 1981,970-974;R. Ruegge, H. Fisher,
Int. J Chem. Kinet. 1986,/8,145-158.A. L. J. Beckwith, V. Bowry, G. Moad,
J. Org. Cheiri. 1988,53,1632-1641.
I211 J Pfenninger. C. Heuberger, W Graf.Helv. Chim.Acta 1980,63,2328-2337.
[221 H. Welde, J. Prukt. Chem. 1877,15,43.
[231 J. E. Forbes, s. z Zard, Tetrahedron Lett. 1989,30,4367-4370;J. E. Forbes,
S . 2. Z a d Tetrahedron 1993,49,8257-8266.
1241 B. Holmberg, J. Prakt. Chem. 1905,71,264.
[25] J E. Forbes; S . Z. Zard, J. Am. Chem. Soc. 1990, 112, 2034-2035.
[26] R. N. Saicic. S. 2 Zard, unpublished observations.
[27] J. Boivin, J. Camara, S . Z. Zard, J. Am. Chem. SOC.1992, 114, 7909-7010.
[28] M.-P. Denieul, 8. Quiclet-Sire, S . Z. Zard, unpublished observations.
1291 F. Mestre; C. Tailhan; S . 2. Zard, Hererocycles, 1989, 28 (special issue),
171- 174.
[30] J. Boivin. C. Tailhan, S . Z. Zard, J. Am. Chem. SOC.1991, ff3, 5874-5876;
K. Banert, H. Hiickstadt, K. Vrobel, Angew. Chem. 1992, 104, 72; Angew.
Chem. Inr. Ed. Engl. 1992, 31, 90-92.
[31] J. Boivin, E. B. Henriet, C. Tailhan. S. Z. Zard, Tetrahedron Letr. 1993. 34,
[32] N. F. Haley, M. W Fichtner, J. Org. Chem. 1980,45, 2959-2962, footnote 3.
[33] Y Gareau, J Chem. Soc. Chem. Commun. 1995, 1429.
[34] J. Boivin. E. B. Henriet, S . Z. Zard, J. Am. Chem. SOC.1994,116,9739-9740.
[35] 0. Mitsunobu, Synthesis. 1981, 1-28; D . L. Hughes, Org. React. 1992. 42.
[36] M. J. Koen, F. Le Guyader, W. B. Motherwell, J. Chem. Soc. Chem. Commun.
1995, 1241-1242.
[37] E. B. Henriet, PhD Thesis, 1995, Ecole Polytechnique. Palaiseau.
[38] J. Boivin. E B. Henriet, S. 2. Zard, Tetruhedron Lett. 1995, 36, 5171-5174.
[39] M. A. Poelert, S . Z. Zard, Synletr 1995, 325-326.
[40] J. Boivin. C. Tailhan, S. 2.Zard, Terrahedron Lerl. 1992. 33, 7853-7856.
[41] D. Wright. S . Z. Zard, unpublished observations.
(421 A. B. Baylis, M. E. Hillman, DBP 2155113, 1972 (Chern. Absrr. 1972. 77,
34 174ql. D. Basavaiah, P. Dharma Rao, R. Suguna Hyma, Tetrahedron 1996.
52,8001 -8062.
[43] M. A. Poelert, W. Roger, S . Z. Zard, Chem. Commun. 1996, 743-744.
[44] M.-P. Denieul, B. Quiclet-Sire, S . Z. Zard, unpublished observations.
Deposition of Data from X-Ray Structure Analyses
In order to make life easier for authors and referees the Cambridge Crystallographic Data
Centre (CCDC) and the Fachinformationszentrum Karlsruhe (FIZ) have unified their procedures for the deposition of data from single-crystal X-ray structure analyses.
Prior to submitting your manuscripts please deposit the data for your compound(s) electronically
at the appropriate data base, that is, at the CCDC for organic and organometallic compounds
and at the FIZ for inorganic compounds. Both data bases will be pleased to provide help (see
our Notice to Authors in the first issue of this year). In general, you will receive a depository
number from the data base within two working days after electronic deposition; please include
this number with the appropriate standard text (see our Notice to Authors) in your manuscript.
This will enable the referees to retrieve the structure data quickly and efficiently if they need this
information to reach their decision.
This is now the uniform procedure for manuscripts submitted to the journals Angewandte
Chemie, Chemische BerichtelRecueil, Chemistry-A European Journal, and Liebigs Annalenl
Angew. Chem. Int. Ed. Engl. 1991,36, 672-685
Без категории
Размер файла
1 604 Кб
chemistry, group, old, xanthate, train, function, new
Пожаловаться на содержимое документа