вход по аккаунту


Di-Grignard Compounds and Metallacycles.

код для вставкиСкачать
Di-Grignard Compounds and Metallacycles
By Friedrich Bickelhaupt*
In memoriam Georg Wittig
Just as the by now famous Grignard reaction can be used to prepare organomagnesium
halides from organic monohalides, organic dihalides can be used under certain conditions
to prepare di-Grignard compounds. Difficulties are encountered in particular when the two
halide functionalities are separated by only a short carbon chain (i.e. by 1-3 carbon atoms):
in such cases side reactions predominate. It has however recently become possible to obtain
such “short” di-Grignard compounds on a preparatively useful scale. This opens up new
perspectives for the synthesis of metallacyclic compounds of other main group and transition metal elements and in particular for metallacyclobutanes. The preparation, structure
and application of a selected number of di-Grignard reagents will be treated in the present
1. Introduction
1.1. Background Perspectives
I should like at the beginning of this article to indicate in
the form of a short personal note the (by no means
straight) line connecting my apprenticeship in Georg Witfig’s laboratory in Tiibingen, and the inspiration he provided, with my own work which is discussed in this review.
During his investigations on the formation and behavior
of ate complexes,[” Wittig had discovered the formation of
a complex between phenyllithium and phenylsodium; he
had shown this interaction to have a stabilizing effect on
phenylsodium, which is otherwise extremely reactive.[”
The “remarkable stability” of phenylsodium in diethyl
ether was interpreted in terms of the formation of diphenyllithium-sodium (or sodium diphenyllithiate) 1 [Eqn.
The fascination arising from this deep red, reactive compoundt5]has never left me since, even though my active interests often involved completely different topics. Apart
from its beautiful red color, the fascination of this compound stems in part from the existence of two highly polarized carbon-metal bonds in one molecule. These lie
close together in space, so that one may expect them to
influence each other to a high degree and thus to produce
a special electronic structure and reactivity. On the other
hand, the well-known synthetic potential of reactive organometallic compounds suggests that bifunctional species
of this type will have many and varied applications in organic and organometallic chemistry. A simple example is
provided by the retro-transformation of 2 into 3 by means
of mercury, or better still mercuric chloride;[51many other
examples could be added.[s-71
1.2. Scope of the Present Review
+ PhNa + Na[LiPhJ
The task assigned to me in the laboratory in Tiibingen
was to extend this concept to other combinations of alkyl
and aryl derivatives of lithium and sodium. This goal was
achieved, but with only a moderate degree of success.[31
When my diploma work was completed, Wittig thus suggested another subject for my doctoral thesis, the preparation of ortho-dilithiobenzene (2) from ortho-phenylenemercury (3)[“-“]and lithium metal in diethyl ether [Eqn.
[*I Prof. Dr. F. Bickelhaupt
Scheikundig Laboratorium, Vrije Universiteit
De Boelelaan 1083, NL-1081 HV Amsterdam (The Netherlands)
990 0 VCH
Verlagsgerellsrhaft mhH, 0-6540 Weinheim. 1987
Organolithium and organomagnesium compounds are
of approximately equal importance in preparative chemistry; they are often used interchangeably for comparable
purposes. Although organolithium reagents are more reactive, this difference in reactivity is often not decisive: more
often it is a question of convenience (the accessibility or
the method of preparation) which determines the class of
compound to be used.
This survey of bifunctional organometallic compounds
will put particular emphasis on those compounds in which
the metal functionalities are separated by a short chain
consisting of 1, 2 or 3 carbon atoms. These types of compound are at present much more readily accessible in the
Grignard series than are the corresponding organolithium
reagents. For this reason, and because polylithium compounds have recently been the subject of several reviews,[*’01 the present article will deal mainly with bifunctional organomagnesium compounds.
A second limitation results from the breadth of this area.
As already indicated, an important goal in the synthesis of
bifunctional organolithium and organomagnesium com-
0570-0833/87/1010-0990 $ 02.50/0
Angew. Chem. I n { . Ed. Engl. 26 (1987) 550-1005
pounds is the synthesis of other bifunctional derivatives
and in particular the preparation of metal-containing cyclic compounds, the metallacycles. As will be shown, this
goal was recognized soon after the discovery by Grignard
of the reagents bearing his name and was intensively pursued. It is thus not surprising that the synthesis of metallacycles by such routes has led to a plethora of reactions and
compounds.“ I ] For this reason I shall mainly concentrate
in the following on results obtained by our own group;
these deal in particular with aliphatic compounds and
arylalkyl compounds.
2. Di-Grignard Compounds
2.1. Preparation
Of the various methods for the preparation of organomagnesium compounds,[””.‘.’21 two in particular are important in the present context. The preparation starting
from organic dihalides corresponds to the famous original
synthesis of the Grignard reagent[’31[Eqn. (c)] and is also
preparatively the most important because of the ready accessibility of the starting materials and because of the (at
least in principle!) simple experimental procedure.
+ Mg
The preparation starting from organic derivatives of
mercury [Eqn. (d)] should also be mentioned, since it yields
very pure Grignard reagents.
+ Mg
+ Mg/Hg
After Grignard had succeeded in 1900 in preparing the
first mono-Grignard reagents [Eqn. (c)], the obvious next
step was to use organic dihalides to prepare the corresponding di-Grignard compounds. Tissier and Grignard in
fact carried out the first investigations in this direction in
the following year. That they[’41and others[’51were at first
unsuccessful was not due to the concept being faulty, but
to the starting materials chosen: 1,2-dibromoethane and
1,3-dibromopropane were the only readily available dibromides, but were at the same time the very two which
were not suited for the desired synthesis. Instead of the
hoped-for di-Grignard compounds 5 [Eqn. (e)] and 7
[Eqn. (01, ethene and cyclopropane were obtained in high
yields. It is very likely[‘61that these reactions occur in two
Angew. Chem. I n ! . Ed. Engl. 26 (1987) 990-1005
2.1.1. General and Historical Remarks
steps, the initially-formed mono-Grignard reagents 4 and
6 very rapidly eliminating magnesium bromide.
Thus, the truism well known to every chemist showed
itself to apply yet again: no product will be obtained without a (suitable) starting material! As in the present case, it
is of little importance whether the required starting material is unknown o r merely difficult to obtain. Higher
homologues, such as the 1,5-dibromopentane (S), with
which the Grignard reaction can readily be carried out,
were in fact already known.“71 But advances towards obtaining di-Grignard compounds were made only after J.
uon Braun had in 1904 shown this compound to be readily
available in pure form from the reaction between phosphorus pentabromide and benzoylpiperidine.[’*]
However, this process did not occur without tension and
a certain amount of drama. J . uon Bruun had certainly immediately recognized the potential of this “synthon” (as
we would say today) and studied its reactivity in various
directions with great success. As early as 1905, in a publication with Steindorff’] he mentioned an investigation of
its reactivity “towards magnesium in absolute ethereal solution. We intend to report in the near future in more detail on the application of the Grignard reaction to 1,5-dihalogen compounds”.
It is not known to me why the publication of these results was delayed. However, the first detailed study of
the preparation and of several reactions of 1 ,5-di(bromomagnesio)pentane (or pentamethylene-l,5-bis(magnesium
bromide)) (9) appeared only in 1907; it was published by
Grignard and Vignon‘20,2’1
+ MgEr2
+ 2 Mg
Grignard made it quite clear that this success was mainly
due to the availability of the starting material 8 : “Malheureusement les dihalogtnures intermtdiaires etaient,
jusqu’a ces dernikres annees, a peu prks inaccessibles, lorsqu’en 1904, M. J. v. Braun“’ decouvrit un elegant proctde
d’obtention d u dichloro- et du dibromopentane, ...rr.[Zol
However, it must have been a great disappointment to
uon Braun, as he reacted swiftly and not without bitterness.[”I He first emphasized the key role of the starting material and then derived his own claim[231from its production: “That I am carrying out such experiments (and, as
the discoverer of the method by which the pentamethylenedihalogen compound first became available in larger
amounts, I surely had the right to use the Grignard reaction, which has by now become common property), . ..”.
He referred to the publication mentioned above of his int e n t i o n ~ “ and
~ ] continued: “In the interests of both sides it
is to be regretted that the references concerned have escaped the notice of the French colleagues, ...” Von Braun
did indeed play a n important part in the discovery and application of the aliphatic a,w-di-Grignard compounds.[241
In addition, he was one of the first to use the new dihalogenoalkanes for the synthesis of a metallacycle, the (oligomeric) p e n t a m e t h y l e n e m e r ~ u r y ; [however,
he used sodium amalgam in its preparation, as did Hilpert and
Griittner and his school, to which the well-
99 1
known E. Krause also belonged, then rapidly developed
the preparation of metallacycles and other organoelement
heterocycles using di-Grignard reagent^.^'^.^^^
The experience gained in the course of these earlier
studies and of many later ones[291can be summarized as
follows. Di-Grignard compounds can be prepared from
the corresponding dihalides (generally the dibromides[’2.291)with some care just as well as can the monoGrignard compounds, when the two halogens are separated by at least four carbon atoms. The necessary precautions involve the reaction being carried out relatively slowly, since otherwise intra- and intermolecular Wurtz coupling occurs, as is also the case for the normal Grignard
reaction. It will be shown in Section 2.2.1 that the “normality” of this type of di-Grignard compound is expressed
not only in its preparation but also in its properties.
In contrast, the Grignard reaction does not succeed, or
the yields are very poor, when 1, 2 or 3 carbon atoms separate the halogens. These three cases will now be dealt
with in more detail.
the other hand, the success of the reaction is due to exactly
these circumstances: all other by-products such as (unsaturated) hydrocarbons and methylmagnesium bromide dissolve in diisopropyl ether and can be removed by simple
Still to be explained is the (intuitively somewhat unexpected) ready formation of the “1,l-di-Grignard reagent”
10 in contrast to the “1Jreagent” 5 and the “1,3-reagent”
7. This is probably due to differences in stability, the carbenoid mono-Grignard 11 (which must be postulated as
an intermediate) being much more stable than 4 and 6.
The a-elimination of 11 to give 12 proceeds more slowly
since it is enthalpically less favorable than the formation of
ethene [Eq. (e)] and cyclopropane [Eq. (01.
2.1.2. Ceminal Di-Crignard Compounds
It is interesting that the simplest members of the group
of “small” di-Grignard compounds are in fact exceptional
in that they can be prepared in relatively good yields and
have thus been known for the longest time. They were first
described by Emschwiiler in 1926,[301and their preparation
was subsequently considerably improved by CainelIi et
a1.;[”’ the most important modifications made by them
were to use magnesium amalgam instead of magnesium
metal and to use a 1 : 1 mixture of ether and benzene instead of pure absolute diethyl ether.
+ CH2:-+
The a-elimination can however not be completely suppressed, and is probably responsible for the formation of
ethene, propene, and c y ~ l o p r o p a n e . [It~ can
~ ’ be reduced to
a minimum by the choice of suitable reaction conditions:
particularly important is that the magnesium content of the
amalgam be as low as possible (0.5-1Yo magnesium). This
results in the main drawback of the synthesis: relatively
large amounts of mercury (which, of course, can be recovered) are required.
Although this procedure for the preparation of 10 is
successful, in our experience it cannot (with one exception) be extended to the use of other 1,l-dibromo compounds. For example, neither 1,l-dibromoethane, 1,l-dibromoneopentane nor benzal bromide give appreciable
amounts of any organomagnesium compounds in an ether/benzene solvent mixture.[351Only trimethylsilyldibromomethane (13) reacts in diisopropyl ether to yield 70% of
the di-Grignard compound 14
even the tin analogue of
13 gives complex product mixtures.‘371
The yields of the methylene di-Grignard reagents were
however still not easily reproducible and generally lay no
higher than 50-60%;[31~321
the by-products were ethene,
propene, cyclopropane and, particularly unwanted for organometallic syntheses, methylmagnesium halide. Better
yields were obtained only in reactions which were carried
out under “Barbier” conditions, i.e. when a dihalogenomethane was reacted with magnesium amalgam and the substrate (for example a ketone) in a “one-pot’’
We were recently able to improve the synthesis further by
using diisopropyl ether as solvent. In this way the methylenedimagnesium dibromide 10 can be obtained in a reproducible yield of 80% and free from unwanted by-product~.‘~’’
It has thus become an interesting synthon for the
anionic introduction of a methylene group.‘341
2.1.3. Yicinal Di-Crignard Compounds
The excellent way in which the reaction proceeds when
this method is used is even more remarkable since it is an
essentially heterogeneous reaction: 10 is practically insoluble in the reaction mixture, just like the amalgam and the
magnesium bromide which is formed as a by-product. On
The preparation of compounds of this type has until
now proven to be the most difficult to realize, particularly
in the aliphatic series. The future prospects are also not
particularly good, since the p-elimination described as an
example in Equation (e) is so strongly favored both by entropy and by enthalpy that the lifetime of a n intermediate
such as 4 is too short for it to react in a second step with
the magnesium surface to give a di-magnesium derivative
such as 5.
Angew. Chem. Inr. Ed. Engl 26 (1987) 990-1005
Success has thus been limited to cases in which it has
been possible to drastically retard the rate of CJ-elimination.
The classical example in the aliphatic series is the preparation of the cis- 1,2-di-Grignard reagent 18 of cyclopropane
by Wiberg and Barf/ey['*] from trans-l,2-dibromocyclopropane ( 1 3 ) . In this case the CJ-elimination from the monoCrignard reagent 16 is made difficult by the fact that it
leads to cyclopropene [Eqn. (g)]: this process doubles the
already high ring strain in cyclopropane.
pounds appear to decompose very readily with elimination
of lithium hydride.['"
+ LiH
From this point of view, the stability of 18 at room temperature is probably due to two factors: firstly the lesser
tendency of magnesium to split off hydride and secondly
the strain already mentioned in the product of magnesium
hydride elimination from 18, which would be a Grignard
derivative of cyclopropene.
It could also be expected that in the aromatic series the
CJ-elimination from o-dihalogenobenzenes 20 would be
made more difficult by the formation of the strained triple
bond of the aryne.
Various aspects of this reaction are remarkable, firstly of
course the fact that 18 is formed at all. The yield is low (ca.
10-15%), the main product being the alkene (i.e. cyclopropene), as in Equation (e). The formation of the cis-compound 18 from the trans-compound 15a is also surprising;
Wiberg and Bartley assigned this structure on the basis of
the carboxylation and deuterolysis of 18, since they obtained cis-products in both cases. They attempted to rationalize the formation of the cis-structure by the formulation 19, but presented no arguments for it.
We have repeated Wiberg's experiments and could in
the main confirm their results.[391In addition we were able
to isolate 18 in a pure form, as it is only slightly soluble in
diethyl ether. It is interesting that its solubility is increased
considerably by the addition of magnesium bromide (see
Section 2.2.2); this allowed us to measure the 'H-NMR
spectrum of 18 and thus to confirm the cis-structure directly. In addition we found that 18 is also formed in 3 comparable yield from the cis-dibromide 15b. Since the cyclopropyl-magnesium bond is generally configurationally stable,I4"I the cis-configuration of 18 is either thermodynamically more stable, which would be somewhat surprising
(see however the discussion in Section 2.2.2), or the stereochemistry is determined at the stage of the configurationally flexible cyclopropyl r a d i ~ a 1 s . l ~The
~ ' immediate precursor for 18 is very likely 17b [Eqn. (g)]. Theoretical calculations show that 17b is more stable than its trans-isomer 17a,[4'1so that the explanation of the favored cis-configuration on the basis of kinetic factors gains credibility.
It appears questionable at present whether a simple aliphatic 1,2-dimagnesium compound such as 5 can in fact
be synthesized. It is also unclear whether it would be thermally stable, since the corresponding organolithium comAngew. Chem. lni. Ed. Engl. 26 (1987) 990-1002
X, Y
The course taken by the reaction depends on the nature
of the halogens X and Y. The first step (20 + 21) only occurs in a satisfactory manner when X = Br o r I.[42.431 The
mono-Crignard compound 21 can be quite stable, especially when Y = C l or Br.[44.451
This has already been utilized for further reactions of 21.[42-461
However, there is a
more o r less distinct tendency for the elimination of magnesium halide to give the aryne (e.g. 23), as has been
shown in particular by Witfig.[441The di-Grignard compounds 22 can be obtained under favorable conditions
from 20 (X = Y = Br,[12.471
X = I, Y = Br[481)in 20-40% yield.
This relatively poor yield can however be of preparative
interest, since one can obtain bifunctional or cyclic products in a one-pot synthesis from a readily available starting
material (20). However, pure 22a can only be obtained via
the dilithium compound 2.[']
2.1.4. 1,3-Di-Grignard Compounds
The chance that a 1,3-di-Grignard compound can be
prepared from the dibromide [Eqn. ( f ) ] is a priori greater
than that for a 1,2-di-Grignard compound [Eqn. (e)]. The
gain in enthalpy from the successfully competing elimination of magnesium bromide is probably approximately
equal in both cases, but the entropy is clearly more favorable for the 1,2-elimination, particularly since the ideal
combination of a strongly electrofugal group (MgBrO) and
a strongly nucleofugal one (BrQ) makes only small demands o n the relative orientation of the groups.
In spite of this, all attempts to prepare 7 were, for a long
time, unsuccessful."sh,'l Zelinsky and Gutt were able to
identify a di-Grignard compound as well as the by-product
propene; however, the product was not 7 but 1,6-bis(bromornagnesio)hexane 27,""l the product of a Wurtz reaction. The main product was always cyclopropane. One can
generalize by saying that the formation of a three-membered ring from a 1,3-dihalide proceeds so well that it has
been developed into a versatile synthesis for cyclopropanes and heterocyclopropanes ; zinc and alkali metals can
be used as well as magnesium.
The first preparation of 7 was achieved by Costa and
Whitesides in 1977 using an indirect method via organoboron and organomercury compounds.[491
C H3Hg(C H 2)3 HgC H,
5C1Hg(CH2),HgCl 3
+ CH,MgBr + n-C3H7MgBr +
Although the concept of this synthesis is very elegant, it
cannot to be recommended for preparative applications
because of its complexity, the low yields, and the large
number of organomagnesium side products ; however,
Costa and Whitesides were able to use 7 to prepare a silacyclob~tane.[~~~
A chance observation in fact finally led to a breakthrough in the direct synthesis of 7. In our laboratory F. A .
Hartog had been studying the Barbier reaction,[501one goal
being to trap unstable Grignard intermediates using carbonyl c o r n p o ~ n d s . [ ~Thus
he obtained 25, which can be interpreted as an adduct between 27 and benzophenone, in
6% yield. This indicated the possibility that the lifetime of
compounds of type 26, to which 6 [Eqn. (f)]also belongs,
is short, yet long enough to permit further intermolecular
+ BrMg(CH2),MgBr +
7 (30%)
27 (15%)
+ (CH,); + MgBr,
24 (0-10%)
It is important for a high yield of 7 (ca. 30%) that the
reaction be carried out slowly in diethyl ether using a large
excess of magnesium (the latter should preferably be purified and activated by double sublimation). Such conditions can be readily achieved when the dibromopropane is
added to the magnesium at room temperature over a period of 2 days in a sealed glass apparatus with an overflow
system. Under these conditions, cyclopropane is still the
main product. A few percent of other hydrocarbons are
also present: propene (1.I%), propane (1 .0%), ethene
(O.6%), methane (0.4%), and ethane (0.1%). This indicates
that the Grignard reaction occurs in a complex manner involving free radicals.[5s1If necessary, 7 can easily be purified by treating it with T H F (see Section 3).
The range of di-Grignard compounds which can be obtained directly has in this case also proven to be small. The
synthesis of the 2,2-dimethyl derivative 28 is preparatively
useful; though the yield is only 10-15%, the reagent is almost free from organometallic by-products and can thus
be used directly for the synthesis of metallacycles; in addition, the starting material is readily available.1561
+ Mg
In contrast, the formation of I,?-di-Grignard reagents
does not occur satisfactorily when a hydrogen atom is
present in position 2, as is shown by the example of 29
(9%). Remarkable in this reaction is the high yield (70%) of
the hydrocarbon 30, which is formed directly during the
BrCH2CH( tBu)CH2Br
BrMgCH2CH( tBu)CH2MgBr
29 (9%)
30 (70%)
1) PhZC=O. 2 ) HO
26 [Cl(CH,),MgBr]
This result, together with the demonstrated ability of 7
removed the psychological barriers which had
for so long prevented attempts to synthesize 7 directly
from 1,3-dibromopropane.[s'~sz1
The first ray of hope was
provided by a signal at 6 2 0 to - 1 ppm in the 'H-NMR
spectrum of the reaction mixture: it could only be due to
7 , but the yield was at first below half a percent. Variation
of the dihalide, the metal and the solvent led to the procedure which gave the best results [Eq. (h)].[531
reaction and not by hydrolysis of an organomagnesium
precursor.[561We still have no completely satisfactory explanation for the formation of 30.["'
Similarly, our expectation that it would be possible by
using the cyclic dibromides 31 and 32 to reduce the formation of three-membered rings, because of the ring strain
generated in such products, and thus favor the formation
of the di-Grignard compounds has regrettably not been
f ~ l f i l l e d . ~The
~ ' ~products 32 and 34 were obtained in low
yields, in each case together with several by-products; in
addition 34 decomposes rapidly to give 35 (see also Section 2.2.2).
On the other hand, we have recently succeeded in obtaining the 1,3-di-Grignard reagent 37 in 96-100% yield by
slow reaction of ortho-bromobenzyl chloride (36) with an
excess of magnesium (sublimed and sieved to give a parAngew. Chem. Inl. Ed. Engl. 26 (1987) 990-1005
B r o B r
2 Mg
32 (< 10%)
B r o B r
34 (< 40%)
high purity by recrystallization. The subsequent reaction,
carried out by shaking the organomercury compound in
T H F with an excess of magnesium metal, is slow but practically quantitative. In mechanistic terms there is probably
an initial reduction to give the diorganomercury compound, which reacts with further magnesium to give the
diorganomagnesium compound; the Schlenk equilibrium
is then set u p between the latter and the magnesium bromide formed [Eqn.
(see Section 2.2).
ticle size < 2 mesh) in THF. It disproportionates immediately according to the Schlenk equilibrium [Eqn. (I) in Section 2.2. I]. The diorganylmagnesium compound 38, a
nearly insoluble solid, is precipitated and can be easily
freed from all impurities by decantation (isolated yield 8590%) [Eqn. (i)] (see also Section 2.2.2).
-7-2 RMgBr
When the substituents R are sterically demanding, the
second step i s sterically hindered and the reaction proceeds only as far as the diorganomercury compound.[s8i
The last step in Equation
the ligand exchange reaction between the diorganomagnesium compound and magnesium bromide, can also be carried out separately: for example, a cyclic organomercury compound can first be converted into the magnesacycle and the latter then treated
with the stoichiometric amount of magnesium bromide
[Eqn. &)I.
2.1.5. Summary of the Previous Results
We can sum u p as follows: it is possible not only to prepare “normal” di-Grignard compounds directly from the
corresponding dihalides; a number of such reagents where
the two magnesium atoms are separated by one o r three
carbon atoms is now known. However, 1,2-di-Grignard
compounds are rareties, and it appears that this will continue to be the case in the foreseeable future.
2.2. Structure and Properties
No X-ray crystal structure determinations have yet been
reported for solid di-Grignard compounds, so that the following discussion refers to solutions, mainly in TH F.
2.2. I . “Normal” Di-Grignard Compounds
2.1.6. Di-Grignard Reagents from Mercury Compounds
The preparation of Grignard reagents from the corresponding organomercury compounds [Eqn. (d)] is normally of no preparative importance, since the mercury
compounds are themselves obtained from the Grignard
compounds (with the exception of the amalgam method).
This cumbersome reaction sequence can however become
necessary when a Grignard reagent of high purity is required for structural studies. This is always the case when
the yield for the reaction of formation from the organic
halide is low; dihalides, as we have seen, very often give
such low yields.
This method has two advantages: firstly the impure
Grignard reagents obtained initially can be converted into
stable organomercury halides which can be obtained in
Chem I n ! . Ed. Engl. 26 (1987) 990-1005
The structure of Grignard reagents in solution is complex, since several parallel equilibria are involved.[’lCi
Apart from a quantitatively unimportant electrolytic dissociation, the so-called Schlenk equilibrium [Eqn. (I)] dominates; all the components of this equilibrium can form associates with themselves or with other components.
+ MgBr,
The position of these equilibria depends on the organic
group R, on the halide and o n the solvent. Our structural
investigations have been concerned mainly with bromides
in T H F as solvent. In this case, association is unimportant
for all three components; they are monomolecular. In the
bifunctional series, the same is true for the di-Grignard
compound 40 and for magnesium bromide but not for the
diorganomagnesium, which is in this case the magnesacycle 39. In dilute solution, magnesacycles are either completely dimerized o r at least dimerized to a large extent to
99 5
[39],. It was not possible to derive the structure and behavior of the di-Grignard compounds 39 until after the structure of [39], was known (see Section 3) [Eqn. (m)].
the cyclic diorganomagnesium compound 42 is so stable
(see Section 3) that the Schlenk equilibrium is shifted completely to the right-hand side (cf. 55, 56, and 57 in Table
2.2.3. “Short” Di-Grignard Compounds
The position of the equilibrium was determined by very
accurate measurements of the osmotic particle number using a stationary isothermal distillation. The results are collected in Table 1 and compared with the values for ethyl
magnesium bromide.
Table I. Thermodynamic parameters of the Schlenk equilibrium in T H F at
25°C [Eqns. (I) a n d (m), upper part].
[kJ mol-’1
[J mol-’ K - ’ 1
Di-Grignard compounds with 1, 2 or 3 carbons between
the two magnesium functionalities differ in several characteristic aspects from the “normal” members of this class.
These differences probably have a common origin in the
relatively short distance between the charges or dipoles of
the two polarized carbon-magnesium bonds. However, we
have no direct information on the structure of 7,10, and
18, since their crystal structures have not been determined
and no association measurements have been carried out.
In general the solubilities of the “short” Grignard reagents are low. In THF, where the Schlenk equilibrium
[Eqn. (I)] is shifted considerably to the right (see Section
2.2. l), disproportionation to the dialkylmagnesium 43 occurs; because of its oligomeric o r polymeric character, the
latter is practically insoluble [Eqn. (n); compare Eqn.
These results show that the mutual influence of the two
organomagnesium functionalities is not large in the thermodynamic ground state, since they d o not deviate appreciably from the values for the monofunctional model compound and show little dependence on the length of the methylene chain. This probably also means that the situation
in solution is the same as that for the mono-Grignard compound. The magnesium usually binds two additional T H F
molecules to become tetracoordinate.[”cl Association, including that with magnesium bromide, is unimportant; we
shall however see in Section 2.2.3 that the situation is very
different for the “short” di-Grignard compounds.
The mono- and di-Grignards also resemble each other in
diethyl ether in that in both cases the Schlenk equilibrium
is shifted completely towards the side of the organomagnesium bromide.[l’c.62.631
2.2.2. Intramolecular Coordination
’[MgZ], + MgBr2
10, 18, 7
Z = CHZ; 18: Z
A ; 7: Z
In the case of 10 this reaction does not go to completion, since magnesium bromide is apparently locked in
structures such as [CH2Mg.CH2(MgBr),], (44) (for the
sake of clarity coordinated solvent molecules have been
In all cases, the insoluble 43 can be redissolved by the
addition of an excess of magnesium bromide. This is not
due to a shift of the Schlenk equilibrium (n) to the right,
since in the case of 18 and 7 exactly two moles of magne-
The situation is quite different when, as in 41, the chain
linking the two organomagnesium functionalities carries
an oxygen atom which is able to form an intermolecular
coordinative bond to the magnesium atom. In these cases
Z = (CH,)P];
0. . Mg
L z l
(CH2)4[651; o - C,H,CH2CH2[65]
Angew. Chem. l n t . Ed. Engl. 26 (1987) 990-1005
sium bromide are used; thus the formation of a complex
such as 45[j9]
must be assumed; analogies are known in the
case of simple organomagnesium compounds.[661
This unusually strong complex formation with magnesium bromide apparently serves to compensate for the
high “concentration” of negative charge in a small space.
As has already been mentioned, the Schlenk equilibrium
in diethyl ether lies far to the side of the organomagnesium
b r o m i d e ~ ; ~ ~10,
* . ‘ 18,
~ ~ and 7 are more soluble in this solvent (where they probably have the structure depicted).
However, their solubilities are still not high, particularly in
the case of 18 (see Section 2.1.3). What is remarkable, and
so far unexplained, is the considerable increase in the solubility of 10 in a 1 :1 mixture of diethyl ether and benzene.
Since the non-polar benzene normally decreases the solubility of the somewhat polar Grignard compounds, one
may speculate that 10 is aggregated [Eqn. (o)]to form the
dimer 46, which has a relatively non-polar outer surfa~e.1~’’
2 S-Mg-CH2-Mg-S
10 .
/ T 2
,My M
, g,
S = Et20
We have also observed a tendency for 7 to form chargecompensated bromine bridges in diethyl
lower temperatures, coalescence effects are observed in the
proton N M R spectra; these can be interpreted as a n equilibrium between the two equivalent structures 47a and 47b
(together 9ovo), and 48 (10%). The ratio of 47 to 48 is temperature-independent, which means that the enthalpy difference is zero; that 47 is favored must thus be due to entropy factors (AS = - 4.3 J mol- K- ’), the ring formation
in 47 apparently being overcompensated by the fixation of
additional solvent molecules S in 48.
Et20 48
Since the two a-methylene groups in 47a or 47b and 48
have different chemical shifts, line shape analysis allowed
us to determine the activation parameters for the equilibria
1 (AH’=2l kJ m o l - ’ , A S f = - 1 0 9 J mol-’ K-I) and I1
Angew. Chem. In!. Ed. Engl. 26 (1987) 990-1005
(AH’=28 kJ mol-I, A S + = - 118 J mol-’ K-I). The negative entropy of activation in particular indicates that the
initial steps in the rearrangement reactions involve attack
by solvent molecules.
Finally, we should comment on a property of 7 which
somewhat reduces its stability and usefulness. It was mentioned in Section 2.1.4 that one of the by-products in the
preparation of 7 by both Whitesides’ and our method is
allylmagnesium bromide (24). Further studies‘52,531
shown that 24 does not arise during the formation of 7,
but is a secondary product formed from 7.
The hydrogens at carbon 2 are hydridically activated
from both sides by hyperconjugation with the two electron-rich carbon-magnesium bonds, so that magnesium
bromide hydride can be split off. Magnesium bromide
functions here as a Lewis acid catalyst, so that the rate of
decomposition of 7 depends on the concentration of
MgBr,. Since the presence of MgBr, is unavoidable in
both syntheses, a certain amount of 24 is always present.
However, when 7 is purified according to Equation (r) and
is free from excess magnesium bromide, it is quite stable in
solution (half-life ca. 120 days). We shall see, however, that
the double hydridic activation can be a great disadvantage
in some applications. The 2,2-dimethyl derivative 28 can
naturally not split off hydride; it is thus quite stable and
also undergoes no side reactions during chemical transformations.
3. Magnesacycles
The preparation and study of magnesacycles is important for two reasons. Firstly, they are of fundamental importance as metal-containing heterocycles: thus, for example, the question can arise as to whether a magnesacyclohexane is stable, and as to its structure. In addition, a
knowledge of the structure of magnesacycles is also essential when the structures of the di-Grignard reagents (see
Section 2.2) are to be studied, particularly in the most important solvent, THF. The dilemma present is that, because of the unavoidable Schlenk equilibrium (Eqns. (I)
and (m)], Grignard compounds can never be obtained in a
pure form in this system. This limitation does not apply to
the other partner of the Schlenk equilibrium, the dialkylmagnesium compound; thus pure [39], can be obtained
from the corresponding mercury compound, as shown in
the first step of Equation (k). A further possibility is the
method discovered by Schlenk himself during his studies
of the equilibrium reaction which bears his name:1681this
involves the removal of the magnesium bromide, formed
according to Equation (I), as an insoluble dioxane complex. This method is faster and more convenient for the
preparative formation of diorganomagnesium compounds,
but is not to be recommended when extremely pure compounds are required for structural studies. In the case of
the magnesacycles there is an additional problem: the
dioxane precipitate contains not only magnesium bromide
but also organomagnesium compounds (either bonded o r
only entrained), so that the yields of pure magnesacycles in
solution are very low.
It should be remembered that all components of the
Schlenk equilibrium [Eqn. (I)] are normally monomeric in
THF. In the case of organomagnesium compounds the
strong nucleophilicity of the T H F molecule towards magnesium usually leads to the formation of a tetracoordinated, saturated species.~"'."9.701
X = R. Hal
It was thus surprising that in the case of magnesacyclohexane an extremely strong tendency to dimerization [cf.
k,, in Eqn. (p)] was observed;i631the monomeric form could
only be detected by measurements in extremely dilute solutions (0.00 l to 0.02 M). The thermodynamic parameters
(Table 2) were determined from the temperature-dependence of the dimerization equilibrium.
[a] Association number n=ci: c,,
cr being the formal concentration when
100%monomer is present and c,,, the panicle concentration determined osmotically. [bj A H o = -48.053.0 kJ mol-', ASo = - 106.02 10.0 J mol-'
K - '. [c] A H D = -48.7 kJ mol- ', ASo = - l18+ 12 J rnol- ' K - '.
Two aspects of this result were extremely puzzling: 1)
Why is magnesacyclohexane SO so unstable compared to
the dimer [SO],, although a six-membered ring structure
with tetracoordinated magnesium could be considered
possible? 2) When SO is so unstable, why does it give only
the dimer without polymerizing further?
The answers are supplied by the X-ray structural analysis of [SO], (Fig. 1).I6l1 The compound is dimeric in the
crystal. As expected, the magnesium is made tetracoordinate by the complexation of two molecules of THF. However, the geometry is not that of an ideal tetrahedron, the
Fig. I. X-ray crystal structure of 1,7-dirnagnesacyclododecane[Sol2; I and I '
are the magnesium atoms, 2, 3, 2' and 3' the oxygen atoms of the four THF
molecules, 4-8 and 4'-8' the carbons of the two pentamethylene chains, and
9-16 and 9'-16' the carbon atoms of the four THF molecules.
C-Mg-C angle being very obtuse (141.5'). According to
one's taste one can either say that the sp-hybridization of
the magnesium (bond angle ISO", see zinc and mercury!) is
only slightly perturbed by the ether oxygens, or that according to Bent's
the magnesium uses much more
s-character in the bond to carbon than in that to the more
strongly electronegative oxygen.
In any case, the result is a large intraannular angle at
magnesium; this cannot be accommodated in the six-membered ring of 50 without introducing strain, but becomes
possible in the twelve-membered ring of [So],. The strain
makes itself felt in the form of the exothermal dimerization
energy of AHD= -48 kJ mol-'. The entropy loss caused
by the dimerization (ASD= - 107 J mol-' K - I ) compensates only partially for this effect (AGD(25"C)= - 16 kJ
mol-I). However, as soon as the driving force of the ring
strain in the six-membered ring has disappeared in the dimerization to the twelve-membered ring, the entropy component opposes any further polymerization. Evaporation
of [39], followed by slow warming in fact yields modifications of 39 [Eqn. (m)] which can be redissolved only with
difficulty on shaking for a long time with THF. Apparently, the removal of the complex-forming T H F has led to an
extensive polymerization (to [39],).
The ring strain in 50 has also been confirmed thermochemically. Model experiments first showed that with respect to one carbon-magnesium bond, the enthalpy of
reaction according to Equation (9) always has the same
value (AH = - 220 kJ mol - I), independent of the group R.
Angew. Chem In{. Ed. Engl. 26 (1987) 990-1005
Deviations from this standard value signalize particular
effects, for example ring strain. Thus, when the position of
the equilibrium 2 50*[50], is known the experimentally
determined heat of reaction can be used to determine the
ring strain in 50, under the reasonable assumption that
[5012 is strain-free. The value obtained for 50 was
A H = -45.4k6.5 kJ mol-I, which within the (in this case
much larger) limits of error agrees with that from the association measurements ( A H = - 48.0 -+ 3.0 kJ mol - ; Table
2). A similar equilibrium position and ring strain was determined for the six-membered benzoannelated magnesacycle 53 (Table 2).
After this detailed discussion of the results for 50, the
other data in Table 2 can be readily explained. It is of
course completely clear that a large C-Mg-C bond angle
in a five-membered ring will generate even more strain
than in a six-membered ring; it is thus not surprising that
[49], in T H F shows no tendency to form the monomeric
five-membered ring 49. 1491, also exists as a ten-membered
ring in the crystalline state.1741We can thus assume that
[Sl],and [5212,which exist solely in the dimeric form, also
contain fourteen- and twenty-membered rings, respectively. Ths complete dimerization of 52, which would like
[49], contain a ten-membered ring if monomeric, is at first
surprising. A detailed analysis of the X-ray structures of
[49], and [50],, however, shows that the two oligomethylene chains in these compounds are sufficiently separated
from each other by the two large magnesium atoms; thus,
they cannot cause a transannular van der Waals repulsion.
In contrast, 52 as a ten-membered ring would contain only
one magnesium atom; therefore, just as for 51, the wellknown ring strain in medium-sized rings plays an unfavorable role.
It is interesting to note that the rapid exchange of bonds
between magnesium and carbon or oxygen[”“] is responsible for the fact that the systems mentioned above can and
must always seek out the thermodynamically most favorable state. Pure carbocycles or hetero- and metallacycles
with stable ring bonds d o not have this possibility, so that
in such systems strain can only be determined thermochemically and not by means of the experimentally simpler
association measurements.
Finally, we should remark on the last four compounds
in Table 2. In contrast to the general trend, they exist in a
purely monomeric form. We have no completely convincing explanation for this in the case of 54; probably the
ring strain is reduced by benzocondensation (i.e. some unfavorable hydrogen interactions are no longer present; see
also [74b]). In the cases of 55, 56, and 57 the monomeric
state is clearly stabilized by intramolecular coordination
(see also Section 2.2.2). Particularly informative is 56,
which without oxygen coordination would have a tenmembered ring as does 52. While the latter is not stable as
the monomer, 56 is, the reason being that it contains not a
ten-membered ring but two six-membered rings.
Anyen’. C‘heni Int. Ed Engf. 26 i f Y 8 7 ) YYO-1005
The heats of the reaction with acetic acid mentioned
above can also be used in conjunction with values for
standard dialkylmagnesium compounds to estimate the
strength of the intramolecular coordination.1651This is
however only possible to a certain degree of approximation, since compounds 55-57 d o not have the idealized
geometry mentioned above; because of the longer Mg-C
and M g - 0 bonds, the rings d o in fact exhibit ring strain,
the value of which is not known. In this manner we arrive
at a lower limit of A H = 9 0 and 87 kJ mol-‘, respectively,
for the strengths of the M g - 0 bonds in 55 and 56.lh5’
value lies considerably higher than that of 68 kJ mol-’ estimated for the magnesium-oxygen bond in Et2Mg.2THF;
it should be remembered that T H F is quite a strong base
toward magnesium.
Little is known about the structures and properties of
the dialkylmagnesium compounds or magnesacycles corresponding to the “short” di-Grignard compounds 10, 18,
and 7. They are in general only slightly soluble and appear
to be oligomeric or even polymeric. This is in fact understandable, since on the basis of experience with larger systems the formation of smaller rings is extremely unlikely
(Table 2).
As already mentioned in Section 2.2.3, 10 can be freed
from most of the magnesium bromide by precipitation and
washing, though it is not possible in this way to obtain
bromine-free methylenemagnesium (58).[32b1However,
Ziegler et al. have prepared 58 as a white insoluble powder
by pyrolyzing d i m e t h y l m a g n e s i ~ m . ‘ ~ ~ ~
n Mg(CH3)>-% nCH4
+ [MgCHJ,
It is also worthy of mention that the bromide- containing 58 obtained by us gives better results in some applications than does 10, for example in the synthesis of a 1,3d i g e r m a c y c l ~ b u t a n (see
e ~ ~ Section
The dialkylmagnesium compound 59 obtained from the
di-Grignard 18 also consists of a mixture of oligomers.
This is suggested by the following observation: while 59 is
almost insoluble in THF, it dissolves in [D,,]HMPA (hexamethylphosphoric acid triamide). The a-protons in the ’HNMR spectrum of 59 occur as a broad signal at 6- - 1.7
(signal width ca. 1 ppm). When the HMPA solution is allowed to stand for several days, a weak unsymmetrical
multiplet is observed which overlaps the broad band; eventually the solution decomposes slowly with a half-life of
ca. 16 days.139.521
We assume that 59 is polymeric or highly
oligomeric in the solid state; when it dissolves in HMPA,
smaller oligomeric units with n > 3 are slowly formed.
The dialkylmagnesium compound 60 corresponding to
the di-Grignard 7 is also fairly insoluble in T H F and can
be obtained in pure form when the impure 7 obtained
from 1,3-dibromopropane is evaporated and the magnesium bromide washed out, together with other impurities
[Eqn. (h)], using small portions of T H F ; large volumes of
T H F cannot be used, since otherwise too much 60 will be
lost as a THF-soluble complex of 7 with MgBr, (see Section 2.2.3). We thus have a preparatively simple method for
the formation of very pure 7 : the pure 60, a dry white
powder obtained by decanting off the THF, is treated with
one molar equivalent of magnesium bromide in diethyl
ether; since in this solvent the Schlenk equilibrium lies on
the side of the Grignard reagent (see Section 2.2.1), 7 is
re-formed quantitatively and dissolves [Eqn. (r)].IS4l
+ impurities
0 7 9
4. Applications of Di-Grignard Compounds
4.1. “Normal” Di-Grignard Compounds
As has already been noted briefly in Sections 1.2 and
2.1.1, the direct synthesis of readily available “normal” diGrignard compounds with four or more carbon atoms between the magnesium functionalities was used early on for
’the synthesis of both acyclic and more particularly cyclic
compounds, especially in organometallic chemistry.’”,52’
Apart from the syntheses already m e n t i ~ n e d , [ * ~the
- * ~pre~
paration of metallacyclic transition metal compounds by
the groups of Whitesides (Ti,[761PtI7’I) and Grubbs (Pt,’781
Ni[791)are worthy of mention as examples of more recent
developments; the same holds for our own work on the
synthesis of heterocycles of main group element^.[^^*^^]
curs, giving the magnesium salt 61 of cyclobutanone hydrate; this in turn gives cyclobutanone 62 on hydrolysis.[821
Although the yield of 62 is only ca. 30% (with respect to
7), this method for its preparation is one of the simplest
and cheapest. It is also simple to execute, since crude 7
[Eqn. (h)] can be used; 61 is the only product in the reaction mixture which is rather insoluble, so that all by-products can be decanted off prior to hydrolysis; 62 can be obtained by simple distillation after the hydrolysis of 61.
The greatest limitation for the use of 7 is unfortunately
to be found in its reactions with carbonyl compounds and
in particular with ketones. Here the double hydridic activation of the P-hydrogens takes the upper hand, and for
example the reaction with benzophenone yields only its reduction product benzhydrol and the tertiary alcohol 1,ldiphenyl-3-buten-1-01, the latter arising from addition of
the ally1 derivative 24 (which is formed as an intermediate)
to benzophenone.
4.2. “Short” Di-Grignard Compounds
The 1,l-di-Grignard reagent 10 has been known since
1926,1301and its application in organic synthesis has already been s t ~ d i e d . ~However,
~ ’ . ~ ~ ~apart from one example,‘*I1its potential for the synthesis of organometallic reagents had not been explored. This is naturally also true for
the recently prepared 1,2- and 1,3-reagents. In the following discussion we shall reverse the sequence used in Sections 2 and 3 and start with the applications of the 1,3di-Grignard reagents 7 and 28, since they precede those of
10 and, as will be shown, have contributed conceptually to
certain developments of the chemistry of 10.
Problems of this type d o not arise with 28, as demonstrated by the reaction with cyclohexanone, which leads to
the expected diol 63.
4.2.1. Organic Compounds
from 1,3-Di-Grignard Compounds
We shall first refer to some applications of 7 and 28 in
organic synthesis; these will show some interesting possibilities, but also clear limitations. 7 reacts readily with
hard nucleophiles such as water and CO,; as expected,
propane and glutaric acid are formed.[531
Glutaric acid is obtained as the main product only by
pouring 7 onto dry ice. When, however, gaseous carbon
dioxide is slowly passed into the reaction vessel, the wellknown[20,241
double reaction with one molecule of C 0 2 oc1000
4.2.2. Acyclic 1,3-Dirnetal Compounds
from 1,3-Di-Grignard Compounds
The reactions of 7 and 28 with chlorotrimethylstannane
and mercuric bromide are very clean, and can be used for
the determination of the yields of the di-Grignards.
The 1,3-bis(bromornercurio)propanes 64 and 65 in turn
offer interesting reaction possibilities, as will be shown using two examples. The two-step reaction with butyllithium
afforded the 1,3-dilithiopropanes 66 and 67,’83’which had
Angew. Chern. Inr. Ed. Engl. 26 (1987) 990-1005
7 : R = H
28: R = M e
64: R = H
their postulated o r demonstrated role as catalytic intermediates in important industrial processes such as olefin metathesis.134,57.87.n81
They have thus been extensively studied,
although until recently no general route to these compounds was available. The proven di-Grignard route to
larger rings suggests itself: as shown by Equation (s), both
main group and transition metal element dichlorides 72
should in principle lead to metallacyclobutanes 73.
65: R = Me
previously been found to be inaccessible by other
- 2 LiRr
7 : R = H
t BuHgCH2CR2CH2Hg tBu
28: R
= Me
64: R = H
2 tRuLi
65: R = Me
(R = H)
- 2 (tBu)#g
66: R = H
67: R = Me
Compound 66 exhibits an even more extreme a-hydride
activation than that described for 7 and splits off lithium
hydride even at -60°C to give allyllithium (68).
In contrast, compound 67 is extraordinarily stable for
an organolithium compound in diethyl ether; according to
calculations by Schleyer et al., this could be due to the existence of a structure with double lithium bridges.18,’0,841
A second interesting application of 64 and 65 lies in
their thermolysis; this occurs with elimination of mercurous bromide, presumably giving trimethylene (69) or its
2,2-dimethyl derivative 70.1s3,851
Interestingly, the product
spectrum of the diradicals obtained in this manner differs
strongly from those which are normally found for 69 and
This applies not only to the relatively low yield of
cyclopropane (or its dimethyl derivative), but also particularly to the unusual formation of 71, which is perhaps due
to a 1,2-methyl migration. A possible explanation could be
that the diradicals are formed in a vibronically excited
64: R = H
69: R = H
65: R = Me
70: R = Me
As is so often the case, this new procedure is not so universally applicable as had been hoped. The greatest,
though trivial, problem is the in some cases inherent instability of 73, e.g. for M=Ni.[891 However, the route via
Equation (s) is useful even in such a case; one can postulate with more confidence than otherwise that 73 is in fact
formed and that the secondary products actually detected
d o arise from 73.
On the other hand, it can happen that the yields of 73
are low although 73 itself is in fact stable.[901The reason is
then to be found in a particularly strong activation of the
j3-hydrogens in the intermediate 74 by certain transition
metals M. While 73 is stable, because the orientation necessary for a-elimination is not possible in the inflexible
four-membered ring:”] 74 can by rotation adopt conformations in which the elimination of 75 is readily possible.
In the case of 28, such a complication cannot occur, so
that better results have been obtained using this reagent.
Table 3 gives an overview of the results so far obtained.
It is outside the scope of the present review to go into
details of the many interesting aspects of the various reacTable 3. Metallacyclobutanes of type 73 (from 1.3-di-Grignard compounds).
6 9 +
7 1 (4%)
4.2.3. Metallacyclobutanes
from 1,3-Di-Grignard Compounds
Metallacyclobutanes are of great interest because of
their structural and bonding properties and because of
Angew Chem. In1
Ed. Engl. 26 (1987) 990-100s
Yield [%]
30 [a1
[a] A different decomposition reaction of the intermediate 74 is observed In
this case 1561. [b] C p * = C5Mes.
tions and compounds; here we must refer to the original
literature (Table 3). We should, however, mention that
many of the metallacyclobutanes in Table 3 (M = Hf, Zr, V,
Re, Sn) have been prepared for the first time using this
method: the hafnium compound has since been obtained
by a different procedure. The gaps in Table 3 make it clear
that much valuable work remains to be done in this area.
A first example is provided by the spiro compounds 76
[Eqn. (t)J.[”’
The chemistry of these compounds cannot be discussed
here; there is however a certain amount of overlap with
that of 10, so that the decision as to which reagent should
be used for a certain reaction is often one of small nuances
and in particular of preparative availability. Developments
in the coming years will show which approach is the most
practical; the recently improved preparation of
certainly put it in a more competitive position.
4.2.5. I , 3- Dimetallacyclobutanes
76a: M = Si (50%)
76b: M = Ge (80%)
The development of new 1,3-di-Grignard reagents has
opened u p other possibilities. Apart from the examples to
be discussed in Section 4.2.4, the reagent 38 [Eqn. (i)]
should be mentioned here; it has been used as the starting
material for the preparation of the new metallacyclobutabenzenes 77[941
and 7SC9’’
and of corresponding derivatives
with Si, Ge, Sn o r Pb.[89h1
79 (87%)
80 (52%)
(co. 5 d)
81 (55%)
It should also be mentioned that the last ten years have
seen the development of the chemistry of the 1,l-dimetal
compounds (“methylene complexes”) on a broad
10.36.95- 1001 .
in this connection the by now classical
Tebbe reagent 82[95.961
and the dizinc reagent 83[971are
worthy of particular mention.
b: M = Sn
We have already referred to the possibility of using the
methylene di-Grignard reagent 10 in organic synthesis in
Section 4.2.L31.32.8‘1
Its use for the preparation of other dimetallomethanes requires only a brief mention; in some
cases such compounds are available via other routes. A
certain disadvantage is the markedly reduced reactivity of
10 in comparison with other Grignard compounds; thus,
while the formation of 79 and 80 occurs readily, 81 is produced only slowly.[32h1
a: M = Ge
4.2.4. The Methylene Di-Crignard Reagent
] Cp’ReOCI,
CH2Mg . CH2(MgBr)z
The reactions of 10 with Me2GeC1, and Me,SnCI, have
so far been successful only in the case of the 1,3-digermacyclobutane 84a ; it is not clear whether the tin analog 84b
was not formed o r whether it merely decomposed very
rapidly.[32h’The use of the only slightly soluble and largely
bromide-free modification 44 gave considerably better
yields of the four-membered ring Ma; the interesting
higher ring-homologues 85 and 86 were also obtained.
1 2%
The reactions which have been studied in the most detail
are those with cyclopentadienyl derivatives of titanium,
zirconium, and hafnium. As the example of the titanocene
dichloride [Eqn. (n)] shows, either one or both chloride
functions can be allowed to react with 10, giving the intermediates 87 and 88, respectively.
The behavior of 87 will be discussed in Section 4.2.6.
Compound 88 can be regarded as a di-Grignard compound in which carbon atom 2 in 7 or 28 has been replaced by the titanocene moiety. It can be used in a completely analogous manner (see Section 4.2.3) for syntheses
of four-membered rings [Eqn. (v)]. The same is probably
true for 91 and 92, which are obtained from zirconocene
and hafnocene dichlorides [Eqn. ( w ) ] . ~ ~ ~ ]
Only two of the nine 1,3-dimetallacyclobutanes described in Equations (v) and (w), 89a11011
and 90a,[’’’’ had
previously been prepared in other ways.
It must be noted that some of the four-membered ring
compounds were obtained in only poor yields. This is for
Angew. Chem.
Ed. Engl. 26 (1987) 990-1005
90a: M‘ = Ti
90b: M‘ = Z r
9Oc: M’ = Hf
89a: M = Si
89b: M = Ge
89c: M = Sn
example the case for 89c, which, however, being a stannacyclobutane is inherently unstable (see Ref. 1921). It also
applies for 94 and 95; in these cases the reasons lie in
practical difficulties such as the solubility of 91 and the
presence of zirconocene dichloride as a n impurity in the
hafnocene analogue. We have also since been able to show
that the published yields can be improved by modification
of the reaction conditions (for example in the case of 95
to 24%13”). A discussion of the interesting
spectral and bonding properties of the 1,3-metallacyclobutanes lies outside the scope of the present survey.1331
It should, however, be mentioned here that the analogy
between the reagents 28 and 88 also applies to the synthesis of the spiro compounds 96,”03] which can be described
as double 1,3-dimetalla-analogues of the double metallacyclobutane 76 [Eqn. (t)].
r M g B r
2 CpzTi
a: M
b: M
= Si,
Ge, c: M
= Sn
versibly by alkenes to give titanacyclobutanes 98 or by carbony1 compounds to give 1-alkenes 99.
The titanium/zinc compound
obtained from the
dizinc reagent 83 and compound 87 obtained by us from
10 [Eqn. (u)] undergo completely analogous reactions.
Free 97 is extremely unstable and is present in only a very
+ R‘R‘C=O
X = C l or I
low steady state concentration; it can, however, be stabilized at low temperatures as 101 by complexation with tertiary phosphanes. Apart from 82[’041and 98,[’O5]compound
90a (obtained according to Equation (v)) has been shown
to be a good starting material for this reaction [Eqn. (y)].
CpzTi, A
p M e Z 82
Compounds such as 82, 87, 100 and 90a can thus be
regarded as “precarbenes” o r as metallaalkenes stabilized
by other metal compounds, Me,AICI, MgBrC1, ZnX, or
Cp2Ti=CH2 (97), respectively, acting as “stabilizers”. Such
an approach is justified not only by the reactivity of these
compounds [Eqn. (y)] but also by theoretical calculat i o n ~ “ and
~ ~ ]by their ‘H- and I3C-NMR spectra; the carbenoid carbon and its attached protons are generally very
strongly deshielded (Table 4).
4.2.6. Metallaalkenes
Metallaalkenes or metal-carbene complexes contain the
structural element L,M=CR2 and are, like the metallacyclobutanes, objects of great interest.[871A point of contact
between the two types of compound can be found in olefin
metathesis,1xxJfor which a mechanism according to Equation (x) is now generally accepted.
Table 4. Chemical shifts of the methylene hydrogen and carbon nuclei of
methylenetitanium compounds.
XTi(Cp2)CHzZnX [bl
CH2Ti(Cp,)CH1TiCp2 [a]
As Tebbe et al.[’’] and Grubbs et al.Ly61
have shown, the
Tebbe reagent 82 provides an easy route to the titanaethene 97; this can, when prepared in situ, be trapped reInt
101, R
Angew Chern
Ed Engl 26 (1987) 990-1005
Me (PMe3)Ti(Cp2)=CHZId]
98, R = H
C H ~ T I ( C ~ ~ ) C H [b]
[a] In [D6]benzene.[b] In [Ds]toluene. [c] In [D,o]EtZO/[D,]benzene 1: I.[d] In
The different N M R behavior of 87, the structure 87a of
which is at a first glance completely analogous to those of
82 and 100, shows that there are still problems to be
solved in this area. Very recent results'371have surprisingly
shown that 87 can be converted into 102 when a solution
of 87 in etherlbenzene is evaporated to dryness and the
residue so obtained is dissolved in toluene.
Compound 102 then exhibited the expected low-field
shift for the methylene protons (Table 4) and showed itself
to be a very good starting material for 97.The unexpected
NMR data for 87 are probably due to the fact that in the
ether/benzene solution, which contains excess magnesium
halide, 87 exists as 87b or 87c and not as 87a (for the sake
of clarity the ether molecules are not shown; both C1 as
well as Br are symbolized by X).
It remains astounding that structures such as 87 and 102
are in fact relatively stable, although the differences appear subtle when one considers that bonds between magnesium and carbon or halogen normally undergo rapid exchange in ethereal
It is also surprising that
these On paper
have wide-reaching consequences for the bonding situation, as must be concluded from the very different NMR
data for the methylene protons (Table 4). Further studies
will undoubtedly shed more light on this important question.
Manv o f the results discussed here were obtained in our
own laboratories; they are due maink to the enthusiasm, the
hard work and the great skill of my co-workers, whose names
can be found in the list o f references. I am very thankful to
all of them, but in particular to Dr. 0.S . Akkerman and Dr.
C. Blomberg, who have been with me for many years, and to
Mr' Gerrit
the Often
work by his immeasurable experimental ability and passed
on oractical exoerience and traditions in the arouu to many
generations of undergraduate and graduate students. I
thank the foundation for Chemical Research in the Netherlands (S.O.N.) for supporting our work with funds from
the Organization for Fundamental Scientific Research
( Z . W.O.).
, . I
" .
Received: February 12, 1987 [A 639 IE]
German version: Angew. Chem. 99 (1987) 1020
[I] G. Wittig, Angew. Chem. 70 (1958) 65.
121 G. Wittig, R. Ludwig, R. Polster, Chem. Ber. 88 (1955) 294.
[3] G. Wittig, F. Bickelhaupt, Chem. Ber. 91 (1958) 865.
[4] L. Vecchiotti, Ber. Drsch. Chem. Ges. 63 (1930) 2275.
151 G. Wittig, F. Bickelhaupt, Chem. Ber. 91 (1958) 883.
[6] While according to 151 the degree of oligomerization of 3 is n =6, later
measurements gave the value n = 3 : a) D. S. Brown, A. G. Massey, D.
A. Wickens, Acta Crystallogr. Sect. 5 3 4 (1978) 1695; b) N. A. A. AlJabar, A. G. Massey, J . Organornet. Chem. 275 (1984) 9.
171 a) H. J. S . Winkler, G. Wittig, J. Org. Chem. 28 (1963) 1733; b ) R. E.
Humphries, N. A. A. AI-Jabar, D. Bowen, A. G. Massey, G. B. Deacon,
J. Organornet. Chem. 319 (1987) 59
[8] P. von R. Schleyer, Pure Appl. Chem. 56 (1984) 151.
191 A. Streitwieser, Jr., Arc. Chem. Res. 17 (1984) 353.
[lo] A. Maercker, M. Theis, Top. Curr. Chem. 38 (1987) I .
[ I I ] For reviews see: a) Houben- Weyl-Miiller: Methoden der Organischen
Chemie. Band X l I l . Thieme, Stuttgart 1970-1978: b) J. P. Collman, L.
S. Hegedus: Principles and Application of Organotransrtion Metal Chemistry, University Science Books, Mill Valley, CA (USA) 1980: c) G.
Wilkinson, F. G. A. Stone, E. W. Abel: Comprehensive Organometallic
Chemistry. Pergamon Press, Oxford 1982.
[I21 F. Bickelhaupt in J. J. Zuckerman (Ed.): lnorganic Reactions and Methods, Yo/. I#. VCH. Weinheim, in press.
1131 V. Grignard, C. R. Hebd. Seances'Acad. Sci. 130 (1900) 1322.
[I41 L. Tissier, V. Grignard, C. R. Hebd. Seances Acad. Sci. 132 (1901)
[IS] a) F. B. Ahrens, A. Spaler, Ber. Dtsch. Chem. Ge.9. 38 (1905) 1296: b) C.
A. Bishoff, ibid. 38 (1905) 2078: c) Z. Zelinsky, J. Gutt, ibid. 40 (1907)
[I61 W. H. Saunders, Jr., A. F. Cockerill: Mechanism of'Elimination Reactions. Wiley, New York 1972.
[I71 a) Geromont, Justus Liebigs Ann. Chem. 158 (1871) 370; b) Freund,
Monatsh. Chem. 2 (1881) 639; c) Gustavson, J. Pr. Chem. [2] 62 (1900)
[IS] J. von Braun, Ber Dtsch. Chem. Ges. 37 (1904) 3210.
1191 J. von Braun, A. Steindorff, Ber. Drsch. Chem. Ges. 38 (1905) 956.
1201 V. Grignard, G. Vignon, C. R. Hebd. Seances Acad. Sci. 144 (1907)
1358; see also 1211.
[21] It is not clear whether the preparation of the 1,9-di-Grignard reagent of
n-nonane was achieved earlier; it is claimed without details by E. E.
Blaise, Houillon, Bull. Soc. Chim. Fr. [3] 31 (1904) 960.
[22] J. von Braun, Ber. Dtsch. Chem Ges. 40 (1907) 4065.
[23] One should note that seven years had passed since the discovery of the
Grignard reagents, and three years since the discovery of 8 !
1241 J. von Braun, W. Sobecki, Ber. Dtsch. Chem. Ges. 44 (191 I ) 1918.
1251 J. von Braun. Ber. Dfsch. Chem. Ges. 46 (1913) 1782:. 47 (1914)~ 490.
[26] S. Hilpert, G. Griittner, Ber. Dtsch. Chem. Ges. 47 (1914) 186.
[27] a) S. Hilpert, G. Gruttner, Ber. Dtsch. Chem. Ges 47 (1914) 177: b) G.
Griittner, M. Wiernik, rbid. 48 (1915) 1473: c) G. Griittner, E. Krause,
ibid. 49 (1916) 437, 2666: d) G. Griittner, E. Krause, M. Wiernik, rbid.
50 (1917) 1549; e) compare also the early example of the synthesis of
1, I-dichlorosilacyclohexane[28].
[28] A, Bygden, Ber. Dtsch. Chem. Ges, 48 (1915) 1236.
1291 M. S. Kharash, 0. Reinmuth: Grrgnard Reactions of Nonmetallic Substances. Prentice-Hall, New York 1954.
[30] G. Emschwiller, C. R. Hebd. Seances Acad. Sci. 183 (1926) 665.
[31] F. Bertini, P. Grasselli, G. Zubiani, G. Cainelli, Tetrahedron 26 (1970)
[32] a) J. W. Bruin, G. Schat, 0. S. Akkerman, F. Bickelhaupt, Tetrahedron
Lett. 24 (1983) 3935; b) J. W. Bruin, G. Schat, 0. S. Akkerman, F. Bickelhaupt, J . Orgonomet. Chem. 288 (1985) 13.
[331 B. J. J. van d e Heisteeg, G . Schat, 0. S. Akkerman, F. Bickelhaupt, J .
Organomet. Chem. 308 (1986) 1.
1341 F. Bickelhaupt, Pure Appl. Chem. 58 (1986) 537.
1351 w. Bruin, unpublished results,
1361 B. J. J. van de Heisteeg. G. Schat. M. A. G. M. Tinaa.
- . 0. S. Akkerman.
F. Bickelhaupt, Tetrahedron Lett. 27 (1986) 6123.
1371 B. J. J. van d e Heisteeg, Dissertation, Vrije Universiteit Amsterdam
[38] K, B, Wiberg, W. J ,
J Am, Chem. Sot. 82 (1960) 6375.
1391 J. W. F. L. Seetz, 0. S. Akkerman, F. Bickelhaupt, Tetrahedron Lett. 22
, 4857.
I401 H. M. Walborsky, Tetrahedron 37 (1981) 1625.
[41] T. Clark, A. J. Kos, P. von R. Schleyer, W. P. Cofino, W. H. de Wolf, F.
Bickelhaupt, J. Chem Sac. Chem. Commun. 1983. 685.
1421 1. T. Millar, H. Heaney, Q. Rev. Chem. Sac. I 1 (1957) 109.
[43] H. Heaney, Organomet. Chem. Rev. l(1966) 27.
[44] G. Wittig, Angew. Chem. 69 (1957) 245.
[45] H. Heaney, F. G. Mann, I. T. Millar, J. Chem. Soc. 1956. 4692: 1957.
[46] F. Bickelhaupt, C. Jongsma, P. de Koe, R. Lourens, N. R. Mast, G. L.
van Mourik, H. Vermeer, R. J. M. Weustink, Tetrahedron 32 (1976)
[47] 0. S. Akkerman, unpublished results.
1481 F. A. Hart, F. G. Mann, J . Chem. SOC.1957. 3939, and references cited
Angew. Chem. lnt. Ed. Engl. 26 (1987) 990-1005
1491 L. C. Costa, G. M. Whitesides, 3. Am. Chem. SOC.99 (1977) 2390.
[50] C. Blomberg, F. A. Hartog, Synthesis 1977. 18.
[51] F. A. Hartog: The Barbier Reaction, A Mechanistic Study. Dissertation,
Vrije Universiteit Amsterdam 1978.
[52] J. W. F. L. Seetz: Bifunctionele organometanluerbindingen met korte
koolstofketens en metallacyclobutonen. Dissertation. Vrije Universiteit
Amsterdam 1983.
1531 J. W. F. L. Seetz, F. A. Hartog, H. P. Bohm, C. Blomberg, 0. S. Akkerman. F. Bickelhaupt, Tetrahedron Lett. 23 (1982) 1497.
[54] J. W F. L. Seetz, B. J. J. van d e Heisteeg, G. Schat, 0. S. Akkerman, F.
Bickelhaupt, J . Organomet. Chem. 277 (1984) 319.
[ S S ] For the free-radical nature of the reaction see: a) H. M. Walborsky, M.
S . Aronoff, J . Organomet. Chem. 51 (1973) 31; b) B. J. Schaart, C.
Blomberg, 0. S . Akkerman, F. Bickelhaupt, Con. J . Cheh. 58 (1980)
932: c ) H. R. Rogers, C. L. Hill, Y. Fujiwara, R. J. Rogers, H. L. Mitchell, G. M. Whitesides, J . Am. Chem. SOC.102 (1980) 217, and references cited therein.
[56] J. W. F. L. Seetz, B. J. J. van d e Heisteeg, G. Schat, 0. S. Akkerman, F.
Bickelhaupt, J . Orgonomet. Chem. 275 (1984) 173.
1571 J. W. F. L. Seetz, B. J. J. van de Heisteeg, G. Schat, 0. S. Akkerman, F.
Bickelhaupt, 3. Mol Catal. 28 (1985) 71.
I581 P. Markies, unpublished results.
[59] M. B. Smith, W. E. Becker, Tetrahedron 23 (1967) 4215.
I601 H. C Holtkamp, G. Schat, C. Blomberg, F. Bickelhaupt, J . Organomet.
Chem. 240 (1982) I .
[61] A. L. Spek, G. Schat, H. C. Holtkamp, C. Blomberg, F. Bickelhaupt, J .
Organomet. Chem. 131 (1977) 331.
[62] M. B. Smith, W. E. Becker, Tetrahedron 22 (1966) 3027.
[63] H. C Holtkamp, C. Blomberg, F. Bickelhaupt, J . Organomet. Chem. 19
(1969) 279.
1641 C . Blomberg, G. Schat, H. H. Grootveld, A. D. Vreugdenhil, F. Bickelhaupt, Justus Liebrgs Ann. Chem. 763 (1972) 148.
[65] F. J. M. Freijee, G. van der Wal, G. Schat, 0. S. Akkerman, F. Bickelhaupt, J. Organornet. Chem. 240 (1982) 229.
[66] J. Toney, G. D. Stucky, J. Organomet. Chem. 28 (1971) 5 .
[67] P. Voorbergen, C. Blomberg, F. Bickelhaupt, J. Organornet. Chem. 40
(1972) 225.
[68] W. Schlenk, W. Schlenk jun., Ber. Dtsch. Chem. Ges. 62 (1929) 920.
1691 A. D. Vreugdenhil, C. Blomberg, Red. Trau. Chim. Pays-Bas 82 (1963)
453, 46 I .
1701 F. W. Walker, E. C. Ashby, J . Am. Chem. Soc. 91 (1969) 3845.
[7 I ] Lower degrees of association are given in the literature for other solvents; it is unclear whether a solvent effect is involved or whether the
reason lies in experimental difficulties in handling these extremely sensitive compounds: 9. Denise, J. Ducorn, F. J. Fauvarque, Bull. SOC.
Chim. Fr. 1972, 990.
[72] F. J. M. Freijee, G. Schat, 0. S . Akkerman, F. Bickelhaupt, J. Organomet. Chem. 240 (1982) 217.
[73] H. A. Bent, Chem. Rev. 61 (1961) 275.
[74] a) M. Vallino, ThCse, Universite de Paris V1 1972: b) on the other hand,
O - C ~ H ~ ( C H ~ ) ~ M is
~ . trimeric
~ T H F in the crystal; this has been explained on the basis of Bayer strain in the magnesahydrindane skeleton: M. F. Lappert, T. R. Martin, C. L. Raston, B. W. Skelton, A. H.
White, J . Chem. SOC.Dalton Trans. 1982. 1959.
[7S] K. Ziegler, K. Nagel, M. Patheiger, Z . Anorg. Allg. Chem. 282 (1955)
[76] J. X. McDermott, G. M. Whitesides, J. Am. Chem. SOC.96 (1974) 947.
[77] J. X. McDermott, J. F. White, G. M. Whitesides, J. Am. Chem. SOC.98
(1976) 6521.
[78] C. G. Bielefeld, H. A. Eick, R. H. Grubbs, Inorg. Chem. 12 (1973)
[79] R. H. Grubbs, A. Miyashita, M. Liu, P. Burk, J . Am. Chem. SOC.100
(1978) 2418.
[SO] C. Jongsma, J. J. de Kok, R. J. M. Weustink, M. van der Ley, J. Bulthuis, F. Bickelhaupt, Tetruhedron 33 (1977) 205.
Angew. Chem. lnt. Ed. Engl. 26 (1987) 990-1005
[Sl] M. M. Koton, T. M. Kiseleva, Zh. Obshch. Khim. 35 (1965) 2036.
[82] J. W. F. L. Seetz, R. Tol, 0. S. Akkerman, F. Bickelhaupt, Synthesis
1983, 721.
[83] J. W. F. L. Seetz, G. Schat, 0. S. Akkerman, F. Bickelhaupt, J . Am.
Chem. SOC. 104 (1982) 6848.
1841 P. von R. Schleyer, A. J. Kos, E. Kaufmann, J . Am. Chem. SOC. I05
(1983) 7617.
[SS] M. A. G. M. Tinga, B. J. J. van d e Heisteeg, unpublished results.
[86] R. G. Bergman in J. K. Kochi (Ed.): Free Radicals, Vol. 1, Wiley, New
York 1973, p. 191.
187) K. H. Dotz, H . Fischer, P. Hofrnann, F. R. Kreissl, U. Schubert, K.
Weiss: Transition Metal Curbene Complexes, Verlag Chemie, Weinheim
[88] K. J. Irvin, Olefin Metathesis, Academic Press, London 1983.
1891 a) R. H. Grubbs, A. Miyashita, J . A m . Chem. SOC.100 (1978) 7418; b)
H. J. R. d e Boer, unpublished results; c) A. Miyashita, H. Shitara, H.
Nohiro, J . Chem. SOC.Chem. Commun. 1985, 850,and references cited
1901 J. W. F. L. Seetz, G. Schat, 0. S. Akkerman, F. Bickelhaupt, Angew.
Chem. 95 (1983) 242; Angew. Chem. I n t . Ed. Engl. 22 (1983) 248; Angew. Chem. Suppl. 1983, 234.
[91] H. J. R. d e Boer, B. J. J. van d e Heisteeg, M. Floel, W. A. Herrmann, 0.
S. Akkerman, F. Bickelhaupt, Angew. Chem. 99 (1987) 88: Angew.
Chem. Int. Ed. Engl. 26 (1987) 73.
[92] J. W. F. L. Seetz, G. Schat, 0. S . Akkerman, F. Bickelhaupt, J . Am.
Chem. SOC.105 (1983) 3336.
[93] B. J. J. van de Heisteeg, M. A. G. M. Tinga, Y. van den Winkel, 0. S .
Akkerman, F. Bickelhaupt, J. Organomet. Chem. 316 (1986) 51.
1941 H. J. R. de Boer, 0. S. Akkerman, F. Bickelhaupt, G. Erker, P. Czisch,
R. Mynott, J. M. Wallis, C. Kriiger, Angew. Chem. 98 (1986) 641 ; Angew. Chem. lnt. Ed. Engl. 25 (1986) 639.
1951 a) F. N. Tebbe, G. W. Parshall, G. S. Reddy, J. Am. Chem. SOC. 100
(1978) 3611; b) U. Klabunde, F. N. Tebbe, G. W. Parshall, R. L. Harlow, J . Mol. Catal. 8 (1980) 37.
I961 a) J. B. Lee, K. C. Ott, R. H. Grubbs, J. Am. Chem. SOC.104 (1982)
7491; b) S. H. Pine, R. Zahler, D. A. Evans, R. H. Grubbs, ibid. 102
(1980) 3270.
I971 a) J. J. Eisch, A. Piotrowski, Tetrahedron Lett. 24 (1983) 2043: b) A. M.
Piotrowski, J. J. Eisch in R. B. King, J. J. Eisch (Eds.): Organometallic
Syntheses, Vol. 3. Elsevier, Amsterdam 1986, p. 16.
[98] W. A. Herrmann, Angew. Chem. 94 (1982) 118;Angew. Chem. lnt. Ed.
Engl. 21 (1982) 117; J . Organomet. Chem. 250 (1983) 319.
I991 J. Hibino, T. Okazoe, K. Takai, H. Nozaki, Tetrahedron Lett. 26 (1985)
[I001 P. Knochel, J. F. Normant, Tetrahedron Lett. 27 (1986) 1039, 1043.
[ l o l l a) W. R. Tikkanen, J. Z. Liu, J. W. Egan, Jr., J. L. Petersen, Organometnllics 3 (1984) 825; b) W. R. Tikkanen, J. W. Egan, Jr., J. L. Petersen,
ibid. 3 (1984) 1646.
11021 K. C. Ott, R. H. Grubbs, J . Am. Chem. Sor 103 (1981) 5922.
[I031 a) M = Si: B. J. J. van de Heisteeg, G. Schat, 0. S . Akkerman, F. Bickelhaupt, Organometallics 5 (1986) 1749; b) M = Ge, Sn: see [37].
11041 a) J. Schwartz, K. I. Gell, J . Organornet. Chem. 184 (1980) C 1: b) F. W.
Hartner, Jr., J. Schwartz, K. 1. Gell, J . Orgonomet. Chem. 184 (1980)
C 1 ; b) F. W. Hartner, Jr., J. Schwartz, S. M. Clift, J . Am Chem SOC.
105 (1983) 640; S . M. Clift, J. Schwartz, J . Orgunornet. Chem. 285
(1985) C 5 .
[I051 L. R. Gilliom, R. H. Grubbs, Organometallrcs 5 (1986) 721.
[I061 B. J. J. van d e Heisteeg, G. Schat, 0. S. Akkerman, F. Bickelhaupt, J.
Orgonomet. Chem. 310 (1986) C25.
[I071 a) M. M. Francl, W. J. Pietro, R. F. Hout, W. J. Hehre, Organometallics
2 (1983) 281,815; b) T. H. Upton, A. K. Rappe, J . Am. Chem. SOC.107
(1985) 1206, and references cited therein; c) F. Volatron, 0. Eisenstein,
ibrd. 108 (1986) 2173, and references cited therein.
[IOS] J. B. Lee, G. J. Gajda, W. P. Schaefer, T. R. Howard, T. Ikarya, D. A.
Straws, R. H. Grubbs, J . Am Chem. Sac. 103 (1981) 7358.
Без категории
Размер файла
1 529 Кб
metallacyclus, compounds, grignard
Пожаловаться на содержимое документа