вход по аккаунту


Novel Reactions of Organomolybdenum and Organotungsten Compounds AdditiveЦReductive Carbonyl Dimerization Spontaneous Transformation of Methyl Ligands into -Methylene Ligands and Selective Carbonylmethylenation.

код для вставкиСкачать
Novel Reactions of Organomolybdenum and Organotungsten Compounds:
Additive -Reductive Carbonyl Dimerization, Spontaneous Transformation
of Methyl Ligands into p-Methylene Ligands, and Selective
Thomas Kauffmann"
Hitherto there was no reaction known
that permits transformations of R'R2CO + 0.5 R1RZR3C-CR1RZR3in one
step. This type of additive-reductive
carbonyl dimerization is now possible
using alkoxy(alkyl)tungsten(v) complexes with aromatic, heteroaromatic,
or cQ-unsaturated aldehydes and ketones. When a corresponding phenyl
complex was employed in a test experiment, it was revealed that an aliphatic
ketone could be used as the substrate in
this reaction. A second interesting type
of reaction is the transformation of CH,
ligands into p-CH, ligands, which occurs during the treatment of MeLi or
Me,AI with molybdenum or tungsten
chlorides (oxidation states VI and v, for
Mo additionally IV) at low temperatures
with liberation of CH,. Here, the question arises as to whether the intermediate involved has a terminal CH, ligand
(Schrock cdrbene complex) or a p-CH,
ligand (CH, bound by a two-electron
three-center bond to two metal atoms).
Of all the p-CH, complexes obtained,
those which were synthesized by the action of MeLi on molybdenum chlorides
can be recommended as reagents for carbonylmethylenation of aldehydes and
ketones. They display high selectivity,
very low basicity, a surprising resistance
to protons, they are readily available,
can be easily modified and, as regards
1. Introduction
The large number of transition metals and the resulting virtually immeasurable number of theoretically possible alkyl derivatives necessitates repeated comparative investigations of alkyl
transition metal corn pound^.[^-^] In 1982 my co-workers and I
began research in this field, and concentrated for a prolonged
period on the elements molybdenum and tungsten (other main
topics see refs. [ 5 , 6 ] ) Preceding
this were reports on reactions of
nucleophilic alkylating reagents with WCI,: Thieler7'and other
authors[81had treated WCI, with Me&, Me,Hg, Et,Sn, or
nBu,B and obtained crystalline complexes of RWCI, (R = Me,
Et, nBu), which rapidly decomposed at room temperature. On
the other hand, Wilkinson et aI.E9]had treated WCI, with MeLi
or Me,AI and synthesized Me,W, which explosively decomposes at room temperature to liberate three equivalents of CH,.
Prof. Dr. T Kauffmann
Organisch-chemisches Institut der Universitat
Corrensstrasse 40, D-48149 Munster (Germany)
Fax. Int. code + (251)83-39772
Orgdnomo~ybdenum-and Organotungsten Reagents, Part 7. Part 6: 111.
Angru. C h m . Inr. Ed. En@ 1997, 36, 1258-1275
their selective behavior, they have been
investigated more thoroughly than other readily accessible carbonylmethylenation reagents of comparable selectivity.
The results of NMR spectroscopic investigations on the structure of the pCH, complexes, and associcated reaction mechanisms are discussed. A survey
of carbonylmethylenation reagents,
which have been reported in the literature, permits comparisons to be made
with carbonylmethylenating molybdenum and tungsten complexes.
Keywords: C-C coupling * carbonylmethylenation p-methylene complexes
. molybdenum tungsten
However, analogous experiments to obtain defined compounds
by alkylation of [Mo,CI,,], remained unsuccessful at first.[lol In
1975 Muetterties" 'I found that on treating WCI, with Me,Zn
there is elimination of a-H to liberate CH, (does not contain D
if treated in deuterated solvents), and a species is produced that
can catalyze olefin metathesis and therefore probably contains
a terminal CH, ligand. In analogy to their syntheses of alkylideneniobium(v) and -tantalum(v) complexes. Schrock and
Clark" 21 attempted to synthesize alkylidene complexes by reaction of neopentyllithium with [Mo,CI,,] or WCl,. However,
they obtained dimers of alkylidyne complexes as a result of
double a-H elimination, and so they turned to indirect synthesis
(by transalkylidenation) of alkylidenetungsten complexes.
Eventually Wilkinson et al." 31 obtained the alkylidene complex
along with [(Me,SiCH,),Mo=CSiMe,l after treatment of [Mo,CI,,] with excess
Me,SiCH,Li. Thus, alkyl derivatives of high-valent Mo and W
repeatedly demonstrated a powerful tendency towards a-H
elimination. The publication by Muetterties," ' I mentioned
above as well as in Section 3.4, can be regarded as the starting
point for the investigations described in Sections 3 and 4. On
Veriug.sgesell.schafimhH, 0-69451 Wrmheim, lY97
0570-0833/9713612-125Y $ 17.50+ 5OjO
T. Kauffmann
account of its novelty the additive-reductive carbonyl dimerization discovered during that work is the first topic to be discussed in this review.
2. Additive- Reductive Carbonyl Dimerization
(ARCD Reaction)
Scheme 2. Additive-reductive carbonyl dimerization (A, reagents see Scheme 1
and text) and two further kinds of reductive carbonyl dimerization together with
reagents (B,C).
2.1. Characterization of the Reaction
During investigations as to whether alkoxy groups inhibit the
transformation of a methyl ligand into a p-methylene ligand (see
Section 3), the known Wv complexes l [ I 4 - l 8 l and 2r'6*181
well as the W"' complexes (MeO)4WC1,['91 and (Me0)WOCI,[201were each treated with two equivalents of MeLi per
tungsten atom, and then with 4-methoxybenzaldehyde
(Scheme 1). Whilst in the case of the W"' complexes the expectR
1 : R = Et
2 : R = nPr
2.2. Preparation and Properties of the Reagents
We synthesized[231chlorides 1 and 2 using the two-step
method of Brubaker et al.['5,161from WCI, via [W,C1,,][271
(Scheme 3A). On account of the unsatisfactory yields from the
9 H3
Scheme 1. Discovery of the additive-reductive carbonyl dimerization [21,22a].
Conditions: 1) 4 MeLi, THF, -78 + 20°C in 1 h, 2) 4-MeOC6H,CH0, 3 h at
66'C; meso:ruc =1.1.
ed carbonylmethylenation product was obtained (see Table 4),
treatment of 1 and 2 both produced good yields of equimolar
amounts of meso- and ra~-2,3-bis(4-methoxyphenyl)butane.[~
This reductive coupling, discovered by J. Jordan,[*'.
can be
used with other aromatic aldehydes as well as aromatic ketones,
conjugated enones, and benzoic acid derivatives. Not only that,
the MeLi can be replaced by other organolithium compounds
and Grignard reagents.[22b,22c,
2 3 1 Thus, a new type of reaction
had been discovered, which-for aldehydes and ketones as the
substrates-is described by the general Scheme 2A. This reaction supplements the reductive carbonyl dimerization to produce glycols[241(Scheme 2 B) and that to alkenes using lowvalent titanium[25a1(McMurry reaction; Scheme 2 C) or lowvalent tungsten!25b. 321 Prior to this, in order to convert 3 + 4,
it was necessary to treat 3 with an organolithium or Grignard
compound, the alcohol obtained by hydrolysis was then isolated, followed by reductive dimerization with (3 TiCI, +
LiAIH,) [261.
[W2C14(OnPr)4 (nPrOH),]
Scheme 3 . Synthesis of reagents 1 and 2
second reaction step, Scheme 3 B presents a three- and four-step
synthesis for 1 and 2, respectively, via WC14[281and 5 from
Cotton et a1.,['7.'81 which they regard as the most favorable
pathway. However, they do not give the yields for each step.
Additionally, there is a third synthesis[291 starting from
[W(CO),], which leads to the key compound 5 in three steps (not
all yields given).
Complexes 6, the actual reagents for the ARCD reaction,
were each prepared from the tetrapropoxy complex 2
(Scheme 4), which is produced in somewhat better yields than 1.
Thomas Kauffmann was born in Reutlingen, Germany, in 1924. He began to study chemistry
in 1947 at the Universitat Wiirzburg, and was a student of Clemens Schopfat the Institutfiir
Organische Chemie der Technischen Hochschule Darmstadt. Since 1965 he has been Professor
of Organic Chemistry at the Universitat Munster (Emeritus Professor in 1990). His main
research interests have focused on sodium hydrazide as a reagent in organic synthesis, hetarynes, homocoupling by organocopper compounds, 1 $anionic cycloaddition, preparative application of the areno-analogy principle, organoelernent-lithium and organotransition metal
reagents ,for organic synthesis, chele- and aniicheleselectivity.
Angen.. Chrm. lnr. Ed. Engl 1997. 36, 1258-1275
Orgdnomolybdenum and -tungsten Reagents
4 R L i or
4 RMgX
THF. -78'C
Table 1. Reaction of 6 a with aldehydes and ketones R'R'CO in a 1 : 1 mole ratio
[23].Conditions: THF, - 78 to 66;C in 1 h. then 3 h a t 66 'C. followed by hydrolysis
with 2~ NaOH at 20-C [23] The unconverted carbonyl compound could not be
recovered from any of the reactions.
6a: R = Me; 6b:R= nBu; 6c:R = Ph
Scheme 4. Synthesis of the reagents 6 [23]. Reaction time for MeLi 15 min. for
alkylLi 30 min. for PhLi and Grignard compounds 45-60 min.
During the reaction of 2 with four equivalents of MeLi there is
a change of color from dark red to green with complete consumption of MeLi (detected by ' H and I3C N M R spectroscopy)
to furnish a tetramethyl complex. This complex was assigned the
structure 6a for the reasons which follow: no gas was evolved
on methylation, which points to the preservation of the oxidation state W". The ' H N M R
displays four singlets of almost equal intensity, these arise from the CH, ligands,
and indicate. as in 2, the presence of bridging and nonbridging
propoxy groups with an intensity ratio of 1 :2. The sharpness of
all the N M R signals indicates that the solution is diamagnetic,
and therefore a W-W bond is present in 6a. It is striking that
the transformation of a methyl ligand into a p-methylene ligand
(see Section 3 ) . which has been repeatedly observed for methyl
derivatives of Mo", Mo"', Wv, and Wv', does not take place
with 6a. This is discernible by the lack of evolution of CH4[2zb1
on warming.
In analogy to MeLi, 2 was also treated with four equivalents
of RLi (R = Et, nPr, nBu, iBu, sBu, Me,SiCH,, and Ph) as well
as with some Grignard reagents.[231With the exception of
MeMgBr as the reagent, a color change from dark red to green
occurred. which had already been observed for MeLi (see
above). Therefore, it can be assumed that the transmetalation
process is substantially complete and the individual complexes,
which are analogous to 6 a, have been formed. Correspondingly,
when the green reaction solutions were treated with aldehydes
and ketones, alcohols of type R'R'CHOH or R'R2R3COH
were either not produced a t all or only in small amounts (see
Fable 3). In those cases in which a small amount of alcohol did
form, this could have been due to only the first phase of the
ARCD reaction having taken place. Experiments to isolate 6a
and analogous complexes (see Table 3 ) were not successful
because of the extreme thermolability of these compounds.
The approximate decomposition temperatures of the proposed complexes 6a-c (Scheme 4) were determined to be
- 40,122ht
- 55,r22c1and - 70 0C,[z2c1respectively (see reference [23]). Rapid decomposition occurs at these temperatures,
and the color of the reaction solution changes from green to red.
Slow decomposition of 6a had already been established
at - 78 'C,r22h.
2 3 1 and, even at -40 "C, there is no evolution of
CH,. Nothing is known about the decomposition reaction.
Entry R'R'CO
2.3. Range of Application and Limits of the Reaction
2.3.1. Compound 6 a as a Reagent
lnr llrl Enxl. 1997, 36, 1258-1275
90 [bl
90 IbI
Table 2. Treatment of 6 a with one mole equivalent of the heteroaromatic or x.@-unsaturated cdrbonyl compound [23]. Conditions as In Table 1 : recovery of unconverted carbonyl compound: 0-14%.
benzylidene acetone
with aromatic ketones according to Scheme 2 A to give the ARC D product, generally with a good yield. The reaction with
aldehydes and ketones containing OH, CI, E OMe, or NMe,
groups reveals a great tolerance towards functional groups. This
makes the reaction attractive for synthetic applications, because
difunctionalized, and therefore readily modified A R C D products, are now accessible. One of the difunctionalized products,
2,3-bis(4-hydroxyphenyl)-2,3-dimethylbutane (Table 1, entry
8), exhibits anti-estrogen properties and can inhibit the growth
of mammary tumors.[301The nitro group is not tolerated.[231An
essential limitation of the range of application of the methylating A R C D reaction with 6a (in contrast to the phenylating
A R C D reaction with 6c) is that the carbonyl group must be
conjugated with an unsaturated group (this was tested for aromatic groups and C-C double bonds; see Tables 1 and 2). If this
is not the case, only the carbonyl group is methylated to yield the
corresponding alcohol.[z31If the conjugated, unsaturated group
is a phenyl unit, only the A R C D product is formed. If, on the
other hand, the carbonyl group is linked to a 2-furyl, 2-thieny1,
o r 2-(N-methyl)pyrrolyl group, or a C - C double bond
(Table 2), isomeric compounds ("rearranged ARCD products")
are also produced in addition to the A R C D product (7,9- 12).
Their structures (8, 13, 14) were elucidated (Table 2, entries 1
and 6) and are given in Scheme 5. In the reactions of furfural
Table I prcsents reactions in which the substrates were added
to the thermolabile reagents at about - 78 "C. As with benzaldehyde, 6a also reacts with other aromatic aldehydes as well as
[a] Not determined. [b] Instead of the ARCD product. its disproportionation products were obtained (see Scheme 6)
3 A1lclij~de.sund Ketones us Substrates
Yield [%I
43 (7) [a]
80 (9)
79 (10)
34 (11)
3 (12)
Yield [YO]
of the ARCD
30 (8)
9 [bl
9 [bl
33 [c]
20 (13) [d]
16 (14) [dl
86 [cl
[a] Two diastereomers in the ratio of about 1 : l . [b] Structure not elucidated.
[c] Separation of the ARCD product and the isomers did not succeed. [d] The given
structure is based on ' H N M R spectroscopy and GC'MS analysis.
T. Kauffmann
amide indicates the limits of the surprisingly high nucleophilic
activity of 6 a towards carbonyl compounds.
2.3.2. Further Compounds of Type 6 as Reagents
ARCD reactions in which the alkyl group is transferred are
also possible with good yields, even if the structural prerequisite
for B-H elimination is present (Table 3, entries 1-6). This conTable 3 Reaction of reagents of type 6. prepared according to Scheme 4, with
aldehydes and ketones R'R'CO in a 1 :1 mole ratio [23]. The yields and degrees of
recovery given in parentheses apply to thecase when 6 was synthesized with RMgX,
the others to when it was synthesized with RLi. Reaction conditions as in Table 1.
Scheme 8. Additive-reductive carbonyl dimerization with heteroaromatic carbony1 compounds as well as with an a,b-unsaturated ketone [23].
and heteroaromatic ketones (Table 2, entries 1-4) there is only
one rearranged product (structure not elucidated) which is only
produced in minor amounts compared to the ARCD product,
so that these ARCD reactions still remain attractive for syntheses. In contrast, the reactions of cinnamaldehyde and conjugated enones (Table 2, entries 5-7) furnished several rearranged
ARCD products. Therefore, these reactions are not recommended for preparative use. In the case of benzophenone and a
further ketone (Scheme 6), a disproportionation product was
obtained instead of the ARCD product. The ARCD product 15
from benzophenone disproportionates on warming to 1,Idiphenylethane (16) and 1,l-diphenylethene (17) .[311
2 Ph2C0
2 6a
Scheme 6. Disproportionation of ARCD products [23]
Yield ["/I
(R1RZR3C), R'R2R3COH [Yo]
30 (82)
65 (66)
62 (89)
57 (53)
89 (61)
78 (75)
67 [b]
4 (0)
15 (0)
16 (0)
14 (12)
0 (0)
3 (0)
0 (0)
15 (0)
6 (0)
0 (0)
0 (0)
1 (0)
[a] Not determined. [b] Instead of the ARCD product its disproportionation products 16 and 17 were each obtained with 67% yield.
siderably widens the range of application of the ARCD reaction. The surprisingly low tendency towards B-H elimination,
for example, in 6b, which is in contrast to n-butylation products
of WOCl,, WOCl, (see Section 3.1), or WCl,[321,corresponds
to the disinclination of 6a to form p-CH, ligands from CH,
ligands by a-H elimination (cf. Section 3.2). The tetraphenyl
complex 6c is also suitable for use as an ARCD reagent
(Table 3, entries 8-10). It is particularly interesting that cyclohexanone, that is a substrate with an isolated keto group, also
produces the ARCD product (Table 3, entry 9). From these
results it can be expected that other carbonyl compounds (aldehydes, ketones, carboxylates, carboxylic acid anhydrides, and
acyl chlorides) containing saturated hydrocarbon groups can
also take part in a phenylating ARCD reaction. Not only that,
it should be possible to employ other reagents in which the
phenyl group of 6c is replaced by other aryl groups. Benzoic Acid Derivatives as the Substrate
Compound 6a reacts with ethyl benzoate, benzoic anhydride,
and benzoyl chloride to give 2,3-dimethyl-2,3-diphenylbutane
(18). Here there is initial formation of acetophenone by nucleophilic substitution, which then undergoes the ARCD reaction
(Scheme 7). The unsuccessful reaction with N,N-diethylbenzPhCOX
1 6 a
2.4. Mechanistic Discussion
The radical mechanism postulated for the ARCD reaction
with aldehydes and ketones is given in Scheme 8.[231This mechanism is supported by the following: a) The reaction only proceeds with the tetramethyl reagent 6 a if the carbonyl group of
Yields of18 for X = OEt, CI, OCOPh, NEt,: 79,9456, and 0%. respectively
Scheme 7. Additive-reductive carbonyl dimerizatlon of benzoic acid derivatives
[23]. Two equivalents of6a were used for the reaction with benzoic anhvdride: side
product: 2-phenyl-2-propanol ( 2 0 % ) . N,N-Diethylbenzamide was recovered in
90% yield.
Scheme 8. Postulated mechanism of additive-reductive carbonyl dimerization
Angew. Chum. Ini. Ed. Engl. 1997, 36, 1288-1275
Organomolybdenum and -tungsten Reagents
the substrate is conjugated with an unsaturated group; this leads
to mesomeric radical stabilization. b) The reaction is also possible with an aliphatic ketone, if the tetraphenyl reagent 6c is
employed. c) The radical mechanism explains the occurrence of
the "rearranged ARCD products" in the reactions given in
Table 2, because the mesomeric radicals which are produced,
can couple on both the primary radical center and in the ally1 or
pentadienyl position.
In a process which conforms with Scheme 8, the homolysis of
the 0 - C bond is aided by mesomeric stabilization of the organic radical, as well as by the mesomeric interaction between the
resulting electron-withdrawing 0x0 ligands and the electron-donating propoxy ligands. The following observations indicate a
process in which the resulting radicals are in close proximity to
one another, as in species 19, and are within the effective domain
3. Spontaneous Transformation of Methyl Ligands
into p-Methylene Ligands in Complexes Containing
High-Valent Molybdenum or Tungsten
3.1. Formation of p-CHI Complexes after Methylation of
Molybdenum and Tungsten Chlorides with MeLi or Me,Al
In 1978 Wilkinson et al[331isolated the complex [(Me,P),Ru(p-CH,),Ru(PMe,),] after treating the quadruply bridged
Ru",Ru"' complex [Ru,(O,CMe),]Cl, or a further Ru",Ru"'
derivative, with excess Me,Mg. They assumed that the p-CH,
ligands had been formed by a-H transfer via p-CH, ligands
The transformation of a p-CH, ligand into a
p-CH, ligand had been discovered only shortly before by
Calvert and S h a ~ l e y [on
~ ~an] 0 s cluster.
- CH4
Scheme 10. Proposed mechanism for the formation of the p-CH, ligands of a Ru
complex [33]. The ligands not taking part in the reaction are omitted.
of the concomitantly formed organometallic species or in a solvent cage. a) In the ARCD reactions with 6a, there were no
cases observed of H-transfer from the solvent T H F onto the
postulated radical intermediate R' R'kCH, (formation of
R'R2CHCH, without the simultaneous formation of
R'R2C=CH,). b) The elimination of an H atom in an a position to the radical center was not observed (formation of
R'R2C=CH, without the simultaneous formation of R'R2CHCH,). c ) Normally, there is a total absence of disproporH ,
of equivalent amounts
tionation of R' R ~ ~ C (nonformation
of R'R2CHCH, and R'RZC=CH,) as well as yields of the
ARCD product of up to 94%. In those cases in which the two
central C atoms of the ARCD product are chiral centers (for
example, Table 2, entry I ) , the ratio of the meso form to the ruc
form is always about 1 : 1, as is usual for a radical process. This
lack of stereoselectivity is a powerful argument against an electrocyclic process via transition state 20. The fact that the ARCD
product 18 from acetophenone is also accessible using the pathway given in Scheme 9, in which the methylating step is carried
out on acetophenone instead of on 2, is also in agreement with
the mechanism given in Scheme 8. However, the low yield and
the side products 21 a and 21 b make this second pathway of
additive reductive carbonyl dimerization appear to be relatively unfavorable.
Ph (Me)&
1 2
18, 24%
21a, 13%
Scheme 9 Additive reductive carbonyl dimerization with reversed course of reaction [23]
A n K w Clic,m. lnt. Ed Engl. 1991. 36, 1258-1275
In our investigations, which were prompted by an observation by J. Sander,[35a1
we found more than twenty examples for
the transformation of CH, into p-CH, ligands according to
Scheme 11A and 11 B:[1.36-431t reatment of MeLi or Me,AI
2 MCH,
( M = Mo,W)
* MCHpAl
+ CH,
( M = Mo,W)
Scheme 11. Transformation of CH, ligands into p-CH, hgands in the reactions
given in Tables 4-6 as well as in Section 4 The ligands not taking part in the
reaction are omitted.
with MoV,Mo", Wv, and Wv'
(see Tables 4-6) in
T H F or Et,O at low temperatures produces methylmolybdenum and -tungsten complexes. On warming these complexes
they liberate CH, and largely convert into thermolabile complexes containing p-CH, ligands in the partial structures 22
and 23, respectively, and can carbonylmethylenate aldehydes
and ketones['*3 6 - 4 2 1 (Tables 4-6 as well as Section 4). In some
cases this ligand transformation could be proved qualitatively
by NMR spectroscopy (Sections 3.2 and 3.3), and for molybdenum chlorides, it could be determined quantitatively by the
measurement of the amount of CH, liberated in combination
with the determination of the number of CH, groups transferred to the aldehyde groups (Table 6). Since at least 0.75-0.96
carbonylmethylenating p-CH, ligands were formed per liberated CH, molecule, the CH, can only be formed at most in
minor amounts by the reduction of the molybdenum chloride.
The reactions according to Scheme 11 not only proceed in
T H F or Et,O but also in chlorobenzene, toluene, cyclohexane,
or in n-pentane, as can be discerned by the liberation of
CH, and the formation of a carbonylmethylenating complex.
The reactivity of the species formed in these solvents is not
T. Kauffmann
Table 4. Reactions for the detection of CH, ligands by methylenation according
RCHO + RCH=CH, in THF. So far, they have only been published in dissertations. MeLi was added to the chloride at about -78 'C, followed by addition of the
aldehyde (0.5 equivalent per molybdenum or tungsten atom). Unless otherwise
stated, the mixture was then warmed to 20'C over a period of 18 h. Corresponding
reactions of aldehydes and ketones with [Mo,CI,,], MoOCI,, MoOCI,. WOCI,,
and WOCI,, in which very good yields were often obtained: see references [36-411
and Section 4.
p-CH, complex
2MoO,CI, + 4MeLi [22d]
4MeLi [22d]
2(MeO)2MoCI, + 4MeLi [22g]
2[WOCI,(PPh,)] + 4 MeLi [22a]
2(MeO)WOCI, + 4MeLi [22a]
2(MeO),WCl, + 4MeLi [22a]
4-MeOC6H,CH0 [a]
4-MeOC6H,CH0 [a]
[d] -78 * 6 6 ' C in 1 h, then 3 h a t 6 6 ' C
Table 5. Detection of CH, ligands by carbonylmethylenation by the reaction
R'R'CO + R1R2C=CH, in T H F [l] and the yield-increasing action of HMPA
(hexamethylphosphoric acid triamide). Conditions: per molybdenum or tungsten
atom oneequivalent R'R'CO. Me,AI was added to the chloride at - 78 '.C followed
by addition of the carbonyi compound; subsequently, the molybdenum complexes
were allowed to warm up from -78 + 20'C over a period of about 18 h, the
tungsten complexes from -78 + 66°C over a period of 3 h. then warmed for
3 h at 66.C. Further examples and recovery of the unconverted carhonyl compound: [I].
p-CH, complex
MoO,CI, + 2 Me,Al
WOCI, + 2Me3A1
WOCI, + 2Me,AI
+ 2Me,AI
R 1R ~ C O
4-MeOCtiH,CH0 ( 0 % ) [el
+ PhCOMe (73%) [el
98 (67 [a1)
58 (87 [bl)
74 (84 [bl)
38 [cl (81 [bl MI)
[a] Yields from the reaction with heptanal. [b] Yields from the addition of two
equivalents of HMPA per WOCI,. [c] 1 0 % Mono- and 24% diolefin. [d] 56%
Mono- and 25% diolefin. [el Molybdenum ch1oride:carbonyl compounds = 1 : 1 . I ,
in parentheses: recovery of R'R'CO.
Table 6. Determination of the number of moles of CH, ( = (CH2),r)per liberated
mole CH, transferred to the aldehyde during carbonylmethylenation.
p-CH, complex
CH, [a1 [bl aldehydebl ( C H A [bl
2-MoOCl, + 4MeLi
2MoOC1, + 4MeLi
2MoCI, + 4MeLi
MoOCI, + 2 Me,AI
MoOCI, + 2Me,AI
MoO,CI, + 2 Me,AI
1.0 [3S]
1.0 [38]
0.5 [22d]
2 0 [I]
2.0 [ l ]
1.9 [l]
1.5 [l]
2 [CJ
2 [d]
1 [el
4 [el
4 [el
4 [el
4 [el
0.79 [35h]
0.75 [22e]
0.48 [22d]
1.60 [35i]
1.84 [35i]
1.50 [35j]
1.44 [35i]
[a] Formed during the formation of the p-CH, complexes. [b] Per mole molybdenum atom. [c] Pyrrole-2.5-dicarboxaldehyde. [d] Thiophene-2.5-dicarboxaldehyde. [el Benzaldehyde
restricted by any relatively strongly bound T H F or Et,O ligands. Thus, they are capable of attacking C-C double bonds
and allylic H atoms. The reactions resulting from this occur
either as a secondary reaction after a carbonylmethylenation
(olefin dimerization, olefin metathesis), or instead of a carbonylmethylenation (cyclization, isomerization). For further
information on these reactions see references [22n, 220, 35d,
35 k, 42, 431.
3.2. Further Details on the pCH, Complexes Formed by
the Reaction with MeLi
Optimization of the reaction of MeLi with molybdenum and
tungsten chlorides to give carbonylmethylenating reagents revealed that the most favorable ratio is two equivalents of MeLi
per molybdenum or tungsten atom. Since it is highly probable
that 1,3-dimolybda- or 1,3-ditungstacyclobutanesare formed
(see below), the individual reagents are represented within
brackets and each contain two molybdenum or tungsten atoms.
The methylation step has been investigated for the reagents
{2 MoOCl, + 4 MeLi},[381 {2 WOCI, + 4 MeLi),[22h1 and
{2 WOC1, + 4 MeLi}:[22h,441
' H N M R spectra were measured
directly after mixing the components in [DJTHF at low temperature (in the first case - 30 "C, otherwise about - 70 "C). The
signal at 6 = -1.98 is absent, which indicates the complete
consumption of MeLi. For each reagent there are two singlets
present (6 = 0.96/0.79, 0.92/0.69, and 0.95/0.73, respectively),
which lie in the range expected for metal-bound methyl groups.
As the temperature is raised these signals disappear, whilst the
intensity of the CH, signal at 6 ~ 0 . 2 3increases and signals of
p-CH, ligands appear. The methylation products, which must
have Lewis acidic properties, are incapable of methylating aldehydes and ketones at low temperature in T H F (formation of
stable reagent-THF complexes?; see references [45,46]). On
warming these primary products CH, is liberated (for amounts
see Table 6). Whilst this process is complete at about 20 "C for
the molybdenum reagents, the tungsten reagents must be
warmed to about 45"C, although partial liberation of CH,
starts below 20 "C. The chronological course of the CH, elimination for {2 MoOC1, + 4 MeLi} as an example, is described in
reference [38].
Because of the high thermal lability of the carbonylmethylenating species (determination of the time-dependent loss of activity in the case of {2 MoOC1, + 4 MeLi):[381)attempts at isolation were unsuccessful. However, NMR measurements lead to
a clear idea of the structure.[22g.22h, 3 7 . 3 8 . 4 3 1 Taking into account the method of formation, the carbonylmethylenating action (see Section 3.l), the low basicity, and the surprising resistance to protons (see Sections4.2 and 4.4; the latter results
exclude compounds with MoCH,Li or WCH,Li bridges), the
complexes must either contain terminal CH, ligands or a p-CH,
ligand (CH, bound to two metal atoms). NMR spectroscopy
helped to decide between these two po~sibi1ities.l~~.The I3C
NMR signals (Table 7) and 'H NMR signals[37,3 8 , 2 2 h , 35b1which all disappear on addition of aldehydes or ketones to the
reaction solution, whilst at the same time signals of terminal
alkenes appear-all lie in the range (Scheme 12)[33,47-501
is typical for p C H , ligands in small rings. Signals of terminal
CH, ligands did not appear. 'H NMR signals, which appeared
suspiciously close to the range characteristic for terminal CH,
ligands (6 = 9-11), can be plausibly explained by hydrogen
bonds, as demonstrated in references [37,38]. A direct comparison is possible by using the complex obtained from
{ 2 WOCI, + 4 MeLi) : whilst its p-CH, ligand produces a
' H NMR signal at 6 = 3.5 (broad),[22h1
the proton signals ofthe
terminal CH, ligands of [(Me,P),Cl,(O)W=CH,] are found at
6 = 12.34 and 11.47.r511
The coupling constants 'J&, which are
caused by the p-CH, groups of the carbonylmethylenating
Anxrw. Chem. I n r . Ed. Ennl. 1997. 36. 1258-1275
Organomolybdcnum and -tungsten Reagents
Table 7. I J c ' N M U data of the p-CH, and p-CHSiMe, ligands ofcomplexes that
were rynthesized following or analogously to Scheme 13 using H,"CLi in T H F or
with Me,SiCH,Li i i i [DJTHF. Measured at 25 'C.
6( "c)[d]
2 WOCI, + 4MeLl [22h, 411 ( = A)
2WOC1, + 4Me,SiCH,Li [35b] ( = B)
~MoOCI,+ 4Mel.i [ 3 8 ] ( = C)
ZWOCI, + 4McLi [22h, 411 ( = D)
[Mo,CI,,,] 4MeLi [37] ( = E)
4Mel.1 [38] ( = F )
166.2, 151.7
169.5. 167.1
150.9, 149.9
185.7, 184.5
176.4, 173.9.
172.7. 168 7
164.9, 162.8,
166.6, 164.2
161.3. 157.7
+ 4MeLi [22g] ( = G )
cis-29. rrons-29
c1.s-30, rruns-30
centrosymmetric 26 with four T H F ligands in axial positions. Two I3C
N M R signals are observed for B-E and each can be explained by the
presence of two cislrruns isomers: the N M R data of B (discussed here as
an outsider) support the complexes cis- and fruns-27. which are centrosymmetric with regard to the equatorial ligands C1 and 0.The Me,Si
groups give rise to the isomerism in an analogous manner to the fBu
groups in comparable Zr complexes.[501In C-E the two I3C N M R
signals can be interpreted by the presence of cis- and rriins-28, cis- and
?runs-29, and cis- and rruns-30, respectively, whose isomerism is caused
25, R =
26,R= H
6 C3CH)
240 - 370
100 - 210
[33,48-501 120-180
2 -9
f r a n s - 2 7 , R = SiMe,
t r a n s - 2 8 , M = Mo
trans-29, M = W
c i s - 2 8 , M = Mo
cis-29, M = W
[a] Measured ;it 74 4 MHL. T M S as reference, proton-broadband-decoupled spectrum By using the DEPT or INEPT procedure [52 b] it was proved that the signals
originate l'rom methylene o r in B from methyne groups.
trans- 30
[( thf)pCI(0)M~/-L\M~(O)CI(thf)g]
Scheme 12 Ranges of I3C and ' H N M R signals of the given types of structure
(R = H or hydrocarbon group)
[(thf )( M e O),CI
(OMe )2(thf)]
molybdenum c ~ m p l e x e s ,381
[ ~ each
~ ~ amount to about 14.4 Hz.
Therefore, they correspond to the coupling constants for geminal protons in four-membered ring systems (10-14 H Z ' ~ ~ " ] ) .
Taking into account the formation with liberation of one equivalent of CH, per molybdenum or tungsten atom, the quotients
(CH,),,/CH, given in Table 6 along with the presence of p C H ,
ligands detected by spectroscopy, implies that the complexes
obtained according to Scheme 13 are 1,3-dimolybda- or 1,3-diA1
2 X3M'CI2
+ 4MeLi
- 2 CH,
Scheme 13 Postulated formation of 1,3-dimolybda- and 1.3-ditungstdcyclobutanes. M = Mo. W , X = CI, I Z 0, OR, CH,.
tungstacyclobutanes or their oligomers. The presence of a
metal-metal bond in complexes 24 is very probable on account
of the diamagnetism (generally very sharp N M R signals) of the
reaction solutions obtained from MoVand Wv chlorides.
An in-depth analysis of the I3C (Table 7) and 'HN M R data substantiates the idea that 1,3-dimetallacyclobutanestructures are present. 1)
13C N M R : A affords only one signal (Table 7). In analogy to 2Si4" (C
and H atoms of t h c /i-CH2 ligands are each equivalent), A could be
by the axial chlorine ligands. The assumption that there are two THF
ligands in cis- and truns-30 is consistent with the fact that the carbonylmethylenating activity of E towards benzaldehyde is drastically reduced
on addition of one equivalent of Ph,As or Ph,P per molybdenum atom
(Table 8). Evidently, the loosely bound T H F ligands arc replaced by
strongly bound hgdnds, which either hinder or prevent the coordination
of the oxygen atom of the aldehyde group to the molybdenum atom.
In this context it is interesting that the carbonylmethylenating activity
of F (= 3 1 ) . which is presumed to contain four THF iigdnds, is
not restricted by the addition of one equivalent of Ph.,As or Ph,P
per molybdenum atom (Table8). More than two 13C N M R signals
were observed for F and G . Evidently, these two each are a relatively
complicated product mixture, which can be explained by the fact that
in the resulting complexes 31 and 32 the CI and 0 and CI and
Table 8. hihibition ofcarbonylmethylendtion reaction PhCHO + I'hCH=CH, in
T H F (-78 + 20'C in 18 h; 0.5 PhCHO per molybdenum atom) hy the addition
of potential ligands at about - 7 8 ' C [22f,64]
E [a]
+ n eqmv hgand
per molybdenum dtom
PhCH=CH, F [b) n equiv. hgand
per molybdenum atom
5 (CH,),NMe
1 (CH,),NMe
1 Ph,N
1 Ph,P
1 Ph,As
19 [dl
3 (CH,),NMe
0.5 Me2N(CH,),NMe2
1 Ph,P
0.5(2) Ph,P(CH,),PPh,
1 Ph,As
79 (50)
[a] Postulated structure of the methylenating complex: ( I s - .ind 1run.s-30.
[b] Postulated structure of the methylenating complex 31. [c] No addition of a
hgand. [d] Ref [35c].
T. Kauffmann
M e 0 ligands, respectively, itre not only in cquatorial but also in axial
positions. I n F the ollgoinerization of the primarily formed complexes 31
must also be considered as the reason for the numerous nonequivalent
pCH, carbon
2) ' H N M R : The ' H N M R data12Zh.35h.37.381
+ 2 Me3AI
has been analyzed for €5, C, E, and F and IS in agrcement with the
+ 2 L
proposed structures ci.y- and /ra/i,s-27, c.i,s- and tru/?.r-28,c.i.s- and /runs-30
and 31, r e s p e c t i ~ e l y . ~ ~ ~ - ~ ~ ~ ~ ~ ~ ]
CDCI,, - 7 0 0 2
Rcactions of [Mo,CI,,], analogous to those in Scheme 13,
were not possible with alkyllithium compounds other than
MeLi: the use of nBuLi or n-hexyllithium suppresses the formation of l ,3-dimolybdacyclobutanes by B-H elimination,r22d1
in the reaction of neopcntyllithium with Mo,Cl, double a-H
elimination occurs instead of singIe.[l2]If Me,SiCH,Li is treated with [Mo,CI,,], an alkylidene complex is formcd together
with an alkylidync complex by single or double a-H elimination,
as described in Section
Accordingly, the mixture of the
rcactioii products olefinize benzaldehydc only to a small degree
(9 "A[j-(trimethylsilyl)styrene)
Furthermore, the reaction of
WOCI, with two equivalcnts of Me,SiCH,Li (THF, 2 0 T ) and
subsequcnt treatment with benzaldehyde produces the carbonylolcfination product only in unsatisfactory yields (21 % cis- and
11 'Yo trcm.s-~~-(trimcthylsilyl)styrenc)
L = [Dl8]
[L(Me)CIAI r(
\ M o \ L o pAICI(Me)L]
+ 2 Me3AI
2 CH,
36,M =
+ 1 oder 2 Me3AI
(Me 0 ) ,Mo CI
Mo, W
37, L =
Scheme 14. Tehbe reagent (33)and the postulated formation of similarly structured
p C H , complexes with molybdenum or tungstcn as the transition metal.
3.3. Further Details on the p-CH, Complexes Formed by
the Reaction with Me,Al
The rcactions of the carbonylmethylenating complexes
formed by reaction of Me,AI with molybdenum and tungsten
chlorides (Tables 5 and 6). via the methylation products detected by NMR spectroscopy,[22k.22"J exhibited the following diffcrenccs to the corresponding reactions with MeLi: a) In many
cases approximately t w equivalcnts of CH, per molybdenum
o r tungsten atom were liberated, and the carbonylmethylenating cornplcxcs formed transferred more than one CH, group
(maximum 1.84) per molybdenum or tungsten atom (see
Table 6). b) Thc solutions formcd from MoV chlorides as well
as from WOCI, by action of Me,AI are paramagnetic. c) Thc
complexcs obtained with Mc,AI are highly sensitive to protons
and are unsuitable for the carbonylmethylcnation of hydroxy
ketones. d) Their carbonylmethylenating activity can be enhanced by addition of a base (see Table 5 ) as found for the Tebbe
reagent (33, Scheme 14)15,. 541 and a Zr-A1 complex of similar
structure.r50'These observations leavc little doubt that complexes containing aluminum have been formed. Accordingly, during
experiments to isolate the complex resulting the reaction of
Mo02CI, and two equivalents of Me,AI, a yellow, amorphous
solid complex containing aluminum precipitated from the T H F
solution at 0 'C. A suspension of this solid in T H F methylenated
b e n ~ a l d c h y d e . ' The
~ ~ ' formation of the carbonylmethylenating
complex from MoO,CI,. two equivalents of Me,AI, and two
equivalents of [D,,]HMPA in CDCI, was monitorcd by
'HN M R spectroscopy"-35J1(in the case of Mo" and Wv chlorides the formation of paramagnetic soIutions prevents such
investigations): on mixing the components at -78°C a singlet
appears at 0 = 0.82, which is assigned to thc dimethyl complex
34, since i t disappears on warming and the same signal
(6 = 0.81)rz2d1was observed during the reaction of MoO,CI,
with two equivalcnts of MeLi at -70 'C in [DJTHF. At 10°C
thc first p-CH, signals appear. After 1 h at 25 ' C the spectrum
exhibits not only very weak Li-CH, signals (6 = 5.7-5.0 and
10.8-10.1), which arc characteristic of 1,3-dimolybdacyclobutanes, but also a very intense singlet at b = 5.34. This lics in thc
region typical for p C H , ligands in small rings (see Scheme 12)
and is in agreement with structure 35. which was postulated on
the basis of the way of formation (liberation of 1.9 CH,) and
reactivity (CH, transfer rate = 1.5; see Table 6).".421 The remaining carbonylmcthylenating Mo- Al and W -A1 complexes,
which were obtained from reactions in reference [I] and Table 5,
are also assigned cyclic structuresI'I (e.g. 36 and 37) on the basis
of the results described above as well as their high aldehyde
versus ketone selectivity (usually > 94:6; see reference [I]).
3.4. Discussion of the Mechanism
The discussion of the mechanism of the ligand transformations described in Sections 3.1 -3.3, which exhibit parallels to
the formation of the Tebbe reagent,[531is headed by the question
as to the identity of the intermcdiate species involved. According to Scheme 15A or 15B it is a terminal CH, hgand, and
according to Schcmc 15 C or D it is a p-CH, Iigand (CH, bound
by a two-electron three-center bond to two metal atoms). The
argument against a complex with a terminal CH, ligand
(Schrock carbene complex) a s the intermediate product steins
from the fact that during the formation of the p-mcthylcncmolybdenum complexes described in Sections 3.1 and 3.2. the
carboxylic acid chlorides, or carboxylates, or bcnzyl chloride
added to the reaction mixture are not attacked. Schrock carbene
complexes arc usually not inert to such clcctrophiles (cf. references [54-561). In addition, thc high aldehyde verswi ketone
selectivity found in competition experiments (see Table 5 and
Section 4.3) does not support the theory of an intermediate
Organomolybdenum and -tungsten Reagents
+ CI-A1
low basicity of these compounds and the corresponding tungsten complexes proved to be advantageous in carbonylmethylenations. In an analogous manner to the Lewis acidic alkylation
reagents RTiX,[581these reagents also offer the possibility of
varying the selectivity within certain iimits by variation of the
electronegative ligands on the molybdenum or tungsten atoms
(reaction of various chlorides according to Scheme 13 or ligand
exchange on the p-CH, complexes).
4.1. Carbonylmethylenation of Base-Sensitive Ketones
- CH,
The low basicity of the complexes obtained according to
Scheme 13 permits carbonylmethylenation of readily enolizable
ketones.[22h.3 5 e 1 Scheme 16 gives an example, and a comparison
M \,M2
Scheme 15. Possible mechanisms for the transformation of CH, ligands into p-CH,
Iigands. A and B - Complexes with terminal CH, ligands as intermediate products.
C and D. Complexes with a p-CH, ligand as the intermediate product. M' = Mo,
W , M 2 = Mo,W.Al;X = C l , O M e . O . F o r X =0,H,C=M1-O-M2isproduced
instead of 38 for mechanism B (bond between M' and 0 is covalent instead of dative
in 38)
Schrock cdrbene complex, because in these experiments p-CH,
complex formation did occur in the presence of the added ketoaldehyde or aldehyde- ketone mixture. Support for the formation of the p-CH, ligand from a primarily formed p-CH,
ligand, in an analogous manner to the p-CH,
p-CH, transformation reported in reference [34], comes from the fact that
the elimination of a proton from a p-CH, ligand shouId be more
favorable than from a CH, ligand. The possibility of facilitating
a-H elimination by primary p-CH, formation was already
pointed out in 1975 by Muetterties.'"] In addition, the intermediate formation of MCH,M or MCH,AI bridges (M = Mo, W)
is very plausible on account of the electron deficiency on the
high-valent molybdenum and tungsten atoms as well as the
AICH,Al bridges in (Me,AI),. It also agrees with the results
from the reaction with MeLi: the tendency to form a carbonylmethylenating complex on going from MoV chlorides to
M o C I , [ ~ ~decreases
(see e.g. Table 6) and was not even observed for MoCl, , /l-MoCI,, and a-MoBr,
The same question arises for the formation of the Tebbe reagent (33): not only
the mechanism following the principle of Scheme 15B, which
was postulated by Grubbs et al.[571,but the mechanism according to Scheme 15 C need to be considered.
4. Application of pMethylenemolybdenum
and -tungsten Complexes as Lewis Acidic, Selective
Carbonylmethylenating Reagents
In contrast to the intensive efforts to optimize the stereoselectivity of carbonylolefinations, relatively little was undertaken to
confer high regioselectivity to this important C- C coupling reaction (see Section 4.9). Of all the carbonylmethylenating pCH, complexes described in Section 3, the 1,3-dimolybdacyclobutane type in particular display high selectivity. Also the
Angrit Climi /nr Ed EngI. 1997. 36. 1258-1275
1.5{2 W 0 C l 3 + 4 MeLi}
95 %
1.5{2 WOC14+ 4 MeLi}
93 %
Scheme 16. Methylenation of a readily enolizahle ketone. Conditions: T H E
-78 4 4 5 ' C in 18 h [22h].
with Ph,P=CH, as methylenating reagent. Whilst Ph,P=CH,
or MeMgBr or even the weakly basic [MeCrCl,(thf),][591lead to
the rapid cleavage of the base-sensitive ketone 40 by a retro-aldo1 reaction, (2 WOCl, + 4 MeLi) and {2 MoOC1, + 4 MeLi}
are capable of methylenating 40, and this with high regioselectivity[22J,391
(Scheme 17).
OH 0
2 ) H+/H*O
l a ) 1 { 2 WOCI, + 4 M e L i }
i b ) 1 { 2 MOOCI, + 4 M e L i }
Scheme 17. Highly regloselective monomethylenation of an extremely base-sensitlve hydroxy diketone. 1 a) T H E - 78 + 45°C in 16 h [22j]. The second possible
monocarbonyl methylenation product was not formed. 1 b) THF, - 70 + 20 "C in
12 h [22~,39].
4.2. Carbonylmethylenation in Protic Media
The carbonylmethylenating activity of {2 MoOCI, + 4 MeLi) ,
(2 MoOCI, + 4 MeLi), and ([Mo,Cl,,,] + 4 MeLi) is restricted
only to a surprisingly small degree by water, ethanol, or other
alcohols. This permits carbonylmethylenation in aqueous or
alcoholic media, which was hitherto not possible (for examples
see Scheme 18 and reference [40a]) and which could be useful
for hydrophilic substances. As was to be expected for these
results, {2 MoOC1, + 4 MeLi} smoothly methylenates 2-, 3-,
{ 2 MoOCI,
T. Kauffmann
+ 4 MeLi}
Scheme 18. Methylenation of aldehydes in protic media [40a] The yield given in
parentheses was obtained when the reagent solution was added at 0°C instead of
at - 7 8 ‘ C This lower yield is attributed to the thermolabillty of the carbonylmethylenating complex.
and 4-hydroxybenzaldehyde (to give the alkene in 89, 86, and
56 % yield, respectivelyrzze1)as well as hydroxy ketones (see
Section 4.4). Even glyceric aldehyde (suspended in THF) is a
suitable substrate (affords alkene in 41 % yield).[3se1It can be
concluded that the mechanism of carbonylmethylenation in
aqueous or alcoholic media involves nucleophilic substitution of
chlorine ligands for hydroxy or alkoxy ligands on the 1,3-dimolybdacyclobutanes. This has already been reported for the
analogous reaction of EtOH with [Mo,CI,,] to give
(EtO),MoCI, under similar conditions,r601and corresponds to
the fact that the carbonylmethylenating activity of the reagents
is usually somewhat weakened in aqueous or alcoholic media.
o ,Ln~
Scheme 20. Carbonylmethylenation according to CA-CE mechanism [37] and the
proposed mechanism [S2] for carbonylmethylenations with open-chain dimetalliomethanes.
per molybdenum atom are often transferred (see for example the
first two experiments given in Table 6), it is reasonable to assume that 2,4-dimetallaoxetanes 41 (Scheme 20 A) also have
carbonylmethylenating activity.r381Accordingly, in the methylenation of propanal with {[Mo,CI,,] $ 4 MeLi) in n-pentane,
MoOCI, (intense IR bands at 950 cm- ’) was detected as the
inorganic end product.[z2g,
4.4. Cheleselectivity
4.4.1. 1,3-Dimolybda- and 1,3-Ditungstacyclobutanes
4.3. Aldehyde Versus Ketone Selectivity
(“Aldehyde Selectivity”)
In intermolecular[37.391 and intramolecular competition experiments (Scheme 19), the carbonylmethylenating 1,3-dimolybdacyclobutanes exhibit an aldehyde selectivity that is
Ph -4’
* 20°C inlEh,THF
+ 4 Me Li}
2 { ZMoOCI,
+ 4MeLi}
89 %
80 %
70 %
In alkylations of keto groups with alkyl transition metal
reagents, the basic groups (HO, MeO, Me,N, MeS) in the CI or
position exercise an accelerating effect on the reaction. This
permits alkylations with a high cheleselectivity (for a definition
of the term see reference [5]). If (2 MoOCI, $ 4 MeLi) or
12 MoOCl, + 4 MeLi) are treated with the substrate pairs in
Scheme 21, the hydroxyl group displays a powerful accelerating
effect. This neighboring group effect, which also permits the
very regioselective monomethylenation of the hydroxy diketone
40 in Scheme 17, is explained[391b y primary esterification of the
hydroxyl group by 1,3-dimolybda- and 1,3-ditungstacyclobutanes containing chlorine ligands, and cleavage of HCl (cf.
Scheme 28). The carbonylmethylenation then follows in-
Scheme 19. Stringent aldehyde-selective monomethylenation of a keto aldehyde
137,391.The second possible monocarbonyl methylenation product was not formed
in each case
higher than that of 1,3-ditungstacyclobutanes as well as that of
the Mo-A1 or W-A1 complexes described in Section 3.3. The
discernible agreement here with the high aldehyde selectivity of
alkyl transition metal reagents (metal e.g.
readily understood if it is supposed that the 1,3-dimolybdacyclobutanes, which represent specific alkyl transition metal
reagents, react with the carbonyl compounds without previous
separation into complexes with terminal CH, ligands according
to the cycloaddition (CA)-cycloelimination (CE) mechan i ~ m [in~ Scheme
20A. This mechanism is similar to that represented in Scheme 20 B, in which the aldehyde-selective geminal dimetalliomethanes (see Section 4.9) carbonylolefinate.
Since in the methylenation of aldehydes-but not of ketoneswith 1,3-dimolybdacyclobutanes,more than 0.5 CH, groups
I ‘I‘
{ 2 MoOC14
+ 4 MeLi}
77%/J2‘ ”
63 %L39’
Scheme 21. Cheleselectivity of 1.3-dimolybdacyclobutanereagents in intermolecular competition reactions. Mole ratio reagent :substrates = 1 . 1 :1; THF, - 78 +
20°C in 18 h. Numbers under the arrows: yields of the carbonylmethylenation
product; in square brackets: selectivity. For details on the recovery of the substrate
see reference [39].
Angew. Chem. Inl. Ed. Engl. 1997, 36, 1258-1275
Organomolybdcnum and -tungsten Reagents
tramolecularly and therefore with a favorable entropy term.
Evidently, a covalent bond between the reagent and the substrate is necessary, because the methoxy group or the dialkylamino groups do not accelerate the reaction when 1,3-dimolybda- and 1,3-ditungstacyclobutanesare used, but rather have a
retarding effect (see Scheme 24). The solvent was varied in competition reactions of {2 MoOCI, + 4 MeLi} with a 8-hydroxy
ketonelketone pair and it was found that the cheleselectivity increases in the order CH,CI, < T H F < MeO(CH,),OMe,
that is with increasing basicity of the solvent (Table 5 in
reference [39]). The reagents 12 WOCl, + 4 MeLi} and
(2 WOCl, + 4 MeLi} are normally less suitable for selective carbonylmethylenation of hydroxy ketones than the corresponding
molybdenum reagents.
4.4.2. Mo- A1 and W - A1 Reagents
In competition reactions of a-methoxyketone/ketone pairs
with {MoO,Cl, + 2 Me,Al} there is a clear preferential
methylenation of the functionalized ketone (for examples see
Scheme 22 and reference [l]), whilst there is n o such preference
+ 2 Me,AI}
1{ MoO,GI,
Ph %NEt2
28 %(37 %)
Scheme 23. Proposed mechanism for carbonylmethylenation and also the alternative 1-methylation of 43 with {MoO,CI, + 2 Me,Al) [I]. Conditions. THF,
-78 + 25°C in 18 h [1.22k]. The yields given in parentheses ;ire those for the
corresponding reaction with {MoO,CI, + 2 Me,AI f 2 HMPA;
addition of a base (HMPA,@'] Ph,P) (cf. Table 5) has a distinct
promoting effect on the carbonylmethylenation reaction with
Mo-A1 and W-A1 reagents, and offers further support for the
ring-opening mechanism. This effect has also been reported for
the Tebbe reagentr541and a Zr-A1 complex of similar struct ~ r e [ ~ and
' ] is in these cases, without a doubt, caused by ringopening.
4.5. Anticheleselectivity
42 (35 %)
61 %
The intermolecular competition reactions given in Scheme 24
with 1,3-dimetallacyclobutanereagents reveal that methoxy,
1 {WOCI,
{2 MoOCI,
(89 Yo)
48 %
+ 2 Me3AI}
5 0 % ~ ~ ~ ' 20%[71:29]
Scheme 22. Cheleselective carbonylmethylenation with {MoO,CI, f 2 Me,Al}-70+20'Cin 12 h:mole
[22k]and{WOCI, f 2 Me3AI}[220].Conditions:THF,
ratio reagent .substrates = 1: 1: 1. The yields of ketone recovered are given in parentheses
discernible for hydroxy ketone/ketone pairs. An amine-N atom
in a /3 or y position to the keto group often drastically retards the
carbonylmethylenation with { WOCl, + 2 Me,Al} . This has
been attributed to the formation of stable complexes.['] On the
other hand, an electron-donating atom in a 6 position to the
keto group has a marked promoting effect on the carbonylmethylenation of 42 with {WOCl, + 2 Me,Al) (Scheme 22) as
well as of MeCO(CH,),COMe and MeCO(CH,),CO,Et with
{MoOCl, 2 Me,Al} .[I1 For the carbonylmethylenation of ketones that contain an electron-donating atom in an a or /3 position to a keto group, it is proposed that there is an active terminal CH, ligand which was formed by ring-opening (Scheme 23).
Clear evidence in support of this mechanism is the a-methylation of 43 * 44,'''also given in Scheme 23, as well as analogous
a - m e t h y l a t i ~ n s . [The
~ ~ ~latter
is an alternative reaction to carbonylmethylenation, and can only be readily explained by a
combination of C-H insertion and reductive elimination. The
{2 MoOCI,
+ 4 MeLll
37 %[391
52 %C22d'
45%c22 "
Scheme 24. Anticheleselectivity of 1.3-dimetallacyclobutane rexgents in intermolecular competition experiments. Reaction conditions and other information as
in Scheme 21
C l i ~ i i iI n / .
Ed EngI. 1997. 36, 1258-1275
amino, dimethylamino, piperidino, and nitro groups retard the
carbonylmethylenation of an aldehyde or keto group in the a or
position (for further examples see reference [39]). Whilst retardation in the case of 45 can be explained by the M effect of
the amino group, there must be other reasons for this for 46
( - M effect of the nitro group) and for the other remaining
cases. The moderate anticheleselectivity observed in competition reactions with ketone/methoxy ketone as well as ketone/
T. Kauffmann
dialkylamino ketone pairs is similar to the anticheleselectivity
observed in the corresponding reactions of transition metal carboxylates and of dimeric ally1 and cyanomethyl transition metal
reagents, which are probably dimeric. In these cases it is attributed to steric hindrance.". 631 This interpretation is suggested for the 1,3-dimetallacyclobutanes with bulky T H F ligands.
In addition, deactivating complexation could also be involved.
In 45 and 46 the inhibiting action of the amino and nitro group,
respectively, is so great that even in individual experiments there
is no carbonylmethylenation. There is little doubt that stable
substrate-reagent complexes have been formed here.[391Apart
from these exceptions, the functionalized aldehydes and ketones
listed in Scheme 24 reacted with the cited reagents in single
experiments to give good yields of the methylenation products.
The reagents given in Scheme 25 can differentiate well between 47 and carvone. The reagent (2 MoOCl, + 4 MeLi} was
particularly selective. The unconverted ketone was recovered in
high yields.1391
Q9 Q4
n 3 1
49 (42%)
4.6. Ketone- Enone Differentiation
(mole ratio 1:1 and 2: 1 , respectively) to 2-vinyithiophene (78 %)
and 2-vinylpyrrole (95 %), respectively.rzze1Therefore, it is
probable that during monomethylenation a complex of type 48
is formed in which the electrophilicity of the remaining aldehyde
group is greatly diminished (cf. the retardation of carbonylmethylenation of 1,2-diketones in Section 4.8). The reagent 31
reacts with di- and triketones 49, 51, 53, and 55 (Scheme 27)
T = 20
58 %
T = 50
60 %
T = 50
65 %
Thiophene- and pyrrole-2,5-dicarboxaldehydecan be monomethylenated with extreme selectivity if less than an equimolar
amount of (2 MoOC1, + 4 MeLi] (31) is used (Scheme 26). In
the case of pyrrole-2,5-dicarboxaldehydethe reaction stops at
the monomethylenated stage, even if an equimolar amount of
reagent is used. The deactivation of the second aldehyde group
by the M effect of the heteroaromatic moiety cannot be the
crucial factor, because thiophene- and pyrrole-2-carboxaldehyde are smoothly methylenated by 12 MoOCI, t 4 MeLi}
0 % (59 %)
0 % (0%)
Scheme 26. Monomethylenation of dialdehydes with {2 MoOCI, + 4 MeLi) (31)in
THF, -78 + 20'C in 18 h [22e,35h,35i]. Yields given: values not in brackets are
valid for n = 0.5, values in brackets for n = 1.
40 %
51 %
5 4 a . 2 5 ~ ~ 54b,4%
<t %
1 3 1 )
10 0% [87:131
4.7. Highly Selective Monomethylenation of Dialdehydes,
Diketones, and Triketones
3 % [95:5]
Scheme 25. Ketone-enone differentiation with 1,3-dirnetallacyclobutane reagents
t35g.391. In square brackets: selectivity.
X= S
X = NH
53 (a.
65 %)
{2 MoOCl3+ 4 MeLi}
(2 WOCI, + 4 MeLi}
{2 MoOC13 + 4 MeLi}
Scheme 27. Monomethylenation of diketones 49 [35h,43] and 53 [22rn,43] as well
asthetriketones51[22m,43]and55[43,67].Conditions:THF,-78 4 2 0 T i n 1 8 h.
Compound 56 was the only product of the reaction of 55, and was only detected
qualitatively The yields in parentheses give the amount of substrate recovered.
with high selectivity to give the monomethylenated products
52 [22m1 (along with traces of a dimethylenated product),
54a,b [z2m1 and 56,[641respectively. In the case of the triketones,
a peripheral keto group is methylenated.r431The reagent employed demonstrates a particular affinity towards hydroxy ketones (see Section 4.4). Therefore, it is proposed that 49 (if solvent-free it is enolized to 76.4%[651),51, and 53 react to give
molybdenum enolates 57,59, and 60, respectively (Scheme 28),
so that only one free keto group remains which can be methylenated. Since diketone 49 was recovered in 43 % yield after treatment with excess reagent and subsequent addition of water,
(second experiment in Scheme 27) then it is probable that not
only 57 was produced, but the endiolate 58 as well, in which
both keto groups are protected against methylenation. Surprisingly, methylenation of the cyclic 1,2,3-triketone 55 does not
take place in the middle position, which is particularly electrophilic, but exclusively at a peripheral
641 This can
be explained by the formation of a stable 1,Zdiketone complex
(61), which probably methylenates a similar type of complex in
which there is still a free keto group (cf. Scheme 29 B) The openchain 1,2,3-triketone PhCOCOCOPh is not methylenated by 31
Angeu. Chem. Inr. Ed. Engl. 1997,36, 1258-1275
Organomolybdenum and -tungsten Reagents
8 PhCOCl
1) 1 {2MoOC13+4 M e L i }
2) Et OH + Et ONa
L, MO( ,L/-CH~)~MOL,
> 4-MeOC6H4CH =CH,
3 56
MO(L,) ( P-CH,),MOL,
Scheme 28. Postulated intermediates for the highly selective rzactions of
Scheme 27.
but degraded by decarbonylation to benzil.[22m1
The extremely
selective monomethylenations in Scheme 26 and 27 illustrate the
extent of the proficiency of a methylenating reagent if it can
form quasi-aromatic complexes with the substrate.
4.8. Tolerated and Nontolerated Functional Groups
4.8.1. 1,3-Dimolybdacyclobutanesas reagents
Representatives of the compound classes RCOC1,[39.661
RCOOOCR,[22d1 R’COOR2,1431 RCONMe,,1431 ArCOCOAr,[22d.
RCH=CH-CH=CH213’] as well as the individual compounds
Ph2C=C=0,‘35Jl PhCN,‘35c’
381 were not permanently altered
PhCH2C1,[641and CH2C12f35f.
on reaction in T H F with the reagents obtained from [Mo,CI,,],
MoOCI,. and MoOCl, according to Scheme 13. The corresponding tungsten reagents have not been well investigated. In
some cases (e.g. aromatic nitro compounds; see Section 4.5)
there may be some fairly stable reagent -substrate complexes
formed, from which the substrate is released unchanged on addition of water. Since acyl chlorides are usually more powerful
electrophiles than aldehydes, their resistance to the molybdenum reagents is surprising.[39,661 Intermolecular competition
reactions reveal that the carbonylmethylenation of 4-methoxybenzaldehyde or acetophenone are not hampered by acyl chlorides, even in excess (Scheme 29A[391). This is similar to the
inertness of Lewis acidic methyl transition metal reagents‘661
such as MeTiC1, and MeNbCI, towards acyl chlorides. However, it does contrast with the behavior of carbonylmethylenating reagents such as Ph,P=CH, and [Cp2Ti=CH2](formation
of [RCOCH2PPh3]C11671
or [RC(=CH2)OTi(Cp)2C1][541)
and is
informative with regards to the mechanism of carbonylmethylenation with molybdenum reagents (see Section 4.3). In addition,
they do not react with carboxylates, benzyl chloride, benzonitrile, and tolan. This differentiates them from [Cp,Ti=CH,] and
other complexes containing terminal CH, ligands which attack
these s u b ~ t r a t e s . [ ~The
~ - ~aromatic
IJdiketones benzil and
9,lO-phenanthrenequinone are not methylenated by one equivalent of {2 MoOCI, + 4 MeLi) or 12 MoOCl, + 4 MeLi) (recovery of the ketone 80-90%), and 9,lO-phenanthrenequinone
does not inhibit the methylenation of benzaldehyde with
Angm Clrem Int Ed En81 1997, 36, 1258-1215
n= 1
n= 2
Scheme 29 Resistance of acyl chlorides [22e,39] and 1.2-diketones [22d, 35 h]
towards carbonylm~thylenatingmolybdenum reagents in THF, - 78 + 20 C in
18 h. The yields in parentheses give the amount of diketone recovered.
{2 MoOCl, + 4 MeLi)[22d1(Scheme 29B). The aliphatic 1,2diketone 62 is also remarkably resistant towards
{ 2 MoOC1, + 4 MeLi} , because there is no carbonylmethylenation with only one equivalent of reagent. Only when two
equivalents are used, does this reaction proceed with low
yields[35h1(Scheme 29C). Since the sole product is the bismethylenated 63, the first methylenation step is evidently slower
than the second. The behavior of the 1,2-diketones, as well as
the formation of benzil from PhCOCOCOPh, mentioned in Section 4.7, leave little doubt that thermodynamically stable 1,2diketone-reagent complexes are formed in which the keto
groups are protected against methylenation. Azomethynes,
epoxides, and nitrosobenzene have proved to be nonresistant
towards carbonylmethylenating 1,3-di1nolybdacyclobutanes.
Anomethynes are “azomethyne-methylenated” in moderate
yields by {[Mo,Cl,,] + 4 MeLi), whilst aromatic epoxides are
deoxygenated and aliphatic epoxides are converted to x-chlorohydrines.135 c .
Nitrosobenzene is inexplicably altered by
{2 MoOCI, + 4 MeLi) .[35c,431
4.8.2. (MoO,CI, + 2 Me,AI)
This reagent is optimal for the methylenation of aldehydes.
However, in T H F at about 2 0 T , it proved to be resistant towards acyl chlorides, ethyl carboxylates, nitriles, phenyl isothiocyanate, azobenzene, and diphenyl ~ u l f o n e . [Nevertheless,
did reduce diphenyl sulfoxide to diphenyl sulfide.[, ’Jl
4.9. Comparison of Reported Carbonylmethylenating
With regard to their aldehyde versus ketone selectivity, low
basicity, relatively high resistance towards protons, and ready
availability, carbonylmethylenating reagents of the 1,3-dimolybdacyclobutane type are clearly superior to strongly basic
carbonylmethylenating reagents. These include the classical
reagents such as Ph3P=CH2,[681Ph2(0)PCH,M (M = Li,
Na),[691and Me3SiCH,Li,i7*1(which are linked to the names of
T. Kauffmann
Table 9. Non- or weakly basic carbonylmethyleiiering reagents reported i i i the literature Category A : wcak. selective nucleophiles (substraies. aldehydes and ketones)
Category B: strong nucleophiles (substrates: additionally carboxylates. carbox;lmidcs. and lactoncs)
[CHJ, + f?Zn/Cui 64 [76-7X] (IZnCHJnl) [h]. Fried-. Hashitnoto reagent
nZii TiCI,; aged), X = Br, 1. 65 [79]. Oshima Lomhardo reagents
Me,SiCH,'l'iCI, 66 [74]
Mc,SiCH,CrCII 67 [74]
i.l-methyleninolybdcnurn complexes [16 40, 42. 431
[CHJ, + n Z n Ti(OiPr),) 68 [SO]
(CH,I, + n Z n Me,AIj 69 [8@]
[Ph,P(O)CH 2'fi(OiPr)n]Li 70 [75]
(Ph,P(0)CH2),CrCI 71 [75]
complexes [41 .43]
/CH,I, + nCrCI,) 72 [ X l ] (L.Crii'CH,Cri"L,,) [b]
33 [53]. Tebbc reagent
Cp,Ti=CH, 78 [54], Grubbs reagent
(CH,Br, + n Z n + TiCI, + TMEDA) [c] 79 1891 (complex with terminal CH, ligand) [h]
{[Cp,XMe,] bci 6 0 - 6 5 ' C ) 80 [YO]
+ Represents good aldehyde versus ketone selectivity
Wittig, Horner, and Peterson) along with Ph,MCH,Li
(M = Gc, Sn, Pb),[7'' Ph2MCH2Li (M = As, Sb, Bi),[71JPhSCH2Li,r721
and PhSeCH,M (M = Li, Na).[731A superior alternative are the noii- or weakly basic, weakly nucleophilic carbonylmethylenating reagents 64-72 listed in Table 9: the
reagents 66, 67, 70, and 71[74.7s1
synthesized by us, along with
similar, modified Peterson'74J and Horncr reagents[7s1 can
methylenate with high aldehyde selectivity and in good yield,
however, they are rclatively difficult to obtain. On the other
hand, the reagents 64,[76-781 65,[79168,[801and 72,[8'1which
also appear to be dimctalliomethanes,[821are serious competition for the reagents of 1,3-dimolybdacyclobutanetype because
of their straightforward preparation, nonexistent or low basicity, and their proven selectivity. However, this is only partly true
for 65.
Compound 64, prepared by Hashimoto et 2 1 1 . ~ ~(THF,
6 h,
40°C) and by Fried et al.'771(Et,0, 4 h, 78°C) by rcaction of
CH,I, with a Zii-Cu couple, methylenates aldehydes (yield
18--61(YO),but does not attack either ketones or methyl benzoate in THE['"] In E t 2 0 it also methylenatcs ketones, but only
in good yields if the reaction is supported by a sterically favorable hydroxyl
781 This permits the highly regioselective methylenation of 73 in high yields (Scheme 30). The reaction 74 +76 (yield not mentioned) demonstrates that the
assisting hydroxyl group may be situated several bonds away
from the keto group in a folded ring system. In analogy to the
hydroxyl-promoted Simmons- Smith reaction of ally1 and hoinoallyl alcohols (see reference [77]) as well as to the hydroxylpromoted carbonylmethylenations described in Section 4.4, it is
proposed that there is a similar covalent fixation of the reagcnt
as in species 75.r781
If reagent 64 is used, the cyclopropanatioii
of thc primarily formed vinyl group can occur as an interfering
side reaction. The Fried- Hashimoto reagent 64, which has incited little interest to date, is certainly a good alternative or
supplement to the molybdenum reagents with an affinity for
hydroxy ketones (Section 4.4). Admittedly, its range of application has not been sufficiently well investigated. The reagents
("Oshima Loinbardo
derived from 64,
have recently been the subject of an exhaustive rcvicw by
Stille.r7ycJThey are suitable for the methylenation of ketones
and display ketone versus carboxylic acid selectivity. However,
[b] Structure proposcd by the authors. [c] TMEDA
First publications
Aldehyde selectivity [a]
= teti-amethylethyleiiediaminc
- dPC
Scheme 30. Hydroxyl group promoted carbonylmethylenation with the FricdHashimoto reagent [77,78].
they are unsuitable for the methylenation of aldehydes, because
reductive coupling of the carbonyl function occurs as a competing
A hydroxyl group in an x position to the keto
group distinctly retards methylenation with 65, X = Br,lE3I
whilst a [{-positioned group docs not.[84'The reagents 6SrSo1
are, in contrast to 65, equally suitable for the methylenation of aldehydes and ketones, and are distinguished by high
aldehyde versus kctone selectivity. If the carbonyl group of dodecanal is protected in situ by Ti(NEt,),, as according to Rcetz
et a1.["], 69 inethylcnates exclusively the 4-dodecanone (95 %
alkcne), which is also present in the reaction mixture.[*']
Whether the chromium reagent 72 introduced by Takai et al.[*']
methylenates with stringent aldehyde selectivity, does not
appear to have been tested. However, because of the high
aldehyde selectivity of the analogous reagents {Me,SiCHBr, + n CrC12)[861and (CHI, +?? CrCl,)[*71this selectivity can be presumed. Using the readily available chromium
reagents of type (alkylCHHa1, + n CrCl,) (77), Takai et al.[**l
were able to undertake the important jump from regio- and
Organomolybdenum and -tungsten Reagents
chemoselective carbonylmethylenation to the corresponding
carbonylalkylidenation. Lj-H Elimination, which destroys the
reagents, apparently does not occur to a high extent at low
temperatures. However, the yield of the alkylidenation (E-selective) of ketones with 77, R = nPr, iPr, tBu is regarded as “rather
low” (in the example given for 77, R = nPr, +cycledodecanone: 15% yield and 75% recovery of the ketone),
whilst this limitation does not apply to aldehydes and the reaction of 77. R = Me with ketones.[”] Carbonylalkylidenations
(good yields; E- and aldehyde selective) are also possible with
the reagents introduced by Knochel and Normant[”]
{allylZnBr + 1-alkenylMgBr + BF,-OEt,j. Because of the
particular method of reagent preparation only specific alkylidene groups are transferable.
The non- or weakly basic, strongly nucleophilic carbonylmethylenating reagents 33 and 78-80, listed in the lower part of
Table 9. can carbonylmethylenate not only aldehydes and ketones but also carboxylates and carboxamides as well as lactones. They can only be regarded as an alternative to 1,3-dimolybdacyclobutanes if the weak basicity of the reagent is
important (readily enolizable aldehydes and ketones as well as
aldehyde and ketones with a-positioned stereocenters as substrates). Because of the high nucleophilicity of these reagents,
high aldehyde selectivity or selectivity by hydroxyl group assistance can be expected to a lesser degree than in reagents of
category A in Table 9. The Tebbe reagent (33),[531
used to methylenate carboxylates and carboxamides as well as
lactones, is not very aldehyde ~ e l e c t i v e . [ ~It* ’is~deactivated by
hydroxyl groups in the substrate and reacts with bases to give
the highly reactive, less selective, carbonylmethylenating
Grubbs reagent 78. This reagent can also be produced by the
thermolysis of ti t a n a c y c l ~ b u t a n e s . [The
~ ~ ’ carbonylmethylenating reagent 79, introduced by Takai et a1.,[8y1
does not differentiate between aldehydes and ketones (for corresponding alkylidenations see reference [89a,b]). The authors assume that the
active species of 79 is a complex with a terminal CH, ligand, and
demonstrated that traces of lead in the zinc used were a prerequisite for the high nucleophilicity of 79.[sy‘1The carbonylmethylenating reagent 80, described by Petasis et al.,190a1does
not appear to methylenate by primary formation of 78, but by
complexation on the carbonyl group, subsequent methyl transfer and elimination of CH,. Apparently, aldehyde selectivity,
which not is expected on the grounds of the high reaction temperature required for 80, was not reported. Corresponding
alkylidenations are possible with [Cp,TiR,] ( R = groups which
d o not contain any P-H atoms).[’ob-dl The ketone-methylenating reagent {[Cp,TiCI,] + Me,Zn),[”] which has not been well
investigated, should probably be assigned to category B in
Table 9.
5. Conclusion and Future Prospects
Thermolabile complexes [R,(nPrO),W(p-OnPr),W(OnPr),R,] (6; R = Me, Et, nPr, nBu, iBu, sBu, Me,SiCH,, Ph) were
obtained in situ by nucleophilic substitution with organolithium
or Grignard compounds from the reported [Cl,(nPrO),W(pOnPr),W(OnPr),Cl,] (2). They react with aromatic, heteroaromatic, or %,/$-unsaturated aldehydes and ketones:
R ’ R 2 C 0 + 0.5 R‘RZR3C-CR’R2R3. Hitherto, this had not
been possible in one step. In an orientating experiment with
R = Ph, such an additive-reductive carbonyl dimerization
(ARCD reaction, for postulated mechanism see Scheme 8) was
also successful on an aliphatic ketone. Complex 6, R = Me,
converts benzoic acid derivatives into Me,PhC--CPhMe, . In
order to make the A R C D reaction interesting for applications in
organic synthesis, it is still necessary to improve the accessibility
of 2 or analogous alkoxy compounds, and to sound out the
possibility of converting aliphatic aldehydes and ketones into
A R C D products with reagents 6, R = Ar.
The CH, complexes prepared at low temperature by action of
MeLi or Me,A1 on chlorides of high-valent molybdenum and
tungsten display a great tendency to eliminate CH, and transform into thermolabile, carbonylmethylenating p-CH, complexes. This transformation of a methyl ligand into a pmethylene ligand has been monitored qualitatively by N M R
spectroscopy, and has also been quantified by a combination of
gas measurement and yield determination during carbonylmethylenation reactions (Table 6). Depending on whether they
were methylated with MeLi or Me,AI, the p-CH2 ligands are
present either in MCH,M or in MCH,Al bridges (M = Mo,
W). The N M R data of some p-methylenemolybdenum and
-tungsten complexes permitted detailed structural proposals.
The p-methylenemolybdenum complexes, easily accessible by
treatment with MeLi, have a high selectivity potential as readily
modified, Lewis acidic carbonylmethylenating reagents that are
surprisingly insensitive towards protons. An examination of the
literature for Section 4.9 revealed that they have been more
intensively investigated as regards their selective behavior than
other carbonylmethylenating reagents of comparable selectivity. The selectivity of the molybdenum complexes exerts itself
above all in reactions with keto aldehydes, hydroxy ketones,
dialdehydes, di- and triketones, and is partly based on the formation of quasi-aromatic cyclic compounds in which one or two
carbonyl groups are blocked. Attention should also be paid to
carbonylmethylenating p-methylenemolybdenum and -tungsten
complexes obtained from reaction with Me,AI, and which are
regarded as bicyclic. They have a particular affinity towards
keto groups which contain a basic atom in the 6 position. Future
research will confirm o r disprove the proposed structures of
26-32 as well as 35-37 of p-CH, complexes and the probable
mechanism of the methyl + p-methylene ligand transformation. Future investigations will show whether particular examples of the molybdenum and tungsten complexes described
here prove to be suitable as selective carbonylmethylenating
reagents in organic synthesis.
I have tried to present the experimental work carried out by m-v
co-workers cited in the references and which was discussed in our
“Friday meetings”. I would like to thank all who participated in
this work,for their great commitment and many intelligent, original contributions. Furthermore, m y thanks go to the Deutsche
Forschungsgemeinschaft, the Volkswagen-Stiftung, and the Fonds
der Chemischen Industrie for their ,financiul support.
Received: March 1. 1996 [A1521E]
German version: Angrw. Chem. 1997. f 0 Y . 1312-1329
Translated b y Gtlhan Scheibelem. Offenbury. Germany
[ t ] T. Kauffmann, M. Enk, P. Fiegenbaum, U. Hansmersmann, W. Kaschube,
M. Papenberg, E. Toliopoulos, S. Welke, Chem Ber. 1994. f27. 127-135.
[2] M. S. Kharasch, P. 0. Tawney, J. Am. Chem. SOC.1941.63, 2308-2316.
131 R. R. Schrock, G. W. Parshall, Chem. Rev. 1976, 76, 243-268.
[4] J. Kochi, Organometallic Mechanisms und Catalysis, Academic Press, New
York, 1978, pp. 230-755.
[5] T. Kauffmann, Sjnthesi.s 1995, 745-755
[6] T. Kauffmann, Angen. Chem. 1996,108.401-408; Angew. Chem. Int. Ed. Engl.
1996,35, 386-403.
[7] K -H. Thiele, Pure Appi. Chem. 1972, 30, 575-585.
[8] C. Santini-Scampucci, J. G . Riess, J. Chem. Soc. Daifon Trans. 1976. 195-200.
[9] A. L. Galyer, G. Wilkinson, J Chem. SOC.D d o n Trans. 1976, 2235-2238:
A. J. Shortland, G. Wilkinson, ibid. 1973, 872-876.
[lo] K.-H. Thiele, U. Dieckmann, 2. Anorg. A&. Chem. 1972, 394, 293-300.
[ l l ] E. L. Muetterties, Inorg. Chem. 1975, 14. 951-953.
[I21 D. N. Clark, R. R. Schrock, 1 Am. Chem. SOC.1978, 100, 6774-6776.
[I31 R. A. Andersen, M. H. Chisholm, J. F. Gibson, W. W. Reichert, I. P. Rothwell,
G. Wilkinson, Inorg. Chem. 1981, 20, 3934-3936.
1141 0. Klejnot, Inorg. Chem. 1965.4. 1668-1670.
[15] D. P. Rillema, W. J. Reagan, C. H. Brubaker, Jr., lnorg Chem. 1969,8, 587590; D. P. Rillema, C H. Brubaker, Jr., ibid. 1969, 8, 1645-1649.
[16] W. J. Reagan, C. H. Brubaker, Jr., Inorg. Chem. 1970, 9,827-830.
1171 F. A. Cotton, D. DeMarco, B. W. S . Kolthammer, R. A. Walton. Inorg. Chem.
1981, 20. 3048-3051.
1181 F. A Cotton, L. R. Falvello, M. F Fredrich, D. DeMarco, R. A. Walton,
J. Am. Chem. SOC.1983, 105. 3088-3097.
1191 L. B Handy, K. G. Sharp, F. E. Brinckman, Jnorg. Chem. 1972,ff. 523-531.
[20] H. Funk, G. Mohaupt, 2. Anorg. A&. Chem 1962,315,204-212.
[21] T. Kauffmann, J. Jordan, K:U. VoB, Angen. Chem. 1991, 103. 1160; Angeu.
Chem. Jnt. Ed. Engl. 1991, 30, 1138- 1139.
1221 Dissertations, Universitat Miinster: a) J. Jordan, 1990; b) K.-U. VoB, 1992;
c) H.-W Wilde, 1992; d) P. Fiegenbaum, 1987; e) R. Wieschollek, 1986; f ) G .
Kieper, 1985; g) M. Papenberg, 1991; h) S. Welke, 1988; i) H. Bocker, 1985;
j) J. Baune, 1988; k) M. Enk, 1988: I) U. Hansmersmann, 1987; m) J. Hilsmann,
1988; n) U. Pucknat, 1990; 0 ) E. Toliopoulos, 1989.
[23] T. Kauffmann, J Jordan, K.-U. VoB, H.-W. Wilde, Chem. Eer 1993, 126.
2083 - 2091.
[24] a) J. March, Advunced Organic Chemi.strj, 3rded, Wiley, New York, 1985,
S. 1 f 10; b) T. Wirth, Angew. CJiem. 1996, 108, 65-67; Angen.. Chem. Int. Ed.
Engl. 1996, 35, 61-63.
(251 a) J. E. McMurry, Ace. Chem. Res. 1974, 7,281 -286; ibid. 1983,16,405-411;
Chem. Rev. 1989,89,1513-1524; A. Fiirstner, Angew. Chem. 1993,105,171197; Angeu. Chem. Int. Ed. Engl. 1993,32,164-189;A. Furstner, A. Hupperts,
J. Am. Chem. Soc. 1995, 117, 4468-4475; b) Y. Fujiwara, R. Ishikawa, F.
Akiyama, S . Teranishi, J. Org. Chem. 1978, 43. 2477-2480; M. Petit, A.
Mortreux, F. Petit, J. Chem. SOC.Chem. Commun. 1984, 341-342.
[26] E. E. van Tamelen, M. A. Schwartz, J. Am. Chem Soc. 1965.87, 3277-3278;
J. E. McMurry, M. G . Silvestri, J Org. Chem. 1975, 40, 2687-2688.
[27] E. L. McCann, T. M. Brown, Inorg. Synth. 1972, 13, 150-154.
(281 M. A. Schaefer King, R. E. McCarley, Inorg. Chem. 1973, 12, 1972-1979.
[29] L. B. Anderson, F. A. Cotton, D. DeMarco, A. Fang, W. H Ilsley, 8.W. S.
Kolthammer, R. A. Walton, J. Am. Chem. SOC.1981. 103. 5078-5086.
[30] R. W. Hartmann, W. Schwarz, A. Heindl, H. Schoenenberger, 1 Med. Chem.
[31] K. Ziegler, W. Deparade, Justus Liebigs Ann. Chem. 1959. 567, 123-141.
[32] K. B. Sharpless, M. A. Umbreit, M. T. Nieh, T C. Flood, I Am. Chem. SOC.
[33] M. B. Hursthouse, R. A. Jones, K. M. A. Malik, G . Wilkinson, J Am. Chem.
SOC.1979, 101,4128-4139.
[34] R. B. Calvert, J. R. Shapley,J Am. Chem. Soc. 1977, 99,8225-5226; ibid. 1978,
100, 6544.
1351 Diploma theses, Universitat Miinster: a) J. Sander, 1983, b) R. Kucznierz,
1988; c) P. Fliegenbaum, 1983; d) U. Pucknat, 1987; e) S. Welke, 1984; f ) M.
Papenberg, 1987; g) C. Neiteler, 1988;h) S. Robbe, 1988;i) W. Kaschube, 1984;
f) M. Enk, 1986; k) J. Jordan, 1987.
I361 T Kauffmann, B. Ennen, J. Sander, R Wieschollek, Angen. Chem. 1983, 95,
237-238; Angew. Chem. Int. Ed. Engl. 1983,22,244.
[37] T. Kauffmann, P. Fiegenbaum, M. Papenberg, R. Wieschollek, J. Sander,
Chem. Ber. 1992, 125, 143-148.
[381 T Kauffrnann, P. Fiegenbaum, M. Papenberg, R. Wieschollek, D. Wingbermiihle, Chem. Eer. 1993, 126, 79-87.
[39] T Kauffmann. J. Baune, P. Fiegenbaum, U. Hansmersmann, C. Neiteler. M.
Papenberg, R. Wieschollek, Chem. Eer. 1993, 126, 89-96.
[401 a) T. Kauffrnann, P. Fiegenbaum, R. Wieschollek, .4ng<ew. Chem. 1984, 96,
500-501. Angew. Chem. In[. Ed. Engl. 1984, 23, 531, b) T. Kauffmann. G.
Kleper. h c l . 1984, 915~502-503 and 1984, 23, 532-533.
14'1 T Kauffmann. R.Abeln, S. Welke, D Wingbermiihle, A n ~ r u Chem.
1986, y8,
927-928; Angew Chem. In!. Ed. Engl. 1986, 25, 909.
[421 T Kauffmann, M. Enk, w. Kaschube, E. Toliopoulos. D Wingbermuhle,
Angen'. Chem. 1986, 98, 928-929; Angen. Chem. In[. Ed. En*/. 1986, 25,
T. Kauffmann
[43] T. Kauffmann in Advances in Metal Curhene Chemistry (Ed.: U. Schubert),
Kluwer, Dordrecht, 1989, S. 359-378.
[44] MoOCI,, WOCI, and MoC1, were employed as bis(tetrahydr0furan)complexes
1451 M. T. Reetz, S. H. Kyung, M. Hiillmann, Tetruhedron 1986, 42, 2931-2935.
[46] T. Kauffmann. T. Abel, M. Schreer, D. Wingbermuhle, Tetrahedron 1987, 43,
2021 -2028.
1471 W. A. Herrmann, Adv. Organomet. Chem. 1982, 20. 159-263.
[48] B.J. J. van de Heisteeg, G. Schef, 0. S. Akkerman, F Bickelhaupt,
J. Organomet. Chem. 1986,308, 1- 10.
[49] K. Isobe, A. Vazquez de Miguel, P. M. Bailey, S . Okeya. P. M. Maitlis,
J. Chem. Sor. Dalton Trans. 1983, 1441-1447.
[50] F. W Hartner, Jr., J. Schwartz, J. Am. Chem. Soc. 1981, 103,4979-4982; F. W.
Hartner, Jr.. J. Schwartz. S . M. Clift, ibid. 1983, 105, 640-641.
[Sl] R. R. Schrock, S. Rocklage, J. H. Wengrovius, G. A. Rupprecht, J. D. Fellmann. I ,4401. Catal. 1980, 8, 73-83; J. H. Wengrovius, R. R. Schrock,
Orgunometa//ics 1982, 1, 148-155.
[52] a) D. H. Williams, 1. Fleming, Spektroskoprsche Methoden zur Strukruruufklirung, 3rded.. Thieme, Stuttgart, 1975, S. 118; b) H.-0. Kalinowski, P.
Berger, S . Braun, '3C-NMR-Spektro.skopiee,1st ed., Thieme, Stuttgart, 1984,
p. 63.
[S3] a) F. N. Tehhe, G. W Parshall, G. S . Reddy. J. Am. Chem. Soc. 1978, 100,
361 1 - 3613; b) E. Negishi, D Choueiry in Enc~dopediaof Reagentsfor Orgunic Synthesis, Vol. 7 (Ed.: L. A. Paquette), Wiley, Chichester. 1995, pp. 51905192; c) S. H. Pine. Org. Relict. 1993.43,l-91; d) S. H. Pine, R. J. Pettit, G . D.
Geib, S . G Cruz, C. H. Gallego, T. Tiierina, R. D.Pine, J. O r , .Chem. 1985,SO.
K. A. Brown-Wensley, S. L. Buchwald, L. Cannizzo, L. Clawson, S. Ho, D.
Meinhardt, J. R. Stille. D. Straus. R. H. Grubbs. Pure Appl. Chem. 1983, 55,
1733- 1744.
R. R. Schrock, Acc. Chem. Res. 1979, 12, 98-104.
A. Aguero. J. Kress, J. A. Osborn. J Chem. Soe. Chem. Commun. 1986. 531 533; ibid. 1985. 793-794.
K. C. Ott, E. J. M. de Boer, R. H. Grubbs, Organometallics 1984,3,223-230.
M. T. Reetz. Organotitanium Reugents in Organic Synthesis, Springer, Berlin,
1986, p. 2.
T. Kauffmann, C. Beirich, A. Hamsen, T. Moller, C. Philipp. D. Wingbermiihle, Chem. Ber. 1992,125,157-162.
D. C. Bradley, R. K. Multani, M. Wardlaw, J. Chem. SOC. 1958, 46471651
Ref. [%I, S. 75
HMPA is regarded as a particularly favourable ligand for Al: T. Mole, E. A.
Jeffrey, Organouluminium Compounds, 1st ed., Elsevier, New York, 1972,
p. 110.
T. Kauffmann, T. Moller, H.-W. Wilde, Chem. Ber. 1994, 127, 2277-2283, and
references therein.
T. Kauffmann, B. Laarmann, D. Wingbermiihle, unpublished results from
Ref. [24a], S. 67.
T. Kauffmann, T. Abel, C. Beirich, G. Kieper, C. Pahde, M. Schreer, E Tohopoulos. R. Wieschollek, Tetrahedron Len 1986, 27, 5355-5358.
H. J. Bestmann, B. Arnason, Chem. Ber. 1962,95, 1513-1527.
G Wittig, Angen. Chem. 1956, 68, 505-508; G Wittig, G. Geissler, Justus
Liebigs Ann. Chem. 1953,580,44-57.
L. Homer, H. Hoffmann, H. G . Wippel, Chem. Ber 1958, 91, 61-63; dd.
1959, 92,2499-2505.
D. J. Peterson, J. Org. Chem. 1968,33,780-784; D. J. Ager, Org. React. 1990,
38. 1-223.
T. Kauffmann, Top. Curr. Chem. 1980,92, 109-147; Angen. Chem. 1982,94,
401-420; Angeu, Chem. Int. Ed. Engl. 1982,21, 410-429; T. Kauffmann, R.
Kriegesmann. A. Hamsen, Chem. Eer. 1982, lf5, 1818- 1824; T. Kauffrnann,
R. Jou0en. N. Klas, A. Vahrenhorst, ihid.1983,116,473-478; T. Kauffmann,
R. Knegesmann, A. Rensing. R. Konig, F. Steinseifer. ibid. 1985, fl8,370379; T.Kauffmann, F. Steinseifer, N. Klas, ibid. 1985, lf8, 1039-1044.
R. L. Sowerby. R M. Coates, J Am. Chem. Soc. 1972, 94,4788-4759.
J. Remion, W. Dumont. A. Krief, Tetrahedron Lett. 1976, 1385-1388.
T Kauffmann, R. Konig, C. Pahde, A. Tannert, Tetruhedron Lett. 1981, 22.
5031 -5034.
T. Kauffmann. P. Schwartze, Chem. Eer. 1986, 119, 2150-2158.
H. Hashimoto, M Hida, S. Miyano. Kogyo Kagaku Zusshi 1966. 69, 2134;
J. Oraunomet. Chem. 1967. 10. 518-520.
1771 P Turnbull. K . Syhora. J.H. Fried, J. Am. Chem Soc 1966, 88, 47644766.
1781 1. T. Harrison, R. J. Rawson, P. Turnbull, J. H. Fried, J. Org. Chem. 1971, 36,
1791 a) K. Takai, Y. Hotta, K. Oshima, H. Nozaki, Tetrahedron Lerr. 1978,24172420; b) L Lomhardo, ibid. 1982,23,4293-4296: c) J. R. Stille in Comprehensive Organometallic Chenristrv 11, Voi 12(Eds.: E. W. Ahel, F. G. A. Stone, G.
Wilkinson (Ed : L. S . Hegedus), Elsevier, Oxford, 1995, p. 577-599.
[801 T Okazoe. J Hibino. K. Takai. H. Nozaki, Tetrahedron Lett. 1985. 26, 55815584.
Angew. Chem. Int. Ed. Engl. 1997, 36, 1258-1275
Organomolybdenum and -tungsten Reagents
[81] T. Okazoe. K. Takai. K. Utimoto, J Am. Chem Sor. 1987, f09, 951-953.
[X?] Review of geminal bimetallic compounds: F. Bertini. P. Grasselli, G. Zubiani,
G. Cainelli. Tetruhedron 1970, 26, 1281-1290.
[83] L. Plamondon, J. D. Wuest, J Org. Chem. 1991, 56, 2076-2081.
[84] K. Shishido. Y.Tokunagd, N. Omachi, K. Hiroya, K. Fukumoto, T. Kametani,
J Cl7rm. So?. Chem. Commun. 1989. 1093-1094.
[85] M. T. Reetz. B Wenderoth. R. Peter. J Chem. Soc. Chem. Commun. 1983,
[X6]K . T a k a . Y Kataoka, T. Okazoe. K. Utimoto, Terruhedron Let?. 1987. 28,
1443- 1446. J. Y0shidd.T. Maekawa, Y Morita, S . hoe, J. Org. Chem. 1992,57,
1321 -~1322.
[87] K. Takai, K. Nitta, K. Utimoto, J Am. Chem. SOC.1986. 108,7408-7410.
[88] P. Knochel, I. F. Nonnant, Tetrahedron Lett. 1986, 27, 1039-1042.
[89] a) T. Okazoe. K. Takai, K. Oshima, K. Utimoto. J. Org. Chem. 1987, S2,
4410-4412; b) K. Takai, Y. Kataoka, T. Okazoe, K. Utmoto, Terruhedron
Lett. 1988,29, 1065-1068; c) K. Takai, T. Kakiuchi. Y.Kataoka, K. Utimoto,
J Org. Chem. 1994, 59,2668-2670.
[90] a) N. A. Petasis, E. 1. Bzowej, J Am. Chem. Soc 1990, /I.?.6392-6394; b) J
Org. Chem. 1992, 57, 1327-1330; c) N. A. Petasis, I. Akritopoulou, Synfetr
1992, 665-667; d) N. A. Petasis, E. I. Bzowej, Tefruhedron Lett. 1993, 34.
[91] J. J. Elsch, A. Plotrowski, Tetrahedron Lett. 1983, 24, 2043 -~2046
Time is money...
Save both by starting a personal subscription
to Angewandte Chernie
At 9Q per page you can't go wrong.
With your own private copy you can read the
"hot" articles before they cool, and waste
less time in front of the copier.
Order now!
An order form can be found on
the last page of this issue.
Angrw. Chdm lnt E d EnXI.
1997, 36, 1258-1275
Без категории
Размер файла
2 456 Кб
methylene, carbonyl, organomolybdenum, reaction, compounds, selective, carbonylmethylenation, organotungsten, ligand, methyl, transformation, spontaneous, dimerization, additiveцreductive, novem
Пожаловаться на содержимое документа