Group 4 metal complexes of benzynes, cycloalkynes, acyclic alkynes, and alkenes

12 Pages·2001·2.17 MB·English

Chem.Rev.1988,88, 1047-1058 1047
Group 4 Metal Complexes of Benzynes, Cycloalkynes, Acyclic Alkynes, and
Alkenes
STEPHEN L. BUCHWALD" and RALPH B. NIELSEN
Department of Chemistty, Massachusetts Institute of Technology, Cambridge, Massachusetts 02 139
Received April 12, 1988 (Revised Manuscript Received June 3, 1988)
Confenfs SCHEME 1
0
I. Introduction 1047
11. Alkyne Complexes 1047
A. Complexes of Benzynes 1047
B. Complexes of Cyclohexyne and Other 1051
Cycloalkynes
Much of this paper will deal with our work at MIT
C. Complexes of Acyclic Alkynes 1053
where we have developed a relatively general route to
111. Alkene Complexes 1055
complexes of a variety of organic fragments bound to
IV. Conclusion 1057
a single metal system, zirconocene.' Our goals in this
V. Acknowledgment 1057
area have been the understanding of structure-reactivity
VI. References and Notes 1057
relationships and the development of reagents for or-
ganic synthesis. Due to space limitations, the studies
I. Infroducfion
of aldehyde, ketone, ketene, and thioaldehyde com-
Transition metals have the ability to stabilize highly plexes carried out by Erker? Floriani? and Grubbs4a nd
in our laboratorylcYhw ill not be discussed. Also, the
reactive organic fragments and also to activate stable
molecules toward selective attack. A large number of related chemistry of group 5 metals (i.e., the niobium-
imine complexes of Pedersen5 or the tantalum-alkyne
transition-metal complexes of small organic molecules
complexes of Schrock6) will not be covered. The
have been prepared, but syntheses of such species are
frequently limited to a specific type of organic fragment chemistry of v4-diene complexes of group 4 metals
(which can also be formulated as metallacyclopent-3-
bound to a particular metal system. Few general stra-
enes) has been extensively studied and reviewed7 and
tegies lend themselves to the preparation of a wide
will also not be dealt with in this paper.
variety of fragments bound to a specific metal or to a
particular fragment bound to a wide variety of metals.
I I. Aikyne Complexes
As a result, few structure-reactivity relationships have
been mapped in organometallic chemistry. This paper
A. Complexes of Benzynes
will discuss recent research into the preparation,
characterization, and reactivity of two-carbon fragments Numerous examples of the reductive coupling of two
(i.e., alkynes and alkenes) bound to a single group 4 (Ti, unsaturated organic molecules to yield metallacyclic
Zr, Hf) metal center, in the hope that the initial outlines compounds using high-energy group 4 metal species
of such structure-reactivity relationships will be dis- such as zirconocene, Cp,Zr (Cp = 775-C5Hs)h, ave been
cernible. reported.8 This is a useful approach for the coupling
Of the many transition-metal complexes of alkynes, of two identical fragments to form symmetrical me-
alkenes, aldehydes, and ketones which have been pre- tallacycles, but several limitations of this approach are
pared, relatively few involve group 4 metals. Strategies apparent. First, intermolecular cross-coupling reactions
for the preparation of later transition-metal complexes (i.e., involving two different alkynes) are not practical.
of such molecules usually involve the direct combination Second, in essentially all of the cases reported, the
of the unsaturated organic fragment and a coordina- coupling of terminal alkynes is precluded due to the
tively unsaturated metal center. However, low-valent, basicity of the organometallic reagents. Third, it has
coordinatively unsaturated group 4 species are unstable proven difficult to effect cross-coupling between alkyne
and difficult to prepare. When such metal centers bind complexes and heteroatom-containing organic sub-
alkynes, alkenes, or thioaldehydes, the high degree of strates.
x-back-bondingf rom the metal to the organic fragment In contrast, there are few reports of the generation
often makes the resulting complex behave more like a and coupling reactions of discrete alkyne or olefin x-
metallacyclopropene, metallacyclopropane, or metal- complexes. One particularly important example is
lathiirane. These complexes have been studied for a Erker's ground-breaking study on the generation and
number of reasons, including intrinsic interest in their trapping of zirconocene complexes of ben~yne.T~h e
structure and reactivity, the desire to model catalytic benzyne complex is formed without generating free
processes such as Fischer-Tropsch chemistry and hy- zirconocene, and subsequent coupling with an excess
drogenation, and because such species may be of of an olefin produces the zirconaindan products shown
practical use as vehicles for olefin and acetylene po- in Scheme 1.
lymerization and for achieving selective transformations The ability to prepare discrete x-complexes of this
of use in organic synthesis. sort and an understanding of their reactivity can po-
0009-2~~5ia810788-1047~0~.500i 1o9 88 American Chemical Society
1048 Chemical Reviews. 1988. Voi. 88. No. 7 Buchwald and NielSen
PaSCHEME 2 F*m
I
SCHEME 3
Stephen L. Buchwald was born in Biwmington. IN. in 1955. He
became hooked on chemistry while a participant in an NSF High
Schwi Summer Research Program during which time he worked
in the laboratories of Professor Gary M. Hienje at Indiana Univ-
.-el"
ersity. Aner high school he matriculated to Brown University where
he received an Sc.B. Degree in chemistry in 1977. While at Brown \
he did undergraduate research with Professors Kalhlyn A. Parker
and David E. Cane and spent a Summer in the laboratories of
Professor Giibwfl Stork at Columbia University. In 1977 he entered
graduate school at HaNard University as an NSF Qaduate Fellow.
At HaNard he worked with Professor Jeremy R. Knowles per-
fwming research to unravel the mechanisms of phosphoryl transfer
reactions. Atter receiving his PhD.. he went west to the Califwnia
Institute of Technology where he spent Z'/* years as a Myron A.
Bantell Postdoctoral Fellow in the laboratwies of Professor Robert
H. Grubbs. In 1984. he moved back to Cambridge to join the SCHEME 4
faculty at MIT. where he is currently the Firmenich Assistant
Professor of Chemistry. He is a Dreyfus Grantee for Newly Ap-
pointed Facum in Chemistry. an American Cancer Society Junior
Facum Research Awardee. an Eli Lilly Grantee. an Alfred P. Slmn
Research Fellow, and a participant in the 1988 Union Carbide
Innovation Recognition Program. While not engaged in chemistry.
his hobbies are sports. food, and cats.
_I-
SCHEME 5
Ralph B. Nieisen was born in Logan, UT, in 1960. and he grew up
in Baton Rouge, LA. He became interested in chemistry while
working for two Summers during high school in the laboratories of
Professor William H. Daly at Louisiana State University. He then
attended Brigham Young University in Provo. UT. where he received
a Bachelor's degree in 1983 and a Master's degree in 1984. At
BYU he worked on the synthesis of functionalized crown ethers
with Professor Jerald S. Bradshaw. Since 1984 he has been
working toward a doctoral degree in Professor Stephen L. Bu-
chwaid's laboratories at MIT. and ha has been aided financially by
graduate fellowships from the National Science Foundation and
Arthur D. Linie Inc. After graduation from MIT he will begin work
in the Polymer Science and Technology Laboratory at Eastman
Kodak in Rochester. NY, should be possible. Milder reaction conditions may be
developed which tolerate sensitive functionality. Also,
tentially overcome several of the difficulties encoun- in some cases the reductive coupling of unstable organic
tered in reductive coupling reactions using low-valent fragments such as henzynes or thioaldehydes might be
metal species such as Cp,Zr. The selective coupling of possible without having to generate the unstable species
two different groups (such as benzyne with an olefin) in the free state.
Group 4 Metal Complexes Chemical Reviews, 1988, Vol. 88, No. 7 1049
As early as 1963,'O unusual reactivity of group 4 di- SCHEME 6
arylmetallocene derivatives was noted. Heating di-
phenyltitanocene with diphenylacetylene leads to the m":
-
ltoasnsi uomf a mmeotlaelclaucley colfe beshnozewnne ainn dS fcohremmaet io2n. of Itth ew tais- Cp,TiCi2 + I Ph-Ph Mg ChTi
somewhat later that the intermediacy of a titanocene-
benzyne complex was proposed by Vol'pin,'l who
showed that the benzyne complex couples carbon di-
oxide and other unsaturated molecules. Initial kinetic SCHEME 7
0
investigations by Teuben and co-workers12e stimated
a value for AG* of 29 kcal/mol for formation of the
RCN (76-99%)
benzyne intermediate, and they also showed that deu-
terium labels are scrambled between the phenyl ligands,
the Cp ligands, and the solvent.
R = Me, Ph, nPr, i-Pr, vinyl,
The elegant study by Erkergao f the reactivity of di-
phenyl- and ditolylzirconocene derivatives demon-
strated the intermediacy of benzyne complexes and SCHEME 8
suggested that the benzynes form by a concerted elim-
ination of benzene and not by a @-hydridee limina-
tion/reductive elimination sequence. As outlined in
Scheme 3, Erker established that the benzyne complex
and substituted benzyne complexes reversibly activate
and toluidene by heating diaryltitanocenes under a
benzene solvent to re-form diarylzirconocene complexes.
nitrogen atmosphere.21
He also showed that photolysis of ditolylzirconocenes
In at least one case, the titanocene-benzyne complex
gives high yields of dimethylbiphenyl products which
appears to have been prepared by a completely different
reflect the substitution pattern of the diarylzirconocene
route. Vol'pin reports22t hat treatment of a mixture of
precursors. He then showed that heating di-p-tolyl-
titanocene dichloride, o-bromofluorobenzene, and di-
zirconocene in benzene solvent followed by photolysis
phenylacetylene with magnesium metal at low tem-
gives 4,4'-dimethylbiphenyl, 4-methylbiphenyl, 3-
perature leads to formation of modest amounts of the
methylbiphenyl, and, after an induction period, un-
benzyne-derived metallacycle shown in Scheme 6. This
substituted biphenyl. However, no 3,4'-dimethylbi-
apparent simultaneous generation of free titanocene
phenyl is formed. All of the observed products are what
and free benzyne is intriguing, but the rather harsh
should be expected from reaction of a 4-methylbenzyne
reaction conditions and the low yield suggest that this
complex with the solvent. However, since the entire
approach has limited synthetic utility.
thermal process is reversible, a mechanism which in-
Our efforts in this area stemmed from a desire to
volves a @-hydridee limination to form a zirconocene
extend Erker's work with the zirconocene complex of
4-methylbenzyne hydrido p-tolyl intermediate is dis-
benzyne to determine the feasibility of coupling other
counted because it predicts that a reversible hydro-
functional groups, especially aldehydes, ketones, and
zirconation should form p-tolyl-m-tolylzirconocene,
nitriles. Of these, only nitriles can be successfully em-
which is not produced.
ployed as substrates. This reaction produces a high
Erker also showed that the benzyne complex reduc- yield of metallacycles as shown in Scheme ?.le
tively couples olefins to form zirconaindan derivativesagb As will be discussed later, we found that we could
This reaction, too, is reversible, as shown. Heating the prepare and isolate a stable trimethylphosphine adduct
metallacycles with alkenes scrambles to olefin frag-
of a closely related zirconocene-cyclohexyne complex.lb
ments. The coupling of cis- and trans-stilbene was This success led us to immediately probe whether we
shown to be stereo~pecificflo~r the formation of cis- and could utilize similar methodology to prepare and isolate
trans-substituted metallacycles. Subsequent reactions the trimethylphosphine adduct of the zirconocene-
of these zirconaindan species with CO have also been benzyne complex. As shown in Scheme 8, this is effi-
studied (Scheme 4).14 ciently accomplished by heating a benzene solution of
In addition to the reactions of benzyne complexes diphenylzirconocene in the presence of a large excess
with olefins and alkynes,15v arious other reactions of in of trimethy1phosphine.la The thermal stability of 2 is
situ generated benzyne complexes have been discovered attested to by the fact that it survives conditions of its
(Scheme 5). Reaction with elemental selenium can formation, refluxing in benzene for 24 h. With this
yield the diselenametallacycles shown.16 Methylene- reaction, 2 can be isolated in -90% yield as an air- and
triphenylphosphorane reacts with the zirconocene- moisture-sensitive solid. The X-ray crystal structure
benzyne complex to give a new phenylzirconocene- of 2 shown in Figure 1 was determined by John Huff-
substituted ylide as the only product, but the titanocene man at Indiana University. A high level of back-
derivative gives a 2:l mixture of the ylide and the me- bonding from zirconium to the n*-orbital of benzyne
tallacyclobutene sh0wn.l' Tungsten hexacarbonyl re- is evidenced by the fact that all the carbon-carbon
acts to form a metallacyclic Fischer carbene.l* Zirco- bonds in the ring of the benzyne fragment are identical
nocene benzyne complexes with substituted Cp ligands in length, within experimental error. Compared to the
have also been shown to undergo intramolecular C-H two other structurally characterized transition-metal-
bond activati~n.'~Ev~e~n ~th e possible activation of benzyne complexes, this is similar to what is seen in
dinitrogen by benzyne complexes has been suggested Bennett's nickel-benzyne complex23b ut contrasts that
to explain the formation of low yields (<2%) of aniline manifested in Schrock's tantalum complex.6
1050 Chemical Reviews, 1988, Voi. 88, No. 7 Buchwald and Nielsen
SCHEME 11
R . R
+ Cp,Zr(Me)Ci 1 5 I 6
P 70% overall yield,
R R
/
R=OMe 98 2
R-Me 100 0
SCHEME 12
Figure 1. X-ray crystal structure of 2. Bond distances (A):
X X
ZAG, 2.267 (5); Zr-C7,2.228 (5);C W7, 1.364 (8); C7-C8, 1.389
(8); C8-C9, 1.383 (9); C9-Cl0, 1.380 (9); ClO-Cll, 1.377 (9);
Cll-C6,1.406 (8);Z r-P, 2.687 (3). Bond angles (deg): C6-2147,
35.33 (19); Zr-C6-C7, 70.8 (3); Zr-C7-C6, 73.9 (3); C6-C7-C8,
122.1 (5); C7-C6-Cll, 120.2 (5).
-
SCHEME 9 X Y Isolated Yield of 10 (based on aryl bromide)
OMe H 50%
0 Me OMe 69%
II OMs OMe 70%
Me H 65% (estimated by 'H NMR)
Although we had no a priori predictions about the
regiochemical outcome of these coupling reactions, we
found that conversion of a 2-substituted phenyllithium
to the benzyne complex, followed by coupling with a
nitrile and hydrolysis, leads almost exclusively to the
p+ non-Friedel-Crafts isomer of the resulting phenone
p"~cl (Scheme 11). We first used 2-lithio-l-methoxybenzene
SCHEME 10 as the substrate and initially attributed the -501 ratio
X X Me X of the regioisomeric metallacycles 5 and 6 obtained by
I
coupling 4 with acetonitrile to a chelation of a lone pair
on the oxygen to the oxophilic zirconocene fragment.24
However, changing the substituent from methoxy to
CHI methyl has little effect on the product ratio, and we now
Y Y
x believe that the regiochemical preferences are steric in
R X
origin. Below are shown the proposed intermediate
complexes immediately before insertion. Since the
R-R' interaction in 7 is far worse than the analogous
R'-H interaction in 7', it follows that 7' is converted into
Unlike the in situ generated benzyne intermediate 1,
complex 2 undergoes clean, high-yield reactions upon
treatment with a number of reagents as shown in
I
Scheme 9.
H
Encouraged by these initial results, we sought to ex-
pand the utility of this chemistry by preparing sub- the regioisomer which predominantes. Thus, high levels
stituted derivatives of 2. Erker described the inability of regioselectivity can be achieved in a predictable
to prepare the di-o-tolylzirconocene precursor to the manner, based on the relative size of the substituents
3-methylbenzyne complex, presumably due to unfa- at carbons 2 and 6. This regiochemical outcome has
vorable steric interaction^.^^ We sought a method that also been previously observed in the coupling reaction
would allow us to prepare zirconocene complexes of of alkynes with the titanocene-2-methylbenzyne com-
highly substituted benzynes and at the same time ~1ex.l~~
eliminate the need for 2 equiv of the substituted aro- The substituted metallacycles obtained by coupling
matic precursor. We found that the addition of a the benzyne complex with a nitrile can be readily con-
lithiated aromatic species to zirconocene methyl chlo- verted to both simple aromatic ketones and iodo-
ride produces 3, which upon heating to 70 "C in benzene aromatic ketones, upon hydrolysis or iodinationlhy-
smoothly loses methane to yield 4 (Scheme lO).'g drolysis.lg Some of our results obtained from coupling
Complex 4, formed in situ, can be coupled with various acetonitrile with benzyne complexes derived from sev-
substrates to form metallacycles, which in turn can be eral aryl halide precursors are outlined in Chart I.
converted into useful organic products without the need If the benzyne complexes are generated under an
for their isolation or purification. atmosphere of ethylene, high yields of zirconaindans 8
Group 4 Metal Complexes Chemical Reviews, 1988, Vol. 88, No. 7 1051
CHART I SCHEME 14
x x x
PBrecursorr W orkup Fegloimer A Regloisomer B Isolated
Yield
H30' 86%
6''
Uary Rl A2 X Y 12:13 Ratio Isolated Yield (based on aryl bromide)
H3Ot 62
5 1 Et Et H H 1oo:o 91% (yield based Cp2ZrPh2)
0 2 TMS Me H H 1OO:O 92% (yield basedCp2ZrPh2)
Mge0 3 Me Me Me H 1OO:O 72%
74 4 TMS Me Me H 1OO:O 69%
H30* 5 Et Et OMe H 1OO:O 71%
6 Et Et OMe H 1OO:O 80%
oBr Br fl0 7 TMS Me OMe H 95:5 7 1 %
6 8 Me Me Me Me 1OO:O 76%
Me 9 Et El Me Me 1OO:O 67%
47 10 TMS Me Me Me 1OO:O 75%
H30' 11 Me Me OMe OMe 1OO:O 60%
OMe Me0 0 Me0 12 TMS Me OMe OMe 1OO:O 69% TMS3SiMe3
OMe SCHEME 15
61
?Me
Me
67
SCHEME 13
X X X
I
OMe
I
R X Y Isolated Yield of 11 (based on aryl bromide) SCHEME 16
Me H H 83% (yield based on isolated metallacycle)
Me Me H 59%
Me OMe H 25%
nPr OMe OMe 52%
are formed, as Erker found for the unsubstituted case.
Treatment of 8, without isolation, with excess iodine
produces diiodide 9, which upon exposure to n-BuLi at
-78 cleanly yields substituted benzocyclobutenes 10
O C
in good overall yield,4l as shown in Scheme 12.
In addition to providing novel routes for the synthesis
a-position of the metallacyclic intermediate,lsbi t is in-
of carbocyclic and substituted aromatic compounds,
teresting to note entries 2,4,7,10,a nd 12 where almost
benzyne-derived metallacycles can be used as precursors complete regioselectivity is seen, both with respect to
to sulfur-containing heterocycles. An efficient
the substituents of the benzyne and with those of the
"metallacycle transfer" reaction of zirconacycles using
alkyne.
dihalides of main-group metals and chalcogens has been
The X-ray crystal structure of the (trimethyl-
reported by Nugent and Fagan.26 Using this approach,
phosphine)zirconocene-benzyne complex indicates that
we have derived general methods for the synthesis of
there was not an undue amount of ring strain present.
benzisothiazoles and benzothiophenes as shown in
This observation encouraged us to prepare a zirconoc-
Schemes 13 and 14. enebenzdiyne complex, as shown in Scheme 15.l' The
Treatment of the azametallacycles derived from ni-
benzdiyne complex was characterized by X-ray crys-
trile coupling with sulfur monochloride yields benziso-
tallography, and each zirconocene-benzyne fragment
thiazoles 11 (Scheme 13). In all examples studied to
of the molecule is structurally quite similar to the sim-
data, only one isomer has been observed. This repre-
ple benzyne complex. The reactivity of this complex
sents a general entry to these heterocycles that have not
is also similar, as shown by the ethylene coupling/io-
been studied in detail, due to a lack of convenient
dination sequence shown to give a mixture of tetra-
methods for their synthesi~.~'
iodides, which can be converted in high yield to the
The generation and in situ coupling of the nascent
tricyclic compound shown.
benzyne complex with a nonterminal alkyne, followed
by treatment of the zirconaindene intermediate with
B. Complexes of Cyclohexyne and Other
freshly distilled sulfur dichloride, provide good to ex-
Cycloalkynes
cellent yields of the benzothiophenes indicated (Scheme
14).% Although one previous report of the coupling of As an outgrowth of our studies on trapping the zir-
trimethylsilyl-substituted alkynes with a benzyne com- conocene-benzyne complex with nitriles, we queried
plex indicated that the silyl substituent prefers the whether our methodology could be extended to allow
1052 Chemical Reviews, 1988, Vol. 88, No. 7 Buchwald and Nielsen
SCHcEME 18 pp
=
Cp&&o . - B u m
n-Bu= alkyne rotation or
n-Bu
disassociation
SCHEME 19
cp(m
(-100% by NMR)
0
CPzZr
'
OF (-100% by NMR)
Figure 2. X-ray crystal structure of 17. Bond distances (A): I
Zr-C1, 2.165 (13); Zr-C2, 2.244 (12); C1-C2, 1.295 (25); C2-C3,
1.532 (18);C l-C6,1.519 (19); Zr-P,2 .689 (3). Bond angles (deg):
Cl-Zr-C2,34.1 (0.5);C l-C2-C3,126.0 (1.2); CWl-C2,125.2 (1.2).
-
SCHEME 17 n
/ \ (61% toennet adtiivaesltye raesosiisgonmeder ,
CHART I1
Based on Starting
Bromocycloalkene: 57% 62% 52% 60% ('H NMR)
d
(66%)
hexyne unit and indicate that this complex is best de-
(58%)
scribed as a metallacyclopropene. The stability of 17
for the synthesis of zirconocene complexes of cyclo- is dependent on this back-bonding, since it stabilizes
alkynes. In our initial experiments, we attempted to both the highly strained cyclohexyne fragment and the
prepare cyclohexenylmethylzirconocene( 14) (Scheme high-energy Cp2Zr unit.
16). However, 14 could not be isolated at room tem- Complex 17 reacts readily with a wide variety of un-
perature, due to its partial decomposition. Since for- saturated organic compounds, including nitriles, ke-
mation of the zirconocenebenzyne complexes requires tones, aldehydes, ethylene, both internal and terminal
temperatures of 60-80 "C to proceed at a reasonable alkynes, and dienes, as shown in Scheme 17. Of par-
rate, it is surprising that the transformation of 14 to 15 ticular interest is the coupling reaction between 17 and
takes place at 25 "C. This was confirmed by allowing l-hexyne. At room temperature a -3:2 mixture of
a solution of 14, prepared at low temperature, to react regioisomers 18 and 18' is formed. Heating the reaction
with butyronitrile at 25 "C to form a quantitative yield mixture for 20 h at 80 "C causes isomerization to only
(lH NMR) of the azametallacycle 16. This transfor- one isomer, tentatively assigned as 18.29 As depicted
mation is also somewhat surprising in that it shows in Scheme 18, isomerization of the metallacycle must
complete selectivity for activation of the sp20 -hydrogen involve carbon-carbon bond cleavage of metallacycle
over the sp3@ -hydrogenso f the cyclohexenyl ligand. We 18'. Since this reaction is relatively rapid at 80 "C, it
were pleased to find that addition of excess tri- implies that the cyclohexyne intermediate is relatively
methylphosphine to 14 at low temperature, followed by strain-free. Related isomerization reactions of metal-
warming to room temperature and recrystallization lacyclopentanes (or pentenes) has been reported in a
from ether, allowed the isolation of the compound 17, number of ~ a s e ~bu,t 9litt~le ~is ~kno~wn~ a~bou~t t he
a stable trimethylphosphine adduct of the cyclohexyne mechanism of this low-energy carbon-carbon bond
complex. This "one-pot" procedure can readily be used cleavage reaction.
to prepare 6-8 gs of analytically pure 17 in -60% Using a similar synthetic strategy, we have been able
isolated yield, based on bromocyclohexene.lb to prepare the additional cycloalkynes in Chart 11, and
The X-ray crystal structure of 17 (Figure 2), deter- we are currently investigating their reaction chemistry.32
mined by John Dewan, has several interesting features. Metallacycles formed by coupling ethylene or 1,3-
First, the Cl-C2 bond length of 1.295 (25) A indicates dienes with 15 undergo chemoselective reactions to in-
a bond order substantially closer to 2 than to 3. sert aldehydes, ketones, and nitriles, specifically into
Moreover, the Cl-C2-C3 bond angle of 126.0 (1.2)", the sp3 carbon-zirconium bond as shown in Scheme
again, is much closer to the 120" value expected for a 19.29 Free cyclohexyne complex 15, generated in an in
sp2-hybridized carbon than to the 180" value expected situ manner in the presence of ethylene or a 1,3-diene,
for a sp-hybridized carbon. Taken together, these is converted into a metallacycle which can be further
values reflect the high degree of back-bonding from the treated with an aldehyde followed by hydrolysis or an
Cp2Zr(PMe3) moiety to the r*-orbital of the cyclo- iodinolysis-hydrolysis workup to produce organic
Group 4 Metal Complexes Chemical Reviews, 1988, Vol. 88, No. 7 1053
SCHEME 20
0
Figure 3. X-ray crystal structure of 22. Bond distances (A):
-1
(40%) Zr-P, 2.6576 (11);Z &6, 2.244 (3); ZA7, 2.211 (3); CW7, 1.286
(5); C7-C8, 1.491 (5). Bond angles (deg): C647-C8, 135.8 (3);
Zr-C6-C7, 71.81 (19); Zr-C7-C6, 74.64 (20).
SCHEME 21
R
SCHEME 23
SCHEME 22
Ph
Ph \ Ph
(80%)
the CO with PMe3 has been reported to enhance the
stability of the alkyne complex.36 The complexes
products (Scheme 20). Overall yields are remarkably
high for a sequence of six or seven chemical steps which Cp*,Ti(2-butyne), Cp*,Ti(diphenylacetylene), and
occur in a single flask. Cp*,Zr(diphenylacetylene) (Cp* = q5-C5Me5h) ave been
briefly mentioned30 as stable, isolable compounds. In
the same report, the complex Cp*,Ti( l-propyne) was
C. Complexes of Acyclic Alkynes
implicated as in intermediate in the thermal decom-
For some time it has been known that free metal- position of bis( l-propeny1)permethyltitanocene. Ne-
locene or its equivalent (i.e., Cp,M, M = Ti, Zr, Hf) can gishi and co-workers recently reported4, the isolation
reductively couple two alkynes to give symmetrically and X-ray structure of Cp2Zr(PMe3()d iphenyl-
tetrasubstituted metallacyclopentadiene products, and acetylene), which was obtained by treating Cp2Zr-
it has been generally that a metallocene- (PMe,), with the alkyne. The compound is structurally
alkyne complex is an intermediate in this process quite similar to Floriani's titanocene complex. Unfor-
(Scheme 21). A remarkably large number of examples tunately, this is the only alkyne complex that can be
of this reaction type using diphenylacetylene have been prepared by using their methodology.
reported, yielding the very stable tetraphenylmetalla- Rausch and co-workers found that reaction of
cyclopentadienes. Methods for generating "Cp,M" in- Cp2Ti(PMe3),w ith acetylene, propyne, phenylacetylene,
clude the following: TiC14 and 4 equiv of sodium cy- or diphenylacetylene leads to formation of metalla-
~lopentadienidep,~ho~t~oc hemically induced reductive cyclopentadienes, and the intermediate Cp2Ti-
elimination of Cp2MR2,2d*t3h4e rmal or photochemical (PMe,) (alkyne) complexes can be observed spectro-
reactions of Cp2M(C0)2135i3r6e duction of Cp2MC1, with ~copically.T~h~is is the first report of such reductive-
sodium na~hthalide,o~r magnesium/mercuric chlo- coupling chemistry using terminal alkynes.
and ligand displacement from Cp,M(diene) We have found that the methods we used to prepare
complexes,2dw hich are generated from Cp2MC12a nd the zirconocene-cyclohexyne complex can be readily
(butadiene)magnesium. This reaction has even been extended to the preparation of zirconocene complexes
documented for a polymer-bound titanocene fragment.% of both mono- and disubstituted acyclic alkynes as
There are a few examples where similar chemistry has shown.ld Using the hydrozirconation methodology for
been extended to other symmetric internal alkynes such alkynes developed by Sch~artz,f4o~ll owed by the ad-
as 2-butyne7 3-hexyne7 or dimethyl acetylenedi- dition of methyllithium, produces vinylmethyl-
carb~xylate.A~l~ky~n~es have also been reported to zirconocene 19. Methane loss occurs smoothly at room
induce reductive elimination of RH from Cp,M(H)R, temperature to produce alkyne complex 20,45 which, in
to give metallacyclopentadiene' products.39 However, the presence of trimethylphosphine, is transformed into
discrete alkyne complexes have been observed or iso- complex 21 (Scheme 23). In this manner the tri-
lated in relatively few cases, Floriani reported the methylphosphine adducts of the zirconocene complex
preparation40 and X-ray crystal structure41 of Cp,Ti- of l-hexyne 22 and of 3-hexyne 23 can be isolated in
(CO)( diphenylacetylene) , which is formed from -70% yield. This result suggests that the previously
Cp,Ti(CO), and the alkyne (Scheme 22). The complex reported alkyne-induced "reductive elimination" of
is thermally unstable and disproportionates to form Cp2Zr(H)MeS is likely to proceed via hydrozirconation
Cp,Ti(CO), and the metallacyclopentadiene. Replacing of the alkyne to form the same vinylmethylzirconocene
1054 Chemical Reviews, 1988, Vol. 88, No. 7 Buchwald and Nielsen
SCHEME 25
/A”- X X
, _._. .
X = StMe3, SnMe,, alkyl, aryl, alkenyl 1
Y= (CH,),, alkyltdene. RN, etc ‘2
k
OXz = MezSnCIz.P hAsClz,P hPCIz.S ,CIz, etc.
Q = Mez% PhAs, PhP, S, etc. SCHEME 26
OSiMed-Bu
intermediate which we observe.
An X-ray crystal structure of the 1-hexyne complex
22, determined by John Huffman, at Indiana Univer-
sity, is shown in Figure 3. The C1-C2 bond distance
is 1.286 (5) A, while the Cl-C2-C3 bond angle of 135.8
(3)’. This should be compared with the corresponding
angles of 126.0 (1.2)’ and 122.1 (5)’ in the cyclohexyne
17 and the benzyne complex 2, respectively. The C-C
“triple bond” distance in 22 is identical, within exper-
imental error, with that found for the cyclohexyne
complex. This suggests that the 135.8 (3)’ value more
accurately reflects the bond order in these complexes
and that the lower value observed for the cyclohexyne
complex is due to some ring strain which prohibits the I 9 2 : 8 ‘
bond angles from splaying further outward. That the
1-hexyne complex 22 is essentially unstrained and still such as arsenic and tin to yield zirconocene dihalide and
displays a zirconacyclopropene structure indicates that the corresponding thiophene, phosphole, or main-group
back-bonding significantly stabilizes the Cp2Zr(PMe3) metallacycle.26
unit, in which the zirconium prefers to be in the +4 Negishi’s related cyclization of enynes initially47a
rather than in the +2 oxidation state. employed Farona’s conditionsaa but later milder con-
Although the structures of several symmetric tetra- ditions were discovered by using a reagent prepared
substituted metallacyclopentadienes have been sub- from zirconocene dichloride and 2 equiv of n-butyl-
stantiated by isolating 1,3-dienesap33bya4n2d 1,4-dihalo- lithium (Scheme 25).47H Cyclization of enynes gives
1,3-dieneo~b~ta~in ed after hydrolysis of halogenation, good yields of the metallacyclopentenes, which can be
the synthetic utility of “Cp,M”-induced reductive cou- hydrolyzed to give exocyclic olefins, can be halogenated
plings of alkynes has been very limited until recently. to give diiodides which are useful precursors to bicyclic
Three problems in particular have made this an unat- cyclobutenes, or can be carbonylated to give cyclo-
tractive reaction for synthetic applications. First, both pentenones. Two apparent shortcomings of this pro-
“Cp2M”a nd the reagents used in its generation (Mg, cedure are that neither terminal enynes nor enynes with
Na, etc.) react with functional groups such as terminal di- or polysubstituted double bonds give cyclized
alkynes, oxygen-containing groups, or alkyl halides. products.47c
Second, the attempted cross-coupling of two different Very recent work by Nugent, Rajanbabu, and Ta-
alkynes (or an alkyne and another unsaturated group) bera using the Cp2ZrC12/n-BuLir egent has extended
usually leads to a mixture of cross- and self-coupled this reaction to some very highly functionalized enynes,
products. Third, even if cross-coupling could be effi- as shown in Scheme 26. The major stereoisomer from
ciently controlled, ambiguities still exist in the regio- the first example was converted in several steps to the
chemical positioning of the metallacyclic product. cyclopentanone shown, which is an intermediate in a
Many of these problems have been overcome to some previous synthesis of the fire-ant queen pheromone
degree in the recent work of Nugent and Fagan26s4a6n d in~ictolide.T~h~e second example shows the successful
Negi~hi:~w ho have developed relatively mild conditions cyclization of an enantiomercially pure, highly oxygen-
for the intramolecular coupling of diynes and enynes, ated enyne obtained in several steps from a sugar.
and from our development of routes to discrete alkyne The Cp2ZrC12/n-BuLir eagent is very useful, and it
complexes for intermolecular cross-coupling,lda nd our is worthwhile to briefly examine how it induces bi-
study of the reversibility of reductive coupling.4a cyclization. Negishi showed that Cp2Zr(n-Bu), is
Nugent’s initial work on the coupling of diynes and formed rapidly at -78 ‘C and that this dialkyl un-
enynes is shown in Scheme 24.46a* The intramolecular dergoes first-order decomposition at 20 OC at a rate
coupling of several diynes and one enyne (typically with independent of the enyne c~ncentrationW.~h~e~n the
a reagent prepared from sodium amalgam, titanocene bicyclization reaction is monitored by NMR, signals for
dichloride, and methyldiphenylphosphine) followed by the metallacyclic products appear as signals for
hydrolysis gave reasonable yields of carbocycles with cp,Zr(n-B~d)i~sa ppear. Since the byproducts of the
only the E-exocyclic olefins as products. Alternatively, decomposition of dibutylzirconocene are n-butene and
the intermediate metallacycles can be treated with 1-butene, Negishi proposes that the cp,Zr(n-B~m)~u st
dihalides of sulfur, phosphorus, or main-group metals undergo @-hydridee limination with loss of butene as
Group 4 Metal Complexes Chemical Reviews, 1988, Vol. 88, No. 7 1055
SCHEME 27
Mechanism as propose3 by Negishi:
SiMe,
(-95%,by NMR)
only isomer
Anernative mechanism: assoclation
SCHEME 30
SCHEME 28
R4
R2 R4 Reagent Yield (based on the
H- nC2HS- S,CI, 69% vinyl(ch1oro)-
H- CH,- S2CI, 65% zirconocene
H- SIMe3-(H-) SCI, 80% precursor)
H- Ph- SCI, 54% "
n-C,H,- mC,Hs- S,CI, 90% "
RC~HS- CH,- SzCI, 63% "
n-C,H,- SiMe3-(H-) SzCI, 75% "
n-C,H,- P h- S&I, 70% "
H- nC2Hs S,CI, 60% (based on Cp,ZrHCI)
H- CHy S~CIZ 53% "
heating this mixture to 80 "C for 24 h in benzene leads
L
to complete conversion to the regioisomer in which the
shown to give Cp,Zr(H)(n-Bu), which rapidly reduc- silyl substituent occupies the a-position of the metal-
tively eliminates butane to generate free Cp,Zr, which lacycle.% The analogous coupling reaction of the in situ
he believes is the active metal species (Scheme 27). generated 3-hexyne complex leads to a 13:l mixture of
However, some evidence suggests that a zirconocene- the two metallacycles as shown in the scheme. Heating
diyne or -enyne complex can be formed without the this mixture to 80 "C leads to a slight diminution of this
need to generate free Cp,Zr. First, simple alkyl hy- ratio.48
drides of zirconocene such as Cp2Zr(H)(Me) do not With this measure of regiocontrol in the coupling
always rapidly reductively eliminate alkane, even in reactions described above, we sought to devise metho-
strongly coordinating solvents such as THF.60 Second, dology to convert these metallacyclic intermediates into
we have reportedld the observation and isolation of useful organic products. Our ultimate goal here is to
Cp,Zr(l-butene)(PMe3) as the major product formed develop one-pot procedures for the conversion of simple
by warming Cp,Zr(n-Bu), in the presence of PMe3, in- precursors into the final organic products without
dicating that butane is eliminated from Cp,Zr(n-Bu), needing to manipulate or isolate organometallic inter-
before butene is lost. Therefore, a more plausible mediates.
mechanism involves the initial loss of butane to give a One example of this approach is the very general
zirconocene-butene complex which undergoes associa- synthesis of thiophenes outlined in Scheme 30.
tive ligand exchange with the enyne. Treatment of the intermediate zirconacyclopentadienes,
The isolated 1-hexyne and 3-hexyne complexes that obtained by the selective intermolecular cross-coupling
we have prepared (and in many instances, the com- of two alkynes, with SC12 (freshly distilled) or S2C12, in
plexes prepared in situ without the need for PMe3) react tetrahydrofuran produces good to excellent yields of the
with unsaturated organic functional groups such as polysubstituted thiophenes.26i28T he overall transfor-
nitriles, alkynes, aldehydes, ketones, and ethylene to mation is to selectively combine two different alkynes
give metallacyclic compounds in high yield, as shown and a sulfur atom to form the thiophene, with good
in Scheme 28.Id control over the regiochemical placement of the four
It would be synthetically interesting to be able to substituents. As can be seen from the examples, shown,
selectively cross-couple two different unsymmetrical di-, tri-, and tetrasubstituted thiophenes can be pre-
alkynes with complete regiocontrol. However, one lim- pared by usieg this methodology. The disubstituted
itation of this chemistry is that neither the 1-hexyne thiophenes are readily available via protodesilylation
or the 3-hexyne complex couples cleanly with a second of initially formed trimethylsilyl-substituted thioph-
terminal alkyne, which would give the largest steric e n e ~T.h~is~ sy nthesis can be carried out conviently as
difference between the ends of the alkyne. Therefore, a one-pot procedure starting with Cp,ZrHCl.
we decided to investigate the coupling of the 1-hexyne
complex with 1-(trimethylsily1)propyne. Unfortunately, III. Alkene Complexes
when carried out at room temperature, the reaction is
not selective, leading to a 1:l mixture of regioisomers Compared to the number of alkyne complexes that
as shown in Scheme 29. However, we found that have been reported, very few simple mononuclear al-
1056 Chemical Reviews, 1988, Vol. 88, No. 7 BuchwaM and Nielsen
SCHEME 31 SCHEME 33
Ph Ph
SCHEME 34
Ph
(50%)
qR
-
RCECR
R = Me, Et, Ph, SiMe3
CPzZr
6545% yield
0
in Scheme 32, the complex reacts with a variety of
substrates, with loss of bound ethylene in some cases
kene complexes of group 4 metals have been prepared. and reductive coupling or C-H bond activation in oth-
One binuclear olefin complex has been reported by ers.
ErkeF'2 in which a styryl fragment and a chloride ligand We were interested in determining whether the same
bridge two zirconocene centers. The styryl ligand is strategies we employed to generate zirconocene-alkyne
a-bound to one metal and v2,a-bound to the other. A complexes could be extended to alkene complexes. In
related binuclear zirconocene complex with an q2-arene our initial efforts, we attempted to prepare the zirco-
ligand has been prepared by a tedious procedure and nocene complex of cyclohexene and to couple it with
in low yield by Pez and Stu~kyT.~hi~s complex con- a nitrile, as shown in Scheme 33. Instead of the me-
tains bridging hydride and naphthyl ligands, and the tallacyclic product, however, the N-metalated imine was
X-ray crystal structure shows that the arene fragment produced in good yield.56 Presumably, this is formed
is a-bound to one metal center. from the exchange of the bound cyclohexene with nitrile
Three examples of mononuclear alkene complexes are and subsequent insertion of the nitrile into the Zr-H
shown in Scheme 31. Compared to the corresponding bond. Hydrozirconation of nitriles is well-precedent-
alkyne complexes, these alkene complexes are more ed,57 and it is interesting to note that ligand exchange
likely to participate in ligand exchange reactions which and insertion are faster than reductive elimination of
lead to loss of free alkene, and they participate in fewer methane from the cis alkyl hydride.
reductive coupling reactions. The zirconocenel-butene With this result in mind, we reasoned that a bound
complex which we have preparedld has already been cyclopropene ligand would be less likely to be displaced,
mentioned in the acyclic alkyne section. Negishi has as it possesses -54 kcal/mol ring strain.% The ready
reported the preparation of the zirconocene-stilbene availability of cyclopropyllithium6ga llowed the gener-
complex Treatment of this complex with ation of cyclopropene complex 22 (Scheme 34). Nitriles
aqueous acid leads to formation of a mixture of 1,2- and internal alkynes are coupled by 22 to give the
diphenylethane and stilbene. Erker has shown that unique bicyclo[3.1.0]metallacycles,29 but terminal alk-
metallacyclopentanes, such as the hafnocene example ynes give mixtures of regioisomeric metallacycles, and
above, decompose to form an intermediate ethylene carbonyl compounds fail to react cleanly. If 22 is gen-
c ~ m p l e x .T~he~se~ c~om~p lexes undergo both ligand erated in the presence of excess trimethylphosphine,
exchange and reductive coupling reactions, depending complex 23 can be isolated. Despite a good deal of
on the particular reaction conditions. experimentation, no reliable route to 23 of acceptable
The most extensively studied group 4 metal olefin purity (>95%) has been found, although occasional
complex is the permethyltitanocene-ethylene complex batches are produced in a pure state, and we have been
reported by Bercaw.30s55T his complex is prepared from able to successfully obtain an elemental analysis for the
C!~*~Ticbly~ r eduction with sodium amalgam under an compound. Compound 23 couples cleanly with ketones
atmosphere of ethylene. The X-ray crystal structure (but not other carbonyl compounds) to form metal-
of this complex indicates that there is roughly equal lacyclic products as shown. The bicyclic metallacycles
contribution of both the Ti(I1) a-ethylene and the Ti- obtained from 22 or 23 can also be converted into a
(IV) metallacyclopropane resonance foxms. As shown number of cis-1,2-disubstituted cyclopropanes via hy-