Novel alkylidenating agents of iron(III) derivatives by base-mediated α,μ-dehydrohalogenation and their chemical trapping by cycloaddition

Article abstract of DOI:10.1002/ejoc.201000281

Studies of the reactions between group 4 metal, chlorides (M = Ti, Zr, Hf) and methyllithium at -78 °C in toluene can lead to methylidene-metal complexes, H₂C=MCl₂, by a sequence of monomethylation, α-carbon lithiation and α,μ-elimination of LiCl. Here study of the preparation of alkylidene derivatives of iron was attempted by the interaction of FeCl₃ with n-butyllithium in various ratios at -78 °C. The presence of any resulting butylidene-iron(III) derivative, nPrCH=FeE (E = Cl, nBu), was probed by adding chemical trapping agents, such as diphenylacetylene, benzonitrile, methyl benzoate and benzophenone. In each experiment the hydrolyzed products were consistent with a cycloaddition reaction of nPrCH=FeE with the trapping agent. The products from, di-phenylacetylene and from, benzonitrile with D₂O workup are uniquely in accord with such a carbene precursor. A 3:1 ratio of nBuLi/FeCl₃ gave the optimal yield of nPrCH=FenBu, ca. 80%, from, the MBu₂FeCl precursor. When a 3:1. reaction mixture was simply brought to 25 °C and hydrolyzed, the purple alkylidene-iron complex decomposed completely to iron metal. A study of a 3:1 interaction of PhCH₂MgCl and FeCl₃ under similar conditions and trapping with diphenylacetylene provided evidence for the formation of PhCH=FeCH₂Ph in ca. 40%. These results support; the hope that alkylidene-iron(III) analogs of the Grubbs reagents may be accessible by this process.

Full text of DOI:10.1002/ejoc.201000281

FULL PAPER
DOI: 10.1002/ejoc.201000281
Novel Alkylidenating Agents of Iron(III) Derivatives by Base-Mediated
α,µ-Dehydrohalogenation and Their Chemical Trapping by Cycloaddition[‡]
John J. Eisch,*[a] Jane U. Sohn,[a] and Edon J. Rabinowitz[a]
Dedicated to Professor Dr. Uwe Rosenthal on the occasion of his 60th birthday
Keywords: Iron(III)–alkylidene complexes / Alkyliron(III) derivatives / α,µ-Dehydrohalogenation / Lithiation / Cyclo-
addition
Studies of the reactions between group 4 metal chlorides (M
= Ti, Zr, Hf) and methyllithium at –78 °C in toluene can lead
to methylidene–metal complexes, H2C=MCl2, by a sequence
of monomethylation, α-carbon lithiation and α,µ-elimination
of LiCl. Here study of the preparation of alkylidene deriva-
tives of iron was attempted by the interaction of FeCl3 with
n-butyllithium in various ratios at –78 °C. The presence of
any resulting butylidene–iron(III) derivative, nPrCH=FeE (E
= Cl, nBu), was probed by adding chemical trapping agents,
such as diphenylacetylene, benzonitrile, methyl benzoate
and benzophenone. In each experiment the hydrolyzed prod-
phenylacetylene and from benzonitrile with D2O workup are
uniquely in accord with such a carbene precursor. A 3:1 ratio
of nBuLi/FeCl3 gave the optimal yield of nPrCH=FenBu, ca.
80%, from the nBu2FeCl precursor. When a 3:1 reaction mix-
ture was simply brought to 25 °C and hydrolyzed, the purple
alkylidene–iron complex decomposed completely to iron
metal. A study of a 3:1 interaction of PhCH2MgCl and FeCl3
under similar conditions and trapping with diphenylacety-
lene provided evidence for the formation of PhCH=Fe–
CH2Ph in ca. 40%. These results support the hope that alk-
ylidene–iron(III) analogs of the Grubbs reagents may be ac-
cessible by this process.
ucts were consistent with
a cycloaddition reaction of
nPrCH=FeE with the trapping agent. The products from di-
in 2005, along with an equal share to Chauvin for his in-
sight into the metallacyclobutane intermediate involved in
metathesis.
Introduction
The formal complexation of a Lewis-acidic carbene 1
with a transition metal compound 2, of variable oxidation
state and having unshared d-electrons, can lead to transition
metal–carbenes 3 [Equation (1)].[2] The broad utility of such
metal–carbenes in organic synthesis, in industrial hydro-
carbon transformations and in specialized polymerizations
has spurred the preparation of diverse metal–carbenes from
almost all transition metals. Their individual applications
in synthesis range from carbonyl olefination with titanium
reagents of the Tebbe type, Cp2Ti=CR2·MmXn,[3,4] through
cyclopropanation with cationic iron complexes of the
Fischer[5] type, [Cp2(CO)2Fe=CR2]+, to olefin metathesis or
ROMP polymerization Grubbs catalysts[6] of ruthenium,
[Cl2(Ph3P)2Ru=CHR], and of the Schrock[7] molybdenum
type, [(RO)(ArN)Mo=CHR]. The scientific impact of such
structurally defined carbene complexes was recognized by
the joint award of the Nobel prize to Grubbs and Schrock
The great versatility of Grubbs catalysts has led us to
explore whether similar metathesis catalysts based on inex-
pensive iron reagents could be developed. Therefore, we
have examined the interaction of iron(III) chloride with
organolithium and organomagnesium reagents in THF at
low temperatures (Scheme 1). Many previous attempts to
isolate or to detect alkyliron(III) intermediates at or near
room temperature, in the absence of special stabilizing li-
gands, have failed.[8] Some success in generating alkyl-
iron(II) species in situ, such as MeFeCl, Me2Fe and nBu4-
FeLi2, has been achieved, and these reagents have proved
valuable in effecting cross-coupling with 1-alkenyl halides.[9]
In such a proposed reaction we expected that the dialkyl-
ated intermediate 4 would be formed readily. But in further
attempted alkylation to form (RCH2)3Fe, steric hindrance
might be expected to favor the carbene 5, by a process we
have found to occur between the early transition metal com-
pounds, TiCl4 and ZrCl4, with CH3Li, namely, α,µ-de-
hydrohalogenation, leading to H2C=MCl2.[10] Therefore, in
the following study we carried out the reactions in Scheme 1
at –78 °C and thereafter proposed to trap chemically iron–
carbenes 5 with reagents such as alkynes, nitriles and
[‡] Organic Chemistry of Subvalent Transition Metal Complexes,
45. Part 44: Ref.[1]
[a] Department of Chemistry, State University of New York at
Binghamton,
P. O. Box 6000, Binghamton, NY 13902-6000, USA
Fax: +1-607-777-4865
E-mail: jjeisch@binghamton.edu
Eur. J. Org. Chem. 2010, 2971–2977
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2971

J. J. Eisch, J. U. Sohn, E. J. Rabinowitz
FULL PAPER
resulted. When such a mixture was gradually brought to
25 °C, a black suspension was formed and coated the mag-
netic stirbar. Since this black solid readily dissolved in
aqueous HCl with evolution of H2, it can be assumed that
it was iron powder.[11c] If any tri-n-butyliron (6) or n-butyl-
iron(III)–propylidene (nBu–Fe=CH–nPr, 7) were formed at
–78 °C, it is evident that they decomposed to radicals and
iron.
(2) Reaction with Diphenylacetylene (8)
Scheme 1.
To the preformed mixture of 1:3 ratio of FeCl3/nBuLi at
–78 °C was added 0.93 mol-equiv. of diphenylacetylene in
hexanes and the reaction mixture treated further and hy-
drolyzed as in section 1. The hydrolyzed products consisted
of (Z)-stilbene (9, 8%), (E,E)-1,2,3,4-tetraphenyl-1,3-buta-
diene (10, 10%), (Z)-1,2-diphenyl-1-hexene (11, 18%) and
(E)-1,2-diphenyl-1-hexene (12, 64%).
Therefore, the formation of isomeric 11 and 12 as the
major hydrolysis products is uniquely in accord with the
presence of iron–carbene 7 in the reaction mixture (Cat-
egory 1). Capture of 7 by cycloaddition with 8 would yield
13. Electrocyclic ring-opening of 13 would produce an
equilibrated conformational mixture of 14 and 15, whose
subsequent hydrolysis would lead to the observed mixture
of 11 and 12. Steric factors would favor the (E)-iron–car-
bene 15 and its hydrolyzed counterpart 12 (Scheme 2). (In
Scheme 2 and successive schemes substituents on chiral
carbon atoms will be shown with solid lines, which indicate
the stereochemistry is undefined.)
carbonyl derivatives by cycloaddition, a reaction of transi-
tion metal carbenes well-precedented with those of
group 4.[10,11a,11b]
Results
General Overview and Empirical Categorizations of the
Results
From the interaction of 1:3 molar ratios of FeCl3/n-but-
yllithium at –78 °C the anticipated products were tri-n-but-
yliron(III) (6), nBu3Fe, with its β-elimination of 1-butene
derivative, nBu2Fe–H or the reduction product nBu2Fe, and
its possible carbene, n-butyliron(III)–butylidene (7), nBu–
Fe=CH–nPr. In summary, three categories of results with
the trapping agents were observed. Category 1 consists of
the reactions of diphenylacetylene (8) or benzonitrile (18)
with the 1:3 FeCl3/nBuLi reaction mixture, with or without
workup with D2O or aqueous DCl: such reactions were
uniquely interpretable in terms of cycloaddition trappings
of 7, as depicted in Schemes 2 and 4. The ratio of 6/7 gener-
ated in the reaction mixture was estimated at 20:80. Cat-
egory 2 consists of the reactions of methyl benzoate (22) or
benzophenone (24), with or without attempted deuter-
iolysis. These results were consistent with cycloaddition
trappings of 7 but lack the detection of deuteriated interme-
diates in high yield expected from Schemes 5 and 6. Cat-
egory 3 are the reactions of benzyl chloride or (E)-stilbene
oxide with the 1:3 FeCl3/nBuLi mixture, which proceeded
well to give 97% of bibenzyl and 3% of toluene or 85% of
(E)-stilbene, respectively. But the nature of the products did
not permit any distinction between the presence of 6 or 7.
Therefore, such reactions will be deferred to future study.
Consideration of certain stereochemical and deuteriolytic
aspects connected with Categories 1 and 2 will be treated in
the Discussion.
Scheme 2.
Reaction of Iron(III) Chloride with n-Butyllithium
1:3 Molar Ratio of FeCl3/nBuLi
The minor products 9 and 10 can be readily explained
by the formation of an iron–hydride linkage by β-elimi-
nation of 1-butene from any 6 (nBu3Fe) present or from 7
itself. Hydroferration of 8 and insertion of 8 into adduct 16
(1) Simple Hydrolytic Workup
When the reaction was conducted in THF at –78 °C with seem to be the minor pathway leading to the formation of
a 1:3 molar ratio of FeCl3/nBuLi, a purple reaction mixture 17 (Scheme 3). (Cf. further consideration in the Discussion.)
2972
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 2971–2977

Novel Alkylidenating Agents of Iron(III) Derivatives
Scheme 3.
Scheme 5.
(3) Reaction with Benzonitrile (18)
In the attempt to further support the role of 7 in the
major path, an identical reaction run was worked up with
Similar to the procedure in section 2, the FeCl3/nBuLi
mixture (1:3) was treated with 0.93 mol-equiv. of benzoni-
trile (18) and worked up by hydrolysis. The sole product
was valerophenone (19) in 98% yield.
1  DCl in D2O with the hope of detecting a deuteriated
1
derivative of 23b. However, by H NMR spectroscopy 23
had little (ca. 10%) deuteriation at the α-CH2 of the butyl
group. Hence, this reaction is placed in Category 2. How-
ever, the tendency of neutral, high-valent σ-bonded alkyl
transition metal species to undergo homolysis above 0 °C
may mean that 23a might not survive at the deuteriolysis
temperature (cf. Discussion). Note the complete homolytic
cleavage of C–Fe bonds in the purple reaction mixture when
brought above 0 °C for hydrolysis (section 1).
When an identical reaction run was worked up with 1 
DCl in D2O (98%), the resulting valerophenone was found
to be 75% monodeuteriated at the α-carbon center (20).
Therefore, this observation is uniquely in accord with the
principal reaction path being via iron–carbene 7 (Scheme 4)
(Category 1).
(5) Reaction with Benzophenone (24)
As in the foregoing procedures, the FeCl3/nBuLi mixture
(1:3) was treated at –78 °C with 0.93 mol-equiv. of benzo-
phenone (24). Hydrolysis of the reaction mixture gave as
products 1,1-diphenyl-1-pentanol (26, 78%) and benzhydrol
(27, 22%).
Again, the main product 26 can be attributed to a cyclo-
addition of 7 to 24 to yield 25. Hydrolysis would then pro-
duce 26 (Scheme 6). An attempt to detect 25 by deuter-
iolytic workup, as in section 4, was also inconclusive. The
1
α-CH2 of the butyl group in 26 showed by H NMR analy-
sis Ͻ20% of CHD. Thus, this experiment falls into Cat-
egory 2, and the similar failure of deuteriation deserves fur-
ther consideration (cf. Discussion).
Scheme 4.
The undeuteriated 19 could be attributed to minor, direct
butylation of 18 via 21.
(4) Reaction with Methyl Benzoate (22)
As in sections 2 and 3, the FeCl3/nBuLi mixture (1:3) was
treated with 0.93 mol-equiv. of methyl benzoate (22).
Workup gave a quantitative yield of 5-phenyl-5-nonanol
(23b). Since the foregoing reactions with 8 and 18 indicated
that over 80% of the reaction products were consistent with
presence of iron–carbene 7, the following path must be the
principal route to 23 (Scheme 5). Because of the involve-
Scheme 6.
The minor product could arise from a pinacol coupling
ment of butyl–iron bonds in a minor reaction path for sec- induced by the small amounts of alkyliron species 6, itself
tions 2 and 3, it can be concluded that 15–20% of the 23b a free radical (Scheme 7). The bracketed structures (*) are
formed resulted directly from a butyliron–alkyl species, suggested, without evidence, as transitory intermediates in
such as 6.
such C–C bond coupling.
Eur. J. Org. Chem. 2010, 2971–2977
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2973

J. J. Eisch, J. U. Sohn, E. J. Rabinowitz
FULL PAPER
teraction the nickel–carbene 35, analogous to 30, is thought
to form.[13] Here again, protolysis of 35 yields 31, 32 and
33 (Scheme 9).
Scheme 9.
Scheme 7.
Discussion
When this reaction was attempted by adding the benzo-
phenone at 25 °C, 98% of 24 was recovered, and only a
trace of tetraphenylethylene was detected.
From the foregoing reaction of iron(III) chloride in THF
with n-butyllithium at –78 °C in a 1:3 molar ratio, all the
results are consistent with formation and reactions of two
organoiron intermediates in about an 80:20 ratio. The
major product has the chemical properties consistent with
an iron–carbene of the structure nBu–Fe=CH–nPr (7),[14]
namely the cycloaddition reaction,[11] whereas the minor
reagent undergoes different reactions with an array of sub-
strates consistent with an n-butyliron(III) intermediate,
most likely of the structure (nBu)3Fe (6).[14] One discordant
observation in interpreting this body of results in terms of
cycloaddition reactions of iron–carbene 7 is the failure to
detect organoiron intermediates 23a and 25 by deuteriation
in high yields of 23b and 26. But similar failure is encoun-
tered with titanium derivatives of the structure 36 (verified
by X-ray analysis.[15]). Direct treatment of 36 in CH2Cl2
with D2O cleaved both C–M bonds with the quantitative
formation of 37 (Scheme 10). However, when 36 was treated
first with excess Et3N, then a 50:50 mixture of completely
protonated (Z) and (E) isomers of 38 was obtained (38a
and 38b). Subsequently, it was shown that when a solution
of 36 in CDCl3 was treated with a pyridine/water mixture,
the (Z) and (E) isomers obtained were 84% deuteriated as
38c and 71% deuteriated as 38d, respectively. These results
clearly indicate an induced homolysis of the pyridine com-
plex 39 before protolysis[16] (Scheme 11).
Reaction of Iron(III) Chloride with Benzylmagnesium
Chloride
When 1 mol-equiv. of iron(III) chloride in THF was al-
lowed to react with 3 mol-equiv. of benzylmagnesium chlo-
ride at –78 °C and the mixture then warmed to 25 °C, a
black suspension was ultimately formed. The black solid,
which coated the magnetic stirbar, dissolved in aqueous
HCl with the evolution of H2 and thus was iron powder.[11c]
The foregoing reaction was carried out again, but this
time the iron(III) chloride was first admixed with 0.93 mol-
equiv. of diphenylacetylene (8) at –78 °C before the 3 mol-
equiv. of benzylmagnesium chloride were added. Similar
further reaction and hydrolysis led to a mixture of bibenzyl
(49%), 2,3-diphenylindene (31, 7%), (Z)-1,2,3-triphenyl-1-
propene (32, 17%) and (E)-1,2,3-triphenyl-1-propene (33,
17%). The origin of 31–33 can be attributed to the genera-
tion of Ph–CH2–Fe=CHPh (28) in the reaction mixture and
its capture by its cycloaddition to 8 with the formation of
29. Electrocyclic ring-opening of 29 would then produce 30.
Protonation of this carbene would yield 31–33 (Scheme 8).
Scheme 10.
Scheme 11.
Scheme 8.
By analogy we propose that iron intermediates 23a and
The unusual protonation of iron–carbene 30 to yield 31, 25 largely undergo homolysis before the deuteriolysis step
32 and 33 finds its exact precedent in the reaction of 1,2,3- with the resulting organic radicals abstracting hydrogen
triphenylcyclopropene (34) with (Et3P)4Ni.[12] From this in- atoms from THF.
2974
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 2971–2977

Novel Alkylidenating Agents of Iron(III) Derivatives
As to the cycloaddition proposed to occur between 7 and Grubbs catalyst, our future research will be focused on gen-
diphenylacetylene, one should consider whether alternative erating such iron–carbenes selectively and in high yield and
modes of forming products 11 and 12 might be competitive. in isolating these carbenes with stabilizing phosphanes or
Both such involve the carbometallation of the CϵC bond, N-heterocyclic carbenes.[17]
either by n-butyllithium itself at –78 °C or by an n-butyliron
reagent, such as 6. Both alternatives can be rejected based
on known reactions. n-Butyllithium adds to diphenylacety-
Conclusions
lene exclusively in an (E) manner, leading only to 12.[19]
The interaction of a 1:3 ratio of FeCl3/n-butyllithium in
Were this so, the origin of 11 would be unexplained. As to
THF at –78 °C forms a purple solid containing about an
the possible addition of an n-butyl–iron bond from 6, in the
80:20 mixture of n-butyliron(III)–butylidene (7)/tri-n-butyl-
iron(III) (6), whose presence was uniquely in accord with
chemical trapping results obtained with diphenylacetylene
or benzonitrile, augmented by deuteriolysis. Further trapp-
whole of transition metal chemistry there appears to be no
example of a butyl–metal bond carbometallating a C=C or
CϵC linkage. What occurs preferentially is the loss of 1-
butene and the hydrometallation of the unsaturated C–C
ing with methyl benzoate or benzophenone gave products
bond, as proposed in Scheme 3.
completely consistent with the presence of 7 but deuter-
A unifying network proposing the likely origin of 6 and
iolysis gave low deuteriation of the products. Because of the
low stability of any C–Fe bond above 0 °C, such organoiron
intermediates most likely underwent homolysis before re-
acting with D2O in the workup. Similar results with the 1:3
interaction of FeCl3 and benzylmagnesium chloride support
the generation of benzyliron(III)–benzylidene (28) in mod-
erate yield.
7 and their possible relationship to each other is offered in
Scheme 12. The actual paths of the depicted reactions can
be delimited by the reaction of FeCl3 with 2 mol-equiv. of
n-butyllithium. Simple butylation of 40 by chloride dis-
placement would lead to 41 (path a), but α-lithiation of 40
would produce 42, and α,µ-elimination would then generate
43 (path b) (Scheme 12). A decisive distinction between
these two pathways was made by adding 1 mol-equiv. of
methyl benzoate (22) to the reaction mixture. Workup pro-
Experimental Section
vided a quantitative yield of 5-phenyl-5-nonanol (23b)
(Scheme 5). Since 41 is the only organoiron reagent present
having two C4 groups available (43 having only one C4
group), the reaction of 40 with nBuLi must have proceeded
only by path a.
Instrumentation, Analysis and Starting Reagents: The n-butyllith-
ium (1.6  in hexane) and the iron(III) chloride (anhydrous, 97%)
were used as received from Aldrich Chemical Company. All reac-
tions were carried out under a positive pressure of anhydrous, oxy-
gen-free argon. All solvents employed with organometallic com-
pounds were dried and distilled from a sodium metal/benzophe-
none ketyl mixture prior to use.[18] The IR spectra were recorded
with a Perkin–Elmer instrument (model 457), and samples were
measured either as mineral oil mulls or as KBr films. The NMR
spectra (1H and 13C) were recorded with a Bruker spectrometer
(model EM-360), and tetramethylsilane (Me4Si) was used as the
internal standard. The chemical shifts reported are expressed on
the δ scale in parts per million (ppm) from the Me4Si reference
signal. The GC/MS measurements and analyses were performed
with a Hewlett–Packard GC 5890/Hewlett–Packard 5970 mass-se-
lective-detector instrument. The gas chromatographic analyses were
carried out with a Hewlett–Packard instrument (model 5880) pro-
vided with a 2-m OV-101 packed column or with a Hewlett–Pack-
ard instrument (model 4890) having a 30-m SE-30 capillary col-
umn. Melting points were determined with a Thomas-Hoover Un-
imelt capillary melting point apparatus and are uncorrected.
Scheme 12.
The further reaction of 41 with the third equivalent of
nBuLi could have led to 6 by path c, involving butylation
of 41 with chloride displacement or produced 7 by path d
through a sequence of α-lithiation and α,µ-elimination, sim-
ilar to path b. Alternatively or competitively, 7 might have
Reactions of Iron(III) Chloride with n-Butyllithium
formed from 6 (paths c and f). Although a compelling General Reaction Procedure and Hydrolytic Workup: All glassware
and needles for the transfer and reaction of liquid samples were
dried in an oven at 120 °C and while cooling were thoroughly
flushed with argon. The standard reaction apparatus was a 125-
mL, two-necked Schlenk flask, provided with a Teflon-coated stir-
bar and having the argon inlet on one neck and a rubber septum
on the other neck. Then 50 mL of the anhydrous, deoxygenated
solvent (generally THF or occasionally hexane was introduced and
afterwards the content of a 5.00 g-bottle of commercial, powdered
and anhydrous iron(III) chloride (97%, 30.0 mmol) was emptied
into the flask under a stream of emerging argon. The reaction sus-
choice cannot yet be made, we favor the simplicity of path d
leading to 7, for whose intermediates cogent evidence has
been adduced with alkyl group 4 compounds.[10]
Similarly, preliminary experiments herein reported with
the organylation of iron(III) chloride by benzylmagnesium
chloride support the generation of intermediate iron(III)–
carbenes, such as benzyliron(III)–benzylidene (28) in mod-
erate yields (Scheme 8). Since such relatively inexpensive
iron(III)–carbenes would be direct structural analogs of the
versatile and highly valuable ruthenium–carbenes of the pension was cooled to –78 °C with magnetic stirring. Thereupon,
Eur. J. Org. Chem. 2010, 2971–2977
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2975

J. J. Eisch, J. U. Sohn, E. J. Rabinowitz
FULL PAPER
2 or 3 mol-equiv. of 1.60  n-butyllithium in hexanes (38 mL or
56 mL) were added dropwise by syringe through the septum. The
resulting dark purple mixture was stirred at –78 °C for 3 h. In most
reactions, the reaction mixture at –78 °C was then treated with
28.0 mmol of the chemical derivatizing reagent dissolved in anhy-
drous, deoxygenated hexanes through the rubber septum with a
gastight syringe. The resulting mixture was warmed to 25 °C over
15 h, during which time the purple reaction color turned black.
Hydrolysis of the mixture with 1  aqueous HCl (or with D2O),
extraction with three 20-mL portions of diethyl ether and combina-
tion of the diethyl ether extracts gave an organic solution, which
was washed with aqueous NaHCO3 and then dried with anhydrous
Na2SO4. Removal of volatiles left an oil in good mass recovery (ca.
95%), which was analyzed directly by NMR and IR spectroscopy.
In any attempted deuteriolytic workup the stirred reaction mixture
was warmed up to 0 °C in an ice bath and treated dropwise from
a gastight syringe, first with 5 mL of D2O (98%) and then with
5 mL of 1  DCl in D2O. After deuteriolysis, a normal hydrolytic
workup was conducted.
strates. On the other hand, iron–carbene 7 should signal its pres-
ence by its characteristic cycloadditions with strained olefins, al-
kynes or nitriles. The reaction of FeCl3 in hexanes with 3 mol-
equiv. of n-butyllithium was carried out at –78 °C according to the
general procedure, and then various runs were treated with
28 mmol of the following reagents.
(1) Methyl Benzoate (22): A quantitative yield of 5-phenyl-5-non-
anol resulted. Vide supra for 1H and 13C NMR spectroscopic data.
(2) Benzophenone (24) at –78 °C: A mixture of 1,1-diphenyl-1-
pentanol (26, 78%) and benzhydrol (22%) resulted, as shown by
1
1H and 13C NMR spectroscopy. H NMR of 26 (CDCl3): δ = 0.86
(t, CH3), 2.13 (m, CH2–CH2), 2.23–2.28 (t, 2 H), 7.25–7.50 (m, 10
H) ppm. 13C NMR (CDCl3): δ = 14.0 (CH3), 23.05 (CH2–CH3),
25.09 (CH2CH2CH3), 32.5 [CH2(CH2)CH3], 126.0–129.0 (4 ArC)
1
ppm. H and 13C NMR spectra of benzhydrol were identical with
the known spectra.
(3) Benzophenone (24) at 25 °C: When the benzophenone was
added to such a reaction mixture brought to 25 °C, ca. 98% of 24
was recovered, and a trace of tetraphenylethylene was formed.
Reactions of Iron(III) Chloride with 2 mol-equiv. of n-Butyllithium
and Subsequent Chemical Treatments: This reaction could occur in
two principally different ways (cf. Scheme 12): either (1) to form
nBu2FeCl (41); or (2) to generate carbene 43 by α,µ-dehydrohaloge-
nation from 42. The reaction of FeCl3 in hexanes with 2 mol-equiv.
of n-butyllithium was carried out at –78 °C according to the general
procedure, and then various runs were treated with 0.28 mmol of
the following reagents. Hydrolytic workup gave the following re-
sults.
(4) Benzonitrile (18): After addition of 18 at –78 °C, usual reaction,
hydrolytic workup and column chromatography, valerophenone
1
(440 mg, 98%) was isolated. H NMR of 19 (CDCl3): δ = 0.93 (t,
CH3). 13.5 (sext, 2 H), 1.71 (quint, 2 H), 2.96 (t, 2 H), 7.42–7.46
(m, 2 H), 7.52–7.56 (m, 1 H), 7.94–7.97 (m, 2 H) ppm. 13C NMR
(CDCl3): δ = 14.09, 22.65, 26.63, 38.48, 128.21, 128.7, 133.0, 137.3,
200.7 ppm. Repetition of the foregoing procedure but workup with
1  DCl in D2O (98% D) gave 20, which was ca. 75% monodeuteri-
ated at the C-2 position (diminution of the 1H triplet area at δ =
2.96 ppm by 38%).
(1) Methyl Benzoate (22): Usual workup gave essentially a quantita-
tive yield (6.30 g) of relatively pure 5-phenyl-5-nonanol (23b). 1H
NMR (CDCl3): δ = 0.82 (m, 6 H), 1.05 (m, OH), 1.25 (m, 6 H),
1.80 (m, 6 H), 7.20–7.45 (m, 5 H) ppm (overlapping multiplets).
13C NMR (CDCl3): δ = 13.93 (CH3), 22.416 (CH2), 25.56 (CH2),
42.683 (CH2), 57.0 (quaternary C), 125.2, 126.3, 127.7 and 133.0
(phenyl C) ppm. Since the formation of 23b from methyl benzoate
(22) requires 2 equiv. of n-butyl groups, the iron intermediate in-
volved here must be nBu2FeCl (41), rather than 43.
(5) Diphenylacetylene (8): The reaction mixture with 8 was stirred
at 20–25 °C for 24 h. The original dark purple mixture had by then
turned dark brown. Workup with 6  aqueous HCl (gas evolution)
and usual processing gave an organic residue, after solvent removal,
1
that was analyzed directly by H and 13C NMR spectroscopy. The
residue was composed of (Z)-stilbene (9, 8%), (E,E)-1,2,3,4-tet-
raphenyl-1,3-butadiene (10, 10%), (Z)-1,2-diphenyl-1-hexene (11,
18%) and (E)-1,2-diphenyl-1-hexene (12, 64%). The hexenes 11 and
12 could be separated from 9 and 10 by column chromatography
(2) Benzophenone (24): After the ketone in hexanes was added and
the mixture stirred at 20–25 °C for 18 h, the color changed from
dark purple to brown. Usual hydrolytic workup gave an organic
residue of 82% of 1,1-diphenyl-1-pentene and 18% of 1,1,2,2-tet-
raphenyl-1,2-ethanediol (27). 1H NMR of 1,1-diphenyl-1pentene
(CDCl3): δ = 0.86 (t, CH3), 1.40 (sext, CH2–CH2), 2.10 (q, CH2),
6.12 (t, CH), 7.15–7.40 (m, 10) ppm. 13C NMR (CDCl3): δ = 23.03
(CH3), 36.95 (CH2), 43.882 (CH2), 64.1 (CH), 126, 127, 128, 128.5,
1
on silica gel with elution by hexanes. H NMR of (Z) isomer (11,
18%) (CDCl3): δ = 6.5 (s, C-1), 2.48 (t, C-3), 1.30 (q, C-4), 0.80
1
(sext, C-5), 0.60 (t, C-6) ppm (key hexane peaks). H NMR of (E)
isomer (12, 64%) (CDCl3): δ = 6.73 (s, C-11), 2.70 (t, C-3), 1.30 (q,
C-4), 0.80 (sext, C-5), 0.60 (t, C-6) ppm (key hexane peaks). The
positions of the C-H signals for the (Z) and (E) isomers are in
relative fields (δ = 6.51 and 6.73 ppm) in reasonable agreement for
literature values in dioxane (δ = 6.38 and 6.64 ppm).[19]
1
141 ppm. H NMR of 1,1,2,2-tetraphenyl-1,2-ethanediol (CDCl3):
δ = 2.4 (s, 2 H), 6.9–7.3 (m, 20 H) ppm.
(3) Benzophenone (24) at 25 °C: After the reaction of FeCl2 in hex-
ane suspension with 2 mol-equiv. of n-butyllithium was conducted
at –78 °C, the purple reaction mixture was brought to 25 °C.
Within 24 h at this temperature the mixture had turned black. Then
benzophenone (5.1 g, 0.28 mmol) in dry THF (15 mL) was added.
After 2 h, usual hydrolytic workup with 1  aqueous HCl caused
gas evolution from a gray solid coating the stirbar (likely Fe). The
layer was found to contain Ͼ95% of benzophenone with only
traces of benzhydrol and tetraphenylethylene.
Reactions of Iron(III) Chloride with 3 mol-equiv. of Benzylmagne-
sium Chloride: Benzylmagnesium chloride was freshly prepared
from benzyl chloride and magnesium turnings in diethyl ether at
0.67 . To 50 mL of anhydrous diethyl ether under argon was
added FeCl3 (97%, 5.0 g, 30.0 mmol) and the mixture stirred and
cooled to –78 °C. Then 100 mmol of the benzyl Grignard reagent
was added dropwise. The resulting partial solution turned dark
brown and almost black as the mixture warmed to 12 °C (stirbar
coated with iron). After 24 h, the reaction was quenched with 1 
aqueous HCl, leading to gas evolution. Usual workup led to the
detection of only bibenzyl in the organic residue. The foregoing
reaction was repeated in that diphenylacetylene (5.34 g, 30 mmol)
was added to the FeCl3 in diethyl ether at –78 °C before the benzyl
Grignard reagent was introduced. As with the 1:3 reaction of the
FeCl3/n-butyllithium reactions, this alkyne was to serve as a chemi-
Reactions of Iron(III) Chloride with 3 mol-equiv. of n-Butyllithium
with Subsequent Chemical Derivatization: By means of appropriate
chemical trapping reagents a clear distinction between the two pro-
posed reaction pathways in Scheme 10, paths a and c vs. paths a
and d, was sought. The generation of nBu3Fe (6) should be indi-
cated by its alkylating or hydrometallating action on organic sub-
2976
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 2971–2977

Novel Alkylidenating Agents of Iron(III) Derivatives
Goering), vol. 36 (“Eisen-Organische Verbindungen”), Part B1,
Springer, Berlin, 1976, p. 1–10.
[9] For an authoritative review of the excellent research on this
theme conducted by Professor Kauffmann’s research group
since 1983, see: T. Kauffmann, Angew. Chem. Int. Ed. Engl.
1996, 35, 386–403.
cal trapping agent for any carbene formed, in this case the
iron(III)–benzylidene 28. Indeed, hydrolytic workup, column
chromatography on silica gel with hexane eluent and 1H NMR
analysis showed the presence of bibenzyl (58%), (Z)-1,2,3-tri-
phenyl-1-propene (32, 17%), 2,3-diphenylindene (31, 7%) and (E)-
1,2,3-triphenyl-1-propene (33, 17%). Quantification of the yields
was achieved by integrating the area of a single CH2 signal in each
product: (Z) isomer (32), δ = 3.70 (d, 2 H) ppm; (E) isomer (33), δ
= 4.08 (s, 2 H) ppm; indene (31), δ = 3.92 (s, 2 H) ppm. The prod-
ucts 31, 32 and 33 have been shown to be the typical protolysis
products of the corresponding nickel–carbene 35 (vide supra).[12,13]
[10] J. J. Eisch, A. A. Adeosun, Eur. J. Org. Chem. 2005, 993–997.
[11] a) Cycloadditions with alkynes: F. N. Tebbe, G. W. Parshall,
G. S. Reddy, J. Am. Chem. Soc. 1978, 100, 3611; b) J. J. Eisch,
A. Piotrowski, Tetrahedron Lett. 1983, 24, 2043; c) ref.[10]; cy-
cloadditions with nitriles: ref.[11b]; d) however, a referee has
pointed out that in the interactions of Grignard reagents with
FeCl2 (as might be operative with those reactions depicted in
Scheme 8), intermetallic complexes, such as [Fe(MgX)2], could
be generated and be responsible for H2 evolution upon proto-
lysis (A. Fürstner, A. Leitner, M. Méndez, H. Krause, J. Am.
Chem. Soc. 2002, 124, 13856–13863). But what the magnetic
properties of such a complex would be and what the relevance
to the FeCl3/nBuLi reaction might be, remain unclear.
[12] a) For the isolation and spectroscopic data of 32 and 33, see:
J. J. Eisch, Y. Qian, M. Singh, J. Organomet. Chem. 1996, 512,
207–217; b) for spectroscopic data of 31, see: Y. Qian, Doctoral
Dissertation, State University of New York at Binghamton,
1996, p. 119.
Acknowledgments
This research has been conducted with the financial support of
grants from the Solvay Corporation of Brussels, Belgium and the
Boulder Scientific Company of Mead, Colorado, as well as with a
Senior Scientist Award to J. J. E. from the Alexander von Hum-
boldt Stiftung of Bonn, Germany. In addition, we are indebted to
our colleagues, Rosko I. Tzolov and Dr. Renuka N. Manchanayak-
age, for significant experimental assistance. This paper is dedicated
to Professor Uwe Rosenthal of the Leibniz-Institut für Katalyse,
Rostock, Germany, who served as the gracious host for J. J. E. dur-
ing a lecture visit to LIKAT in the spring of 2006.
[13] Although dark green 35 has not been isolated analytically pure,
its 13C NMR spectrum exhibits a sharp peak at δ = 251 ppm,
characteristic of nickel(0)–carbenes; furthermore, 35 reacts
with diphenylacetylene to produce 1,2,3,4,5-pentaphenylcyclo-
pentadiene.[12a]
[14] Since iron(III) halides and alkyliron(III) compounds are Lewis
acids, such derivatives as 6 and 7 can be expected to exist in
these THF reaction mixtures as complexes with Lewis bases
present, such as 6·THF and/or 6·LiCl.
[15] J. J. Eisch, A. M. Piotrowski, S. K. Brownstein, E. J. Gabe,
F. L. Lee, J. Am. Chem. Soc. 1975, 97, 7219.
[1] J. J. Eisch, P. O. Fregene, Eur. J. Org. Chem. 2008, 4482–4492.
[2] F. Z. Dörwald, Metal Carbenes in Organic Synthesis Wiley-
VCH, New York, 1999.
[3] F. N. Tebbe, G. W. Marshall, G. S. Reddy, J. Am. Chem. Soc.
1978, 100, 3611–3613.
[4] J. J. Eisch, A. Piotrowski, Tetrahedron Lett. 1983, 24, 2043–
2046.
[16] J. J. Eisch, J. E. Galle, A. Piotrowski in Transition Metal Cata-
lyzed Polymerizations – Alkenes and Dienes, Harwood Aca-
demic Publishers, 1983, pp. 799–823.
[17] J. J. Eisch and R. N. Manchanayakage, studies in progress.
[18] A detailed description for conducting organometallic reactions
in a safe and reproducible manner is given in: J. J. Eisch, Orga-
nometallic Syntheses, Academic Press, New York, 1981, vol. 2,
pp. 1–84.
[5] M. Brookhart, M. B. Humphrey, H. J. Kratzer, G. O. Nelson,
J. Am. Chem. Soc. 1980, 102, 7802–7803.
[6] R. H. Grubbs, P. Schwab, J. W. Ziller, J. Am. Chem. Soc. 1996,
118, 100–110.
[7] R. R. Schrock, J. S. Murdzek, G. C. Bazan, J. Robbins, M. Di-
Mare, M. O’Regan, J. Am. Chem. Soc. 1990, 112, 3875–3886.
[8] For a survey of such abortive attempts and the corresponding
literature citations up to 1975, consult: Gmelin Handbuch der
Anorganischen Chemie, Ergänzungswerk zur achten Auflage
(Eds.: R. J. Meyer, E. H. E. Pietsch, A. Kotowski, M. Becke-
[19] J. E. Mulvaney, Z. G. Garlund, S. L. Garlund, J. Am. Chem.
Soc. 1963, 85, 3897–3898.
Received: March 1, 2010
Published Online: April 14, 2010
Eur. J. Org. Chem. 2010, 2971–2977
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2977