Insights into the formation of symmetrical trimers of dialkylated ketenes starting from acid chloride precursors

Article abstract of DOI:10.1007/s00706-007-0689-z

Application of known dialkyl ketene di- and trimerization to more complex precursors could readily open the route to highly functionalized symmetrical cyclobuta-1,3-diones and cyclohexa-1,3,5-triones. We report herein the results on three substrates containing either a C=C double bond or a protected glycol moiety as illustrative functionalized groups. The nature of the substituents is found to be crucial: while cyclopentenyl and more constrained dioxolanocyclopentenyl precursors efficiently dimerize, a diallylic derivative fails. At the millimolar scale, methoxide-catalyzed trimerization shows limited reproducibility, even for the reported substrate tetramethylcyclobuta-1,3-dione. However, systematic studies, including the use of microwaves, demonstrate that formation of symmetrical trimers is favored under solvent-free conditions and conventional heating, which allowed us to isolate and characterize trispiro[4.1.4.1.4.1]octadeca-2,9,15-triene-6,12,18-trione.

Full text of DOI:10.1007/s00706-007-0689-z

Monatshefte fu¨r Chemie 138, 1011–1018 (2007)
DOI 10.1007/s00706-007-0689-z
Printed in The Netherlands
Insights into the Formation of Symmetrical Trimers of Dialkylated
Ketenes Starting from Acid Chloride Precursors
Pierre-Lo¨ıc Saaidi1, Gabriel Doridot1, Erwann Jeanneau2, and Jens Hasserodt1;ꢀ
1
´
Laboratoire de Chimie, UMR CNRS 5182, Ecole Normale Superieure de Lyon, France
´ ´
Centre de Diffractometrie Henri Longchambon, Universite Claude Bernard Lyon 1, Villeurbanne, France
2
Received April 6, 2007; accepted (revised) April 12, 2007; published online July 20, 2007
# Springer-Verlag 2007
Summary. Application of known dialkyl ketene di- and tri-
triketone precursor. However, it is well-established
merization to more complex precursors could readily open
the route to highly functionalized symmetrical cyclobuta-
1,3-diones and cyclohexa-1,3,5-triones. We report herein the
results on three substrates containing either a C¼C double
bond or a protected glycol moiety as illustrative functionalized
groups. The nature of the substituents is found to be crucial:
while cyclopentenyl and more constrained dioxolanocyclo-
pentenyl precursors efficiently dimerize, a diallylic derivative
fails. At the millimolar scale, methoxide-catalyzed trimeriza-
tion shows limited reproducibility, even for the reported sub-
strate tetramethylcyclobuta-1,3-dione. However, systematic
studies, including the use of microwaves, demonstrate that
formation of symmetrical trimers is favored under solvent-free
conditions and conventional heating, which allowed us to iso-
late and characterize trispiro[4.1.4.1.4.1]octadeca-2,9,15-tri-
ene-6,12,18-trione.
that such species do only exist if no ꢀ protons are
present, thus avoiding enolization and thus the estab-
lishment of extended ꢁ systems or even aromatic
systems [2]. By contrast, in the absence of such pro-
tons, cyclohexane-1,3,5-triones can be quite stable
and are in a number of cases synthetically accessible
(Scheme 1). Compounds 2 and 5 can be formally
regarded as the symmetric homo-trimerization prod-
ucts of the corresponding ketenes.
Yet isolation of ketenes [3, 4] is a thoroughly dis-
tinct science from that of the preparation of diones
and triones as depicted in Scheme 1. In effect, tri-
meric dimethylketene 2 can be made by treatment at
100ꢁC of the dimer 1 or its polymer with bases in a
solvent free manner [5]; sodium methoxide has
proved most efficient [6]. A more extensive investi-
gation, conducted by Clark [7], led to the conclusion
that, starting from the unsymmetrical dimer 12, the
most vigorous basic conditions always favor the
symmetric trimer 2 over all other oligomerization
products. Compound 2 also forms unexpectedly dur-
ing attempts to polymerize dimethylketene. Under
specific conditions it forms considerable amounts of
the trimer 2 (up to 42% in AlBr3=toluene) [8]. How-
ever, using Lewis or protic acids and letting the mix-
ture reach room temperature leads reliably to dimer
1 formation.
Keywords. Microwave-assisted synthesis; Ketenes; Cycload-
ditions; Strained molecules; Spiro compounds.
Introduction
In our quest to create new, highly functionalized, tri-
dentate concave ligands for a variety of applications
in coordination, supramolecular, and bio-organic
chemistry, we focussed our attention on C3-sym-
metric cis,cis-configured tri-functionalized cyclo-
hexane derivatives [1]. One way to simultaneously
or even selectively introducing three functional groups
on a cyclohexane system can be envisaged via a 1,3,5-
The most systematic study of symmetric ketene
trimer formation appears to be the study by Erickson
ꢀ
Corresponding author. E-mail: jens.hasserodt@ens-lyon.fr

1012
P.-L. Saaidi et al.
acid chloride (9) according to well-known literature
procedures [15] starting from dimethyl malonate.
Several related conditions were explored to opti-
mize the conversion of 9 to dimer 6. Although there
exists a more recent protocol [16], we opted for
the one originally described by Erickson et al. [9]
as a guide in optimizing yields of 6. While working
under ‘‘diluted’’ concentrations of 9 (<0.3 molꢂ dmꢃ3)
gives poorly reproducible yields ranging from 16
to 52%, highly concentrated acid chloride in THF
using TEA is the method of choice (70% yield). Even
a one-pot procedure starting from carboxylic acid
10 and using oxalyl chloride followed by TEA gave
a very satisfactory yield of 55% (Scheme 2).
The attempt to obtain yields of target trimer 7
comparable to those observed in the literature [9] for
5 proved disappointing (Table 1, Scheme 3). Short
periods of heating of 6 either in neat form or dis-
solved in DMF in the presence of catalytic amounts
Scheme 1
et al. [9]. They succeeded in preparing 5 and several
other saturated cyclic compounds in a manner com-
parable to that of 2. Their work expanded on the
observation that dimethylketene dimer 1 could be
transformed into its trimer 2 very efficiently by
sodium methoxide catalysis (97% yield) [6]. Litera-
ture indicates that this method is sufficiently re-
producible to be of service as a reliable source of
symmetric ketene trimers [10–12], provided that the
spiro unit does not contain a too strained ring sys-
tem, i.e., a cyclopropane unit [6, 9, 13].
We therefore decided to explore the preparation
of trimer 7 from dimer 6 that shows enhanced anal-
ogy to 4 in that it does not carry any heteroatoms on
its substituents, displays the same ring size, while
only differing in the presence of two isolated double
bonds. The latter may serve later on to efficiently
functionalize any tridentate concave system, such
as 8 that may be obtained in the same manner as 3
from 2 [14].
Scheme 2
Results and Discussions
Scheme 3
In order to obtain ketene dimers like 1 and 4, one
usually prepares the corresponding acid chlorides
and dehydrochlorinates them in ether solution with
a nitrogen base leading directly to the dimerized
product. Literature mentions in several instances the
isolation of an acylated quaternary ammonium spe-
cies as the initial intermediate (see Ref. [9] and ref-
erences therein). In following this synthesis scheme,
we set out to prepare 3-cyclopentene-1-carboxylic
Table 1. Trimerization attempts of 6
Entrya
Conditionsb
6
7
11
1
2
3
30 min, 170ꢁC
26.3%
3.3%
2.5%
2.4%
–
10 min, reflux DMF
5 min, reflux DMF
–
–
1.8%
5.8%
a
b
Isolated yields except entry 1; 0.2 eq. MeONa

Symmetrical Trimers of Dialkylated Ketenes
1013
of sodium methoxide as described by Erickson et al.,
did not lead to appreciable amounts of 7 (Table 1).
On the other hand, simple nucleophilic opening of
the strained cyclobutadione ring by methoxide at-
tack, postulated for the trimerization mechanism [6],
led to the formation of ꢂ-ketoester 11. Large amounts
of the starting material 6 were left unaccounted for
with the applied work-up procedure, and presumably
polymerized. In view of the high similarity and thus
the impossibility to separate trimer 7 from dimer 6
by silica gel chromatography, it was crucial for the
identification and characterization of trimer 7 that
the conditions in entries 2 and 3 led to complete con-
sumption of 6. To give one an appreciation of the
similarity of the NMR spectra of both compounds 6
short reaction times appeared to favor formation of
7, change of the solvent or even working without one
did not improve the results. We thus turned to the
application of microwaves as the source of energy
for its known capacity to stem side reactions or even
wholly altering reaction paths [17]. Use of a mono-
modal research instrument that allowed for the
execution of test reactions in sealed vials gave the
results as listed in Table 2 (entries 4–6). Base catal-
ysis appears to be essential, while either neat condi-
tions or working in DMF equally favor formation of
7. Also, reaction times should be as short as possible,
a feat that is scarcely possible with classical heating
devices. However, in no case did we observe a rever-
sion of the dimer=trimer ratio.
1
and 7, they are depicted in Fig. 1. While H spectra
We thus decided to elucidate the factors favoring
trimerization by exploring varying conditions in the
conversion of dimer 1, since its efficient trimeriza-
tion is well established [5–7], and also because 1 is
commercially available. The simpler NMR spectra
of the dimethyl system allowed us to analyze a num-
ber of conditions employing conventional or MW
heating. It was thus possible to assess the ensemble
of all products formed from dimer 1, basically con-
sisting of four compounds: unreacted starting mate-
rial 1, trimer 2, unsymmetrical dimer 12 [18], and
unsymmetrical trimer 13 [19] (Scheme 4). Reactions
conducted in a classical oil bath (Table 3, entries 1–
3) pointed to the habitual conditions favoring trimer
formation, i.e., methoxide catalysis with dimer 1
either introduced in neat form or dissolved in DMF.
Curiously, in the case of 1, formation of any methyl
ester was not detected. Use of DMF which allows
complete consumption of 1 leads to higher amounts
of lactone derivatives 12 and 13 compared to sol-
vent-free conditions. Entries 1 and 2 illustrate how
poorly reproducible this base-catalyzed transfor-
mation really is. In fact, the reaction has to take
off in a violent fashion to give excellent results
which is unfortunately not reliably observed dur-
ing repeated attempts using apparently identical
conditions.
in CDCl3 are virtually identical, the 13C spectra dif-
fer in all signals except that for the olefinic carbons.
With an authentic sample of 7 in hand, we were
able to conduct a more thorough search for condi-
tions favoring trimerization by using gas chromato-
graphic analysis (Table 2). While neither long nor
Fig. 1. 1H (a) and 13C (b) NMR spectra of 6 (down) and 7
(up) in CDCl3
Table 2. Trimerization of 6: selective conditions, analyzed by
GC
Entry
Conditions (MeONa catalytic)
6:7
1
2
3
4
5
6
Oil-bath heating, 5 min, 130ꢁC, DMF
Oil-bath heating, 80 min, reflux toluene
Oil-bath heating, 3 min, 115ꢁC, DMSO
MW heating, 5 min, 180ꢁC, neat
MW heating, 1 min 15 sec, 220ꢁC, neat
MW heating, 45 sec, 180ꢁC, DMF
5000:1
1200:1
1000:1
760:1
175:1
15:1
Scheme 4

1014
Table 3. Trimerization of 1, analyzed by H NMR
P.-L. Saaidi et al.
1
obtained after aqueous work-up (Scheme 5). Since
strictly anhydrous conditions were maintained dur-
ing the reaction, we assume that the reaction was put
on hold at the initial stage of the acylated ammo-
nium intermediate; this apparently does not continue
to react either towards proton abstraction, thus initi-
ating ketene formation, or attack by any enolate as
proposed by Erickson et al. [9].
We made one more attempt to benefit from our
stock in 10 to test whether the internal C¼C bond
present in dimer 6, representing the only difference
compared to published compound 4, is responsible
for poor conversion to the trimeric form. A dihy-
droxylated analogue of 6 would still be a very at-
tractive entry into functionalized cyclohexatriones.
In order to form dioxolane 19 starting from 10, the
2-step procedure developed by David et al., was
applied [22]. Unfortunately, dihydroxylation using
KMnO4 under basic conditions furnished glycol 18
in poor yields. We thus turned to asymmetric dihy-
droxylation [23] which gave rise to diastereoisomer
18 in a satisfactory 65% yield. Its relative stereo-
chemistry (trans) was determined by X-ray crystal
structure analysis of the corresponding dioxolane
derivative 19 (Scheme 6).
Entry Conditions (MeONa catalytic)
2a
12a
13a
1b
2b
Oil-bath heating, 30 min, 170ꢁC 35%
1% 5%
9% 15%
Oil-bath heating, 5 min,
reflux DMF
Oil-bath heating, 2 h,
reflux toluene
MW heating, 1–5 min, 180ꢁC
MW heating, 1–3 min,
180ꢁC DMF
76%
40%
28%
3
–
48%
4c
5d
1%
4%
trace 61% 21%
a
Remaining starting material 1 represents the only other
species detected otherwise mentioned; b average results over
3 runs; best yield in 2: 98%; c average results over 3 runs; best
yield in 2: 84%; d average results over 3 runs; presence of an
acyclic trimer
This impression is not significantly altered when
switching to microwave-enhanced experimentation
(Table 3, entries 4–5). Base catalysis is still essen-
tial, while working in DMF is actually favoring
unsymmetrical dimerization and trimerization prod-
ucts (12 and 13). This confirms the notion that
microwaves effectively alter reaction paths, but un-
fortunately, in this case, not in favor of formation of
the trimer 2. Again, working with 1 in neat form is
not reproducible in our hands. This behavior finds its
parallel in the observation of Egret et al., that cation-
ic polymerization of dimethylketene gives ‘‘disperse
yields . . . , even if similar conditions of polymeriza-
tion are used’’ [8].
The one-pot procedure previously employed for
the synthesis of 6 allowed us to convert 19 into the
cyclobutadione 20 (Scheme 7). Surprisingly, only
In spite of the poor reproducibility of the conver-
sion of 1 into 2, we then decided to test whether ex-
tension of the ‘‘open-chain’’ substituents in 1 could
lead us to highly functionalized cyclohexatriones.
We thus prepared 2-allylpent-4-enoic acid (15) ac-
cording to a literature procedure [20], and subjected
it to the conditions previously employed for the syn-
thesis of cyclobutadione 6. Surprisingly, no dimer-
ization product 16 [21] could be detected at all.
Instead, carboxylic acid 15 and anhydride 17 were
Scheme 6
Scheme 5
Scheme 7

Symmetrical Trimers of Dialkylated Ketenes
1015
one diastereomer of 20 was detected. Its 13C NMR
spectrum, showing two distinct carbonyl peaks, furn-
ished clear evidence for the cis orientation of the two
dioxolane rings. The pentacyclic ring system of 20
appears to suffer from elevated strain. As a conse-
quence, it turned out to be impossible to purify 20 by
column chromatography; neither did it tolerate a
recrystallization attempt. Also, 20 loses one protect-
ing group in CDCl3 over the course of a few hours.
Hence, no attempt was made to trimerize it.
solvents except tetrahydrofuran (THF), dichloromethane
(DCM), and triethylamine (TEA) were used without purifi-
cation. THF, DCM, and TEA were introduced into the reac-
tions freshly distilled from sodium-benzophenone, calcium
chloride, and calcium hydride. Extra dry dimethylforma-
mide (DMF) was purchased from Acros. The relevant follow-
ing compounds were prepared according to known literature
procedures: cyclopent-3-enecarboxylic acid (10) [15], cyclo-
pent-3-enecarbonyl chloride (9) [24], and 2-allylpent-4-enoic
acid (15) [20]. NMR spectra were recorded on a Bruker
1
Avance 200 (operating at 200.13 MHz for H and 50.32MHz
1
for 13C). For H and 13C, chemical shifts (ꢃ) are reported in
This contribution has presented studies on the
potential exploitation of published ketene dimeriza-
tion and trimerization procedures for the synthesis
of heavily functionalized, C2 and C3 symmetrical
cyclobutadiones and cyclohexatriones. While forma-
tion of ketene dimers worked well for the two more
constrained substrates, dimer-to-trimer conversion
did not appear to be a reliable reaction at the here
applied millimolar scale, even for the reported te-
tramethylcyclobuta-1,3-dione (1). However, explo-
ration of a large variety of conditions allows us to
communicate the following observations: 1. the use
of DMF, while aiding in substrate consumption and
improving reaction reproducibility, tends to favor
the formation of unsymmetrical dimers and trimers;
2. switching from conventional heating to mild
microwave-assistance even enhances this trend, thus
resulting in a more efficient access to the unsymmet-
rical trimer of dimethyl ketene 13 (61%, starting
from commercially available dimer 1) than the one
previously reported (37% from dimer 12 [19]). Even
if dehydrochlorination of acid chlorides has been
reported for the formation of a number of symmet-
rical ketene dimers, the nature of the ꢀ substituents
plays a critical role as illustrated by the failure to
transform 14 into dimer 16. Finally, functionalized
symmetrical ketene dimer 6 was synthesized in a
one-pot procedure from the free acid, providing good
yields on a 3-g scale. Target cyclohexatrione 7, an
unsaturated symmetrical ketene trimer, was isolated,
albeit in poor yield, and fully characterized. The low
tendency of 6 to convert to 7, in contrast to its satu-
rated analogs, may be explained with the apparent
higher thermodynamic stability of dimer 6, or the
corresponding polymeric ketenes, over trimer 7.
ppm relative to TMS, using the solvent line as secondary
internal reference (s: singlet, d: doublet, t: triplet, q: quartet,
qt: quintet, m: multiplet, br: broad). All reactions were fol-
lowed by TLC on Kieselgel 60 F254 with detection by UV light
and=or with aqueous 10% permanganate solution and heating.
Kieselgel 60 (Merck) was used for flash column chromatogra-
phy. The microwave heating was carried out in closed vials
with a CEM-Discover monomode microwave apparatus at
300 W under the conditions (temperature, time) given below.
After completion of the reaction, the vessel was cooled down
rapidly to 60ꢁC. Varian CP-3800 gas chromatograph equipped
with a flame ionization detector, a Varian CP-8400 autosam-
pler and a CP-Sil5CB capillary column (30 m, 0.32 mm inter-
nal diameter, 0.25ꢄm film thickness) was employed. Nitrogen
with a flow rate of 2 cm3=min was used as gas carrier. The
oven was programmed from 50ꢁC with 0 min hold and
50ꢁC=min increment to 170ꢁC with 1.8 min hold. After a
second temperature ramp (50ꢁC=min) was applied from 170
to 200ꢁC, the oven was maintained at 200ꢁC during 9.7 min.
Melting points were measured on a Perkin-Elmer DSC7 micro-
calorimeter. The IR spectra were recorded on a Mattson 3000
FT-IR spectrometer between 3500 and 500 cmꢃ1 using KBr
disks (w: weak, m: medium, s: strong, sh: shoulder). Mass
spectra were acquired on a ThermoFinnigan LCQ Advantage
ion trap instrument, detecting positive ions (þ) or negative
ions (ꢃ) in the ESI mode. Samples (in methanol:DCM:water,
45:40:15, v:v:v) were infused directly into the source (5mm3=
min) using a syringe pump. The following source parameters
were applied: spray voltage 3.0–3.5kV, nitrogen sheath gas
flow 5–20 arbitrary units. The heated capillary was held at
200ꢁC. X-Ray crystal data were collected at room tem-
perature on a Nonius Kappa CCD diffractometer using the
COLLECT software [25]. Reduction of the data was carried
out with DENZO [26]. Frame scaling and unit-cell param-
eters refinement were made through SCALEPACK [26].
The crystal structures were solved by direct methods with
SIR97 [27] and successive difference Fourier map analyses.
Refinement on F was carried out with CRYSTALS [28]. The
hydrogen atoms were found in a difference Fourier map
and included in the refinement using soft restraints on bond
lengths and angles to regularize their geometry (C–H in
˚
˚
the range 0.93–0.98A and O–H ¼ 0.82 A) and isotropic
atomic displacement parameters (U(H) 1.2–1.5 times Ueq
of the adjacent atom). CCDC-638272 and CCDC-638273
contain the supplementary crystallographic data for com-
Experimental
˚
pounds 6 (C6H12O2; unit cell parameters: a ¼ 8.689(5)A,
Reagents were purchased from commercial suppliers and
used without further purification otherwise mentioned. All
˚
˚
b ¼ 10.063(5)A, c ¼ 11.821(5)A; space group Pbca) and 19

1016
P.-L. Saaidi et al.
˚
(C9H14O4; unit cell parameters: a ¼ 5.582(5)A, b ¼
Trispiro[4.1.4.1.4.1]octadeca-2,9,15-triene-6,12,18-trione
(7, C18H18O3) and Methyl 1-(cyclopent-3-enecarbonyl)
cyclopent-3-enecarboxylate (11, C13H16O3)
ꢁ
ꢁ
˚
˚
6.284(5)A, c ¼ 14.678(5)A; ꢀ ¼ 83.177(5) , ꢂ ¼ 82.283(5) ,
ꢅ ¼ 74.235(5)ꢁ; space group P-1). These data can be obtained
free of charge from the Cambridge Crystallographic Data
Centre via www.ccdc.cam. ac.uk=data_request=cif.
Typical Trimerization Procedure under Conventional Heating
Under Argon, dimer 6 (402mg, 2.13mmol) and MeONa
(23 mg, 0.41 mmol) were placed in a 10-cm3 2-neck round-
bottomed flask connected to a reflux condenser. Addition of
3 cm3 anhydrous DMF resulted in a purple color. Heating at
170–175ꢁC was stopped after exactly 5 min and the crude
reaction mixture was cooled down to 0ꢁC under Argon. The
resulting dark brown oil was transferred to a separatory
funnel using 10cm3 EtOAc and 15cm3 diluted NH4Cl. After
separation, the aqueous layer was extracted with 2ꢄ10cm3
EtOAc. The combined organic phases were washed with 10 cm3
brine, dried (Na2SO4), filtered, and concentrated in vacuo.
Purification of the dark oil by flash chromatography (n-hexane:
EtOAc ¼ 10:1; pure EtOAc) afforded 9.5 mg 7 as white
crystals (0.034 mmol, 2.4%), 27 mg 11 as a light yellow
oil (0.123 mmol, 5.8%), and 240 mg unknown more polar
fractions.
Compound 7: Rf ¼ 0.32 (n-hexane:EtOAc, 5:1); GC reten-
tion time: 10.4min; IR (KBr): ꢆꢀ¼ 3060 w, 2949 w, 2926 m,
2854 w, 1699 s, 1435 w, 1340 m, 1294 sh, 1261 m, 1217 m,
1096 m, 1063 m, 1022 m, 968 m, 951 w, 818 m, 804 sh, 694
sh, 690 m, 660 m, 644 w cmꢃ1; 1H NMR (CDCl3, 200 MHz):
ꢃ ¼ 2.89 (s, 12H) 5.60 (s, 8H) ppm; 13C NMR (CDCl3,
50 MHz): ꢃ ¼ 206.6 (C), 127.3 (CH), 66.9 (CH2), 41.8 (C)
ppm; HR-MS (CI): m=z calcd. for C18H18O3 [Mþ Hþ]
283.1334, found 283.1331.
Dispiro[4.1.4.1]dodeca-2,9-diene-6,12-dione (6, C12H12O2)
Typical Dimerization Procedure Starting from Acyl Chloride 9
Cyclopent-3-enecarbonyl chloride 9 (3.11 g, 23.8 mmol),
placed in a 100-cm3 flame-dried Schlenk-flask, was diluted
with 40cm3 THF under Argon. Triethylamine (2.3mol ꢂ dmꢃ3
,
4.80 cm3 TEA, 15 cm3 THF) was added to the yellow solu-
tion over 1 h under vigorous stirring. After only a few
drops, a white solid material appeared while the colour
slowly turned to orange. The reaction mixture was allowed
to react overnight. After dilution with 70 cm3 Et2O, the
brownish solution was filtered through a fritted funnel and
washed with 50 cm3 1 N HClaq. After extraction of the aqu-
eous layer with 3ꢄ25 cm3 Et2O, the combined organic
phases were washed with 30 cm3 brine, dried (Na2SO4),
filtered, and concentrated in vacuo. Purification by flash
chromatography (n-hexane:EtOAc¼ 15:1–10:1–8:1–3:1) gave
1.56g 6 as a white powder (8.29 mmol, 70%). X-Ray quality
monocrystals were obtained from recrystallization in
a cyclohexane=EtOAc mixture. Rf ¼ 0.34 (n-hexane:EtOAc,
5:1). GC retention time: 4.0 min; mp 124ꢁC; IR (KBr): ꢆꢀ¼
3072 w, 2910 m, 2841 w, 1745 s, 1718 sh, 1620 w, 1441 m,
1434 m, 1329 w, 1294 m, 1273 w, 1225 s, 1153 w, 1099w, 1038
1
m, 953 m, 814 w, 661 s cmꢃ1; H NMR (CDCl3, 200 MHz):
ꢃ ¼ 2.88 (s, 8H) 5.59 (s, 4H) ppm; 13C NMR (CD2Cl2,
50MHz): ꢃ ¼ 214.0 (C), 128.1 (CH), 78.0 (CH2), 39.8 (C)
ppm; HR-MS (EI): m=z calcd. for C12H12O2 [Mþ] 188.0357,
found 188.0835.
Compound 11: Rf ¼ 0.45 (n-hexane:EtOAc, 5:1); GC re-
tention time: 4.86 min; IR (KBr): ꢆꢀ¼ 3059 m, 2999 w,
2949 m, 2920 m, 2850 m, 1743 s, 1712 s, 1622 w, 1435
m, 1352 m, 1340 m, 1263 s, 1219 s, 1190 m, 1132 m, 1109
sh, 1084 m, 1066 m, 1016 w, 901 w, 881 w, 843 w, 793 w,
694 cmꢃ1
;
1H NMR (CDCl3, 200 MHz): ꢃ ¼ 5.62 (s, 2H)
One-pot Procedure Starting from Cyclopent-3-enecarboxylic
Acid 10
5.60 (s, 2H), 3.74 (s, 3H), 3.37 (qt, 1H, J ¼ 8.3 Hz) 2.96 (s,
4H), 2.54 (d, 4H, J ¼ 8.3 Hz) ppm; 13C NMR (CDCl3,
50MHz): ꢃ ¼ 208.7 (C), 173.5 (C), 128.7 (CH), 127.8 (CH),
65.5 (C), 52.6 (C), 45.7 (CH3), 39.0 (CH2), 38.7 (CH2) ppm;
HR-MS (EI): m=z calcd. for C18H18O3 [Mþ] 220.1099, found
220.1097.
Cyclopent-3-enecarboxylic acid 10 (7.01 g, 62.4 mmol),
placed in a 250 cm3 flame-dried 2-neck round-bottomed flask,
was diluted with 70 cm3 DCM under Argon. Addition of oxa-
lyl chloride (6.65 cm3, 74.7 mmol) followed by 3 drops of
DMF at 0ꢁC gave rise to strong gas evolution. After 1.5 h at
0ꢁC, the temperature was raised to ambient. After ca. 2.5 h
the system was purged with Argon for 30 min and concen-
trated in vacuo (T ¼ 20ꢁC, P ¼ 50 mmHg). Two successive
additions of 20cm3 DCM followed by evaporation under
reduced pressure helped to get rid of excess oxalyl chloride.
After dilution of the brown–green oil in 100 cm3 THF, TEA
(12.5 cm3, 89.7 mmol), diluted in 7.5 cm3 THF, was added
at room temperature over 1 h and the reaction mixture was
stirred overnight. It was then filtered and the solid rinsed
with Et2O (ca. 180 cm3) until complete decoloration. The
mother liquor was washed with 50 cm3 1 N HClaq. After ex-
traction of the aqueous layer with 3ꢄ25 cm3 Et2O, the com-
bined organic phases were washed with 30 cm3 brine, dried
(Na2SO4), filtered, and concentrated in vacuo. Purification
by flash chromatography (n-hexane:EtOAc ¼ 15:1–10:1; pure
EtOAc; EtOAc=MeOH ¼ 9:1) gave 3.22 g 6 as a white powder
(17.1mmol, 55%).
Typical Trimerization Procedure under Microwave Heating
An 8-cm3 reactor vial containing a magnetic stirring bar was
charged with dimer 6 (502 mg, 2.67 mmol) and MeONa
(20 mg, 0.37 mmol). It was then placed inside a 50 cm3
Schlenk tube which was further purged with Argon. After
addition of 2 cm3 of dry DMF, the reactor vial was sealed and
irradiated with microwave according to the conditions (tem-
perature and time) mentioned in Table 2. The crude mixture
was transferred to a separatory funnel containing 15 cm3
EtOAc and 15cm3 of diluted NH4Cl. The layers were sepa-
rated and the organic phase was washed with 10 cm3 brine,
dried (Na2SO4), filtered, and concentrated in vacuo. The dark
oil was passed through a short pad of silicagel using a 5:1
cyclohexane=EtOAc eluant mixture. After evaporation of the
solvents, the crude material was dissolved in DCM (7mg=cm3)
and analyzed by GC.

Symmetrical Trimers of Dialkylated Ketenes
1017
Argon. Addition of oxalyl chloride (3.60 cm3, 41.3 mmol)
followed by 3 drops of DMF at 0ꢁC gave rise to strong gas
evolution. After 1.5 h at 0ꢁC, the temperature was raised
to ambient until no more gas evolution was detected (ca.
2 h). The system was purged with Argon for 30 min and
concentrated in vacuo. The crude orange oil was vacuum
distilled to afford 4.14 g acyl chloride 14 as a light yel-
low oil (26.1 mmol, 80%). Bp ¼ 85–89ꢁC at 50 mmHg
[Ref. [29] 64ꢁC (10 torr)]; IR (KBr): ꢆꢀ¼ 3082 m, 2993 w,
2983 m, 2945 w, 2918 w, 1795 s, 1711 w (traces of car-
boxylic acid), 1643 m, 1443 m, 993 m, 922 s, 897 m, 856 w,
2,2,4,4,6,6-Hexamethylcyclohexane-1,3,5-trione (2), 3,3-
Dimethyl-4-(propan-2-ylidene)oxetan-2-one (12), and 3,3,5,5-
Tetramethyl-6-(propan-2-ylidene)-dihydro-3H-pyran-2,4-
dione (13)
Typical Trimerization Procedure under Conventional Heating
Under Argon, dimer 1 (400 mg, 2.85mmol) was placed in a 10-
cm3 2-neck round-bottomed flask connected to a reflux con-
denser. After addition of 1.5cm3 anhydrous DMF, the reaction
mixture was refluxed for 3 min with a pre-heated oil bath
(T ¼ 170–175ꢁC). Addition of MeONa (20 mg, 0.37 mmol) to
the yellow solution initiated a strong exothermic process. The
reaction was stopped after exactly 5 min by pouring the con-
tents of the flask into an ice-bath. The crude orange solution
was transferred into a separatory funnel containing 10 cm3
EtOAc and 15cm3 diluted NH4Cl. After separation, the
aqueous layer was extracted with 2ꢄ10cm3 EtOAc. The com-
bined organic phases were washed with 10 cm3 brine, dried
(Na2SO4), filtered, and concentrated in vacuo. The orange oil
was analyzed by 1H NMR to determine the relative percentages
of compounds 1, 2, 12, and 13. When no dimer 1 was detected,
flash column chromatography (n-pentane:Et2O, 15:1) afforded
compounds 2, 12 and 13 in pure form.
1
631 w cmꢃ1; H NMR and 13C NMR in CDCl3 identical to
Ref. [30].
2-Allylpent-4-enoic anhydride (17, C16H22O3)
Acyl chloride 14 (2.00 g, 12.6mmol) placed in a 50cm3
flame-dried Schlenk-flask was diluted with 20cm3 THF under
Argon. At room temperature, TEA (1.25 mol ꢂ dmꢃ3, 2.60 cm3
TEA, 15 cm3 THF) was added to the yellow solution over 1 h
under vigorous stirring. After only a few drops, a white solid
material appeared while the colour slowly turned to orange.
The reaction mixture was allowed to react overnight at room
temperature under Argon. The reaction mixture was filtered
through a fritted funnel and the remaining brown solid was
washed with 35 cm3 Et2O. After addition of 25 cm3 1 N HClaq,
the layers were separated. The aqueous phase was extracted
with 3ꢄ15cm3 Et2O, and the combined organic extracts
were washed with 15cm3 brine, dried (Na2SO4), filtered, and
concentrated in vacuo. The crude orange oil was purified
by silicagel column chromatography (cyclohexane:EtOAc¼
10:1–8:1–5:1; pure EtOAc) to afford 0.97 g 17 as a color-
less oil (3.70 mmol, 59%) and 0.50 g 15 as a colorless oil
(3.60 mmol, 29%).
Compound 2: Rf ¼ 0.37 (n-hexane:EtOAc, 5:1); IR, 1H, and
13C NMR spectra were found to be identical with the ones
described in Ref. [8].
1
Compound 12: Rf ¼ 0.57 (n-pentane:Et2O, 10:1); IR, H,
and 13C NMR spectra were found to be identical with the
ones described in Ref. [8]; HR-MS (EI): m=z calcd. for
C9H14O4 [Mþ] 140.0837, found 140.0836.
Compound 13: Rf ¼ 0.38 (n-pentane:Et2O, 10:1); IR (KBr):
ꢆꢀ¼ 2985 s, 2939 s, 2873 m, 1765 s, 1720 s, 1674 m, 1468 s,
1384 s, 1290 s, 1225 m, 1198 m, 1134 s, 1084 s, 1036 m, 933 w,
1
876 w, 825 w, 769 w, 586 w cmꢃ1; H NMR spectrum was
found to be in good agreement with Ref. [7] (neat); 1H NMR
(CDCl3, 200 MHz): ꢃ ¼ 1.88 (s, 3H), 1.84 (s, 3H), 1.44 (s, 6H),
1.39 (s, 6H) ppm; 13C NMR (CDCl3, 50 MHz): ꢃ ¼ 209.5 (C),
171.4 (C), 143.0 (C), 118.2 (C), 51.7 (C), 48.7 (C), 25.1 (CH3),
21.7 (CH3), 20.8 (CH3), 18.6 (CH3) ppm; HR-MS (EI): m=z
calcd. for C9H14O4 [Mþ] 210.1256, found 210.1252.
Compound 17: Rf ¼ 0.46 (cyclohexane=EtOAc, 5:1); IR
(KBr): ꢆꢀ¼ 3080 m, 3003 w, 2981 m, 2941 sh, 2918 m, 2848
w, 1815 s, 1745 s, 1643 m, 1443 m, 1417 w, 1363 w, 1300 w,
1230 w, 1038 s, 997 s, 960 m, 920 s, 849 w, 810 w, 638 w cmꢃ1
;
1H NMR (CDCl3, 200 MHz): ꢃ ¼ 5.77 (m, 4H), 5.10 (m, 8H),
2.60 (m, 4H), 2.37 (m, 8H) ppm; 13C NMR (CDCl3, 50MHz):
ꢃ ¼ 170.0 (C), 134.1 (CH), 117.6 (CH2), 45.6 (CH), 34.8
(CH2) ppm; HR-MS (CI): m=z calcd. for C16H22O3 [Mþ Hþ]
263.1647, found 263.1645.
Typical Trimerization Procedure under Microwave Heating
An 8-cm3 reactor vial containing a magnetic stirring bar was
charged with dimer 1 (704mg, 5.02mmol) and MeONa (20 mg,
0.37mmol). The vial was then placed inside a 50 cm3 Schlenk
tube which was further purged with Argon. After addition
of 2 cm3 dry DMF, the reactor vial was sealed and irradiated
with microwave according to the conditions (temperature and
time) mentioned in Table 3. The crude mixture was transferred
to a separatory funnel containing 15 cm3 EtOAc and 15cm3
diluted NH4Cl. The layers were separated and the organic
phase was washed with 10 cm3 brine, dried (MgSO4), and
Compound 15: Rf ¼ 0.1 (cyclohexane=EtOAc, 5:1); 1H NMR
and 13C NMR in CDCl3 identical to Ref. [30].
(ꢅ)-(3R,4S)-3,4-Dihydroxycyclopentanecarboxylic acid (18)
Procedure adapted from Ref. [23]. A 250-cm3 round-bottomed
flask, equipped with a magnetic stirring bar, was charged with
88cm3 tert-butyl alcohol, 88cm3 water, and 24.7g AD-
mix-ꢀ. Stirring at room temperature for 15min produced two
clear phases, the lower one appearing red. Carboxylic acid
10 (2.00 g, 17.8mmol) was added at once and the resulting
slurry was stirred vigorously for 3 days. Sodium hydrosulfide
(26.5 g, 0.47mol) was added and stirring was continued for an
additional 4 h. The gray slurry was transferred to a separatory
funnel containing 80cm3 2 N HCl and 220 cm3 EtOAc. The
layers were separated, and the red aqueous phase was further
extracted with 4ꢄ220 cm3 EtOAc, while the pH was main-
1
concentrated in vacuo. The orange oil was analyzed by H
NMR to determine the relative percentages of compounds 1,
2, 12, and 13.
2-Allylpent-4-enoyl chloride (14)
Carboxylic acid 15 (4.60 g, 3.82mmol) placed in a 100 cm3
flame-dried Schlenk flask was diluted with 35 cm3 DCM under

1018
P.-L. Saaidi et al.: Symmetrical Trimers of Dialkylated Ketenes
tained at 2 after each extraction by addition of 2 N HCl. The
combined organic extracts were dried (Na2SO4) and concen-
trated in vacuo to afford 1.69g diol 18 as a white powder
(11.57 mmol, 65%) with traces of an unknown product. No
attempt was made to determine the stereochemistry of 18 at
References
[1] Saaidi PL, Jeanneau E, Bouchu D, Hasserodt J (2007)
Polyhedron 26: 1191
[2] Osman MA, Seibl J, Pretsch E (1977) Helv Chim Acta
60: 3007
[3] Tidwell TT (1994) Ketenes. Wiley, Chichester, p 52
[4] Tidwell TT (2006) Eur J Org Chem 3: 563
[5] Pregaglia GF, Binaghi M (1963) J Org Chem 28: 1152
[6] Erickson JL, Kitchens GC (1962) J Org Chem 27: 460
[7] Clark RD (1967) J Org Chem 32: 399
[8] Egret H, Couvercelle JP, Belleney J, Bunel C (2002) Eur
Polym J 38: 1953
1
that stage. H NMR (DMSO-d6, 200 MHz): ꢃ ¼ 12.00 (br s,
1H) 4.41 (br s, 2H), 3.86 (m, 2H), 2.88 (m, 1H), 1.79 (m, 4H)
ppm; 13C NMR (DMSO-d6, 50 MHz): ꢃ ¼ 177.7 (C), 72.8
(CH), 38.7.0 (CH), 34.4 (CH) ppm.
(ꢅ)-(3R,4S)-Isopropyliden-3,4-dioxycyclopentanecarboxylic
acid (19)
Procedure adapted from Ref. [31]. A 100-cm3 round-bottomed
flask was charged with a solution of diol 18 (3.6g, 24.6 mmol)
in 4 cm3 DMSO, and treated with (þ)-camphor-10-sulfonic
acid (100 mg, 0.43mmol). Cyclohexane (30 cm3) and 2,2-
dimethoxypropane (34 cm3, 274 mmol) were added and the
volatile fractions (bp ¼ 51–70ꢁC) were distilled off. The re-
maining orange oil was concentrated in vacuo and purified
by column chromatography (cyclohexane:EtOAc¼ 2:1–1:2;
pure EtOAc) to afford 4.49g of carboxylic acid 19 as a white
crystalline powder (24.1mmol, 98%). X-Ray quality mono-
crystals were obtained from recrystallization in CHCl3.
Rf ¼ 0.15–0.34 (cyclohexane:EtOAc, 1:1); mp 121ꢁC; IR
(KBr): ꢆꢀ¼ 2981 sh, 2974 s, 2669 sh, 2624 sh, 1703 s,
1468 m, 1450 m, 1429 m, 1377 m, 1319 m, 1279 s, 1253 s,
1219 s, 1205 s, 1178 m, 1132 m,1097 s, 1045 s, 984 m, 964 m,
[9] Erickson JL, Collins FE, Owen BL (1966) J Org Chem
31: 480
[10] Wulf K, Klages U, Rissom B, Fitjer L (1997) Tetrahe-
dron 53: 6011
[11] Krapcho AP, Waller FJ (1972) J Org Chem 37: 1079
[12] Krapcho AP, Waller FJ (1970) Tetrahedron Lett 3521
[13] Hoffmann HMR, Walenta A, Eggert U, Schomburg D
(1985) Angew Chem Int Ed 24: 607
[14] Dale J (1965) J Chem Soc 389
[15] Depres JP, Greene AE (1998) Org Synth 75: 195
[16] Turro NJ, Leermakers PA, Vesley GF (1973) Org Synth
5: 297
[17] Loupy A (2002) Microwaves in Organic Synthesis.
Wiley-VCH, p 61
[18] Hasek RH, Clark RD, Elam EU, Martin JC (1962) J Org
Chem 27: 60
1
899 m, 849 w, 832 m, 717 m, 575 w, 519 m cmꢃ1; H NMR
(CDCl3, 200 MHz): ꢃ ¼ 11.71 (br, s, 1H), 4.67 (m, 2H), 3.02
(tt, 1H, J ¼ 12.0 Hz, J ¼ 6.0 Hz), 2.14 (dd, 2H, J ¼ 6.0 Hz,
J ¼ 14.2Hz), 1.71 (m, 2H), 1.42 (s, 3H), 1.26 (s, 3H) ppm;
13C NMR (CDCl3, 50MHz): ꢃ ¼ 180.8 (C), 109.1 (C), 79.9
(CH), 40.6 (CH), 36.8 (CH2), 26.0 (CH3), 23.6 (CH3) ppm;
HR-MS (CI): m=z calcd. for C9H14O4 [Mþ Hþ] 187.0970,
found 187.0974.
[19] Clark RD (1967) J Org Chem 32: 399
[20] Xu B, Stephens A, Kirschenheuter G, Greslin AF, Cheng
XQ, Sennelo J, Cattaneo M, Zighetti ML, Chen AS, Kim
SA, Kim HS, Bischofberger N, Cook G, Jacobson KA
(2002) J Med Chem 45: 5694
[21] Staudinger H, Schneider H, Schotz P, Strong PM (1923)
Helv Chim Acta 291
[22] David S, Lepine MC, Lubineau A (1972) Bull Soc Chim
Fr 3580
2,4-Dispiro(isopropyliden-20,30-dioxycyclopentyl)cyclobutane-
1,3-dione (20, C18H24O6)
[23] Sharpless KB, Amberg W, Bennani YL, Crispino GA,
Hartung J, Jeong KS, Kwong HL, Morikawa K, Wang
ZM, Xu DQ, Zhang XL (1992) J Org Chem 57: 2768
[24] Hodgson DM, Thompson AJ, Wadman S, Keats CJ
(1999) Tetrahedron 55: 10815
[25] Nonius. COLLECT. Nonius BV, Delft (1997–2000)
The Netherlands
[26] Otwinowski Z, Minor W (1997) In: Carter CW Jr, Sweet
RM (eds) Methods in Enzymology, Vol. 276. Academic
Press, New York, p 307
[27] Altomare A, Burla MC, Camalli M, Cascarano GL,
Giacovazzo C, Guagliardi A, Moliterni AGG, Polidori
G, Spagna R (1999) J Appl Cryst 32: 115
[28] Betteridge PW, Carruthers JR, Cooper RI, Prout K,
Watkin DJ (2003) J Appl Cryst 36: 1487
[29] Wojtowski R (1964) Roczniki Chem 38: 319
[30] Bouhadir KH, Zhou JL, Shevlin PB (2005) Synth Com-
mun 35: 1003
Same one-pot procedure as described for compound 6 starting
with carboxylic acid 19 (675 mg, 3.62mmol). After work-up
with diluted citric acid, the crude mixture was dried under
high vacuum to afford 352 mg orange powder containing
60–85% of dimer 20 (0.63–0.89mmol, 35–50%) according
to 1H NMR. 1H NMR (CD2Cl2, 200 MHz): ꢃ ¼ 4.71 (m, 4H),
2.31 (d, 4H, J ¼ 14.4Hz), 2.05 (m, 4H), 1.61 (s, 6H), 1.26 (s,
6H) ppm; 13C NMR (CD2Cl2, 50MHz): ꢃ ¼ 213.3 (C), 207.3
(C), 111.5 (C), 81.8 (CH), 80.3 (C), 39.3 (CH2), 26.1 (CH3),
24.4 (CH3) ppm; HR-MS (ESI): m=z calcd. for C18H24O6Na
[Mþ Naþ] 359.1471, found 359.1471.
Acknowledgements
The authors thank D. Pitrat and N. Crowther for experimental
support, G. Vives (CEMES, Toulouse) for kind assistance with
microwave instrumentation, and D. Bouchu (UCBL, Lyon) for
mass spectral analysis. Financial support from the CNRS and
the French Ministry is gratefully acknowledged.
´
[31] Drouin J (2005) Manipulations commentees de chimie
organique, 3rd edn. Ecole Normale Superieure de Lyon:
´
`
Librairie du Cedre, p 247