Scandium triflate-catalyzed intramolecular Friedel-Crafts acylation with Meldrum's acids: insight into the mechanism

Article abstract of DOI:10.1016/j.tet.2009.05.058

The intramolecular Friedel-Crafts acylation of arenes with Meldrum's acid derivatives catalyzed by Sc(OTf)₃ was reported as a mild and general entry into functionalized 1-indanones. Mechanistic investigations were undertaken to determine the rate-determining step in the acylation sequence using Meldrum's acid, as well as to examine the role of the Lewis acid catalyst. Enolizable Meldrum's acid derivatives react via an acyl ketene intermediate under thermal conditions, while quaternized Meldrum's acid derivatives are thermally stable and only act as effective Friedel-Crafts acylating agents in the presence of a Lewis acid catalyst. The acylation was postulated to proceed through direct acylation of a Lewis acid-activated carbonyl. In the catalytic Friedel-Crafts acylation of Meldrum's acids, triflic acid appeared to be the active catalytic species, with Sc(OTf)₃ serving as a very mild and convenient reagent for its delivery.

Full text of DOI:10.1016/j.tet.2009.05.058

Tetrahedron 65 (2009) 6682–6695
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Scandium triflate-catalyzed intramolecular Friedel–Crafts acylation
with Meldrum’s acids: insight into the mechanism
y
*
Eric Fillion , Dan Fishlock
Department of Chemistry, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
The intramolecular Friedel–Crafts acylation of arenes with Meldrum’s acid derivatives catalyzed by
Sc(OTf)3 was reported as a mild and general entry into functionalized 1-indanones. Mechanistic in-
vestigations were undertaken to determine the rate-determining step in the acylation sequence using
Meldrum’s acid, as well as to examine the role of the Lewis acid catalyst. Enolizable Meldrum’s acid
derivatives react via an acyl ketene intermediate under thermal conditions, while quaternized Meldrum’s
acid derivatives are thermally stable and only act as effective Friedel–Crafts acylating agents in the
presence of a Lewis acid catalyst. The acylation was postulated to proceed through direct acylation of
a Lewis acid-activated carbonyl. In the catalytic Friedel–Crafts acylation of Meldrum’s acids, triflic acid
appeared to be the active catalytic species, with Sc(OTf)3 serving as a very mild and convenient reagent
for its delivery.
Received 30 March 2009
Received in revised form 21 May 2009
Accepted 21 May 2009
Available online 28 May 2009
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
reactivity of Meldrum’s acid in a multiple carbon–carbon forming
process to complete the first total synthesis of taiwaniaquinol B.7,8 In
Meldrum’s acid (2,2-dimethyl-1,3-dioxane-4,6-dione) (1) and its
derivatives have established synthetic utility as acylating agents for
heteroatomic nucleophiles.1 However, the high acidity of Meldrum’s
acid and its propensity to enolize in the presence of Brønsted bases
have precluded carbon–carbon forming processes through the ad-
dition of carbon nucleophiles.2 Based on the premise that neutral
order to broaden the scope of the Friedel–Crafts acylation with
Meldrum’s acid derivatives and improve its efficiency, a clear un-
derstanding of the reaction mechanism is needed.
Over the years, the unique reactivity of Meldrum’s acid with het-
eronucleophiles has prompted a number of mechanistic studies. 1,3-
Dioxane-4,6-dione ring cleavage by hydrolysis under neutral, acidic,
and basic conditions has been investigated thoroughly. Hydrolysis of
2,2-dimethyl-1,3-dioxane-4,6-dione (1), 2,2,5,-trimethyl-1,3-di-
oxane-4,6-dione (2), and 2,2,5,5-tetramethyl-1,3-dioxane-4,6-dione
(3) with dilute hydrochloric acid results in the formation of malonic
acid, 2-methyl and 2,2-dimethyl derivatives, respectively, in addition
to an equimolar amount of acetone.9 Pihlaja and Seilo thoroughly
studied the kinetics and mechanisms of this transformation.10 It was
concluded that two reactions, one uncatalyzed and the other cata-
lyzed, were taking place concurrently, the catalyzed reaction being
predominant. The acid-catalyzed hydrolysis was determined to pro-
ceed through an AAc2 mechanism, which involves attack of water on
the protonated carbonyl group. Similarly, a bimolecular acyl–oxygen
fission pathway, initiated by attack of a water molecule at the unac-
tivated carbonyl group, was proposed for the uncatalyzed reaction.
Notably, the reaction is autocatalytic when starting with a neutral
solution owing to the acidity of the resulting malonic acid.
non-basic p-nucleophiles would add to Meldrum’s acid derivatives
in the presence of a Lewis acid to activate the carbonyls, our group
has demonstrated that Meldrum’s acids are powerful acylating
agents in metal triflate-catalyzed intramolecular Friedel–Crafts re-
actions (Eq. 1).3–6
O
O
O
R3
M(OTf)
n
(1)
O
R1
R1
R2
R
3O
R2
The acylation method has been applied to the synthesis of a va-
riety of benzocyclic ketones and represents the first practical appli-
cation of the addition of
p-nucleophiles to Meldrum’s acids under
mild and catalytic reaction conditions. We have exploited the unique
On the contrary, Meldrum’s acid (1) and its 5-methyl derivative
2 are inert to hydrolysis in alkaline solution due the formation of
the corresponding enolate anions, consequently reducing the
electrophilicity of the carbonyl toward hydroxide ion.11 Non-eno-
lizable analog 3, however, react instantaneously with 2 equiv of
* Corresponding author. Tel.: þ1 519 888 4567 x32470; fax: þ1 519 746 0435.
E-mail address: efillion@uwaterloo.ca (E. Fillion).
Current address: Roche Palo Alto LLC, Technical SciencesdChemical Synthesis,
y
3431 Hillview Ave, Palo Alto, CA 94304, USA.
0040-4020/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2009.05.058

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6683
aqueous NaOH, while titration with a methanol solution of KOH
furnishes the potassium salt of methyl 2,2-dimethylmalonate.12
Hydrolytic decomposition studies were found to be in agreement
with a BAc2 mechanism, in which ring opening is rate-determining
following formation of the tetrahedral intermediate.13
formation of 2-spirocycloalkane-1,3-dioxane-4,6-diones 11 and 4,
respectively, through substitution of the isopropylidene fragment
in the 1,3-dioxane ring. Similar reactivity was observed for com-
pound 2. A mechanism was put forward in which acylation of the
enol of cyclohexanone and cyclopentane by Meldrum’s acid is fol-
lowed by ring closure of the resulting carboxylic acid onto the vinyl
ether intermediate.
Based on Zav’yalov’s and Polansky’s observations, Matoba and
Yamazaki reported high-yielding alcoholysis and aminolysis of 1.17
Monomethyl malonate derivatives were obtained by reacting
Meldrum’s acid (1) with diazomethane in alcohols or piperidine.
The authors’ major contribution is the suggestion that 6-methoxy-
2,2-dimethyl-1,3-dioxin-4-one (5) undergoes cycloreversion at
room temperature to produce ketene methyl ester (12), the key
electrophilic species. Then, acyl ketene condensation with alco-
hols or piperidine furnishes malonate derivatives. Sato and co-
workers further confirmed the Matoba and Yamazaki proposal by
preparing a series of 6-methoxy- and 6-siloxy-1,3-dioxin-4-ones
and studying their reactivity.18,19 Infrared spectroscopy revealed
facile cycloreversion and acyl ketene formation, which were fur-
ther trapped with alcohols. The retro hetero-Diels–Alder of 6-
methoxy- and 6-siloxy-1,3-dioxin-4-ones is rate-determining step
as deduced from the comparable rates of reactivity of t-BuOH and
EtOH.20
Analternative ring-opening mode, which is exclusive to enolizable
Meldrum’s acids, has also been explored thoroughly. In this context,
enolizable Meldrum’s acid refers to a Meldrum’s acid derivative with
at least one proton at carbon 5, while non-enolizable derivatives
containnoacidicprotonatthe5-position. Upontreatmentwithexcess
diazomethane, Meldrum’s acid (1) and 5-alkyl derivatives undergo
ring cleavage at room temperature to form acetone and methyl mal-
onate or 2-substituted methyl malonates, respectively.14 As first
described by Zav’yalov, reaction of 1 and 1,5-dioxaspiro[5.5]unde-
cane-2,4-dione (4) with diazomethane in undistilled Et2O lead to
dimethylmalonate in addition to acetone or cyclohexanone. 6-
Methoxy-1,3-dioxin-4-ones 5and 6werepostulatedasintermediates.
Carrying out the reaction of 1 with diazomethane in distilled Et2O
allowed for the detection by IR and UV spectroscopy of 6-methoxy-
4H-1,3-dioxin-4-one (5). Polansky and co-workers reinvestigated this
transformation by reacting 2,2-dimethyl-1,3-dioxane-4,6-dione (1),
2,2-diphenyl-1,3-dioxane-4,6-dione (7), and 2-phenyl-1,3-dioxane-
4,6-dione (8) with diazomethane in the presence of MeOH or EtOH.15
Dimethyl malonate or ethyl methyl malonate was generated, re-
spectively, plus acetone, benzophenone or benzaldehyde. Un-
symmetrical malonate formation was explained by attack of the EtOH
molecule at the carbonyl group of 6-methoxy-1,3-dioxin-4-ones 5, 9,
and 10 with concurrent ring opening.
Neutral hydrolysis and alcoholysis of Meldrum’s acid (1) and 5-
alkyl derivatives to the corresponding malonic acids and carboxylic
acid derivatives under thermal conditions, occasionally accompanied
by loss of CO2, isawell-established synthetic protocol.21,22 Sato and co-
workers studied the alcoholysis of 1, and demonstrated its high sus-
ceptibility to nucleophilic reagents at 80–110 ꢀC.18 The authors
concluded that the high temperature promoted the formation of acyl
ketene 13, generated through cycloreversion of tautomeric
1
3
O
O
O
O
O
O
O
O
O
O
O
O
O
5
6-hydroxydioxinone 14. Reacting 1 with L-menthone in refluxing
toluene to furnish Meldrum’s acid 15 was further evidence of the
formationofketene 13, whichactedasadieneina hetero-Diels–Alder
reaction. Of note, the [4þ2] heterocycloaddition was performed
under a slight vacuum to remove the acetone by-product. The need
to remove acetone is suggestive of a reversible retro hetero-Diels–
Alder step.
1
2
3
O
O
O
O
O
O
O
O
O
O
O
O
O
O
OMe
OMe
OMe
A large body of literature has been published on the flash vac-
uum pyrolysis of Meldrum’s acid derivatives.23 Meldrum initially
observed the thermal instability of 2,2-dimethyl-1,3-dioxin-4,6-
dione (1) above its melting point (97 ꢀC) at atmospheric pressure;
the resulting liquid turning yellow and then brown with carbon
dioxide evolution and resinous product formation.11 Ott revisited
Meldrum’s observations on the pyrolysis of 1, under partial vac-
uum, and reported the generation of acetone, carbon dioxide, acetic
acid, and trace amount of carbon suboxide.24 The group of Brown
investigated the flash vacuum pyrolysis of 1 at 430 ꢀC, to furnish
ketene, acetone, and carbone dioxide as established by infrared
spectroscopy.25 Brown and co-workers proposed three fragmen-
tation mechanisms involving either a fully concerted reaction, or an
initial partial charge separation to form a zwitterionic species. From
the latter, a concerted reaction would lead to ketene, or expulsion of
acetone to form a four-membered cyclic anhydride, which would
then fragment into ketene and carbon dioxide.
Ott also studied the pyrolysis of 2,2,5,5-tetramethyl-1,3-dioxan-
4,6-dione (3) to dimethylketene, carbone dioxide, and acetone.24
The Brown group confirmed that excellent yield of dimethylketene
could be obtained via flash vacuum pyrolysis of (3) at 520 ꢀC.26
In stark contrast with the large number of experimental studies,
theoretical investigations on the reactivity of Meldrum’s acid are
scarce.27 Recently, Pinkerton and co-workers established a correla-
tion between chemical bonding and structure–reactivity in Mel-
drum’s acid by combining electron density obtained from X-ray
diffraction data and theoretical electron density calculations.28
4
6
5
Ph
Ph
Ph Ph
O
O
O
O
O
O
H
7
8
9
O
•
Ph Ph
O
O
O
O
O
O
OMe
O
H
OMe
O
O
O
O
10
11
12
Me
O
•
i-Pr
O
O
OH
OH
O
O
O
13
14
15
Figure 1. Meldrum’s acid derivatives and reactive intermediates.
Ziegler and co-workers further illustrated the anomalous re-
activity of Meldrum’s acid and its derivatives.16 Thermolysis of 1 in
the presence of cyclopentanone and cyclohexanone resulted in the

6684
E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
At the outset of these studies, little information on the in-
teraction of Meldrum’s acid with Lewis acids was available.29
2. Results and discussion
A
single Lewis acid-catalyzed acylation procedure with Meldrum’s
acid had been described in the literature prior to our work. Rigo and
co-workers reported the reaction of silylated nucleophiles (amines,
alcohols, and lactams) with Meldrum’s acid at room temperature as
a high-yielding entry into malonic silyl esters.30 Lewis acid was
required exclusively for electron-deficient and hindered nucleo-
philes and no mechanism was put forward to explain the enhanced
reactivity of Meldrum’s acid under these conditions.
2.1. Nature of the Friedel–Crafts acylation precursor versus
cyclization efficiency
In the course of investigating the scope of the Lewis acid-cat-
alyzed intramolecular Friedel–Crafts acylation with Meldrum’s
acid, a reactivity divergence was observed for substrates 16, 18,
and 20 (Eqs. 1–3).4 Benzyl Meldrum’s acid (16) did not provide 1-
indanone (17) when refluxed for 2 h in CH3NO2 in the presence of
Derived from the studies described above, Scheme 1 summa-
rizes three plausible reaction pathways for the Lewis acid-catalyzed
intramolecular Friedel–Crafts acylation of benzyl Meldrum’s acids,
leading to 1-indanones. In all three processes, the metal triflate
initially engages in coordination with one of the Meldrum’s acid
carbonyl and triggers the acylation process.
a catalytic quantity of Sc(OTf)3, but gave exclusively de-
composition of the starting material (Eq. 2). Slow addition of the
substrate via syringe pump to a solution of the catalyst gave a low
13% yield of 17. Increasing substitution at the benzylic position
affected the acylation and indanone 19 was obtained in 56% yield
after 30 min (Eq. 3). Despite the steric hindrance of the neo-
pentylic carbonyl groups, dibenzyl Meldrum’s acid (20) was an
effective substrate and gave access to 2-benzyl-1-indanone (21) in
good yield (Eq. 4).
reaction pathway a
reaction pathway b
reaction pathway c
Sc(OTf)3
Sc(OTf)3
Sc(OTf)3
O
O
O
O
O
O
O
O
O
Sc(OTf)3
O
O
O
O
O
O
(12 mol %)
(2)
Ar OH
Ar
Ar
O
O
CH3NO2
reflux, 2 h
(0-13%)
retro Diels-Alder acylation
-acetone
ring-opening
16
17
O
•
O
O
OSc(OTf)3
Sc(OTf)3
O
X
O
O
Sc(OTf)3
O
O
O
O
H
(12 mol %)
O
(3)
Ar OH
acylation
Ar
O
CH3NO2
reflux, 30 min
(56%)
-acetone
-CO2
-acetone
-CO2
18
19
O
O
O
acylation
-CO2
•
COOH
O
X
Ar
O
Sc(OTf)3
X
O
O
(10 mol %)
(4)
Scheme 1. Proposed reaction pathways.
O
CH3NO2
reflux, 50 min
(66%)
O
20
21
In reaction pathway a, which applies exclusively to enolizable
Meldrum’s acid derivatives, the Lewis acid induces tautomerization
of the benzyl Meldrum’s acid derivative to the corresponding 6-
hydroxydioxinone that cycloreverts to furnish an acyl ketene
intermediate.31 Acyl ketene arylation generates a 1-indanone-2-
carboxylic acid that further decarboxylates to give the observed 1-
indanone. Activation of the acyl ketene via Lewis acid coordination
with either the C]O bond of the ketene moiety or the carboxyl
group potentially enhances its electrophilicity.
This series of acylation reactions showed that, while
p
-nucleo-
philicity remains constant, varying substitution at the benzylic
position of the -nucleophile or the 5-position of Meldrum’s acid
p
had a major impact on the carbon–carbon bond-forming process.
This difference in reactivity was an incentive to examine the acyl-
ation mechanism thoroughly and to pinpoint the substituents’ role
in the catalytic Friedel–Crafts acylation with Meldrum’s acids.
The Meldrum’s acid could participate directly in the acylation
process as depicted in Scheme 1 (reaction pathway b). In this
mechanism, the Lewis acid-activated carbonyl is attacked by the
2.2. Determination of the rate-determining step for the
acylation with enolizable Meldrum’s acids under thermal
conditionsdcompetition studies on decomposition of benzyl
Meldrum’s acid derivatives 16 and 22
tethered
p-nucleophile. Following acylation, subsequent loss of
acetone and CO2 provides the indanone.
It was also postulated that, analogously to flash vacuum pyro-
lysis, upon activation by a Lewis acid a concerted or stepwise ring
opening reaction generates a ketene with concomitant formation of
acetone and CO2 (Scheme 1, reaction pathway c). Arylation of the
ketene, which may be activated by the Lewis acid, completes the
catalytic cycle.
In order to make this methodology as synthetically useful as
possible, and to enable its expanded application to more complex
systems, a thorough understanding is necessary. In this manu-
script, we present detailed mechanistic studies to validate or
refute any of the three proposed mechanistic pathways depicted
in Scheme 1 for enolizable and non-enolizable Meldrum’s acids,
and gain insights on the factors that influence the Friedel–Crafts
acylation process.
Our investigations began with monosubstituted Meldrum’s acid
derivatives, which are able to undergo enolization. It was initially
determined that Meldrum’s acids monosubstituted at carbon 5
(Fig. 1 for numbering) are thermally unstable and sufficiently re-
active to participate in the acylation process and be transformed to
1-indanones without the requirement of a Lewis acid catalyst if
they contain sufficiently electron-rich nucleophiles (Eqs. 5–7).
Considering 5-(3,5-dimethoxybenzyl) Meldrum’s acid (22), its di-
rect thermal conversion to 1-indanone 23 proceeded smoothly
with excellent conversion of starting material (90% conversion was
attained after 6 h at reflux in CH3NO2 as determined by 1H NMR of
the crude reaction mixture), but only in a modest 55% isolated yield

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6685
after flash chromatography (Eq. 5). The cyclization of analogous
MeO
MeO
gem-dimethyl substrate 24 was about four times slower (26.5 h
compared with 6 h when no benzylic substituents were present)
but was less prone to side reactions, and 5,7-dimethoxy-3,3-di-
methyl-1-indanone (25) was isolated in 92% yield at 93% conver-
sion (99% yield based on starting material recovery). Similarly to
the Lewis acid-catalyzed reaction (Eq. 2), the thermal cyclization of
16 was unsuccessful (Eq. 7).
O
O
O
O
O
Me3OBF4
O
O
iPr2NEt
CH2Cl2
0 °C, 30 min
MeO
MeO
R
R
R
R
OMe
22; R = H
28; R = -(CH2)4-
MeO
MeO
O
O
•
O
reflux
O
OMe
MeO
MeO
MeO
MeO
O
R
R
R
R
OMe
O
O
O
CH3NO2
27; R = H (52%)
29; R = -(CH2)4- (67%)
(5)
O
reflux, 6 h
90% conversion
(55%)
MeO
MeO
Scheme 2. Synthesis of acyl-1-indanones via intramolecular Friedel–Crafts acylation
22
23
with acyl ketenes.
MeO
MeO
hetero-Diels–Alder reaction controls the rate of the reaction, then
the overall starting material consumption rate would be unaffected
O
O
O
CH3NO2
(6)
by the
p-nucleophilicity of the arene.
O
reflux, 26.5 h
93% conversion
(92%)
MeO
MeO
Me
Me
Me Me
O
MeO
24
25
O
O
O
O
O
O
CD3NO2
+
O
O
mesitylene
(1.0 equiv)
reflux
MeO
O
O
16 (1.0 equiv)
22 (1.0 equiv)
CH3NO2
Decomposition
(7)
O
16
O
MeO
O
O
CH3NO2
(8)
No reaction
O
reflux, 24 h
MeO
Me
26
O
If formation of the carbon–carbon bond was rate-determining,
the Thorpe–Ingold effect would have operated and substrate 24
would have reacted at a faster rate than 22.32,33 Therefore, the
rate-determining step for the thermal acylation must be the
formation of a transient reactive intermediate, either an acyl
ketene (Scheme 1, pathway a), or a ketene (Scheme 1, pathway c).
As illustrated in Eq. 8, non-enolizable Meldrum’s acid 26 was
inert when refluxed in nitromethane for 24 h; no decomposition
was observed and the substrate was quantitatively recovered.
This result suggests that pathway b (Scheme 1) is not operational
under thermal conditions.
Ketenes have rarely been exploited as electrophiles in intra-
molecular Friedel–Crafts acylations,34 and no precedent on Friedel–
Crafts acylation with acyl ketenes was found in the literature.
Investigation of reaction pathway a (Scheme 1) proceeded by
methylation of Meldrum’s acid 22 with Meerwein’s salt to produce
Figure 2. Comparative relative consumption of enolizable substrates 16 and 22 under
thermal conditions.
The acquisition of kinetic data was problematic for enolizable
Meldrum’s acid derivatives. GC or GC–MS data could not be acquired
since thermal decomposition of the substrates occurred in the inlet
port and on the column. Proton nuclear magnetic resonance data of
crude reaction mixtures was complex, and it was difficult to observe
the 1-indanone products from starting materials and by-products. It
was possible, however, to monitor substrate disappearance by ob-
serving the hydrogen at the 5-position of Meldrum’s acids and the
benzylic protons. An equimolar amount of 16 and 22 was placed in
a sealed NMR tube and dissolved in CD3NO2. The mixturewas heated
at 100 ꢀC and the percent consumption monitored by 1H NMR based
on mesitylene, an internal standard.
As shown in Figure 2, the rate of consumption for 3,5-dime-
thoxybenzyl substrate 22 was slightly faster than for 16, but this
apparent rate difference was very small with respect to the relative
nucleophilicity of the competitors. Decomposition was a first order
process.35 Therefore, the rate-determining step for both substrates
was the same, and since the nucleophilicity of 22 is about ten
thousand orders of magnitude that of 16,36 this step cannot be the
acylation of the acyl ketene intermediate or Meldrum’s acid.
an
6-methoxydioxinone (Scheme 2). Intramolecular arylation of the
acyl ketene intermediate was anticipated to produce a -keto ester
a-oxoketene intermediate via cycloreversion of the resulting
b
incapable of decarboxylation. Gratifyingly, 1-indanone-2-methyl
ester (27) was formed in 52% yield demonstrating the electrophilic
character of acyl ketenes in Friedel–Crafts acylation reactions.
Similarly, disubstituted substrate 28 gave 1-indanone 29.
Additional mechanistic insights supporting pathway a were
obtained by the competing reactions of 16 and 22 (Eqs. 5 and 7). It
was expected that if nucleophilic attack of the
p-nucleophile onto
the Meldrum’s acid carbonyls ( -complex formation) or acylation
s
of the acyl ketene was rate-determining, the consumption of the
starting materials would be substantially different (Fig. 2). On the
contrary, if the tautomerization of the Meldrum’s acid or the retro

6686
E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
Since the rate of decay is equal for each substrate and in-
dependent of -nucleophilicity, the rate-determining step must
either be within the enolization or the retro hetero-Diels–Alder
(Scheme 3). The rate of each of these steps would then be the same
If enolization of Meldrum’s acid was rate-determining, then
p
a primary kinetic isotope effect should exist if the proton at the 5-
carbon of Meldrum’s acid was replaced with deuterium. Using the
3,5-dimethoxybenzyl Meldrum’s acid substrate 22 and deuterated
analog 5-d1-22, the rate equation for production of acetone from
each substrate by thermal decomposition was determined by 1H
NMR observations in a sealed tube (Fig. 3a and b). The plot of
ln(conversion) versus time was linear in both cases (Fig. 3c and d),
indicating their apparent first order nature, and the rate equation
was extracted from the linear best-fit trend line.
for both substrates, assuming that the pKa of the
a-protons are
nearly identical.
O
O
O
From the data presented in Figure 3, kH is 20.3ꢁ10ꢂ3 sꢂ1 and kD
is 18.4ꢁ10ꢂ3 sꢂ1, giving an isotope effect kH/kD of 1.10. For mecha-
nisms in which the C–H bond is broken in the rate-determining
step, deuterium isotope effects of about five in acid- and base-
catalyzed processes would be expected.37 Therefore, in the thermal
decomposition of Meldrum’s acid derivatives to product acyl ke-
tene and acetone, the rate-determining step does not involve a C–H
(or C–D) bond breaking process at the 5-carbon, and thus enoli-
zation of the substrate is not rate-determining.
O
H
H
O
O
O
O
O
enolization
O
H
O
O
enolization
Our experimental data is consistent with Sato’s observations on
the mechanism of Meldrum’s acid ring opening by alcohols.18 As
stated earlier, Sato’s group established that the rate of alcoholysis of
Meldrum’s acid and 5-alkyl derivatives is independent of the nu-
cleophilicity of the alcohol. Morever, Sato mentioned the need to
remove acetone while trapping the ketene via Diels–Alder reaction,
which is suggestive of a reversible retro hetero-Diels–Alder step.
O
O
•
O
O
O
O
retro
hetero-Diels-Alder
OH
OH
Scheme 3. Plausible rate-determining steps of Meldrum’s acids thermal
decomposition.
Figure 3. Effect of deuterium substitution on rate constantdevaluation of isotope effect. (a) Acetone production from 5H-substrate 22 (3.2 h). (b) Acetone production from 5D-
substrate 5-d1-22 (3.2 h). (c) Natural logarithm (ln) of proton data. (d) Natural logarithm (ln) of deuterium data.

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6687
In the course of the Friedel–Crafts reaction, no trace of [4þ2]
2.3. Sc(OTf)3-catalyzed intramolecular Friedel–Crafts
acylation of enolizable Meldrum’s acid derivatives
cycloaddition product was ever observed between arylketone
product and residual Meldrum’s acid substrate. Nonetheless, the
increasing presence of acetone in the reaction might slow the rate
of progress as the cycloreversion equilibrium is shifted to the left.
This would only be an issue if the retro hetero-Diels–Alder reaction
was in fact the rate-limiting step.38 If the rate-determining step for
the thermal decomposition of benzyl Meldrum’s acid derivatives
was the retro hetero-Diels–Alder reaction, then an increase in ini-
tial acetone concentration should decrease the rate of substrate
decomposition, since this step is in equilibrium.
We have previously demonstrated that carrying out the Friedel–
Crafts acylation reaction with enolizable Meldrum’s acids in the
presence of a Lewis acid catalyst significantly accelerates the re-
action, and that side reactions are minimized.4 It was found that
while Sc(OTf)3 did greatly accelerate the rate of reaction (from 6 h
to 1 h for 22), the yield increase was actually modest, from 55% to
73% using 12 mol % of catalyst (Eq. 9 vs Eq. 5). This result demanded
an explanation of the role of Sc(OTf)3 in the Friedel–Crafts acylation
process.
Sealed NMR tube experiments conducted with 3,5-dimethoxy-
benzyl Meldrum’s acid (22) in the presence of different initial
amounts of acetone (4 mol equiv and 20 mol equiv) were con-
ducted, and the consumption of Meldrum’s acid starting material
was determined by integration. The results are presented in Fig-
ure 4. Indeed, the addition of acetone did affect the rate of substrate
decomposition and thus acyl ketene formation, but the effect was
relatively small. Even with the addition of 20 mol equiv of acetone,
the half-life of the reaction only doubled from about 30 min to
about 60 min. Of note, this experiment was conducted in a sealed
tube in which the acetone is trapped, although some portion would
exist as a vapor in the headspace of the NMR tube at elevated
temperature.39
MeO
MeO
O
Sc(OTf)3
O
O
O
(12 mol %)
(9)
O
CH3NO2
reflux, 60 min
(73%)
MeO
MeO
22
23
The sealed-tube NMR experiments performed for the thermal
acylation using 16 and 22 were repeated, but now using 10 mol %
and 30 mol % of Sc(OTf)3. The temperature was decreased to 85 ꢀC
since the reaction was considerably faster with catalyst, and ac-
quisition of the initial data points was difficult. The data is depicted
in Figure 5. The results clearly demonstrate a significant impact of
MeO
O
O
O
O
CD3NO2, 100 °C
mesitylene
(1.0 equiv)
+
O
MeO
the catalyst in the initial few minutes of the reaction, and more p-
(n equiv)
22
nucleophilic substrate 22 was consumed about twice as quickly as
16. This suggests that the addition of Sc(OTf)3 catalyst causes the
rate-determining step to be nucleophile dependent, and therefore
must be the acylation step itself. After this initial divergence,
however, the rates of consumption of the remaining material ap-
pear to be very similar, implying consumption or destruction of the
catalyst and or product inhibition,41 either by acetone and/or 1-
indanone and reversion of the reaction to thermal decomposition
described in the previous section.
n = 4 and 20
It has been reported that lanthanide triflates could significantly
catalyze the enol formation of 1,3-dicarbonyl compounds.42 While
this is indeed likely with Meldrum’s acid, the enolization step is not
rate-determining for this acylation reaction, so accelerating this
step would not have an impact on the overall rate, as observed with
the Lewis acid-catalyzed reaction. In addition, there would also not
be a preference for one substrate over another.
Consequently, the direct effect of Sc(OTf)3 on the reactivity of
acyl ketenes was examined to determine if the Lewis acid catalyst
affects this potential reactive species. The reactivity of known acyl
ketene 30, a stable and easily isolable compound, in intermolecular
Friedel–Crafts acylation was investigated. This ketene is unlikely to
undergo intramolecular Friedel–Crafts acylation due to the forma-
tion of a four-carbon arylketone.
Acyl ketene 30 was refluxed with the electron-rich
p-nucleo-
phile 1,3-dimethoxybenzene in CH3NO2. In the absence of a catalyst,
no product was observed. Ketene 30 was thermally unstable and
decomposed, giving an unidentifiable complex mixture by analysis
of the crude 1H NMR spectra (Eq. 10). However, in the presence of
a catalytic amount of Sc(OTf)3, arylketone 31 was isolated in 51%
yield (Eq. 11). Therefore, in the intermolecular acylation case, the
acyl ketene is insufficientlyelectrophilic for Friedel–Crafts acylation,
but when Sc(OTf)3 is used, the acylation proceeded in good isolated
yield. Similarly, intermolecular acylation of 5-phenyl Meldrum’s
acid 32 with 1,3-dimethyoxybenzene also failed in the absence of
catalyst (Eq.12). The catalyzed version did, however, give product 33
in 34% yield (Eq.13), which supports acyl ketene activation by Lewis
acid as key in the carbon–carbon bond-forming events.
Figure 4. Effect of acetone concentration on substrate decomposition.
For the thermal decomposition of enolizable benzyl Meldrum’s
acids, the rate-determining step appears to be the retro hetero-
Diels–Alder reaction of the tautomerized Meldrum’s acid. In-
creasing the number of benzylic substituents decreases the rate of
the retro hetero-Diels–Alder reaction, but results in a more stable
acyl ketene species that is less prone to side reactions and leads to
an increased yield of the corresponding 1-indanone product.40
Formation of a ketene intermediate under mild reaction conditions
(pathway c, Scheme 1) was discarded; experimental evidence will
be provided and discussed in the next section.

6688
E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
MeO
OMe
O
O
O
Meldrum's acid
decomposition
(12)
O
CH3NO2, reflux, 30 min
32
MeO
OMe
MeO
OMe
O
O
O
O
(13)
O
Sc(OTf)3 (10 mol %)
CH3NO2, reflux, 30 min
(34%)
32
33
For the Lewis acid-catalyzed Friedel–Crafts acylation, the
mechanism for the enolizable substrates is complicated by the
continued presence of the thermal pathway that was discussed
earlier (pathway a, Scheme 1). Under strict thermal conditions,
the rate-determining step is the retro hetero-Diels–Alder reaction
step, but in the presence of Sc(OTf)3, alternative reaction path-
ways appear to predominate the initial stage of the reaction,
which could perhaps relate to pathways b and c as illustrated in
Scheme 1. This underlying thermal instability of the substrates is
a possible contributor to the lower yields generally observed for
enolizable substrates. The Lewis acid-catalyzed acylation was
therefore better studied with the thermally stable quaternized
Meldrum’s acids, which will be presented in the next section.
2.4. Quaternized Meldrum’s acid derivatives in the catalytic
intramolecular Friedel–Crafts acylation reaction
One distinct advantage of the quaternized Meldrum’s derivatives
is their thermal stability in the absence of Lewis acid catalyst (Eq. 8).
This feature greatly simplifies their analysis since it eliminates the
underlying background reaction due to the retro hetero-Diels–Alder
reaction observed in the enolizable cases, and allows for the direct
monitoring of reaction progress by gas chromatography.
As depicted in Scheme 1, two out of the three proposed mech-
anisms do not proceed through tautomerization of Meldrum’s acid,
namely direct acylation (pathway b) and ketene formation (path-
way c), and might be operational for the Friedel–Crafts acylation
with quaternized Meldrum’s acids.
We then set out to study if the Friedel–Crafts acylation proceeded
through the intermediacy of a ketene when a quaternized Meldrum’s
acid was treated with a Lewis acid. It was postulated that the Lewis
acid-catalyzed decomposition of 5-methyl-5-phenyl Meldrum’s acid
(34) would lead to methyl phenylketene (35), a known stable com-
pound (Eq. 14).43
Figure 5. Comparative decomposition of substrates 16 and 22 under catalytic
conditions.
MeO
OMe
O
•
O
•
ketene
decomposition
Sc(OTf)3
(10)
O
O
OMe
(10 mol%)
CH3NO2, reflux, 30 min
O
+
(14)
O
O
O
postulated
decomposition
pathway
CH3
CH3
30
34
35
5-Methyl-5-phenyl Meldrum’s acid (34) was then heated with
a catalytic amount of Sc(OTf)3 in dry deuterated chloroform in
a sealed NMR tube and the 1H and 13C NMR spectra acquired at
regular time intervals. The species that quantitatively arose as il-
lustrated in Eq. 15 is anhydride 36 as a 1:1 mixture of di-
astereoisomers. After 48 h, complete and clean conversion to 36
had occurred, with the production of acetone.
MeO
OMe
MeO
OMe
O
•
O
(11)
OMe
OMe
Sc(OTf)3 (10 mol %)
CH3NO2, reflux, 30 min
(51%)
O
O
31
30

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
MeO
6689
OMe
Sc(OTf)3
CH3
O
CH3
O
O
OCH3
(10 mol%)
CH3
O
O
(15)
O
O
O
O
O
(1.1 equiv)
CDCl3
50 °C, 48 h
CH3
O
(17)
O
O
OCH3
Sc(OTf)3 (10 mol %)
CH3NO2, reflux, 1 h
(83%)
CH3
+
39
36
34
34
After unsealing the NMR tube and treatment of the reaction
mixture with excess methanol, GC–MS analysis revealed methyl
hydratropate (37) and 2-phenyl propanoic acid (38) (Scheme 4).
CH3
O
O
(1.1 equiv)
O
(18)
34
Sc(OTf)3 (10 mol %)
CH3NO2, reflux, 1 h
(62%)
40
Sc(OTf)3
CH3
O
CH3
O
O
(10 mol%)
O
O
+
O
O
CDCl3
50 °C, 48 h
The above results strongly suggest that a ketene or Lewis acid-
activated ketene is not the reactive species in the acylation reaction
with quaternized Meldrum’s acids, and that the direct acylation
mechanism likely is operational (Scheme 1, reaction pathway b). In
the next section, studies on the direct activation mechanism will be
disclosed.
CH3
O
36
34
MeOH
CH3
CH3
O
O
OH
+
CH3
O
2.5. Study of the direct acylation mechanism through
comparison of Lewis acid reaction profiles
37
38
Scheme 4. Reaction of proposed anhydride 36 with methanol.
Insights into the mechanism of activation of Meldrum’s acid by
Lewis acids were obtained by comparing the reaction profiles for
different Lewis acids in the Friedel–Crafts acylation with quater-
nized Meldrum’s acids. As illustrated in Scheme 5, the direct acyl-
ation was postulated to proceed either through direct acylation of
a Lewis acid-activated carbonyl, or via the formation of a triflic
anhydride reactive intermediate in the case of metal triflate cata-
lysts, followed by acylation.
The formation of anhydride 36 is difficult to rationalize con-
sidering the strictly anhydrous reaction conditions. The scandium
catalyst was pre-dried under high vacuum and handled in a glove
box under nitrogen, and the chloroform-d1 was distilled from CaH2
and stored in a Schlenk tube. Theoretically, one molecule of 34
could act as a nucleophile and attack another Lewis acid-activated
molecule of 34. A series of decarboxylations, followed by hydroly-
sis, would produce 36. However, a molecule of water is required for
the conversion to the observed anhydride 36, which suggests that
H2O is still associated with Sc(OTf)3 despite the extreme measures
taken to dry the catalyst.
In spite of this unusual observation, phenyl methyl Meldrum’s
acid did not provide ketene 35 as judged by 1H NMR. To confirm
that phenyl methylketene 35 was not forming transiently and
transforming into anhydride 36, it was treated under the same
conditions, but after 2 h, complete decomposition had occurred
with no recognizable peaks by 1H or 13C NMR. Since direct obser-
vation methods were unsuccessful, indirect methods were
attempted. Methyl phenylketene 35 was combined with 1,3-
dimethoxybenzene in CH3NO2 and a catalytic quantity of Sc(OTf)3.
After refluxing for 1 h, complete decomposition of the starting
material was observed, and none of the desired product was
detected (Eq. 16). It is conceivable that under these conditions the
ketene undergoes predominantly side reactions compared with the
analogous acyl ketene 30 that was successfully acylated under
similar conditions above (Eq. 11).
direct acylation pathway
acyl triflate pathway
O
O
R
O
O
R
Sc(OTf)3
Sc(OTf)3
O
O
O
O
Ar
Ar
ring-opening
OSc(OTf)2
acylation
O
O
OTf
Ar
O
O
OH
O
O
R
R
-acetone
-CO2
-acetone
M
O
R
O
O
R
acylation
-CO2
TfO
O
Ar
M = Sc(OTf)2
MeO
OMe
O
•
Scheme 5. Proposed mechanisms of Friedel–Crafts acylation with quaternized Mel-
(1.1 equiv)
drum’s acids.
Decomposition
(16)
CH3
Sc(OTf)3 (10 mol %)
CH3NO2, reflux, 1h
35
These mechanisms were proposed based on the observations repor-
ted for the intermolecular Friedel–Crafts acylation of arenes catalyzed
by Bi(OTf)3 and Bi(NTf2)3. It was found that benzoyl chloride un-
dergoes exchange with Bi(OTf)3 and Bi(NTf2)3 to produce acyl tri-
The failure of ketene 35 to undergo Friedel–Crafts acylation
under these conditions is in stark contrast to the intermolecular
reactivity of 34 under the same reaction conditions (Eqs. 17 and 18).
Both 3,5-dimethoxybenzene and furan served as effective in-
fluoromethanesulfonate
and
acyl
termolecular
in 83% and 62% yields, respectively.
p
-nucleophiles, providing the arylketones 39 and 40
bis(trifluoromethanesulfonyl)amide intermediates, respectively,
while benzoic anhydride was directly activated by the metal center.

6690
E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
These two mechanisms may also be operational for Meldrum’s acid
electrophiles.44
about half of the substrate within a minute of the reaction be-
ginning. On the other hand, Mg(NTf2)2 was amongst the best
catalysts examined, being comparable to Sc(OTf)3. In general,
metal triflamides are considered to be more Lewis acidic than
their triflate counterparts.44,49,50 This data with the magnesium-
based catalysts is especially revealing when one considers the
possible mechanisms of acylation in Scheme 5. If the acyl triflate
pathway was followed, then Mg(NTf2)2 would produce an acyl
bis(trifluoromethanesulfonyl)amide species which, as an acylat-
ing agent, would be less powerful than the acyl triflate formed
with Mg(OTf)2 based on the leaving group ability of TfOꢂ versus
Tf2Nꢂ.51 This would result in a faster reaction with Mg(OTf)2 as-
suming that acylation is the rate-determining step, not Meldrum’s
acid ring opening, which was not observed in this reaction.
Therefore, these results support the direct acylation pathway in
Scheme 5 since the acylation rates correlate with the Lewis acidity
of the catalysts.
The reaction profiles for a number of Lewis and Brønsted acids in
the intramolecular Friedel–Crafts acylation with quaternized Mel-
drum’s acid was then determined, using (3,5-dimethoxybenzyl)-5-
methyl Meldrum’s acid (26) to produce 1-indanone 41 (Fig. 6).45
The data presented is remarkable in that all the Lewis acids dis-
played extremely similar reaction profiles. All of the reactions were
MeO
MeO
O
Lewis acid
(10 mol %)
O
O
O
CH3
O
CH3NO2
100 °C
MeO
MeO
H3C
26
41
O
Catalyst
MeO
MeO
O
O
O
MeO
MeO
(10 mol %)
CH3
O
CH3NO2
100°C, 20 min
(19)
H3C
42
43
Catalyst = BF3•OEt2 (90%)
Catalyst = Sc(OTf)2 (85%)
Catalyst = TfOH (86%)
BF3•OEt2
O
O
(10 mol %)
(20)
No reaction
O
CH3NO2
reflux
O
20
Insights into the direct acylation mechanism were further
obtained with BF3$Et2O. In addition to 26, it was observed that
BF3$Et2O was effective in promoting the Friedel–Crafts acylation
of less electron-rich Meldrum’s acid 42 to provide, under catalytic
conditions, indanone 43 in 90% yield (Eq. 19). On the other hand,
dibenzyl Meldrum’s acid derivatives 20 was inert to BF3$Et2O (Eq.
20). These observations suggested that the direct acylation
mechanism (Scheme 5) is operational in the intramolecular
Friedel–Crafts acylation with Meldrum’s acids. To the best of our
knowledge, there is no report of acid fluoride formation from
acid chlorides and anhydrides using BF3$OEt2. Moreover, if an
acyl fluoride was to form, Meldrum’s acid substrate 20 would be
observed to decompose if the strength of the nucleophilic portion
was insufficient for a productive Friedel–Crafts acylation. Analysis
of the 1H NMR of the crude reaction mixture showed no reaction,
and 20 quantitatively recovered when treated with BF3$Et2O
(Eq. 20).
In this section, it was proposed that the acylation of quaternized
Meldrum’s acid derivatives could occur either via an activated
carboxylic acid derivative or by direct acylation of a Lewis acid-
activated carbonyl of Meldrum’s acid itself. The adoption of the
direct acylation pathway as a functional mechanism is consistent
with the observations presented, and at a synthetic level, provides
an operationally useful model for reaction design and product
prediction.
Figure 6. Relative reaction progress for different Lewis and Brønsted acids. In all cases,
10 mol % catalyst was used. Point to point line connections are provided solely for the
purpose of clarity.
very fast, with the bulk of substrate conversion occurring within
the first 5 min of the reaction.
In addition, some useful details that can influence the applica-
tion of this methodology in practical synthetic applications can be
gleaned from this data. Scandium triflate was utilized for three
different experiments. [Sc(OTf)3 (warmup)] allowed the reaction
solution to warm up to 100 ꢀC in the presence of catalyst, and so the
initial 2 min show a much slower progress than when then catalyst
was added to a preheated solution of substrate [Sc(OTf)3 (hot)]. It is
clearly advantageous, therefore, to add catalyst to a preheated so-
lution of quaternized derivative. The final scandium experiment
[Sc(OTf)3 w/ind] examined the catalyst inhibition by benzocyclic
ketone produced in the reaction. 1-Indanone was used as a conve-
niently available surrogate for indanone product 41 without com-
plicating the GC analysis (a distinct new peak was produced). While
this reaction still proceeded rapidly and with near complete con-
version within 20 min, the initial rate was much slower than all of
the other conditions examined. Clearly the Lewis acid was partially
sequestered by the 1-indanone (and likely acetone as well) that was
produced in the course of the reaction. This experiment with 1-
indanone (4 equiv) was an extreme demonstration of this
phenomenon.46
2.6. Catalyst inhibition by amine bases
The most profound observation was in the comparison be-
tween magnesium triflate Mg(OTf)2 and magnesium bistriflamide
Mg(NTf2)2.47,48 In this controlled experiment at precisely 100 ꢀC, it
was found that Mg(OTf)2 provided the lowest overall conversion
of the Lewis acids examined, even though it rapidly converted
The role of metal triflate catalysts in the acylation of alcohols
with benzoic anhydride has been studied.52 It was revealed that the
acylation is actually promoted by TfOH, which is the true catalytic
species, when the reaction is catalyzed by metal triflates. It was

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6691
Table 2
proposed that an acyl triflate intermediate is formed from the
Effect of DTBMP on Sc(OTf)3 Lewis acidity
acylating agent and the metal triflate, and upon acylation, triflic
acid is produced, which then assumes the role of active catalyst.
Catalyst inhibition could be obtained by addition of various pro-
portions of 2,6-di-tert-butyl-4-methylpyridine (DTBMP), a hin-
dered organic base that is known to not interact with metal
catalysts.53 Sc(OTf)3 as a source of TfOH was postulated in other
transformations.54 Of note, the efficiency (Eq. 19) and reactivity
O
O
Ph Ph
Dd
rt, CH3NO2-d
3
Sc(OTf)3
+
Sc(OTf)3
O
O
(100 mol%)
Ph Ph
47
Mixture
-Dibenzyl-
d
C]O (ppm)
d
19F3C (ppm)
a
,a
d
-valerolactone (47)
176.8
d
d
d
MeO
MeO
O
O
Sc(OTf)3
Sc(OTf)3
d
ꢂ78.0
ꢂ78.2
ꢂ79.2
ꢂ78.8
ꢂ79.3
ꢂ79.3
(10 mol %)
47þSc(OTf)3
190.5 (s)
190.8 (s)
178.6 (vb)
176.9 (b)
176.9 (s)
13.7
14.0
1.8
0.1
0.1
O
No reaction
(21)
47þTfOH
CH3NO2, 100 °C
47þSc(OTf)3þDTBMP (1 equiv)
47þSc(OTf)3þDTBMP (2 equiv)
47þSc(OTf)3þDTBMP (3 equiv)
O
BnN
44
b¼broad; s¼sharp; vb¼very broad.
Sc(OTf)3
MeO
MeO
O
O
tube containing an internal standard (mesitylene) and the reaction
heated at 105 ꢀC. After a fixed time interval (1 or 2 h) the reaction
was observed by 1H NMR to determine the substrate conversion by
integration.
It was demonstrated that a ratio of 3:1 base/Lewis acid deacti-
vated the catalyst sufficiently and no trace of product was formed.
These results are virtually identical to the results obtained by
(10 mol %)
No reaction
(22)
O
CH3NO2, 100 °C
O
N
45
O
´
Marko in a comparable experiment examining the acylation of al-
TMSOTf
MeO
cohols catalyzed by metal triflates, and also correlated with the
results obtained with Meldrum’s acids 44 and 45.
The Lewis acidity of Sc(OTf)3 in the presence of DTBMP was
directly assessed using a,a-dibenzyl-d-valerolactone 47, a mimic of
(120 mol %)
(23)
45
CH3NO2
100 °C, 1.5 h
(62%)
MeO
46
N
quaternized Meldrum’s acid since the latter decomposes in the
presence of a Lewis acid. An equimolar amount of catalyst and
substrate were combined in a sealed NMR tube and the chemical
shift of the lactone carbonyl examined.55,56 With increased loading
of DTBMP, the downfield shift was diminished, and essentially
disappeared with 2 equiv of DTBMP (Table 2), as evidenced by the
chemical shift of the lactone carbonyl. The 19F NMR also indicated
that upon combination with increasing amounts of DTBMP, the
triflate moiety was more like that of a free triflate anion than in its
scandium complex state.
Scandium triflate has been reported to be a water-tolerant Lewis
acid but no systematic work has yet been done to shed light on the
nature of the catalytic species and mechanism of the reactions.57 It
was proposed that Sc(OTf)3 increased the proticity of MeOH, and
behaves as a protic acid in the presence of MeOH.54c,d A similar
mechanism has been postulated for Yb(OTf)3 and water.54a,58
Based on the results discussed above, it appears as though
Sc(OTf)3 is practically challenging to obtain in a truly anhydrous
form. Consequently, in the catalytic Friedel–Crafts acylation of
Meldrum’s acids, TfOH is the active catalytic species, with metal
triflates acting as a very mild and convenient method for its delivery.
profile (Fig. 6) of TfOH in the catalytic intramolecular Friedel–Crafts
acylation was identical to Sc(OTf)3.
When substrates 44 and 45 were submitted to the standard
reaction conditions, no trace of indanone was obtained and the
starting material was quantitatively recovered (Eqs. 21 and 22). A
superstoichiometric amount of Lewis acid was necessary for the
reaction to proceed and yield 46 (Eq. 23). These observations sug-
gested that Sc(OTf)3 was either deactivated by the amino groups, or
that the amino groups scavenged TfOH, the true catalytic species.
The apparent inhibition of catalytic activity by sp2- and sp3-
hybridized nitrogen in these quarternarized Meldrum’s acids was
further examined by intermolecular competition experiments with
substrate 42 to provide 43. Running the same reaction with various
proportions of pyridine or 2,6-di-tert-butyl-4-methylpyridine
(DTBMP) revealed that low amounts of pyridine were tolerated, but
a threshold was reached such that no substrate conversion was
obtained (Table 1). The experiment was performed in a sealed NMR
Table 1
Catalytic acylation of 42 in the presence of amine
3. Summary
O
Sc(OTf)3
MeO
MeO
O
O
O
MeO
MeO
(15 mol %)
Benzyl Meldrum’s acid derivatives are potent acylating agent in
the intramolecular Friedel–Crafts acylation. Under thermal condi-
tions, without catalyst, monosubstituted Meldrum’s acids enolize,
then undergo a rate-limiting retro hetero-Diels–Alder reaction to
form an acyl ketene intermediate that reacts intramolecularly with
Me
O
base, 105 °C
Me
CH3NO2-d
3
42
43
Entry
Amine
Loading (mol %)
Time (h)
Conversion (%)
1
2
3
4
5
6
7
8
9
Pyridine
Pyridine
Pyridine
Pyridine
Pyridine
DTBMP
DTBMP
DTBMP
DTBMP
0
15
30
45
100
15
30
30
45
1
1
1
1
1
2
1
2
18
100
100
55
0
p
-nucleophiles, and, upon decarboxylation, an arylketone is formed.
For both quaternized and enolizable Meldrum’s acid de-
rivatives, the addition of a Lewis acid is capable of activating the
Meldrum’s acid carbonyls. This complex is either attacked di-
rectly, or undergoes a ligand transfer from the Lewis acid to
generate an activated carboxylic acid derivative. Enolizable Mel-
drum’s acids can still enolize and undergo retro hetero-Diels–
Alder to generate an acyl ketene intermediate, which can interact
0
100
50
75
0

6692
E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
with the Lewis acid in this dichotomous manner. Quaternized
derivatives, in their Lewis acid-activated form, immediately un-
dergo acylation without the possibility of enolization (and further
decomposition), resulting in faster and high-yielding reactions to
generate 2-substituted 1-indanones. Quaternized derivatives are
quite poorly Lewis basic, such that the addition of even small
quantities of a Lewis base inhibits catalyst activity for this class of
substrates. This is particularly true for Sc(OTf)3, which appears to
behave as a TfOH delivery device based on the results presented
here.
was previously reported. This procedure was applied to Meldrum’s
acids 16, 24, and 26.
4.4. Thermal intramolecular Friedel–Crafts acylation with
acyl ketenes
4.4.1. 5,7-Dimethoxy-1-indanone-2-carboxylic acid methyl
ester (27)
Meldrum’s acid derivative 22 (100 mg, 0.34 mmol) and trime-
thyloxonium tetrafluoroborate (Meerwein’s salt) (70 mg,
0.47 mmol) were combined in a round-bottomed flask equipped
with a reflux condenser, magnetic stir bar and a rubber septum
under a dry nitrogen atmosphere. At 0 ꢀC, CH2Cl2 (5 mL) was added
4. Experimental section
4.1. General
via syringe with stirring, followed by diisopropylethylamine (65 mL,
0.38 mmol). The resulting suspension was stirred for 30 min at 0 ꢀC
and then heated at reflux for 1.5 h, at which time no starting material
was observed by TLC. The crude reaction mixture was quenched with
aqueous 10% HCl then diluted with additional CH2Cl2. The layers
were partitioned and the aqueous phase was extracted with CH2Cl2
(3ꢁ). The combined organic fractions were dried over MgSO4, then
filtered and concentrated. The yellow oil was purified by flash
chromatography (2:1 EtOAc/Hex) on silica gel to provide 4 mg (52%)
All reactions were carried out in flame-dried glassware under
a dry nitrogen atmosphere. Nitromethane was distilled from CaH2.
Scandium triflate was dried under high vacuum (0.5 mmHg) for 2 h
at 180 ꢀC and stored in a dry-box. Unless indicated otherwise, all
other reagents were used as received from commercial sources. 1H
NMR spectra were referenced to residual 1H shift in CDCl3
(7.24 ppm). CDCl3 (77.0 ppm) was used as the internal reference for
13C NMR spectra. Reactions were monitored by thin-layer chro-
matography (TLC) on commercial silica pre-coated plates with
a particle size of 60 Å. Developed plates were viewed by UV lamp
(254 nm), and with p-anisaldehyde stain. Flash chromatography
was performed using 230–400 mesh silica gel. Melting points are
uncorrected. High resolution mass spectra were run at the Uni-
versity of Waterloo with a source temperature of 200 ꢀC, mass
resolution of 9000, and electron energy of 70 eV. The synthesis of
Meldrum’s acids 16, 18, 20, 22, 24, 26, 28, 42, 44, and 45 was pre-
viously described in the literature.4a
of the desired
103–104 ꢀC; lit.104–105 ꢀC (EtOAc/Et2O); 1H NMR (CDCl3, 300 MHz)
6.48 (1H, br s), 6.28 (1H, br s), 3.88 (3H, s), 3.86 (3H, s), 3.74 (3H, s),
b-keto ester 27 as an oil that solidified on standing. Mp
d
3.66 (1H, dd, J¼8.2, 3.8 Hz), 3.41 (1H, dd, J¼17.2, 3.4 Hz), 3.20 (1H, dd,
J¼17.3, 8.3 Hz); 13C NMR (CDCl3, 75 MHz)
d 195.1, 170.0, 167.6, 160.0,
158.8,117.6,101.6, 97.8, 55.8, 53.7, 52.7, 30.2. HRMS (EI): m/z calcd for
C13H14O5 (Mþ) 250.0841, found 250.0839.
4.4.2. Methyl 40,60-dimethoxy-30-oxo-20,30-dihydrospiro-
[cyclohexane-1,10-indene]-20-carboxylate (29)
Meldrum’s acid derivative 28 (200 mg, 0.55 mmol) and
trimethyloxonium tetrafluoroborate (Meerwein’s salt) (114 mg,
0.77 mmol) were combined in a round-bottomed flask equipped
with a reflux condenser, magnetic stir bar and a rubber septum
under a dry nitrogen atmosphere. At 0 ꢀC, CH2Cl2 (5 mL) was added
via syringe with stirring, followed by diisopropylethylamine
(0.13 mL, 0.75 mmol). The resulting suspension was stirred for
30 min at 0 ꢀC then heated at reflux for 1 h. An additional portion of
Meerwein’s salt (114 mg, 0.77 mmol) was added and the reflux
continued for an additional 30 min. The crude reaction mixture was
quenched with aqueous 10% HCl then diluted with additional
CH2Cl2. The layers were partitioned and the aqueous phase was
extracted with CH2Cl2 (3ꢁ). The combined organic fractions were
dried over MgSO4, then filtered and concentrated. The yellow oil
was purified by flash chromatography (1:1 Hex/EtOAc) on silica gel
4.2. Lewis acid-catalyzed intramolecular Friedel–Crafts
acylation with Meldrum’s acid derivatives
4.2.1. General procedure
Reactions were typically performed on 200 mg of substrate. In
a flame-dried round-bottomed flask equipped with a magnetic stir
bar, a reflux condenser and under inert atmosphere, was placed the
substrate and Sc(OTf)3 (10 mol %) (or other solid catalyst; liquid
catalysts were administered to a solution of the substrate in
CH3NO2). Distilled CH3NO2 was added in one portion by syringe,
and the resulting suspension immediately placed into an oil bath
preheated to 100 ꢀC. The reaction mixture was maintained at this
temperature and monitored by TLC until complete consumption of
starting material was observed. The reaction mixture was then
concentrated under vacuum and directly subjected to flash chro-
matography, using a small quantity of CH2Cl2 to assist column
loading. The catalytic Friedel–Crafts acylation of 16, 18, 20, 22, 26,
42, and 45 was previously described in the literature.4a
to provide the desired
b
-keto ester 29 as 112 mg (67%) of white
6.46 (1H,
solid. Mp 168–170 ꢀC (Et2O); 1H NMR (CDCl3, 300 MHz)
d
d, J¼1.7 Hz), 6.28 (1H, d, J¼1.6 Hz), 3.88 (6H, s), 3.67 (3H, s), 3.56
(1H, s), 2.04–2.09 (1H, m), 1.77–1.16 (9H, m); 13C NMR (CDCl3,
75 MHz)
d 196.0, 170.0, 168.3, 167.4, 159.6, 116.6, 99.4, 97.4, 64.1,
4.3. Thermal intramolecular Friedel–Crafts acylation with
Meldrum’s acid derivatives
55.8, 55.7, 52.0, 46.6, 41.8, 31.9, 25.3, 23.6, 22.7; HRMS (EI): m/z
calcd for C18H22O5 (Mþ): 318.1467, found: 318.1470.
4.3.1. General procedure
4.5. Thermal intramolecular Friedel–Crafts acylation with
enolizable substrates
Reactions were typically performed on 200 mg of substrate. In
a flame-dried round-bottomed flask equipped with a magnetic stir
bar, a reflux condenser and under inert atmosphere, was placed the
substrate and distilled CH3NO2. The resulting suspension was
placed into an oil bath preheated to 100 ꢀC. The reaction mixture
was maintained at this temperature and monitored by TLC until
complete consumption of starting material was observed. The re-
action mixture was then concentrated under vacuum and directly
subjected to flash chromatography, using a small quantity of CH2Cl2
to assist column loading. The thermal Friedel–Crafts acylation of 22
4.5.1. Competition experiments (Fig. 2)
5-(3,5-Dimethoxybenzyl) Meldrum’s acid (22) (10 mg,
0.034 mmol) and 5-benzyl Meldrum’s acid (16) (8.0 mg,
0.034 mmol) were placed in a sealable Schlenk NMR tube, to which
was added 4.7 mL of mesitylene and CD3NO2 (0.5 mL). The tube was
sealed under a dry nitrogen atmosphere, then the t¼0 data ac-
quired on a 300 MHz NMR. Integration of the mesitylene peak was
defined as nine protons, and integrations of the two benzylic

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6693
protons of the two substrates were recorded according to this
standard. The NMR tube was placed in an oil bath maintained at
precisely 100 ꢀC. At the given time points, the tube was removed
and placed in an ice/water bath for 1 min, then dried. The new NMR
data for that time point was acquired. The NMR tube was kept
tightly sealed throughout the experiment, and the reaction times
refer only to the time that the tube was heated in the oil bath. This
experiment was performed in triplicate. Time in minutes; ‘benzyl’,
‘dimethoxy’, and ‘acetone’ are the standardized integration values
for each signal in each run; %benzyl, %dimethoxy, and %aceto-
ne¼100(valuet/value0) for each experiment.
farthest upfield was recorded according to this standard. The NMR
tube was placed in an oil bath maintained at precisely 100 ꢀC. At the
given time points, the tube was removed and placed in an ice/water
bath for 1 min, then dried. The new NMR data for that time point
was acquired. The NMR tube was kept tightly sealed throughout the
experiment, and the reaction times refer only to the time that the
tube was heated in the oil bath. This experiment was repeated in
triplicate (for each acetone amount). Acetone-free data was used
from the kinetics experiment above. Time in minutes; ‘methyl’ is
the standardized integration value for that signal in each run; %
S.M.¼100(methylt/methyl0) for each experiment.
4.6. Consumption of enolizable Meldrum’s acid derivative
under thermal conditions
4.9. Catalytic intramolecular Friedel–Crafts acylation with
enolizable substrates
4.6.1. Determination of rate constant (kH) (Fig. 3)
4.9.1. Competition experiment (Fig. 5)
5-(3,5-Dimethoxybenzyl) Meldrum’s acid (22) (10.0 mg,
0.034 mmol) was placed in a sealable Schlenk NMR tube, to which
was added 5 mL of mesitylene and CD3NO2 (0.5 mL). The tube was
5-(3,5-Dimethoxybenzyl) Meldrum’s acid (22) (10 mg,
0.034 mmol) and 5-benzyl Meldrum’s acid (16) (8 mg, 0.034 mmol)
were placed in a sealable Schlenk NMR tube containing Sc(OTf)3
sealed under a dry nitrogen atmosphere, then the t¼0 data acquired
on a 300 MHz NMR. Integration of the mesitylene peak was defined
as nine protons, and integration of the Meldrum’s methyl peak
(three protons) farthest upfield was recorded according to this
standard. The NMR tube was placed in an oil bath maintained at
precisely 100 ꢀC. At the given time points, the tube was removed and
placed in an ice/water bath for 1 min, then dried. The new NMR data
for that time point was acquired. The NMR tube was kept tightly
sealed throughout the experiment, and the reaction times refer only
tothe time that the tube was heated in the oil bath. This reactionwas
performed in triplicate. Time in minutes; ‘acetone’ and ‘methyl’ are
the standardized integration values for the acetone and Meldrum’s
acid methyl signal in each run; %conv¼100(acetone/(aceto-
neþmethyl)) for each experiment; ln plot¼ln(100ꢂ%conv).
(2 mg, 0.1 equiv) or (6 mg, 0.3 equiv), to which was added 4.7 mL of
mesitylene and CD3NO2 (0.6 mL). The tube was sealed under a dry
nitrogen atmosphere, then the t¼0 data acquired on a 300 MHz
NMR. Integration of the mesitylene methyl peak was defined as
nine protons, and integrations of the two benzylic protons of the
two substrates were recorded according to this standard. The NMR
tube was placed in an oil bath maintained at precisely 85 ꢀC. At the
given time points, the tube was removed and placed in an ice/water
bath for 1 min, then dried. The new NMR data for that time point
was acquired. The NMR tube was kept tightly sealed throughout the
experiment, and the reaction times refer only to the time that the
tube was heated in the oil bath. Time in minutes; ‘benzyl’, ‘dime-
thoxy’, and ‘acetone’ are the standardized integration values for
each signal in each run; %benzyl, %dimethoxy, and %aceto-
ne¼100(valuet/value0) for each experiment.
4.7. Consumption of deuterium labeled enolizable Meldrum’s
acid derivative under thermal conditions
4.10. Intermolecular Friedel–Crafts acylation with acyl
ketenes, ketenes, and Meldrum’s acid derivatives
4.7.1. Determination of rate constant (kD) (Fig. 3)
5-(3,5-Dimethoxybenzyl)-5-deutero Meldrum’s acid (5-d1-22)
(>95% incorporation) (10.0 mg, 0.034 mmol) was placed in a seal-
able Schlenk NMR tube, to which was added 5 mL of mesitylene and
4.10.1. 3-Oxo-2-phenyl-acrylic acid methyl ester (30)
Phenyl Meldrum’s acid59 32 (337 mg, 1.53 mmol) was dissolved
in dry CH2Cl2 (5 mL) and cooled to ꢂ70 ꢀC in a Schlenk tube under
a dry argon atmosphere. An excess of dry ethereal diazomethane
solution (generated under ethanol free conditions) was added and
the reaction mixture stirred at this temperature for 30 min. The
reaction mixture was then maintained at a temperature between
ꢂ50 and ꢂ40 ꢀC as the solvent was removed under high vacuum
(0.3 Torr) with vigorous stirring. The resulting yellow oil was
warmed to room temperature under vacuum and then purged with
argon. The acyl ketene 30 was obtained as 195 mg of yellow oil
(72%) and used immediately without purification; 1H NMR
CD3NO2 (0.5 mL). The tube was sealed under a dry nitrogen atmo-
sphere, then the t¼0 data acquired on a 300 MHz NMR. Integration
of the mesitylene peak was defined as nine protons, and integration
of the Meldrum’s methyl peak (three protons) farthest upfield was
recorded according to this standard. The NMR tube was placed in an
oil bath maintained at precisely 100 ꢀC. At the given time points, the
tube was removed and placed in an ice/water bath for 1 min, then
dried. The new NMR data for that time point was acquired. The NMR
tube was kept tightly sealed throughout the experiment, and the
reaction times refer only to the time that the tube was heated in the
oil bath. This reaction was performed in triplicate. Time in minutes;
‘acetone’ and ‘methyl’ are the standardized integration values for
each signal in each run; %conv¼100(acetone/(acetoneþmethyl)) for
each experiment; ln plot¼ln(100ꢂ%conv).
(300 MHz, CDCl3)
d 7.40–7.25 (5H, m), 3.81 (3H, s); IR (CDCl3, liquid
cell) ketene 2130, ester 1721 cmꢂ1
.
4.10.2. Methyl 3-(2,4-dimethoxyphenyl)-3-oxo-2-
phenylpropanoate (31)
To solution of 1,3-dimethoxybenzene (76 mg, 0.55 mmol) in dry
CH3NO2 (1 mL) was added dry Sc(OTf)3 (27 mg, 0.055 mmol), and
the resulting suspension brought to reflux in an oil bath under
argon atmosphere. A solution of acyl ketene 30 (97 mg, 0.55 mmol)
in CH3NO2 (1 mL) was added in one portion to the refluxing re-
action mixture via syringe. After 30 min, the reaction mixture was
cooled and concentrated by rotary evaporation, and then purified
by flash chromatography (4:1 Hex/EtOAc) to provide 88 mg (51%) of
31 as a pale yellow oil. If the reaction was performed in the absence
of Sc(OTf)3 then no product or starting material was obtained. 1H
4.8. Relative effect of initial acetone concentration on
reaction rate (Fig. 4)
5-(3,5-Dimethoxybenzyl) Meldrum’s acid (22) (10.0 mg,
0.034 mmol) was placed in a sealable Schlenk NMR tube, to which
was added 5
ml of mesitylene and CD3NO2 (0.5 mL). Acetone (10 mL
or 50 L) was added to the tube, which was then sealed under a dry
m
nitrogen atmosphere, and the t¼0 data acquired on a 300 MHz
NMR. Integration of the mesitylene methyl peak was defined as
nine protons, and integration of the Meldrum’s methyl peak
NMR (300 MHz, CDCl3)
d
7.86 (1H, d, J¼8.8 Hz), 7.24–7.30 (5H, m),

6694
E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6.50 (1H, dd, J¼8.8, 2.1 Hz), 6.37 (1H, d, J¼2.1 Hz), 5.64 (1H, s), 3.83
(3H, s), 3.81 (3H, s), 3.71 (s, 3H); 13C NMR (75 MHz, CDCl3)
193.2,
170.1, 165.1, 160.4, 133.9, 133.8, 129.7, 128.3, 127.6, 119.5, 105.7, 198.1,
64.1, 55.5, 55.2, 52.3; IR (NaCl) ester 1734, ketone 1668 cmꢂ1; EI m/z
(rel int.) 314 (3), 179 (4), 166 (10), 165 (100); HRMS (EI) calcd for
C18H18O5: 314.1154, found 314.1159.
catalysts [(TfOH 5 mL, 10 mol %), BF3$OEt2 (9 mL, 10 mol %)], the
d
catalyst was added by microsyringe and timing of the reaction
immediately begun. For solid catalysts [Sc(OTf)3 (25 mg, 10 mol %),
Mg(OTf)2 (16 mg, 10 mol %), Mg(NTf2)2 (27 mg, 10 mol %)] the solid
was added rapidly through the available neck of the round-bot-
tomed flask and then the flask immediately resealed and timing
begun.
4.10.3. 1-(1,4-Dimethoxyphenyl)-2-phenyl-1-ethanone (33)
Three experiments were performed using scandium triflate. One
is as described above, but in another the catalyst was added to the
flask first, before the stock solution was added. In this experiment
the solution needed to warm up to the oil bath temperature while
being exposed to catalyst. In the third experiment the stock solu-
tion was used to dissolve 1-indanone (252 mg, 400 mol %), and the
catalyst added after the initial warming period.
Samples were taken every 30 s for the first 4 min, then every
minute for the next 6 min, then every 2 min for the next 10 min, for
a total reaction time of 20 min. This distribution ensured a large
sampling rate at the beginning of the reaction. At each time point,
A solution of 1,3-dimethoxybenzene (140 mg, 1.01 mmol) and
Sc(OTf)3 (50 mg, 0.10 mmol) in dry CH3NO2 (5 mL) was brought to
reflux under a dry nitrogen atmosphere. 5-Phenyl Meldrum’s acid
(32) (223 mg, 1.01 mmol) was added in one portion and the re-
action mixture refluxed for 30 min. The dark brown mixture was
cooled, concentrated, and then purified by flash chromatography
(4:1 Hex/EtOAc) to provide 88 mg (34%) of 33 as a pale yellow oil. If
the reaction was performed in the absence of Sc(OTf)3 then no
product or starting material was obtained. 1H NMR (300 MHz,
CDCl3)
2.1 Hz), 6.43 (1H, d, J¼2.1 Hz), 4.26 (2H, s), 3.87 (3H, s), 3.82 (3H, s);
13C NMR (75 MHz, CDCl3)
197.7, 164.5, 160.6, 135.7, 133.1, 129.6,
d
7.79 (1H, d, J¼8.7 Hz), 7.27–7.19 (5H, m), 6.5 (1H, dd, J¼8.7,
100
syringe, and immediately injected into 10
100 L of methylene chloride. At end of experiment, each sample
mL aliquots were withdrawn using a clean and dry disposable
d
mL of distilled Et3N in
128.2, 126.4, 120.8, 105.2, 98.3, 55.5, 55.4, 50.0; IR (NaCl) C]O
1665 cmꢂ1; EI m/z (rel int.) 256 (1), 179 (3), 166 (10), 165 (100);
HRMS (EI) calcd for C16H16O3: 256.1099, found 256.1106.
m
was filtered through a short plug of silica gel and eluted with ethyl
acetate. The samples were then run on a gas chromatograph
equipped with a flame ionization detector, with a temperature
gradient of 50 ꢀC for 3 min, then 50 to 270 ꢀC over 10 min. The
indanone product peak eluted at 11.7 min, and the Meldrum’s acid
derivative starting material at 12.9 min. In the following tables of
raw data, the ‘time’ column is the reaction time elapsed in minutes,
the ‘product’ column is the integration value for 1-indanone, ‘sm’ is
the Meldrum’s acid derivative 68, and the ‘%conv’¼100ꢁ‘prod’/
(‘prod’þ‘sm’).
4.10.4. 1-(2,4-Dimethoxyphenyl)-2-phenyl-1-propanone (39)
To a refluxing solution of 1,3-dimethoxybenzene (128 mg,
0.93 mmol) and 5-methyl-5-phenyl Meldrum’s acid60 (34) (200 mg,
0.85 mmol) in CH3NO2 (5 mL) under nitrogen was added Sc(OTf)3
(42 mg, 0.09 mmol). The reaction mixture was allowed to stir for 1 h
and then cooled and concentrated. Purification by flash chroma-
tography (5:1 Hex/EtOAc) provided 191 mg (83%) of 39 as a pale
yellow oil. 1H NMR (300 MHz, CDCl3)
d
7.65 (1H, d, J¼8.7 Hz), 7.25–
7.11 (5H, m), 6.43 (1H, dd, J¼8.7, 2.2 Hz), 6.34 (1H, d, J¼2.2 Hz), 4.75
Acknowledgements
(1H, q, J¼6.9 Hz), 3.79 (3H, s), 3.78 (3H, s), 1.46 (3H, d, J¼6.9 Hz); 13C
NMR (75 MHz, CDCl3) d 163.9,159.8,142.0,132.9,128.2,128.0,126.3,
This work was supported by Boehringer Ingelheim (Canada)
Ltd., AstraZeneca Canada Inc., the Natural Sciences and Engineering
Research Council of Canada (NSERC), the Canadian Foundation for
Innovation, the Ontario Innovation Trust, and the University of
Waterloo. D.F. thanks the Government of Ontario for a fellowship
121.2, 105.0, 98.2, 55.3, 55.2, 51.1, 19.2; IR (NaCl) C]O 1663 cmꢂ1; EI
m/z (rel int.) 270 (<1), 179 (8), 166 (10), 165 (100), 122 (7), 77 (6);
HRMS (EI) calcd for C17H18O3: 270.1256, found 270.1258.
4.10.5. 1-(2-Furyl)-2-phenyl-1-propanone (40)
´
(OGSST). The authors thank Professor Monica Barra, University of
In a Schlenk tube containing 5-methyl-5-phenyl Meldrum’s acid
(34) (200 mg, 0.85 mmol) and Sc(OTf)3 (42 mg, 0.09 mmol) in
CH3NO2 (5 mL) under nitrogen was added furan (64 mg,
0.94 mmol). The Schlenk tube was sealed and immediately placed in
an oil bath at 105 ꢀC for 1 h. The resulting black reaction mixturewas
cooled and concentrated, and then purified by flash chromatogra-
phy (5:1 Hex/EtOAc) to provide 106 mg (62%) of 40 as a yellow oil.1H
Waterloo, for helpful discussions.
References and notes
1. For reviews on Meldrum’s acid, see: (a) Chen, B.-C. Heterocycles 1991, 32, 529–
´
597; (b) Strozhev, M. F.; Lielbriedis, I. E.; Neiland, O. Ya. Khim. Geterotsikl. Soedin.
1991, 579–599; (c) McNab, H. Chem. Soc. Rev. 1978, 7, 345–358.
2. Meldrum’s acid derivatives are potent electrophiles when disubstituted at the 5-
position (quaternized Meldrum’s acids) and react with organolithium reagents in
an inter- and intramolecular fashion, see: (a) Dauben, W. G.; Kozikowski, A. P.;
Zimmerman,W.T.TetrahedronLett.1975,515–517;(b)Mahidol,C.;Pinyopronpanit,
Y.; Radviroongit, S.; Thebtaranonth, C.; Thebtaranonth, Y. J. Chem. Soc., Chem.
Commun.1988,1382–1383.
3. (a) Fillion, E.; Dumas, A. M. J. Org. Chem. 2008, 73, 2920–2923; (b) Fillion, E.;
Dumas, A. M.; Hogg, S. A. J. Org. Chem. 2006, 71, 9899–9902; (c) Fillion, E.;
Dumas, A. M.; Kuropatwa, B. A.; Malhotra, N. R.; Sitler, T. C. J. Org. Chem. 2006,
71, 409–412.
NMR (300 MHz, CDCl3)
m), 7.12 (1H, br d, J¼3.6 Hz), 6.43 (1H, dd, J¼3.6, 1.6 Hz) 4.47 (1H, q,
189.4,
d
7.50 (1H, dd, J¼0.8, 0.8 Hz), 7.50–7.15 (5H,
J¼7.0 Hz), 1.50 (3H, d, J¼7.0 Hz); CDCl3 (75 MHz, CDCl3)
d
152.2, 146.3, 140.8, 128.8, 127.9, 127.0, 117.8, 112.2, 47.9, 18.3; IR
(NaCl) C]O 1674 cmꢂ1; EI m/z (rel int.) 200 (41), 105 (100), 95 (61),
77 (18); HRMS (EI) calcd for C13H12O2: 200.0837, found 200.0840.
4.11. Relative reaction progressions for various Lewis Acids in
the intramolecular Friedel–Crafts acylation of quaternized
Meldrum’s acid derivatives (Fig. 6)
4. (a) Fillion, E.; Fishlock, D.; Wilsily, A.; Goll, J. M. J. Org. Chem. 2005, 70, 1316–
1327; (b) Fillion, E.; Fishlock, D. Org. Lett. 2003, 5, 4653–4656.
5. Alimitednumberofintramolecularelectrophilicaromaticsubstitutionreactions
of areneswith5-alkylideneMeldrum’sacidderivatives promoted byprotic acids
havealsobeendescribed,see:(a)Campaigne,E.;Frierson,M.R.J.Heterocycl.Chem.
1979,16, 235–237; (b) Van Allan, J. A.; Reynolds, G. A. J. Heterocycl. Chem.1972, 9,
669–673.
6. Protic acid-promoted intramolecular Friedel–Crafts acylation with acyl Mel-
drum’s acid derivatives were reported. Mechanistic evidences suggested a pro-
tonated acyl Meldrum’s acid as the acylating agent, instead of a ketene or
protonated ketene species, see: (a) Xu, F.; Armstrong, J. D., III; Zhou, G. X.; Sim-
mons, B.; Hughes, D.; Ge, Z.; Grabowski, E. J. J. J. Am. Chem. Soc. 2004, 126, 13002–
13009; Formation of 2-(3-oxoacyl)pyrroles through Friedel–Crafts acylation of
A 0.095 M stock solution containing of 5-(3,5-dimethoxy-
benzyl)-5-methyl Meldrum’s acid (26) (2.94 g, 9.55 mmol) in
CH3NO2 (100 mL) was prepared. In a glove box, the solid catalyst
was weighed into a dry screwcap vial. A dry two-necked round-
bottomed flask equipped with a reflux condenser and a stir bar was
placed into an oil bath preheated to 100 ꢀC under a dry nitrogen
atmosphere, and the 5 mL of the stock solution was added via sy-
ringe. After 5 min, the t¼0 sample was taken. For the liquid
´
´
pyrrole with acylated Meldrum’s acids, see: (b) Cantın, A.; Moya, P.; Miranda, M.
A.; Primo, J.; Primo-Yu´ fera, E. J. Agric. Food Chem. 1998, 46, 4748–4753.

E. Fillion, D. Fishlock / Tetrahedron 65 (2009) 6682–6695
6695
7. Fillion, E.; Fishlock, D. J. Am. Chem. Soc. 2005, 127, 13144–13145.
8. For a review of the use of Meldrum acids in the synthesis of natural products,
see: Ivanov, A. S. Chem. Soc. Rev. 2008, 37, 789–811.
9. Davidson, D.; Bernhard, S. A. J. Am. Chem. Soc. 1948, 70, 3426–3428.
10. Pihlaja, K.; Seilo, M. Acta Chem. Scand. 1968, 22, 3053–3062.
11. Meldrum stated that the treatment of 2,2-dimethyl-1,3-dioxin-4,6-dione with
excess of alkali ‘was found to cause disruption into acetone and malonic acid’.
However, no experimental data was provided, see: Meldrum, A. N. J. Chem. Soc.
1908, 93, 598–601.
reactions was determined to be a reversible, unimolecular loss of acetone to
provide the reactive intermediate. It was also concluded that since the de-
composition of 2,2,6-trimethyl-4H-1,3-dioxin-4-one is reversible, acetone is
a competitor with other trapping reagents.
39. Similar inhibition by the 1-indanone product was postulated but not experi-
mentally demonstrated.
40. The stability of acyl ketenes, generated through retro hetero-Diels–Alder of
Meldrum’s acids, is enhanced by increasing the size of the substituent at the 5-
position, see Ref. 18.
12. Pihlaja, K.; Seilo, M. Acta Chem. Scand. 1969, 23, 3003–3010.
13. Pihlaja, K.; Seilo, M. Finn. Chem. Lett. 1976, 123–124.
41. Inhibition of Yb(III)[C(SO2CF3)3]3 by Friedel–Crafts acylation product, see:
Barrett, A. G. M.; Bouloc, N.; Braddock, D. C.; Chadwick, D.; Henderson, D. A.
Synlett 2002, 1653–1656.
14. Zav’yalov, S. I. Izv. Akad. Nauk SSSR, Ser. Khim. 1961, 2185–2189.
15. Bihlmayer, G. A.; Schuster, P.; Polansky, O. E. Monatsh. Chem. 1966, 97, 145–149.
16. Ziegler, E.; Junek, H.; Kroboth, H. Monatsh. Chem. 1976, 107, 317–324.
17. Matoba, K.; Yamazaki, T. Chem. Pharm. Bull. 1983, 31, 2955–2956.
18. Sato, M.; Ban, H.; Kaneko, C. Tetrahedron Lett. 1997, 38, 6689–6692.
19. For a review, see: Sato, M.; Iwamoto, K. J. Synth. Org. Chem. Jpn. 1999, 57, 76–83.
20. The competitive reaction of an equimolar mixture of t-BuOH and EtOH with
Meldrum’s acid at reflux furnished a 1:3.8 ratio of the corresponding esters,
see: (a) Tararov, V. I.; Korostylev, A.; Ko¨nig, G.; Bo¨rner, A. Synth. Commun. 2006,
36, 187–191; The competitive reaction of an equimolar mixture of t-BuOH and
n-pentanol with acetylketene revealed a 8.2:1 selectivity favoring the less
sterically demanding 1ꢀ alcohol, see: (b) Birney, D. M.; Xu, X.; Ham, S.; Huang,
X. J. Org. Chem. 1997, 62, 7114–7120.
42. Examples of tautomerization of 1,3-dicarbonyl compounds by lanthanide tri-
´
´
flates, see: (a) Fillion, E.; Carret, S.; Mercier, L. G.; Trepanier, V. E. Org. Lett. 2008,
10, 437–440; (b) Zhang, J.; Blazecka, P. G.; Angell, P.; Lovdahl, M.; Curran, T. T.
Tetrahedron 2005, 61, 7807–7813; (c) Yang, D.; Li, J.-H.; Gao, Q.; Yan, Y.-L. Org. Lett.
2003, 5, 2869–2871; (d) Yang, D.; Ye, X.-Y.; Xu, M.; Pang, K.-W.; Cheung, K.-K.
J. Am. Chem. Soc. 2000, 122, 1658–1663; (e) Yang, D.; Ye, X.-Y.; Xu, M. J. Am. Chem.
Soc. 1999, 121, 5579–5580.
43. Ketene 35 was prepared independently from 2-phenyl propionyl chloride and
triethylamine, and its 1H and 13C NMR data acquired, see: Baigrie, L. M.; Seiklay,
H. R.; Tidwell, T. T. J. Am. Chem. Soc. 1985, 107, 5391–5396.
44. For mechanistic studies on Bi(OTf)3- and Bi(NTf2)3-catalyzed Friedel–Crafts
´
acylation with acid chlorides and anhydrides, see: (a) Picot, A.; Repichet, S.; Le
21. Hydrolysis by refluxing Meldrum’s acid derivatives in H2O/DMF, see: (a) Fillion,
E.; Wilsily, A. J. Am. Chem. Soc. 2006, 128, 2774–2775; (b) Kno¨pfel, T. F.; Carreira,
E. M. J. Am. Chem. Soc. 2003, 125, 6054–6055; H2O/pyridine, see: (c) Kadam, A.
J.; Desai, U. V.; Mane, R. B. J. Labelled Compd. Radiopharm. 1999, 42, 835–842;
H2O/2-pentanone, see: (d) Trost, B. M.; Brennan, M. K. Org. Lett. 2007, 9, 3961–
3964; Alcohols as solvent, see: (e) von Gassmann, P.; Hagmann, L.; Keller-
Schierlein, W.; Samain, D. Helv. Chim. Acta 1984, 67, 696–705; (f) Brooks, D. W.;
de Lee, N. C.; Peevey, R. Tetrahedron Lett. 1984, 25, 4623–4626; (g) Chorev, M.;
Rubini, E.; Gilon, C.; Wormser, U.; Selinger, Z. J. Med. Chem. 1983, 26, 129–135;
Reaction with phenols, see: (h) Junek, H.; Ziegler, E.; Herzog, U.; Kroboth, H.
Synthesis 1976, 332–334.
22. Formation of amide under neutral conditions has been also reported, see: (a)
Kno¨pfel, T. F.; Zarotti, P.; Ichikawa, T.; Carreira, E. M. J. Am. Chem. Soc. 2005, 127,
9682–9683; (b) Watanabe, T.; Kno¨pfel, T. F.; Carreira, E. M. Org. Lett. 2003, 5,
4557–4558; (c) Carotti, A.; Carrieri, A.; Cellamare, S.; Fanizzi, F. P.; Gavuzzo, E.;
Mazza, F. Biopolymers 2001, 60, 322–332.
Roux, C.; Dubac, J.; Roques, N. J. Fluorine Chem. 2002, 116, 129–134; (b) Re´pichet,
S.; Le Roux, C.; Dubac, J.; Desmurs, J.-R. Eur. J. Org. Chem. 1998, 2743–2746.
45. At fixed time intervals, aliquots were drawn from the reaction mixture and
immediately quenched in a triethylamine/methylene chloride solution. These
samples were then filtered through plugs of silica, and then injected into a GC
equipped with a flame ionization detector. The progression of the reaction was
judged by the simple percent conversion as determined by the direct compar-
ison of integration areas of the starting material with the product. Since no in-
ternal standard was used in the analysis, percent yields were not determined. All
other reaction and analysis conditions were identical between each Lewis acid.
46. For a large-scale synthesis, the removal of the acetone by-product by reduced
pressure could be expected to accelerate the reaction, but this was not explored
in this study.
47. Mg(NTf2)2 was prepared according to: Earle, M. J.; Hakala, U.; McAuley, B. J.;
Nieuwenhuyzen, M.; Ramani, A.; Seddon, K. R. Chem. Commun. 2004, 1368–
1369.
23. For reviews, see: (a) McNab, H. Aldrichimica Acta 2004, 37, 19–26; (b) Gaber, A.
E. M.; McNab, H. Synthesis 2001, 2059–2074.
48. Mg(OTf)2 and Mg(NTf2)2 has been shown to be effective catalysts for the
Friedel–Crafts acylation of 42, see Ref. 4a.
24. Ott, E. Justus Liebigs Ann. Chem. 1913, 401, 159–177.
49. Mathieu, B.; Ghosez, L. Tetrahedron Lett. 1997, 38, 5497–5500.
50. For examples of Friedel–Crafts acylation catalyzed by lanthanide bis(tri-
fluoromethanesulfonate)amides, see: (a) Kawamura, M.; Cui, D.-M.; Shimada, S.
Tetrahedron 2006, 62, 9201–9209; (b) Mikami, K.; Kotera, O.; Motoyama, Y.;
Sakaguchi, H.; Maruta, M. Synlett 1996, 171–172.
51. The Brønsted acidity of TfOH is superior to that of Tf2NH, see: (a) Guthrie, J. P.;
Can, J. Chem. 1978, 56, 2342–2354; (b) Foropoulos, J., Jr.; DesMarteau, D. D.
Inorg. Chem. 1984, 23, 3720–3723.
25. Brown, R. F. C.; Eastwood, F. W.; Harrington, K. J. Aust. J. Chem.1974, 27, 2373–2384.
26. Baxter, G. J.; Brown, R. F. C.; Eastwood, F. W.; Gatehouse, B. M.; Nesbit, M. C.
Aust. J. Chem. 1978, 31, 1757–1767.
27. For ab initio studies on the formation of ketenes via pyrolysis of Meldrum’s acid
derivatives, see: (a) George, L.; Wong, M. W.; Wentrup, C. Org. Biomol. Chem.
2007, 5, 1437–1441; (b) Matsui, H.; Zu¨ckerman, E. J.; Katagiri, N.; Kaneko, C.;
Ham, S.; Birney, D. M. J. Phys. Chem. A 1997, 101, 3936–3941.
28. Chopra, D.; Zhurov, V. V.; Zhurova, E. A.; Pinkerton, A. A. J. Org. Chem. 2009, 74,
2389–2395.
29. For the complexation of Me2AlCl with oximinosulfonate Meldrum’s acid de-
rivatives, see: Renslo, A. R.; Danheiser, R. L. J. Org. Chem. 1998, 63, 7840–7850.
30. (a) Couturier, D.; Rigo, B. J. Heterocycl. Chem. 1999, 36, 1073–1075; (b) Rigo, B.;
Fasseur, D.; Cauliez, P.; Couturier, D. Tetrahedron Lett. 1989, 30, 3073–3076.
52. Dumeunier, R.; Marko´ , I. E. Tetrahedron Lett. 2004, 45, 825–829.
53. (a) Balaban, T. S.; Unuta, C.; Gheorghiu, M. O.; Balaban, A. T. Tetrahedron Lett.
1985, 26, 4669–4672; (b) Brown, H. C.; Kanner, B. J. Am. Chem. Soc. 1966, 88,
986–992; (c) Brown, H. C.; Kanner, B. J. Am. Chem. Soc. 1953, 75, 3865.
54. For a discussion on the formation of protic acid from Yb(OTf)3 and H2O, see: (a)
Kobayashi, S. Synlett 1994, 689–701; See also: (b) Rosenfeld, D. C.; Shekhar, S.;
Takemiya, A.; Utsunomiya, M.; Hartwig, J. F. Org. Lett. 2006, 8, 4179–4182; For
the generation of protic acid from Sc(OTf)3 and MeOH, see: (c) Castellani, C. S.;
Perotti, A.; Scrivanti, M.; Vidari, G. Tetrahedron 2000, 56, 8161–8166; (d) Cas-
tellani, C. S.; Carugo, O.; Leopizzi, C.; Perotti, A.; Invernizzi, A. G.; Vidari, G.
Tetrahedron 1996, 52, 11045–11052.
31. For a review on
a-oxoketenes, see: Wentrup, C.; Heilmayer, W.; Kollenz, G.
Synthesis 1994, 1219–1248.
32. Substrate 22 should be slightly more nucleophilic than 24. For comparison,
toluene reacts 1.8 times faster than tert-butyl benzene in nitration reactions
with Bu4NNO3–TFAA in CH3NO2. In addition, attack at the ortho position is
hindered as the para/ortho ratio is 0.6 for toluene and 6.7 for tert-butyl ben-
zene, see: Masci, B. Tetrahedron 1989, 45, 2719–2730.
55. Quantitative evaluation of Lewis acid strength by 13C NMR, see: (a) Farcasiu, D.;
Lukinskas, P.; Ghenciu, A.; Martin, R. J. Mol. Catal. A: Chem. 1999, 137, 213–221;
(b) Farcasiu, D.; Stan, M. J. Chem. Soc., Perkin Trans. 2 1998, 1219–1222.
56. NMR studies of lanthanide triflimidates, see: Barbier-Beaudry, D.; Dormond, A.;
Duris, F.; Bernard, J. M.; Desmurs, J. R. J. Fluorine Chem. 2003, 121, 233–238.
57. It was observed that dried Sc(OTf)3, dried followed published procedures, but
rapidly weighed outside the dry-box readily absorbed water. Therefore, it
seemed essential to avoid exposure of the catalysts to moisture and preferably
weigh it under an inert and dry atmosphere. Despite these precautions, the
results described in this manuscript (see Scheme 4) suggest that a non-negli-
gible amount of H2O remains associated with Sc(OTf)3 despite the extreme
measures taken to dry the catalyst.
33. For a review on gem-disubstituted effect, see: Jung, M. E.; Piizzi, G. Chem. Rev.
2005, 105, 1735–1766.
34. Intramolecular arylation of ketenium ions, see: (a) Zhang, L.; Kozmin, S. A. J. Am.
Chem. Soc. 2004,126,10204–10205; Intramolecular Friedel–Crafts acylationwith
chromium–carbene complex-derived ketenes catalyzed by ZnCl2, see: (b) Bueno,
A. B.; Moser, W. H.; Hegedus, L. S. J. Org. Chem. 1998, 63, 1462–1466.
35. The plot of the log(data) versus time was linear.
36. Mayr, H.; Kempf, H.; Ofial, A. R. Acc. Chem. Res. 2003, 36, 66–77.
37. March, J. Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, 4th
ed.; Wiley-Interscience: New York, NY, 1992; pp 585–586.
38. Related studies of the thermal decomposition of 2,2,6-trimethyl-4H-1,3-dioxin-
4-one have been reported with both kinetics experiments, see: (a) Clemens, R. J.
J. Am. Chem. Soc. 1989, 111, 2186–2193; and ab initio calculations, see: (b) Birney,
D. M.; Wagenseller, P. E. J. Am. Chem. Soc. 1994, 116, 6262–6270; This compound
decomposes to acetylketene and reacts just as described for acyl ketenes. These
dioxinones do not require the initial enolization step, but still involve a retro
hetero-Diels–Alder and acylation step. The rate-determining step in these
58. For Yb(OTf)3-catalyzed aldol reactions with formaldehyde in water, the for-
mation of TfOH was negligible and found not to be an active catalytic species,
see: Kobayashi, S.; Hachiya, I. J. Org. Chem. 1994, 59, 3590–3596.
59. Crooy, P.; De Neys, R.; Eliaers, J.; Livegns, R.; Simonet, G.; Vandevelde, J. Bull.
Soc. Chim. Belg. 1997, 86, 995–1002.
60. Chen, B.-C.; Lue, P. Org. Prep. Proced. Int. 1992, 24, 185–188.