Acylation mechanisms of DMSO/[D6]DMSO with Di-tert-butylketene and its congeners

Article abstract of DOI:10.1002/ejoc.201100936

Dimethyl sulfoxide (DMSO) and tBu₂C=C=O in diglyme require heating to about 150 °C to furnish the Pummerer-type product tBu ₂CHCO₂CH₂SCH₃ through a novel mechanistic variant. The "ester enolate" tBu₂C=C(O ^-)-O-S⁺(CH₃)₂ arising through the reversible addition of DMSO (step 1) to C-1 of tBu₂C=C=O must be trapped through protonation (step 2) at C-2 by a carboxylic acid catalyst to form tBu₂CH-C(=O)-O-S⁺(CH₃)₂ so that the reaction can proceed. The ensuing cleavage (step 3) of the O-S bond and one of the C-H bonds in the-S(CH₃)₂ group (E2 elimination, no ylide intermediate) results in the formation of tBu₂CHCO ₂^- and H₃CS-CH₂⁺, whose combination (step 4) generates the final product. With a mixture of DMSO and [D₆]DMSO competing for tBu₂C=C=O in diglyme, the small value of the kinetic H/D isotope effect (KIE) k_H/k_D = 1.26 at 150 °C indicates that the cleavage of the C-H/C-D bonds (step 3) does not occur in the transition state with the highest free enthalpy. Therefore, the practically isotope-independent steps 1 and 2 determine the overall rate. The alternative slow initial protonation at C-2 of tBu₂C=C=O generating the acylium cation tBu₂CHC≡O⁺ can be excluded. Preparatory studies were undertaken to compare the mechanistic behavior of tBu₂C=C=O with that of two related acylating agents: (i) The anhydride (tBu₂CHCO)₂O affords the same Pummerer-type product more slowly, again with an unexpectedly small KIE of 1.24 at 150 °C, which indicates that the overall rate is limited here by the almost isotope-independent initial O-acylation of DMSO in the addition/elimination (AE) mechanism. (ii) The acyl chloride tBu₂CHCOCl affords ClCH ₂SCH₃ through a more common mechanistic variant involving neither the ketene nor the acylium cation tBu₂CHC≡O ⁺: The modestly enhanced k_H/k_D value of 2.4 at 55 °C shows that the C-H/C-D bond fissions contribute to the overall rate in cooperation with the retarded initial O-acylation. Deuterium labeling was quantified through ¹H and ¹³C NMR integrations of deuterium-shifted signals.

Full text of DOI:10.1002/ejoc.201100936

FULL PAPER
DOI: 10.1002/ejoc.201100936
Acylation Mechanisms of DMSO/[D6]DMSO with Di-tert-butylketene and Its
Congeners[‡]
Rudolf Knorr*[a]
Dedicated to the memory of Professor Hans Behringer
Keywords: Kinetics / Isotope effects / Ketenes / Reaction mechanisms / Sulfoxides / NMR spectroscopy
Dimethyl sulfoxide (DMSO) and tBu2C=C=O in diglyme re- the overall rate. The alternative slow initial protonation at C-
quire heating to about 150 °C to furnish the Pummerer-type
product tBu2CHCO2CH2SCH3 through a novel mechanistic
variant. The “ester enolate” tBu2C=C(O–)–O–S+(CH3)2 aris-
ing through the reversible addition of DMSO (step 1) to C-1
of tBu2C=C=O must be trapped through protonation (step 2)
at C-2 by a carboxylic acid catalyst to form tBu2CH–C(=O)–
O–S+(CH3)2 so that the reaction can proceed. The ensuing
cleavage (step 3) of the O–S bond and one of the C–H bonds
in the –S(CH3)2 group (E2 elimination, no ylide intermediate)
2 of tBu2C=C=O generating the acylium cation tBu2CHCϵO+
can be excluded. Preparatory studies were undertaken to
compare the mechanistic behavior of tBu2C=C=O with that
of two related acylating agents: (i) The anhydride
(tBu2CHCO)2O affords the same Pummerer-type product
more slowly, again with an unexpectedly small KIE of 1.24 at
150 °C, which indicates that the overall rate is limited here
by the almost isotope-independent initial O-acylation of
DMSO in the addition/elimination (AE) mechanism. (ii) The
acyl chloride tBu2CHCOCl affords ClCH2SCH3 through a
more common mechanistic variant involving neither the ket-
ene nor the acylium cation tBu2CHCϵO+: The modestly en-
hanced kH/kD value of 2.4 at 55 °C shows that the C–H/C–D
bond fissions contribute to the overall rate in cooperation
with the retarded initial O-acylation. Deuterium labeling was
quantified through 1H and 13C NMR integrations of deute-
rium-shifted signals.
–
+
results in the formation of tBu2CHCO2 and H3CS–CH2
whose combination (step 4) generates the final product. With
mixture of DMSO and [D6]DMSO competing for
,
a
tBu2C=C=O in diglyme, the small value of the kinetic H/D
isotope effect (KIE) kH/kD = 1.26 at 150 °C indicates that the
cleavage of the C–H/C–D bonds (step 3) does not occur in
the transition state with the highest free enthalpy. Therefore,
the practically isotope-independent steps 1 and 2 determine
Introduction
Di-tert-butylketene (tBu2C=C=O, 2 in Scheme 1) is a ki-
netically sluggish electrophile,[1–4] because the electronic
structure of its carbonyl center (C-1) demands the ap-
proaching nucleophile to enter from close to the ketene sp2
plane and thus to collide with one of the two bulky tert-
butyl (tBu) groups. Therefore, only very potent nucleophiles
such as phenyllithium,[5] tBuLi,[6,7] Me3SiLi,[8] or anhydrous
KOH in dimethyl sulfoxide (DMSO) (slow at 55 °C[9]), but
not tBuMgCl,[9] are able to add to 2 under mild condi-
tions.[1,2] With a convenient preparation[9] of 2 through the
elimination of HCl from tBu2CClCH=O (1) in hand, it ap-
peared attractive to explore whether DMSO, which is a
highly polar and sterically sensitive[10] nucleophile,[11] would
be able to add to 2.
Scheme 1. Preparation and consumption of the ketene 2.
It may be recalled[12] that tBu2C=C=O (2) does not di-
merize and is completely stable in a closed vessel. Almost
unavoidable trace amounts of adventitious water in even
carefully dried DMSO solvent would add slowly (at above
95 °C)[9] to 2, producing the acid tBu2CHCO2H (3) which
can slowly consume a second equivalent of 2 with forma-
tion of the anhydride (tBu2CHCO)2O (5). Therefore, the
present study involved reduced concentrations (0.4–3.8 m)
of DMSO in anhydrous (H3COCH2CH2)O (diglyme),
which is stable[9] against 2 at 150 °C for many hours in the
[‡] Sterically Congested Molecules, 23. Part 22: Ref.[9]
[a] Department Chemie der Ludwig-Maximilians-Universität
Butenandtstrasse 5-13 (Haus F), 81377 München, Germany
E-mail: rhk@cup.uni-muenchen.de
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201100936.
Eur. J. Org. Chem. 2011, 6335–6342
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6335

R. Knorr
FULL PAPER
absence of pyridine. Under these conditions, both 2 and 5 lecular competitive KIE indicates that the C–H/C–D bond
afforded the hitherto unknown (methylthio)methyl ester 4, fission (step 2 in Scheme 2) did not occur in the energeti-
which is the typical product expected from a Pummerer cally highest transition state, which is then reasonably as-
acylation[13] of DMSO by an anhydride such as 5. It will signed here to step 1 in which DMSO is required to attack
now be shown that these two modes of formation of 4 occur one of the strongly shielded acyl groups of 5. This change-
through novel mechanistic variants.
over of the rate-limiting step establishes a novel mechanistic
variant of the Pummerer acylation in which the overall re-
action rates of DMSO and [D6]DMSO (but not the faster
C–H/C–D bond fissions) are almost equally slow. The large
amount of acid 3 observed in situ is informative: If the
Results and Discussion
A. DMSO with the Anhydride (tBu2CHCO)2O (5)
–
sulfonium ylide tBu2CHCO2–S+(CD3)CD2 had been
formed[13,20] from [D6]6, it would be expected to be rapidly
protonated by 3 to give tBu2CHCO2–S+(CD3)CHD2 and
finally [D4]4 or a less highly deuteriated product; but this
was never detected. Consequently, the sulfonium ylide is
formed in step 2 either irreversibly[21] (faster decay to afford
[D5]8) or not at all (faster E2 elimination of [D6]6). Further-
more, deuterium was not incorporated into the α-position
of the di-tert-butyl parts of 4 and [D5]4 at 150 °C. It is note-
worthy that this reaction temperature is higher by at least
70 °C than in the case of acetic anhydride reacting with
DMSO.[22] This suggests that step 1 in Scheme 2 is, in fact,
slowed down by steric repulsion in 5, so that the C–H/C–D
bond fission in step 2 can no longer co-determine the over-
all reaction rate.
Scheme 2 displays a simplified mechanistic version of the
Pummerer acylation of sulfoxides by anhydrides,[13,14] for-
mulated for 5 with [D6]DMSO. The first intermediate, [D6]-
6, may transfer one of its deuterons (step 2) to the carboxyl-
ate 7 in an intermolecular E2 elimination[14] reaction, form-
ing the perdeuteriated thionium cation [D5]8 (2-thionia-
propene). [D5]8 and 7 are believed to recombine in the final
step 3 with production of the ester [D5]4. A large primary
H/D kinetic isotope effect (KIE, up to a kH/kD ratio of ca.
4 at 150 °C)[15] would be expected for the E2 elimination in
step 2, as supported by known experimental kH/kD ratios
for the acetylation of other sulfoxides by acetic anhydride
at 120 °C (kH/kD = 2.9)[16] or at 76 °C (supposed kH/kD
Ն
4.5).[14a] These values indicate that the C–H (and C–D)
bond fission (step 2) contributes to limiting the overall reac-
tion rate in these cases (but not in the case of 5, as detailed
below). The following competition experiment compares
this usual Pummerer scenario with the kinetic behavior of
anhydride 5.
B. DMSO with the Acyl Chloride tBu2CHCOCl (9)
The closely related acyl chloride 9[9] is now presented as
a contrasting example that tends toward a larger primary
KIE. The acylation of DMSO by 9 will again give 6 in
step 1 of Scheme 3. The mutual interconversion of 6 and
(H3C)2S–Cl+ (10) in step 2 may be visualized[23] as involving
a sulfurane with a tetracoordinate sulfur atom; the process
would be terminated above 0 °C[24a] presumably through an
E2-type elimination (step 3) within the ion pair[23] of 10 +
7, producing the acid 3 together with the thionium chloride
8 and finally (step 4) chloro(methylthio)methane[24a,24b]
(11). Steps 1 and 2 were formulated as reversible processes
in analogy with the main path suggested[23] for the rapid
Scheme 2. Pummerer acylation of [D6]DMSO with anhydride 5.
A diglyme (0.59 mL) solution of 5 (0.282 mmol), DMSO acetylation of DMSO by acetyl chloride, where an experi-
(0.500 mmol), and [D6]DMSO (2.11 mmol) in a tightly mental kH/kD estimate of approximately 5 at 30 °C had
closed NMR tube was heated at 150 °C for 15 h until 5 had been assigned to the C–H/C–D bond-breaking step 3, in
completely disappeared.[17] The in situ 13C NMR analysis rough accord with the values cited above[14a,16] for acetic
revealed the formation of 0.104 mmol (37%) of the two es- anhydride. The following competition experiment is
ters 4 and [D5]4, along with roughly 0.4 mmol of acid 3. thought to slightly modify that mechanistic picture for the
The two side-products (H3C)2S (0.023 mmol) and (D3C)2S present case of 9.
(0.086 mmol) arose through deoxidation reactions of the
A diglyme (0.55 mL) solution of the acyl chloride 9
two dimethyl sulfoxides, but the detached oxygen atoms (0.35 mmol), DMSO (0.528 mmol), and [D6]DMSO
were not transferred to the sulfur atoms of 4 or [D5]4, al- (2.100 mmol) in a tightly closed NMR tube was heated at
though such a transfer would be possible[18] at 150 °C.
55 °C until 9 had vanished (within 6.5 h). The product dis-
The differential diagnosis had to be carried out with the tribution was measured in situ through 13C NMR integra-
isolated product mixture in a more concentrated solution. tions that revealed the following millimolar quantities:
After separation from the deuterium-free acid 3 (66 mg, 0.308 of acid 3, 0.048 of [α-D]3, 0.064 of 11 (18%), 0.110
68%), the crude esters (33 mg) were obtained in a millimo- of [D5]11 (31%), and 0.021 of 12 (but [D5]12 was not deter-
lar ratio 4/[D5]4 of 0.023:0.078 (36% yield), which corre- mined). The product ratio of 11/[D5]11 translates[19] into a
sponds[19] to a kH/kD ratio of 1.24. This very small intermo- rate ratio kH/kD of 2.4, which amounts to a modest fraction
6336
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 6335–6342

Mechanisms of Ketene Acylation
analysis in situ and is, in fact, unable to react with DMSO
at 55 °C, as will be detailed in Section C. (ii) The acylium
cation 13 (Scheme 4) is stable at –60 °C but decays “upon
warming”[28] with loss of carbon monoxide to give the cat-
ions tBu+ (20) and EtMe2C+ (18). No simple chemical
means of detecting 13 are apparently known; so it was
planned to use benzene to trap the cations shown in
Scheme 4, because all four conceivable products (14, 15, 17,
and 21) of these Friedel–Crafts reactions had already been
prepared by other routes. For an initial setup of this novel
test system, AlCl3 (0.5 equiv. or more) was added to a ben-
zene solution of the acyl chloride 9, and the reaction was
monitored by 1H NMR analysis at approximately 20 °C.
Under these conditions, neither 13 nor the expected acyl-
ation product 14[5] could be detected during the slow con-
Scheme 3. Pummerer reaction of DMSO with the acyl chloride 9.
of the maximum KIE (ca. 6.0)[15] at 55 °C. This result ap- sumption of 9, which indicates that the decarbonylation of
pears to indicate that the C–H/C–D bond fissions in step 3 13 to generate 16 is faster[29,30] than the acylation of ben-
contribute perceptively (though weakly) to the overall rate, zene. The subsequent rapid decay of 16 (no formation of
so that the preceding steps 1 and 2 should be at least par- 17[31]) to give the expected[28] tBu+ (20) was recognized
tially reversible. This would mean that, in contrast to the through the detection of 21[32] as the main product. Simul-
acetylation mentioned above,[23] the present acylation of taneous protonation[28] of the second fragment 19 led to
DMSO with 9 in diglyme, [D8]dioxane, or [D8]tetra- formation of the tert-amyl cation 18, which was trapped as
hydrofuran (THF) (all comparably slow at 55 °C) is re- 15.[33] In the absence of the Lewis acid catalyst AlCl3, how-
tarded by an elevation of the free-enthalpy barrier for step 1 ever, 9 was stable in benzene solution at 70 °C for at least
to a level similar to that of the isotope-dependent step 3.
5 h. Likewise, 9 in the more polar diglyme (as a 70 vol.-%
None of the intermediates (6, 8, or 10) in Scheme 3 could component of the solvent mixture) did not react with ben-
be detected in situ (1H and 13C NMR analysis), but the zene (23% by volume, 12 equiv.) at 70 °C during 6 h; subse-
development of HCl in parallel with the emergence of 11 quent addition of DMSO (3.5 equiv.) led to the formation
1
and acid 3 was recognized through an increasing H NMR of ClCH2SCH3 (11) and acid 3 as the main products (but
shift of free DMSO (δ = 2.48 ppm) toward the value for neither 15 nor 21) at 55 °C within 7 h. Thus, the results
DMSO·HCl[24a] (δ = 2.87 ppm). The much slower subse- obtained with the test system outlined in Scheme 4 clearly
quent conversion of 11 into the sulfoxide 12[24b] is known[18] exclude a spontaneous ionization of 9 in step 1 of Scheme 3
to be mediated by HCl. Hydrolysis of some part of 9 by to generate 6 via the acylium ion 13; this leaves the AE
adventitious moisture is a probable source of HCl, and re- mechanism, as claimed above.
sidual HOD in [D6]DMSO should similarly give rise to
DCl. The occurrence of some [α-D]3 mentioned above can
be understood through a very slow reaction of DCl (first
column of Scheme 3) with acid 3 (as formed in step 3), in
analogy with the labeling technique reported[25,26] for a re-
lated system. During storage of the mixture for some weeks
at ambient temperature without workup, the sulfoxide 12
vanished slowly, with appearance[27] of H3CS–SO2-
CH3.[24a,24c] The ester 4 is probably not generated under the
reaction conditions shown in Scheme 3 from the intermedi-
ates 6 or 8 (in contrast to Scheme 2), presumably due to a
shortage of bases such as 7 in this HCl-producing system;
Scheme 4. Decarbonylation of the acylium cation 13 and trapping
this notion gained support in the following way: A sample
(5 mg) of pure ester 4 was added to this aged solution and
required significantly longer than 4 h at 55 °C for its cleav-
age by HCl to give 11 (2nd column of Scheme 3). Therefore,
with benzene.
even modest amounts of 4 would have been detected by 1H C. DMSO with the Ketene tBu2C=C=O (2)
NMR analysis in situ, which was, however, not the case.
Despite steric repulsion, the conversion of 9 into 6 (step 1 Scheme 5 illustrates the competition between DMSO
of Scheme 3) is expected to occur through nucleophilic ad- (0.527 mmol) and [D6]DMSO (2.100 mmol) for the ketene 2
dition of DMSO to 9 and subsequent elimination of a chlo- in diglyme (0.55 mL) at 150 °C. Steps 1H and 1D generate
ride anion (the usual AE mechanism), as judged on the ba- the zwitterionic “ester enolates” 22 and [D6]22, respectively,
sis that the following two alternative modes can be dis- which will be trapped by H+ or D+ (step 2) to provide the
missed: (i) The ketene 2 could not be detected by NMR series 6–[D7]6 in line 2 of Scheme 5. Such trapping can be
Eur. J. Org. Chem. 2011, 6335–6342
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6337

R. Knorr
FULL PAPER
expected to be irreversible, because the isolated product ester
The low value of the KIE (kH/kD close to 1) demon-
4 does not incorporate deuterium on heating at 110 °C with strates that the C–H/C–D bond-breaking step 3 is not in-
an excess of solid KOH in [D6]DMSO for 2 h (continued volved in the overall reaction rate; thus, steps 1 and 2 re-
heating of this mixture to 150 °C for 5.5 h cleaved 4 and main as candidates for this role.[21] In fact, significant rate
afforded the non-deuteriated potassium salt of 3). Thus, the enhancements were observed with increasing concentra-
possibility of α-H/α-D exchange reactions at di-tert-butylace- tions of both DMSO (step 1) and acid 3 (step 2).[34]
A
tyl groups (reversal of step 2) can be dismissed, as similarly strong acceleration (first t1/2 ≈ 20 min at 150 °C) was also
observed[9] with the pivaloin ester of 3. In step 3, a base such observed when the reaction was performed with benzoic
as the carboxylate 7 is required (also in Scheme 2) for the β- acid (0.7 equiv.) in addition to 3 (as generated from 2 by
elimination reaction of intermediate 6 and its isotopologues moisture), wherefrom a 65:35 mixture of 4 and the
(Scheme 5), which will furnish the series of four ion pairs 7 + known[35] ester PhCO2CH2SCH3 was obtained, the latter
8, the final collapse of which (step 4) should be a fast process probably arising through either benzoylation of DMSO by
+
affording the product series 4–[D6]4.
a mixed anhydride or trapping of H3CSCH2 (8) by
–
The experimental progress was halted when ketene 2 was PhCO2 . Thus, the practically isotope-independent steps 1
almost completely consumed (ca. 6 h), so that the slower and 2 co-determine the overall reaction rate; it follows that
acylation of DMSO by the side-product (anhydride 5)[17] the transition state of step 2 must be higher in enthalpy
could not seriously interfere with the measurements. An than that of the isotope-dependent step 3, so that step 2 be-
alkaline workup procedure separated the free acids 3 and comes irreversible, as expected above.
[α-D]3 (37Ϯ2% α-D) from the four esters 4 (33% α-D). A
In a corroboration of the latter results, the acylation of
conspicuously different H/D ratio of 24:76 was found for DMSO could be completely suppressed through initial ad-
the pairs of esters (4 + [α-D]4)/([D5]4 + [D6]4) through dition of KOtBu (0.3–1.2 equiv.), which neutralized any ac-
NMR integrations of the OCH2SCH3/OCD2SCD3 parts ids in the reaction vessel. Ketene 2 remained stable in this
(see Table S1 in the Supporting Information). This ratio re- non-acidic environment at 149 °C for more than 16 h and
sembles that of the starting mixture of DMSO/[D6]DMSO also for a further 10 h after subsequent addition of suf-
(20:80) and yields[19] a kH/kD value of 1.26 for the overall ficient amounts (in excess over KOtBu) first of tert-butyl
reaction. A more detailed analysis of the deuterium par- alcohol and then of 2,6-di-tert-butylphenol. This inhibition
titioning afforded a ratio of 16:8:51:25 for 4/[α-D]4/[D5]4/ was terminated through the final addition of acid 3
[D6]4 (bottom line of Scheme 5); this analysis was assisted (2 equiv.), which promoted the production of ester 4 to-
by results obtained by mass spectrometry, as presented in gether with a small amount of anhydride 5. The initial inhi-
the Experimental Section.
bition of step 2 demonstrates that conceivable alternative
Scheme 5. Competition between DMSO and [D6]DMSO for ketene 2.
6338
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 6335–6342

Mechanisms of Ketene Acylation
steps, such as an intramolecular or bimolecular proton to rate-limiting general acid catalysis in pure water at 25 °C,
transfer from S-methyl to C-α either within or between mo- which convinced the authors[25] of “a modest steric barrier”
lecules of the “ester enolate” 22 (implying a transient sulf- for the approach of H3O+ from regions perpendicular to
onium ylide character), do not occur under the reaction the ketene sp2 plane. This raises the question of whether the
conditions.[20] It also demonstrates the absence of proton results described in Section C might also be compatible
transfer from DMSO to C-α of 22, in agreement with the with the alternative mechanism depicted in Scheme 6. In
conclusion that such a transfer can be excluded, because this approach, initial protonation of 2 by the acid 3 in
products bearing –SCH2D or –SCHD2 functions were not step 1Ј, followed by addition of DMSO to 13 in step 2Ј
observed (as verified in Figure S1 of the Supporting Infor- would result in the formation of the same intermediate 6
mation) in the above DMSO/[D6]DMSO competition ex- as shown in Scheme 5. Considering the evidence[25,44] of a
periments with ketene 2 at 150 °C after many hours.
decreasing proton-transfer efficiency to C-2 of ketene 2 in
Similar mechanistic traits were observed when the small a less polar environment than pure water, a solvent such as
nucleophile methanol was used in place of DMSO. In the diglyme, used in the present work, might not appear to be
presence of solid NaOCH3 (1.4 equiv.) in diglyme, ketene 2 of great promise for promoting the generation of 13 by the
remained totally unreactive towards methanol (11 equiv.) at very weak[45] acid 3. Instead, the concomitant[25] consump-
110 °C for more than 2 h, whereas a parallel run with acid tion of 2 by formic or acetic acid may be viewed as being
3 (1.4 equiv.) in lieu of NaOCH3 at 110 °C was complete caused by an initial nucleophilic addition[46] of the acid at
within 125 min to produce the known[36] methyl ester, C-1 of 2, combined with a proton transfer to C-2 producing
tBu2CHCO2CH3 (yield 60%), together with anhydride 5 anhydrides; under this supposition, formation of an anhy-
(23%) (this methyl ester could not be prepared through a dride such as 5 cannot serve as unequivocal evidence for
direct esterification of 3 with methanol in diglyme at 110 °C the involvement of an acylium intermediate such as 13. In
over more than 10 h). These results are compatible with ini- fact, attempts to detect 13, as described in Section B, with
tial nucleophilic addition of methanol to C-1 of 2, followed benzene (11 equiv.) in a diglyme solution of 2 and acid 3
by C-2 protonation[37–39] by acid 3, in analogy with the ad- (1.3 equiv.) at 147 °C (ca. 1 h) furnished only 5, and no
dition of methanol to dimethylketene at 25 °C as cata- trace of the aromatic products (15 and 21 in the right col-
lyzed[40] by weak carboxylic acids. The results are, however, umn of Scheme 6) derived from the carbocations 18 and 20
at variance with the addition reactions of alcohols to less as generated through decarbonylation[28] of 13. Because the
encumbered ketenes under acid-free conditions, in which decarbonylation was fast[30] at room temperature (and
the alcohol concentrations were included in the rate equa- should be much faster at 147 °C), one has to conclude that
tions as 2nd or higher reaction orders.[37,38,41] The higher step 1Ј does not occur in diglyme with benzene in the ab-
reaction temperatures (110 °C or more) and the inability of sence of DMSO.
ketene 2 to add methanol in diglyme without acid catalysis
illustrate the mechanistic consequences of steric shielding in
2. This inability disappeared upon replacement of diglyme
by the more polar[42] solvent DMSO: While ketene 2 still
did not react with methanol (36 equiv., 7 vol.-%) in DMSO
at 23 °C within at least 2 h, the subsequent addition of solid
NaOCH3 (which is barely soluble) started the production
of tBu2CHCO2CH3,[36] presumably via the ester enolate
tBu2C=C(O–)OCH3 and its subsequent protonation by
methanol, so that the reaction appears to be catalyzed by
methoxide. This clean conversion was complete after 5 h,
Scheme 6. Mechanistic alternatives of the Pummerer acylation of
DMSO with ketene 2 (steps 1–4 are depicted in Scheme 5).
which indicates a temperature reduction of much more than
87 °C in comparison with the total kinetic inactivity of Na-
OCH3 in diglyme at 110 °C as described above. An even
The possibility remains that the highly polar co-solvent
stronger acceleration was discovered with the homogeneous DMSO,[42] even when used in a smaller amount than dis-
solution of KOCH3 (18 equiv., in place of NaOCH3), ketene cussed in Section C, might still promote the proton-transfer
2, and methanol (18 equiv.) in DMSO at 23 °C, which again step 1Ј. In this case, the intermediacy of 13 could perhaps
produced only the methyl ester[36] in less than 48 min.[43] be unveiled through its monomolecular decarbonylation re-
Thus, the addition of methanol to 2 is acid-catalyzed in action (Scheme 6), especially if the competing bimolecular
diglyme solution at 110 °C, but base-catalyzed in DMSO at trapping of 13 (step 2Ј) could be repressed by employing a
23 °C.
deliberately small concentration of DMSO. Thus, ketene 2
in diglyme containing benzene (21 equiv.), DMSO
(1.1 equiv., 0.3% by volume), and acid 3 (up to 0.6 equiv.)
were heated at 149 °C for 11 h, affording a trace of the
Pummerer ester 4 along with mainly the anhydride 5 but,
D. DMSO with the Acylium Cation tBu2CH–CϵO+ (13)?
The hydration of ketene 2 through protonation at C-2 to again, none of the aromatic compounds derived from
generate the intermediate 13 (step 1Ј in Scheme 6) is subject carbocations 18 or 20. Hence, the conjectured step 2Ј must
Eur. J. Org. Chem. 2011, 6335–6342
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6339

R. Knorr
FULL PAPER
be faster than the rapid[30] decarbonylation of 13, under the
proviso that step 1Ј takes place. If so, the consumption of
13 by DMSO in a faster step 2Ј would leave the slow step 1Ј
as the sole[47] rate-limiting step in a very slow overall reac-
tion. As a consequence, the employed concentration of
DMSO would not influence the overall reaction rate (ki-
netic reaction order zero), which contradicts the experimen-
tal observations described in Section C; this contradiction
disproves the above proviso and, with it, the acylium
DMSO (0.300 mL, 4.22 mmol), [D12]cyclohexane (0.03 mL), and a
trace of tetramethylsilane (TMS). The tightly closed tube was
heated at 148 °C for 15 h with occasional NMR analysis. The ke-
tene 2 disappeared within 10 h, rapidly generating some anhydride
5 (from 3) and then the ester 4 (1:2.6); the major part of 5 was
also transformed more slowly into 4 and acid 3. The mixture was
dissolved in Et2O and shaken with aqueous 1 m NaOH. Acidifica-
tion of the aqueous layer afforded acid 3 (27 mg, 21%). The organic
layer was washed until neutral, dried with Na2SO4, and distilled at
105–115 °C (bath temp.)/3 mbar to yield the liquid ester 4 (101 mg,
1
mechanism, at least as the main pathway in Scheme 6. In- 57%). H NMR (CDCl3, 400 MHz): δ = 1.11 (s, 18 H), 2.22 (s, 1
1
H, tert-CH), 2.24 (s, 3 H, SCH3), 5.10 (s, 2 H, OCH2S) ppm. H
NMR (diglyme + DMSO): δ = 1.10, 2.18, 2.20, 5.11 ppm. 13C
NMR (CDCl3, 100.6 MHz): δ = 15.62 (SCH3), 30.70 [2ϫ C(CH3)
3], 35.08 [2ϫ C(CH3)3], 64.42 (br., tert-CH), 67.31 (sharp, OCH2S),
174.22 (C=O) ppm; assigned through 13C/1H correlation (HSQC)
spectroscopy. 13C NMR (100.6 MHz, diglyme + DMSO): δ =
15.20, 30.75, 35.28, 64.3 (br.), 67.3, 173.77 ppm. IR (film on dia-
stead, the results of Section C remain consistent with the
rate-determining steps 1 and 2 in Schemes 6 and 5.
Conclusions
The results presented here are compatible with the usu-
ally assumed mechanistic variants of Pummerer acylations
by anhydrides and by acyl chlorides, except for the lower
than usual kinetic isotope effects kH/kD, which indicate that
the role of the rate-determining steps has shifted toward
the starting (isotope-independent) events. A second novel
variant concerns an initiating DMSO addition to the ketene
tBu2C=C=O, followed by an intermolecular proton-trans-
fer step from carboxylic acids that contributes to the overall
reaction rate, whereas the subsequent steps do not contrib-
ute. The latter mechanistic variant seems to be without pre-
cedence,[4,48,49] because the previously reported reactions of
two other stable ketenes with sulfoxides were apparently not
conducted in this way: (i) The initiating zwitterionic inter-
mediate formulated[50] for Cl2C=C=O reacting with α,β-un-
saturated sulfoxides[48,51] is thought to rearrange through a
(formal) [3,3]-sigmatropic migration; (ii) Ph2C=C=O (re-
ported[52] to be unsusceptible to protonation at C-2) and
acidified DMSO follow a different route to produce benzilic
acid;[53] the corresponding product tBu2C(OH)CO2H ex-
pected from ketene 2 is known[54] but could not be found
in the present investigations.
Neither methanol nor DMSO furnish stable adducts
with the ketene tBu2C=C=O in acid-free diglyme solution,
in contrast to the alcohol-catalyzed alcohol addition reac-
tions to some less encumbered ketenes reported in the lit-
erature. This remarkable dependence on carboxylic acid ca-
talysis of acylations of DMSO or methanol by tBu2C=C=O
in diglyme is not due to C-2 protonation (acylium route) as
the first step. The base-catalyzed acylation of methanol or
pivaloin by tBu2C=C=O is possible in DMSO solution.
mond, ATR): ν = 2958, 2910, 2873, 1731 (s), 1476, 1370, 1349,
˜
1220, 1109 (s), 1047, 969 cm–1. MS (EI, 70 eV): m/z (%) = 232 (18)
[M+], 155 (85) [tBu2CHCO+], 57 (100) [tBu+]. C12H24O2S (232.39):
calcd. C 62.02, H 10.41, S 13.80; found C 62.30, H 10.37, S 12.98.
Chloro(methylthio)methane (11):[24a,24b] 1H NMR (diglyme,
400 MHz): δ = 2.253 (s, 3 H, CH3), 4.88 (s, 2 H, CH2) ppm. 13C
NMR (diglyme, 100.6 MHz): δ = 14.66 (CH3), 52.90 (CH2) ppm.
1
(1,1-Dimethylpropyl)benzene (15): H NMR (CDCl3, 400 MHz):[33]
δ = 0.68 (t, 3J = 7.4 Hz, 3 H, CH3-3), 1.28 (s, 6 H, 2ϫ 1-CH3),
3
1.65 (q, J = 7.4 Hz, 2 H, CH2), 7.16 (disturbed t, ca. 1 H, p-H),
7.29 (disturbed m, ca. 2 H, 2ϫ m-H), 7.32 (disturbed m, ca. 2 H,
2ϫ o-H) ppm. 13C NMR (CDCl3, 100.6 MHz): δ = 9.12 (CH3-3),
28.44 (2ϫ 1-CH3), 36.88 (CH2), 37.87 (C-1), 125.41 (p-C), 125.93
(2ϫ o-C), 127.96 (2ϫ m-C), 149.44 (i-C) ppm; assigned through
HSQC spectroscopy.
1
tert-Butylbenzene (21): H NMR (CDCl3, 400 MHz):[32a] δ = 1.32
(s, 9 H), 7.16 (disturbed t, ca. 1 H, p-H), 7.30 (disturbed m, ca. 2
3
H, 2ϫ m-H), 7.39 (dm, J = 7.5 Hz, ca. 2 H, 2ϫ o-H) ppm. 13C
NMR (CDCl3, 100.6 MHz):[32b]
δ = 31.36 [C(CH3)3], 34.65
[C(CH3)3], 125.24 (2ϫ o-C), 125.29 (p-C), 128.05 (2ϫ m-C), 151.09
(i-C) ppm; assigned through HSQC spectroscopy.
Analytical Procedures for Deuterium Quantification: The most di-
1
rect analytical method would employ quantitative H NMR spec-
troscopy to determine the degree of deuterium labeling. With refer-
ence to the 18-proton integrals of the tBu2 singlets of compounds
3, 4, and 11, integral deficiencies between the NMR signals could
indicate the H/D ratio. However, this simple technique would only
be reliable for those substances for which integrals are not dis-
turbed by impurities. Therefore, a different technique had to be
chosen for the small batches of unpurified samples (prepared in
situ, or obtained as the crude material after workup) as investigated
in the present work. In this situation, 13C NMR integration is the
method of choice in view of its large spectral dispersion (less prob-
able disturbances) and of the various deuterium-induced 13C NMR
chemical shifts Δ = δ(isotopologue) – δ(unlabeled) that offer mul-
tiple chances of integrating separate signals for the labeled and un-
labeled components.
As depicted in Scheme 7, several large 13C NMR upfield shifts Δ,
expressed in units of ppb (1:109), could be used for the integrations
of signal pairs as induced by the CH/CD, CH2/CD2, or CH3/CD3
pairs in the following mixtures: 3/[α-D]3 (63:37, from ketene 2), 4/
[D5]4 (22:78, from anhydride 5), (4 + [α-D]4)/([D5]4 + [D6]4) (24:76,
from ketene 2), and 11/[D5]11 (37:63, from the acyl chloride 9). The
strongly broadened C-α signal (a conformational problem)[9] of 4
Experimental Section
Chemicals: Improved preparations of tBu2C=C=O (2), tBu2CH-
CO2H (3), anhydride 5, and acyl chloride 9 have already been de-
scribed, together with the solvent dependence of their 1H NMR
spectra.[9] Pure commercial samples of DMSO, [D6]DMSO, di-
glyme, and [D8]THF were dried with molecular sieves in two stages.
(Methylthio)methyl 2-tert-Butyl-3,3-dimethylbutanoate (4): An
NMR tube (5 mm) was charged with crude di-tert-butylketene (2;
ca. 116 mg, 0.75 mmol), anhydrous diglyme (0.400 mL), dry
6340
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 6335–6342

Mechanisms of Ketene Acylation
and its isotopologues was unsuitable for these analyses. Table S1
(see the Supporting Information) presents the detailed results and
illustrates their reliability through consistent integral ratios as mea-
sured at two or more molecular positions.[19]
of which had the expected intensity. Consequently, the [M + 1]+
peak of 4 could not contain an α-D contribution and must have
been composed of M + 13C + H+ contributions alone. The M+/
[M + 5]+ = 4/[D5]4 intensity ratio of 25:75 for this sample emerged,
which is in agreement with the value of 22:78 derived from NMR
analysis. Next, the product mixture obtained with ketene 2 was
found (always after corrections for 13C satellites) to generate the
mass fragments tBu2CHCO+ and tBu2CDCO+ in a 66:34 ratio,
which is also consistent with the above NMR ratio of α-H/α-D
=67:33. In addition, the NMR-basedratio(4 + [α-D]4)/([D5]4+[D6]-
4) = 24:76 was also found for both the M+/[M + 5]+ = 4/[D5]4 and
the [M + 1]+/[M + 6]+ = [α-D]4/[D6]4 fractions; these latter two
equalities require the same proportion (33%) of α-D labeling to be
present in both (4 + [α-D]4) and ([D5]4 + [D6]4). Consequently, the
first of these two sum terms (the 24% portion) is distributed by
67:33, so to become (16% + 8%), while the second sum term (the
76% portion) divides into (51% + 25%), which are the portions
reported in Scheme 5 (last line). It may be noticed that due to the
usual uncertainties about mass peak assignments, peak superposi-
tions, and M + H+ contributions, mass spectrometry alone (with-
out NMR analysis) of these unpurified samples would not have
provided these results with a comparable reliability.
Scheme 7. Deuterium-induced 13C and 1H NMR isotope shifts Δ
= δ(isotopologue) – δ(unlabeled) in ppb (averaged values); C-α sig-
nals in 4 and [D5]4 were too broad.
Supporting Information (see footnote on the first page of this arti-
cle): Table S1 of deuterium-induced 13C and 1H NMR shifts; gene-
ral procedure for determining intermolecular competitive kinetic
H/D isotope effects (with Figure S1).
Accurate 13C integral ratios were obtained through the application
of the method of inverse-gated broad-band decoupling, which em-
ploys {1H} decoupling only during the short acquisition period of
the pulsed Fourier transform technique, used to suppress the unde-
sired nuclear Overhauser enhancements of the integrals of proton-
bearing carbon atoms. In addition, the relaxation delay period was
set long enough (20–60 s) for a practically complete 13C nuclear
spin relaxation toward the thermal equilibrium before the next
pulse. 1H NMR integrals do not normally require such a time-
consuming procedure and can be recorded under standard condi-
tions. The α-H signals of a mixture of 4 and [D5]4 are separated
by Δ = –1.3 ppb (Scheme 7), as induced by the deuterium atoms of
the OCD2S group over a distance of five single bonds. Although a
Acknowledgments
Support of this work by the Chemistry Department of the LMU
München is gratefully acknowledged. I thank the two reviewers for
their constructive comments.
[1] H. R. Seikaly, T. T. Tidwell, Tetrahedron 1986, 42, 2587–2613.
[2] T. T. Tidwell, Acc. Chem. Res. 1990, 23, 273–279.
[3] T. T. Tidwell in Science of Synthesis (Ed.: R. L. Danheiser),
Thieme, Stuttgart, 2006, vol. 23, pp. 569, 613, 628, 638.
[4] T. T. Tidwell, Ketenes II, 2nd ed., Wiley, Hoboken, New Jersey,
2006.
[5] E. Schaumann, W. Walter, Chem. Ber. 1974, 107, 3562–3573.
[6] D. Lenoir, H. R. Seikaly, T. T. Tidwell, Tetrahedron Lett. 1982,
23, 4987–4990.
[7] L. M. Baigrie, D. Lenoir, H. R. Seikaly, T. T. Tidwell, J. Org.
Chem. 1985, 50, 2105–2109.
[8] L. Gong, R. Leung-Toung, T. T. Tidwell, J. Org. Chem. 1990,
55, 3634–3639.
[9] R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer, Eur. J. Org.
Chem. 2010, 6651–6664.
[10] D. P. Bauer, R. S. Macomber, J. Org. Chem. 1975, 40, 1990–
1992.
[11] The O-nucleophilicity parameters of DMSO are N = 9.75 in
acetonitrile solution and N = 11.3 in DMSO, according to:
T. B. Phan, C. Nolte, S. Kobayashi, A. R. Ofial, H. Mayr, J.
Am. Chem. Soc. 2009, 131, 11392–11401, on p. 11400.
[12] M. S. Newman, A. Arkell, T. Fukunaga, J. Am. Chem. Soc.
1960, 82, 2498–2501.
[13] For an instructive recent introduction to the Pummerer acyl-
ation reaction, see: K. S. Feldman, Tetrahedron 2006, 62, 5003–
5034, Scheme 2 therein.
1
complete separation of this H signal pair could not be achieved,
its integral ratio of approximately 24:76 for (4 + [α-D])4/([D5]4 +
[D6]4) equals the 13C integral ratio mentioned above for that mix-
ture (these two sums have to be taken because the 13C NMR re-
gions of the –CD2SCD3 parts of [D5]4 and [D6]4 are practically
indistinguishable, as shown in Scheme 7; likewise, the NMR regions
of the –CH2SCH3 parts of 4 and [α-D]4 are practically identical).
For the α-H portions (in 4 + [D5]4) in relation to all α-D compo-
nents ([α-D]4 + [D6]4), the γ-H (in tBu) signal pair with Δ = –1.5
ppb in Scheme 7 (integral ratio 71:29) supports the overall α-H/α-
D = 67:33 ratio as measured by 13C NMR analysis for (4 +
[D5]4)/([α-D]4 + [D6]4) in the same experiment (Table S1 in the
1
Supporting Information). Thus, deuterium-induced H NMR iso-
tope shifts can be recommended for a quick, perhaps semiquanti-
tative, preview before starting a long-term 13C NMR spectroscopic
measurement.
In the presence of unlabeled 4, these NMR methods cannot differ-
entiate [D6]4 from a mixture of [D5]4 plus [α-D]4, so that the pres-
ence of [α-D]4 and the precise portion of [D6]4 cannot be guaran-
teed at this point. Therefore, auxiliary mass-spectrometric analyses
with chemical ionization (isobutane) at temperatures at or below
45 °C were carried out as follows: A first run with 4 + [D5]4 (ob-
tained from the anhydride 5) confirmed the absence of α-D in the
fragment tBu2CHCO+ (the acylium cation 13),[55] the 13C satellite
[14] a) G. E. Wilson, C. J. Strong, J. Org. Chem. 1972, 37, 2376–
2380, Scheme 1 therein; b) S. Wolfe, P. M. Kazmaier, Can. J.
Chem. 1979, 57, 2388–2396.
[15] K. B. Wiberg, Chem. Rev. 1955, 55, 713–743, Table 1 therein.
[16] S. Oae, M. Kise, Tetrahedron Lett. 1968, 9, 2261–2265.
Eur. J. Org. Chem. 2011, 6335–6342
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6341

R. Knorr
FULL PAPER
[17]
[33] For 1H NMR (CCl4), see: S. Satoh, T. Taguchi, M. Itoh, M.
Tokuda, Bull. Chem. Soc. Jpn. 1979, 52, 951–952, compound
3b therein.
Ketene 2 was not generated from 5 under these conditions; this
follows from the absence in Scheme 2 of the tBu2CD deriva-
tives, which arise from 2 (Scheme 5) in the same environment.
[34] Determinations of kinetic reaction orders were deemed imprac-
ticable: Small concentrations of the catalyst 3 would continu-
ally decrease through its side reaction with ketene 2 leading
to the anhydride 5, whereas higher concentrations of 3 would
consume a major part of 2 at the expense of the formation of
ester 4 from 2.
[35] W. A. Pryor, H. T. Bickley, J. Org. Chem. 1972, 37, 2885–2893.
[36] J. K. Crandall, S. A. Sojka, J. B. Komin, J. Org. Chem. 1974,
39, 2172–2175.
[37] A. Tille, H. Pracejus, Chem. Ber. 1967, 100, 196–210.
[38] J. Jähme, C. Rüchardt, Tetrahedron Lett. 1982, 23, 4011–4014.
[39] C. E. Cannizzaro, K. N. Houk, J. Am. Chem. Soc. 2004, 126,
10992–11008.
[40] P. J. Lillford, D. P. N. Satchell, J. Chem. Soc. B 1968, 889–897.
[41] D. P. N. Satchell, M. J. F. Satchell, Z. Naturforsch., B 1991, 46,
391–392, and earlier work cited therein.
[18]
[19]
[20]
For a literature survey, see: C. M. Hull, T. W. Bargar, J. Org.
Chem. 1975, 40, 3152–3154.
The mathematical formula usually employed to calculate kH/
kD from these data is reported in the Supporting Information.
For the feasibility of generating the sulfonium ylide in similar
systems, see: A. H. Fenselau, J. G. Moffat, J. Am. Chem. Soc.
1966, 88, 1762–1765.
A small kH/kD value would also result when the C–H and C–
D bond fissions were reversible (elimination via a kinetically
more stable sulfonium ylide whose subsequent O–S bond fis-
sion would limit the overall rate); this would eventually lead to
D/H incorporation at sulfoxide functions and thus to a de-
creased degree of labeling of the methylthio groups derived
therefrom. However, the corresponding –SCH2D or –SCHD2
groups were never detected in the present work, as exemplified
in Figure S1 of the Supporting Information.
a) L. Horner, P. Kaiser, Justus Liebigs Ann. Chem. 1959, 626,
19–25; b) J. H. Jones, D. W. Thomas, R. M. Thomas, M. E.
Wood, Synth. Commun. 1986, 16, 1607–1610.
M. Cocivera, V. Malatesta, K. W. Woo, A. Effio, J. Org. Chem.
1978, 43, 1140–1145.
[21]
[42] “A relatively high ionizing power” was confirmed on p. 11394
of ref.[11]
[22]
[43] At this point, the previously[9] proposed addition of the potas-
sium alkoxide of HOCH(tBu)C(=O)tBu (pivaloin) to 2 in
DMSO solution was now established through the in situ obser-
vation (1H NMR analysis) of the emerging pivaloin ester[9]
tBu2CHCO2CH(tBu)C(=O)tBu at ambient temperature within
approximately 3 h.
[23]
[24]
For NMR spectroscopic data, see: a) J. R. Gauvreau, S. Poi-
gnant, G. J. Martin, Tetrahedron Lett. 1980, 21, 1319–1322; b)
L. C. Ducati, M. P. Freitas, C. F. Tormena, R. Rittner, J. Mol.
Struct. 2006, 800, 45–50; c) S. W. Bass, S. A. Evans, J. Org.
Chem. 1980, 45, 710–715.
[44] S. H. Kabir, H. R. Seikaly, T. T. Tidwell, J. Am. Chem. Soc.
1979, 101, 1059–1060.
[45] The pKa = 7.04 for 3 in H2O/H3COH (1:1) at 40 °C, as reported
by: M. S. Newman, T. Fukunaga, J. Am. Chem. Soc. 1963, 85,
1176–1178.
[25] A. D. Allen, T. T. Tidwell, J. Am. Chem. Soc. 1987, 109, 2774–
2780.
[26] Compound [α-D]3, with a deuterium content of 23%, was reco-
vered from a solution of unlabeled 3 in D2SO4 at 70 °C after
53 min; compare Equation (5) in ref.[25]
[27] R. Rätz, O. J. Sweeting, J. Org. Chem. 1963, 28, 1612–1616.
[28] G. A. Olah, M. Alemayehu, A. Wu, O. Farooq, G. K. S. Prak-
ash, J. Am. Chem. Soc. 1992, 114, 8042–8045.
[29] As a low-speed example, p-xylene became acylated by an acyl-
ium reagent Ar–CϵO+ within a few hours at room tempera-
ture, apparently without decarbonylation according to Figure 7
in: F. Effenberger, J. K. Eberhard, A. H. Maier, J. Am. Chem.
Soc. 1996, 118, 12572–12579.
[46] This interpretation is based on the discovery that less encum-
bered ketenes add weak carboxylic acids at 25 °C with rates
that increase with higher pKa values, that is, with decreasing
acidities, as reported by: a) J. M. Briody, P. J. Lillford, D. P. N.
Satchell, J. Chem. Soc. B 1968, 885–889; b) N. L. Poon, D. P. N.
Satchell, J. Chem. Res. Synop. 1983, 182–183.
[47] This conclusion implies that the protonation step 1Ј should be
effectively irreversible in the acidic environment that contains
the acid 3. In support of this notion of a very slow deproton-
ation of 13, the base-free deprotonation of acetylium
(H3CCϵO+) at 150 °C to generate H2C=C=O was reported to
require 20 h; see: G. A. Olah, E. Zadok, R. Edler, D. H. Adam-
son, W. Kasha, G. K. S. Prakash, J. Am. Chem. Soc. 1989, 111,
9123–9124.
[30] For comparisons with faster Friedel–Crafts processes, the de-
carbonylation of Alk–CϵO+ was apparently either (a) faster
than, (b) similarly fast, or (c) slower than the acylation of ben-
zene by Alk–CϵO+ to give Alk–COPh; see: a) M. E. Grundy,
W.-H. Hsü, E. Rothstein, J. Chem. Soc. 1958, 581–586 for
tBuCH2CH(tBu)–CϵO+; b) G. A. Olah, W. S. Tolgyesi, S. J.
Kuhn, M. E. Moffatt, I. J. Bastien, E. B. Baker, J. Am. Chem.
Soc. 1963, 85, 1328–1334 for tBu–CϵO+; c) G. A. Olah, M. B.
Comisarov, J. Am. Chem. Soc. 1966, 88, 4442–4447, on pp.
4446–4447 for adamantyl-1-CϵO+.
[48] S. K. Bur, A. Padwa, Chem. Rev. 2004, 104, 2401–2432.
[49] O. DeLucchi, U. Miotti, G. Modena, Org. React. 1991, 40,
157–405.
[50] J. P. Marino, M. Neisser, J. Am. Chem. Soc. 1981, 103, 7687–
7689.
[51] J. P. Marino, G. Cao, Tetrahedron Lett. 2006, 47, 7711–7713.
[52] N. L. Poon, D. P. N. Satchell, J. Chem. Soc. Perkin Trans. 2
1983, 1381–1383.
[53] I. Lillien, J. Org. Chem. 1964, 29, 1631–1632.
[54] W. Adam, A. Alzérreca, J.-C. Liu, F. Yany, J. Am. Chem. Soc.
1977, 99, 5768–5773, on p. 5772.
[55] Acylium cations are the thermodynamically favored isomers in
both the gas phase and in solution, as noted on page 2778 of
ref.[25]
[31] K. K. Wang, Z. Wang, P. D. Sattsangi, J. Org. Chem. 1996, 61,
1516–1518, on p. 1518.
[32] a) For 1H NMR (CDCl3), see: R. J. Abraham, M. Canton, M.
Reid, L. Griffiths, J. Chem. Soc. Perkin Trans. 2 2000, 803–812,
Table 2 therein; b) for 13C NMR (CDCl3), see: C. W. Fong,
Aust. J. Chem. 1980, 33, 1291–1300, Table 1 therein; c) for 13C
NMR (C6D6), see: L. Ernst, Tetrahedron Lett. 1974, 15, 3079–
3080.
Received: June 28, 2011
Published Online: September 13, 2011
6342
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 6335–6342