Shorter and easier syntheses of Di-tert-butylketene and related gem-Di-tert-butyl compounds

Article abstract of DOI:10.1002/ejoc.201000888

The ketene tBu₂C=C=O is prepared from tBu₂C=O in three steps (performable as a two-stage operation) through elimination of HCl from the intermediate product tBu₂CCl-CH=O. The acid tBu ₂CH-CO₂H, obtainable in two, three, or four preparative stages from tBu₂C=O, adds slowly to the ketene to produce the anhydride (tBu₂CH-CO)₂O. Elemental lithium together with ClSiMe₃ converts tBu₂CCl-CH=O into tBu₂C=CH- OSiMe₃, which is a durable precursor of tBu₂CH-CH=O, making this aldehyde easily and cheaply available from tBu₂C=O. By exclusion of alternative mechanistic possibilities, the reduction of tBu ₂CCl-CH=O by tBuMgCl is shown to involve at least one single-electron transfer, leading to the enolate tBu₂C=CH-OMgCl, which can be converted into tBu₂CH-CH=O (three steps from tBu₂C=O) or into tBu₂C=CH-OSiMe₃. Hydride transfer from NaBH ₄ to tBu₂CCl-CH=O affords tBu₂CCl-CH ₂OH, the transformations of which provide an entertaining set of S_N1-type reactions. Several other examples of carbenium-type behavior are encountered in this gem-tBu₂ system; they are attributed to steric congestion, which also impedes bond rotations in the anhydride and in two esters. A convenient route to tBu₂CH-C≡N (five steps from tBu₂C=O) uses the conversion of tBu₂C=CH-OSiMe₃ into tBu₂CH-CH=NOH. The slow thermal (Z)/(E) equilibration of tBu₂CH-NH-CH=O reveals the ranking of ecliptic repulsions as H ₃C < tBu < tBu₂CH. Copyright

Full text of DOI:10.1002/ejoc.201000888

FULL PAPER
DOI: 10.1002/ejoc.201000888
Shorter and Easier Syntheses of Di-tert-butylketene and Related
gem-Di-tert-butyl Compounds[‡]
Rudolf Knorr,*[a] Karsten-Olaf Hennig,[a] Bernhard Schubert,[a] and Petra Böhrer[a]
Dedicated to the memory of Professor Gert Köbrich
Keywords: Aldehydes, α-chloro / Electron transfer / Oxiranes, chloro / Ketenes / Strained molecules / Steric hindrance
The ketene tBu2C=C=O is prepared from tBu2C=O in three
steps (performable as a two-stage operation) through elimi-
nation of HCl from the intermediate product tBu2CCl–CH=O.
The acid tBu2CH–CO2H, obtainable in two, three, or four
preparative stages from tBu2C=O, adds slowly to the ketene
to produce the anhydride (tBu2CH–CO)2O. Elemental lithium
together with ClSiMe3 converts tBu2CCl–CH=O into
tBu2C=CH–OSiMe3, which is a durable precursor of tBu2CH–
CH=O, making this aldehyde easily and cheaply available
from tBu2C=O. By exclusion of alternative mechanistic pos-
sibilities, the reduction of tBu2CCl–CH=O by tBuMgCl is
shown to involve at least one single-electron transfer, leading
to the enolate tBu2C=CH–OMgCl, which can be converted
into tBu2CH–CH=O (three steps from tBu2C=O) or into
tBu2C=CH–OSiMe3. Hydride transfer from NaBH4 to
tBu2CCl–CH=O affords tBu2CCl–CH2OH, the transforma-
tions of which provide an entertaining set of SN1-type reac-
tions. Several other examples of carbenium-type behavior
are encountered in this gem-tBu2 system; they are attributed
to steric congestion, which also impedes bond rotations in
the anhydride and in two esters. A convenient route to
tBu2CH–CϵN (five steps from tBu2C=O) uses the conversion
of tBu2C=CH–OSiMe3 into tBu2CH–CH=NOH. The slow
thermal (Z)/(E) equilibration of tBu2CH–NH–CH=O reveals
the ranking of ecliptic repulsions as H3C Ͻ tBu Ͻ tBu2CH.
Ǟ 15 Ǟ 14 (three preparative stages). This approach starts
with the generation of the highly unstable[11,12] carbenoid
dichloromethyllithium (LiCHCl2, 16); a tetrahydrofuran
(THF) solution of dichloromethane must be kept at below
–100 °C during the slow addition of precooled (–100 °C) n-
butyllithium (nBuLi), followed by the slow introduction of
precooled (–90 °C) ketone 1.
We show that the methods of this last route can be sim-
plified and shortened (three steps in Յ3 stages) for an easier
approach to 14 and to other gem-di-tert-butyl (tBu2C) com-
pounds.
Introduction
Di-tert-butylketene (14, Scheme 1) is noted for its very
low reactivity:[1] unlike other ketenes,[2] it does not dimerize
and adds neutral water very slowly to produce the acid 11.
The following routes (Scheme 1) to this persistent ketene,
starting either from di-tert-butyl ketone[3] (1) or its imine[3]
3, have been published.
I) The oldest and most popular route furnished a 16%
yield of 14 in six steps:[4] namely 1 Ǟ 4 Ǟ 7 Ǟ 10 Ǟ 11 Ǟ
13 Ǟ 14. However, occasional complaints about problems
with the preparations of 7,[5] 8,[6] and 14[7,8] led several
groups[5–8] to recommend improved versions of this route.
II) A method[9] that employs the imine 3 directly pro-
vided a higher yield (Ͻ32%) of 14 in eight steps: namely 3
Ǟ 2 Ǟ 5 Ǟ 6 Ǟ 9 Ǟ 12 Ǟ 11 Ǟ 13 Ǟ 14. However, this
pathway has only rarely been used, presumably because of
certain difficult techniques and unusual reagents.[10]
III) An alternative strategy[10] gave a 51% yield of the
ketene 14 in four steps: namely 1 + 16 Ǟ (17 Ǟ 18 Ǟ) 19
Results and Discussion
A. Preparation and Properties of 2,2-Di-tert-butyl-3-
chlorooxirane (18)
The easier approach to the alkoxide 17 (Scheme 2) con-
sists of the generation of dichloromethyllithium (16) in
THF through deprotonation of CH2Cl2 in the presence of
di-tert-butyl ketone[3] (1), a strategy[13] that appeared to us
to be an argument against the use of carbanionic bases that
might add across the C=O double bond of 1.[14a] Because
the recommended[13b] base lithium dicyclohexylamide can
give annoying problems with the poorly soluble dicyclohex-
ylammonium salt during the aqueous workup, we replaced
[‡] Sterically Congested Molecules, 22. Part 21: R. Knorr, E. C.
Rossmann, M. Knittl, Synthesis 2010, 2124–2128.
[a] Department of Chemistry, Ludwig-Maximilians-Universität,
Butenandtstr. 5-13, 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.201000888.
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6651

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
Scheme 2. Formation, deprotonation, and ionization of 2,2-di-tert-
butyl-3-chlorooxirane (18).
Scheme 1. Routes to di-tert-butylketene (14). [a] This work.
it with lithium diisopropylamide[13b] (LDA), which was
added dropwise to a THF solution of 1 and CH2Cl2 stirred in hot glacial acetic acid and isolated the product 29 of de-
at –70 °C: LiCHCl2 was formed and efficiently trapped by tert-butylation plus ring expansion. A closer examination
the ketone 1,[14b] as shown through final quenching with of this tetrahydrofuran-3-one derivative 29 gave evidence of
acids at –70 °C or 0 °C and subsequent isolation of the impeded rotation of the 4-tert-butyl substituent (but not of
chlorooxirane 18 (86%) as the only product. In contrast 1Ј-tBu): the three methyl groups of 4-tBu become nonequiv-
with previous evidence with less congested systems,[13b,15,16] alent below the coalescence temperatures of roughly –30 °C
it was thus not possible to prevent the alkoxide 17 from (13C NMR at 100 MHz) and –50 °C (1H NMR at
cyclizing to give 18.[17]
400 MHz), as documented in Table S1 in the Supporting
Dichloromethane must always be in excess in relation to Information, corresponding to the free enthalpy barriers of
LDA during the above preparation of the chlorooxirane 18: ΔG# ≈ 10.8 and 11.0 kcalmol–1 at 243 K and 223 K, respec-
residual LDA (and also lithium 2,2,6,6-tetramethylpiperid- tively. The lower barrier (9.9 kcalmol–1) reported[21] for tBu
ide, LiTMP) would lithiate 18 to give the carbenoid 21, rotation in the locally similar environment of 30 may be
which forms the known[18] 3-oxetanone 25, perhaps via the due to an alleviating cog-wheel rotational mode that is not
primary product 23 of a “dimerization” of 21 (an unproven possible in 29.
possibility). Indeed, 25 was isolated as the only product
The chlorooxirane 18 can be expected[22] to generate the
(89%) from ketone 1 after treatment with 1 equiv. of formylcarbenium ion 24 directly (Scheme 2) rather than to
CH2Cl2 and 4 equiv. of LDA in a one-pot synthesis at ionize first to the oxiranyl cation 22. As a tight ion
Յ0 °C. The very slow photolytic decomposition of 25 in pair,[22,23] 24 collapses to yield the α-chloro aldehyde[10] 19
CCl4 solution furnished neither the ketene 14 nor an oxir- but may also rearrange to give 26, which produces the un-
ane (through decarbonylation), although these two types of saturated aldehyde 27 as a troublesome side-product. It was
fissions had been observed[19,20] with other tetrasubstituted somewhat difficult to avoid the formation of 27 under the
3-oxetanones. A small amount of the unknown acid reported[10] conditions, and the slow fractionating distil-
tBu2C=CH–CO2H (28) could be identified and might have lation of crude 19/27 mixtures proved to be an ugly task
been formed via a photoproduct such as tBu2C=C=C=O. because of excessive foaming with generation of more alde-
In another attempt to generate 14, we treated 25 with HCl hyde 27.[24] Chromatographic separation (19 elutes first!) or
6652
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664

Syntheses of Di-tert-butylketene and Related gem-Di-tert-butyl Compounds
a quick trap-to-trap distillation under 20 mbar into a dry- compounds soon started to react with each other, yielding
ice cooled collector can avoid the formation of more 27 after workup the pivaloin ester 36 along with residual ke-
from 19. For a more convenient preparation of 19 contain- tene 14 and pivaloin (38), but no acid 11. An authentic
ing very little 27 (hence requiring no purification, yield specimen of 36 was obtained through short heating (150 °C)
80%), we recommend that the rearrangement of chlorooxir- of pivaloin[28] (38) with the acyl chloride 13 and pyridine in
ane 18 be conducted with dilute (ca. 10%) solutions in dry diglyme (MeOCH2CH2OCH2CH2OMe), whereas the base-
toluene[10] at 110 °C, which yielded 19 together with 27 free acylation of 38 with ketene 14 at 150 °C in diglyme
(4%) in 45 min. A more dilute sample of 18 (5% by volume) required many hours to furnish 36. This indicates that the
gave 19 almost free of 27 at 95 °C within 1 h, whereas a ketene 14 reacts exclusively with the pivaloin anion 34 (not
more concentrated (30%) solution yielded 19 containing 27 with the pivaloin 38) at room temperature, such that the
(13%.)
ester enolate 35 should be an intermediate. Indeed, the for-
This heterolysis of 18 occurred even at ca. 20 °C in a mation of 35 and its protonation by the solvent (DMSO
methanol solution containing 2 equiv. of NaOCH3; it containing KOH and an equilibrium portion of H2O) were
obeyed first-order kinetics with t1/2 = 105 (Ϯ3) min (or ca. established through a run with KOH (17 equiv.) in the sol-
85 min without NaOCH3) but did not incorporate the vent [D6]DMSO, which furnished the dideuteriated pivaloin
methoxy group. Under these conditions, however, the inter- ester [3,2Ј-D2]36 with 86% D at C-2Ј. The label at C-3 is
mediate 24 produced mixtures of 19 with up to 82% of 27. due to H/D isotope exchange under these strongly basic
Comparable mixtures of 19 and 27 were formed much more conditions, as was shown by subsequent treatment of [3,2Ј-
rapidly (Յ70 min at room temp.) from 18 in a suspension D2]36 with KOH in unlabeled DMSO (at 95 °C for 8 h),
of powdered KOH in dimethyl sulfoxide (DMSO), again which afforded 37 with 81% D at C-2Ј (D removed from
without incorporation of the strong nucleophiles existing in C-3). As well as showing 13C NMR line broadening as an
this system. The release of repulsive tBu2 interactions in 18 indication of conformational freezing, 36 is also remarkable
1
upon ionization to 24 appears to be important for these for the substantial H NMR chemical shift difference be-
higher rates in comparison with the related 2-tert-butyl-3- tween its diastereotopic 2Ј-tBu groups, in view of the fact
chlorooxirane, which is stable[25] in aqueous methanol at that these protons are situated in a distance of six single
room temperature for more than 24 h.[26]
bonds away from the center of chirality.
B. Di-tert-butylketene (14) and some Derivatives
An unexpected shortcut in route III (Scheme 1) allowed
the direct conversion of the α-chloro aldehyde 19 into the
ketene 14 by use of the bases LiTMP or (better) LDA in
THF solution. This very unusual mode[27] of HCl elimi-
nation afforded up to 70% yields of isolated, spectroscopi-
cally pure 14, which may be used without purification. Even
a one-pot preparation of 14 from the chlorooxirane 18 was
possible when the solution of 19 in toluene (section A) was
not worked up but treated with LiNR2 in THF; the yields
(45–55%) of subsequently isolated material depend strongly
on consideration of the high volatility of 14, particularly
during the evaporation of toluene at reduced pressure.
Our interpretation (Scheme 3) of this surprising process
as an E2-type elimination postulates an unusually increased
population of the Cl,O-cis conformation 19Ј (with dihedral
angles Cl–C–C–H in the vicinity of 180°) at the expense of
other conformations (19). A mechanistic alternative with
heterolysis of 19 to generate the formylcarbenium ion 24
(Scheme 2) would be recognized through formation of the
rearranged product 27, which was not observed: 19 did not
solvolyze in acid-free methanol (Ͼ6 d at room temp.) and
proved stable in base-free DMSO at 148 °C for at least
8 h.[24] With the intent of employing a more convenient and
cheaper base in place of LDA, we treated 19 with a suspen-
Scheme 3. 2-tert-Butyl-2-chloro-3,3-dimethylbutanal (19): reactions
with bases.
The pathway leading to the pivaloin anion 34, as pro-
sion of solid KOH in DMSO and observed the disappear- posed in Scheme 3, starts with the nucleophilic addition of
1
ance of 19 by H NMR (Scheme 3): the ketene 14 and the KOH to the aldehyde function of 19 (and possibly also 19Ј)
pivaloin anion 34 (molar ratio ca. 2:1) were formed at room to give 33; at the same time, HCl elimination from 19Ј pro-
temperature over a period of several hours, and these two moted by KOH (in place of LiNR2) affords the ketene 14.
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6653

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
In analogy with similar additions of hydroxide[29] or meth- required more than 63 h for the acylation of both the sol-
oxide[30,31] to other α-chloro aldehydes, the adduct 33 can vent and the acid 11 to furnish the ester 41 (yield 47%) and
be expected to expel its chloride anion with cyclization and the anhydride 40 (yield 21%). Neither 41 nor 40 were
subsequent deprotonation; the resulting epoxy alkoxide 32 cleaved by NaOH in dioxane (2 m, at 85–95 °C, 4 d) or by
will rearrange to the alkoxide 31 of α,α-di-tert-butyl-α-hy- solid KOH in CCl4 (at 70 °C, 7 d).
droxyacetaldehyde. The ensuing tBu migration in 31 toward
the aldehyde function with formation of the pivaloin anion
34 has precedent[32] in a similar aryl migration at ca. 70 °C
C. Further Transformations of 18 and 19
in a sterically less congested system. The final fusion of 34
and ketene 14 occurs substantially more rapidly than the
addition of 14 to KOH suspended in DMSO, as shown by
the isolation of residual 14 (without acid 11). This was con-
firmed in a separate run in which ketene 14 and solid KOH
in DMSO at 55 °C reacted slowly (in Ͼ38 h) to yield the
acid 11.
Di-tert-butylacetaldehyde (8 in Scheme 5) is difficult to
purify and to handle in small batches because of its propen-
sity toward autoxidation. Its traditional four-step synthe-
sis[4] (21%, Scheme 1) involves certain complications,[6] and
only small amounts of 8 were obtained through the isomeri-
zation[4,36] of 2,2-di-tert-butyloxirane (43), an epoxide that
had been prepared in three steps from the ketone 1 with
total yields of 44%[36] or 86%.[37] The one-pot preparation
of 8 (54%) from the ketone 1 with the relatively expensive
reagent diethyl isocyanomethanephosphonate[38] might be
limited in its attraction on a larger scale.[39]
Under more convenient conditions than those of route I
in Scheme 1, the acid 11 could now be prepared through
hydration of the ketene 14 with concentrated HCl at room
temperature (bottom line of Scheme 3, yield up to 80%);
even highly contaminated samples or THF solutions of 14
(as obtained from 18 or 19 without workup) were suitable.
The acyl chloride 13 could be produced from 11 under
milder conditions than reported[4] with a small excess of
thionyl chloride in CCl4 solution at 70 °C (13 h) or 55 °C
(43 h); after the final evaporation of CCl4 and residual
SOCl2 down to a pressure of 50 mbar at 60 °C for 3 min,
the volatile liquid 13 could be used without purification.
Di-tert-butylacetic anhydride (40, Scheme 4)[33,34] was ac-
cessible through the acylation of the acid 11 with the ketene
14 when these two reagents were either employed as such
or were generated in situ from the potassium or (better)
lithium salt (39) of 11, which is sufficiently basic to elimin-
ate HCl from the acyl chloride 13 in anhydrous toluene.
Scheme 5. Dechlorination of 18 and 19 by nBu3SnH.
In a search for convenient procedures, we dechlorinated
The initial formation of the ketene 14 from 13 and 39 could 18 and 19 by use of the following reductants. Tributyl-
be observed (1H NMR in situ) in THF or diglyme solutions stannane (2 equiv.) reduced the chlorooxirane 18 (Scheme 5)
at 55 °C. The ensuing slower formation of 40 could be mon- in a benzene solution (0.52 m) under UV irradiation [with
itored[35] in diglyme at 95 °C or above; however, hot diglyme 2,2Ј-azobis(2-methylpropionitrile) as an initiator, 51 h] to a
is not a recommendable solvent, because it will produce the 1:1 mixture of the oxirane 43 (identified through its 1H and
ester 41 through a slow acylation by 13 or 14 (but not by 13C NMR[36] data) and the aldehyde 8. With lower concen-
the anhydride 40) under these conditions. With pyridine as trations of tributylstannane, the transient oxiranyl radical
an alternative base instead of 39, 13 formed a transient por- 42 would have been trapped less efficiently before isomeriz-
tion (up to 26%) of ketene 14 and at 150 °C in diglyme ing to the planar and “relatively persistent”[40] formyl-sub-
stituted radical 44. In a cleaner process, the α-chloro alde-
hyde 19 and tributylstannane generated 44, which furnished
8 as the only product under similar conditions (except for
the much shorter reaction time of 2 h without an initiator).
Nevertheless, we abandoned this approach because of the
usual difficulties of removing stannyl byproducts.
The oxiranyl radical 42 should also be a first intermedi-
ate in the reduction of 18 (Scheme 6) through electron
transfer from magnesium turnings (ca. 4 equiv. of Mg) in
the presence of ClSiMe3 (ca. 6 equiv.) in boiling THF. Upon
escape of 42 from the metallic surface before a second elec-
tron transfer can occur, the isomerization of 42 to give 44
takes place in the nonreducing solution. Its considerable
lifetime[40] enables 44 to approach the Mg0 surface for a
second electron transfer, producing the enolate 45. With si-
Scheme 4. Reactions of di-tert-butylacetyl chloride (13).
6654
lylation of 45 in situ to yield the silyl enol ether 46, a simple
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664

Syntheses of Di-tert-butylketene and Related gem-Di-tert-butyl Compounds
aqueous workup furnished 8 and 46 in a 1:5 ratio. Although
such a conversion of 46 to 8 might be completed intention-
ally, this route was deemed unreliable because of difficulties
in keeping the Mg0 surface sufficiently active: in spite of the
presence of ClSiMe3 and the use of alternatively activated
Mg0, the repeatedly encountered delays in the reduction
process allowed the decay modes of 18 to interfere, produc-
ing 19 and 27 (as depicted in Scheme 2).
Scheme 7. Metal-based reduction of 18 with Li0; e– = electron.
The clean conversion of 19 into 8 or its precursors 46
and 45 (top lines of Scheme 8) at 20 °C (1H NMR spec-
troscopy in situ) by treatment with tBuMgCl (2.7 equiv.) in
THF is also possible within 35 min. Remarkably, tBuMgCl
did not add to the aldehyde group of 19 and acted neither
as a base (pathway A) to afford the ketene 14 (not detected
in situ although it would be stable for Ͼ12 d) nor as a hy-
dride donor (pathway B) to furnish the β-chloro alkoxide
51. As outlined below in reaction 12 in Scheme 9, 51 would
be converted into the aldehyde 8, which was not detected
although it is stable under these conditions. Instead, the 1H
NMR spectra revealed the formation of the Mg enolate 45
(δH ≈ 1.15 and 1.36 ppm) shown in Scheme 8, together with
an equivalent amount of isobutene. The constitution of 45
was established through derivatization with ClSiMe3
Scheme 6. Metal-based reductions of 18 with Mg0 and of 19 with
Li0 and ClSiMe3; e– = electron.
The α-chloro aldehyde 19 (Scheme 6) was readily reduced
in THF at room temperature by metallic lithium (3 equiv.),
the surface of which was most efficiently kept active by the
presence of ClSiMe3 (1.5 equiv.) in situ. This is the simplest
and most productive way of preparing the silyl enol ether
46 via the radical 44 and the Li enolate 47 (δH = 6.82 ppm).
It is essential to add NEt3 (2 equiv.) prior to an aqueous
workup, so as to avoid a premature hydrolysis of 46 through
the hydrochloric acid formed from residual ClSiMe3. As a
less air-sensitive precursor of 8, 46 can be stored until 8 is
prepared through intentional hydrolysis of 46 with aqueous
acid.
(3.2 equiv.), which afforded the silyl enol ether 46 (δH
=
6.21 ppm in situ) as the only product. The addition of NEt3
(2 equiv.) to neutralize the acid arising from residual
ClSiMe3 is essential for 46 to survive the subsequent aque-
ous workup. Alternatively, residual tBuMgCl was identified
through carboxylation (solid CO2) to give pivalic acid,
showing that the excess of ClSiMe3 had not reacted with
tBuMgCl.
Under similar electron-transfer conditions from elemen-
tal Li0 in THF, the chlorooxirane 18 reacted by a mechan-
istic modification of the Mg0 route of Scheme 6 but fur-
nished a different product (compound 49 in Scheme 7). The
stronger reducing agent Li0 (4 equiv.) appears to perform a
faster second electron transfer before the oxiranyl radical
42 can escape from the metallic surface and isomerize to
44: no traces of the aldehyde 8 or (with ClSiMe3) of its silyl
enol ether 46 could be detected. Instead, 42 was presumably
reduced to the O,Li-carbenoid 48 (or a ring-opened iso-
mer), which appears to dimerize[41] by nucleophilic attack
on a second carbenoid 48 that responds by ring opening,
perhaps in the sense of an intermediate 50 (an unconfirmed
explanation). As usual for a 2-(1-lithioalkyl)oxirane, 50
would be expected[42] to open its oxirane ring, producing
the allylic diol 49. The (E) configuration of 49 was estab-
lished through the NMR coupling constant 3JH,H = 15.7 Hz
1
as extracted from the 13C satellites of the olefinic H NMR
singlet.
Scheme 8. Reductive enolization of 19 by treatment with tBuMgCl
in THF. See the text for discussions of routes A–F.
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6655

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
The simplest mechanistic interpretation in terms of a Cl/
Mg interchange reaction (pathway C) leading directly to the
Mg enolate 45 must be dismissed, because the expected by-
product tert-butyl chloride (tBuCl, δH = 1.59 ppm) was not
detected in situ even though it would have been stable under
the reaction conditions in the presence of tBuMgCl over a
period of at least 5 d, as demonstrated in a separate run.
The remaining mechanistic possibility (pathway D), a sin-
gle-electron transfer (SET) from tBuMgCl to 19, leads to
the aforementioned “rather persistent”[40] radical 44. The
concomitant tert-butyl radical (tBu·) might be expected to
transfer a hydrogen atom to 44 (branch E in Scheme 8) with
formation of the aldehyde 8, which would have to be depro-
tonated by tBuMgCl to explain the enolate 45. However, 8
was not detected in situ even though it is stable against
tBuMgCl under the reaction conditions for a much longer
time: a separate experiment (bottom line of Scheme 8)
showed that tBuMgCl did not deprotonate 8 to afford 45
but generated the tBu addition product 52 and the hydride
transfer product 10 very slowly over the course of 5 d. Con-
sequently, the radical 44 appears to have been converted
into the enolate 45 (rather than into 8) with consumption
of a second equivalent of tBuMgCl (branch F). This pro-
cess may be envisioned as a second SET from tBuMgCl
with release of another equivalent of tBu·. (Note that two
tBu· radicals can be expected to disproportionate with for-
mation of isobutane and isobutene.) Alternatively, the first
tBu· to be generated might transfer a hydrogen atom to the
oxygen atom of 44 with generation of isobutene and
tBu2C=CH–OH, whereupon a quick deprotonation of this
enol by tBuMgCl would give isobutane and 45. The simple
strategy of in situ NMR spectroscopy can thus help to ex-
clude some of the mechanistic alternatives in this system.
Scheme 9. NaBH4 reduction of 19 and ensuing transformations; dg
= diglyme. See the text for discussions of steps 1–13.
This interfering formation of oxirane 43 can be avoided
by use of dry diglyme as the solvent; it dissolves enough
tert-Butyllithium (tBuLi) in either THF or benzene re- NaBH4 and does not permit the β-chloro alkoxide 53 to be
acted more rapidly but less cleanly with 19 to furnish the released from the conjectured hydridoborate 54. In a satu-
Li enolate 47 (δH = 6.82 ppm in THF; later trapped as 46) rated solution of NaBH4 (2 equiv.), step 1 was sufficiently
which could arise from a Cl/Li interchange reaction (similar fast to allow the chlorohydrin 55 at 20 °C to be isolated
to pathway C in Scheme 8) or by the D/F (but not the D/ after roughly 4 h (step 4) before more than traces of the side
E) route shown in Scheme 8; aqueous workup afforded the product tBu2CH–CH2OH (10) could be detected (1H NMR
aldehyde 8. There was no unequivocal evidence for the ana- in situ). The non-appearance of oxirane 43 is due to the O-
logues of pathways B (hydride transfer) or A (to generate protecting BH3 group in 54 (or a similar hydridoborate)
ketene 14 and possibly its addition product[43] to tBuLi).
and not to the changed solvent per se: this was established
Steric congestion does not prevent the hydride transfer by treating the chlorohydrin 55 in diglyme with NaOCH3
from NaBH4 to the aldehyde function of 19 (step 1 in (step 5), which afforded only 43 (via 53). When NaBH4 was
Scheme 9). This system is remarkable for its peculiar prod- used as the base in place of NaOCH3 (step 6), no traces of
uct dependence on the solvent and the temperature: the for- 43 could be detected. Instead, the O-protected intermediate
mation variously of 10 or of 43 or of 55 (and thence 8) from 54 (generated by steps 6 or 1) reacted in the presence of
19 in Scheme 9 can be controlled as follows. The hydrido- residual NaBH4 on heating to 150 °C for 3 h (step 7), poss-
borate anion 54, as generated in alkalized methanol ibly via the SN1-type intermediate 56 that can be expected
(step 1), is converted at 20 °C (step 2) into the β-chloro alk- to shift a hydride anion to C-2 in step 8 either from the
–
oxide 53 (1H NMR in situ) and then cyclized (step 3) at a OBH3 group or from C-1 (with a subsequent second hy-
somewhat slower rate to produce the oxirane 43 within dride transfer from a B–H function to C-1): the dechlorin-
20 h. With use of a large excess (4 mol-equiv.) of NaBH4 in ated alcohol tBu2CH–CH2OH (10) was obtained (19%
methanol, with the goal of generating 54 within 1 h or less, yield) after workup. Proton abstraction from C-1 of 56 or
the intended preparation (workup step 4) of the rather un- HCl elimination from 54 in the basic milieu as an alterna-
stable chlorohydrin 55 was unsatisfactory, even though the tive to step 8 can be dismissed, because the resulting O-
very volatile byproduct 43 could be easily removed by protected BH3 enolate of aldehyde 8 would be unable to
evacuation.
form the final product 10.
6656
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664

Syntheses of Di-tert-butylketene and Related gem-Di-tert-butyl Compounds
The expulsion of chloride anion from the chlorohydrin chloro-3,3-dimethylbutanal (63), as identified through its
55 (step 9) displays a characteristic solvent dependence: the 1H NMR spectrum,[47] along with tBuBr (δ = 1.78 ppm)
first half-reaction times[44] (t1/2) for the disappearance of 55 and CH3Br (δ = 2.58 ppm). Neither PCl5 nor PPh3 in hot
are ca. 3 h at 55 °C in methanol and ca. 1 h at 149 °C in CCl4 reacted with 19, but F3C–CO2H in CDCl3 solution at
diglyme solution, to be compared with the reported[45] t1/2 70 °C converted 19 slowly into 63, presumably by de-tert-
= 1.1 h for the SN1-type ionization of 58 to generate 59 at butylation (61 Ǟ 62).
25 °C in aqueous ethanol (20:80). The expected SN1 inter-
mediate 57 does not cyclize to give 43 but appears to prefer
D. Descendants of Di-tert-butylacetaldehyde (8)
the obvious possibility of a stabilizing hydride shift from C-
1 to C-2 (step 10), leading to the dechlorinated aldehyde 8
as the only product (1H NMR in situ). In view of such
properties, the oily key compound 55 cannot be purified by
a slow fractionating distillation, because that would gener-
ate HCl and 8 by steps 9 and 10.
The aldehyde 8 is almost instantaneously deoxygenated
by the catechyl phosphorus tribromide 64 (Scheme 11),
which is approved[48] for the preparation of 1,1-dibromides
such as the desired tBu2CH–CHBr2. However, the repulsive
strain inherent in the gem-tBu2 fragment appears to lead 8
away from the intended course, producing instead the de-
tert-butylated alkenyl bromide 65 as the pure (Z) isomer,[49]
which was slowly converted into the (E) isomer 66 under
the reaction conditions.
The Mg β-chloro alkoxide 51, generated (step 11,
Scheme 9) at 20 °C from the chlorohydrin 55 on treatment
with tBuMgCl (2.7 equiv.) in THF in Ͻ2 min, would ap-
pear to have the options for at least four transformations:
to 8, 10 (by tBuMgCl), the oxirane 43, and tBu2C=CH–
1
OMgCl (45). Perhaps surprisingly, the H NMR signals at
δ = 1.24 (broad) and 1.22 ppm (broader) assigned to 51
(two species?) disappeared rapidly at 20 °C (step 12, first
t1/2 ≈ 14 min)[44] with formation of the aldehyde 8 as the
only product observable in situ over the first 97 min. As is
already familiar from the bottom line of Scheme 8, residual
tBuMgCl and 8 reacted subsequently very slowly (at 20 °C,
Ͼ4 d) in step 13 to provide 52 and 10 but no trace of the
oxirane 43. The enolate tBu2C=CH–OMgCl (45) cannot
have been the precursor of 8, in view of the absence of acids
in the presence of residual tBuMgCl. The O-protecting
group MgCl has thus saved 51 from cyclization (to give 43)
and has facilitated step 12 (at 20 °C in THF) in preference
to step 7 (at up to 150 °C) and steps 9/10 (at 149 °C in di-
glyme), perhaps because of intramolecular assistance to C–
Cl bond heterolysis by the MgCl cation.
Scheme 11. Some transformations of 2-tert-butyl-3,3-dimethylbut-
anal (8).
The aldehyde function of 19 is not prone to autoxidation,
yet is readily brominated. The product tBu2CCl–CO–Br
The propensity of the aldehyde 8 to autoxidize is an in-
(15, shown in Scheme 1)[10] was obtained here (Scheme 10) convenient property that interferes with the preparation of
on treatment with PBr5 in boiling CCl4 (yield 21% in 14 h); the oxime 67 (Scheme 11). Although we were able to con-
it was identified through comparison with the published[10] firm the reported[6] yield of 67 (72% in 18 h), we preferred
13C NMR chemical shifts. In a second futile attempt di- to employ the silyl enol ether 46 under milder conditions
rected towards 1,1,2-trihalogeno derivatives 60, we heated that furnished almost pure, undistilled 67 (87% in only 2 h).
19 with dibromomethoxymethane[46] and obtained only 2- This significant acceleration suggests that a proton transfer
reaction from HO–NH3+AcO– to 46 generated 68 as an ac-
tivated form of aldehyde 8. The usual hydrolysis of 68 in
aqueous ethanol to give 8 was overcome here by the faster
addition of HO–NH2. The previously unreported dehy-
dration of oxime 67 by the usual simple treatment with
SOCl2 (1.1 equiv.) in warm benzene afforded the nitrile 69
(yield 74% after distillation). With a total of five steps (or
Ͻ5 preparative stages), this preparation of 69 from ketone
1 is shorter and more convenient than the traditional syn-
thesis.[4] Our attempts to prepare nitrile 69 through some of
the recommended one-pot procedures, variously from the
acid 11 with chlorosulfonyl isocyanate and then DMF[50]
(decomposition), or from aldehyde 8 with hydroxylamine-
O-sulfonic acid[51] (no conversion), or with HONH3+Cl– in
Scheme 10. 2-tert-Butyl-2-chloro-3,3-dimethylbutanal (19) in reac-
tions with electrophiles.
DMF at reflux[52] (basic side products), were all unsatisfac-
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6657

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
tory. The shortcut treatment of 8 with HONH3+Cl– in hot Cl (13, Scheme 4) reacted with tBu2CH–CO2Li (39) via
formic acid[53] furnished some strongly contaminated nitrile tBu2C=C=O (14) plus tBu2CH–CO2H (11) rather than by
69, accompanied by N-(di-tert-butylmethyl)formamide (70) addition/elimination at 95 °C.
in a (Z)/(E) ratio of ca. 74:26 after distillation. Such a
The ketene 14 did not react with LDA or with tBuMgCl
Beckmann rearrangement competing with the dehydration (at room temp., Ͼ12 d), yet it can acylate the following nu-
of an aldoxime is only rarely observed.[54] The spontane- cleophiles with synthetically acceptable rates at widely dif-
ously crystallizing[9] (Z) isomer was recognized through its fering temperatures: the alkoxide tBu–CH(OK)–CO–tBu
3
small JH,H NMR coupling constant of 2.0 Hz [for the cis (34) of pivaloin in DMSO at room temp. slowly (Scheme 3),
H–(O=)C–N–H fragment] and its upfield 13C NMR signal a suspension of KOH in DMSO at 55 °C much more slowly,
3
of CH-3 (δ = 62.0 ppm), in relation to JH,H = 11.6 Hz pivaloin (38) in diglyme at 150 °C within 21 h (Scheme 3),
(trans) and δ = 68.9 ppm for CH-3 of the previously[9] not tBu2CH–CO2H (11) in toluene at 95 °C in 3 h (Scheme 4),
reported (E) isomer. The slow formation of (E)-70 in CDCl3 or H2O in the presence of concd. HCl at room temp. within
solution is consistent with observations on other sterically 2 h. Because the uncatalyzed addition of 14 to adventitious
congested amides.[55] The (Z)/(E) equilibrium mixture of traces of H2O in hot DMSO or hot diglyme can be a dis-
59:41, obtained after 6 h at 20 °C, might be due to a turbing side-reaction in small-scale runs, it is advisable to
stronger ecliptic repulsion of the oxygen atom by the employ less polar aprotic solvents (such as toluene) which
tBu2CH group in relation to other N-substituents, such as are kept dry more easily. Electrophiles tend to de-tert-butyl-
tBu [78–82% (Z)][56,57] and H3C [92% (Z)][57] in other for- ate gem-di-tert-butyl compounds (Schemes 2, 10, and 11). A
mamides.
very simple ET reduction of tBu2CCl–CH=O (19) employs
lithium metal (Scheme 6) to produce tBu2C=CH–OSiMe3
(46), which is a useful precursor of the activated form 68 of
aldehyde 8, as shown by its fast and clean condensation
Conclusions
+
reaction with HO–NH3 AcO– in aqueous ethanol
The gem-di-tert-butyl compounds studied in this work (Scheme 11) to provide the oxime 67 and from this the ni-
are remarkable for some unusual properties and reactivity trile 69.
patterns. The surprisingly volatile key substance tBu2CCl–
CH=O (19) appears to arise (Scheme 2) through a hetero-
Experimental Section
lytic C–O bond fission that is accelerated by steric over-
crowding to such an extent that the nonpolar toluene can
be recommended as a most suitable solvent.
NMR Spectroscopy in situ: The progress of many of the reactions
described above was observed in situ by H NMR spectroscopy at
1
The unusual elimination of HCl from tBu2CCl–CH=O
(19) by LDA (Scheme 2) or (much more slowly) by KOH
in DMSO (Scheme 3) to produce the ketene tBu2C=C=O
(14) is interpreted as an E2-type elimination, because we
were unable to ionize 19 by an uncatalyzed heterolysis reac-
tion (types E1 or SN1) of the C–Cl bond in hot DMSO
(148 °C) or in methanol. As soon as the aldehyde group of
19 was disabled through the addition of the small nucleo-
philes OH– (Scheme 3) or H– (Scheme 9) or through an
electron transfer (ET, Schemes 6 and 8), the expulsion of
Cl– from the tBu2CCl group became possible and exhibited
a solvent dependence (Scheme 9, step 9) that points to a
certain carbenium character before product formation.
However, a bulkier nucleophile such as tBuMgCl at room
temp. required 5 d (bottom lines of Schemes 8 and 9) for
its addition (52) and hydride transfer (10) to the aldehyde
function of tBu2CH–CH=O (8). A comparably strong
200 MHz and 20–23 °C, with use of NMR tubes (outer diameter
5 mm) that were either tightly stoppered and sealed with stretchable
plastic foil or were closed with a glass stopper in a gas-tight
ground-glass joint. Occasional final in situ analyses by 1H
(400 MHz) and 13C NMR (100.6 MHz) spectroscopy served to
confirm the NMR signal assignments of the primary products be-
fore workup. Most of these runs were carried out in nondeuterated
solvent (0.5–1.5 mL) containing [D12]cyclohexane (as the “lock
substance”, at least 0.03 mL), a trace of TMS (δ = 0 ppm), and the
reagents (Ն0.1 mmol). Thanks to the strong 18-proton singlets of
the starting and final compounds, the gem-tBu2 system proved well
suited for studies in the solvents diglyme, THF, methanol, DMSO,
toluene, and also benzene or CCl4. However, the δH values of the
tBu2CH groups depend slightly on the solvent, so detailed δH col-
lection was essential for the differential in situ analyses and is re-
produced in part below (excluding those solutions that are used for
characterization in the pertinent paragraphs).
δH Values (ppm): tBuMgCl/tBu2Mg 0.90 and 0.88 in THF, consis-
retardation is to be expected with tBu2CCl–CH=O (19) tent with ref.[58] tBu2CH–CO2H (11) 1.10/2.11 in diglyme, 1.13/2.18
in diglyme containing pyridine, 1.07/2.06 in DMSO, 1.06/2.15 in
toluene, 1.14/2.15/11.3 in CCl4; tBu2CH–CO–Cl (13) 1.17/2.90 in
diglyme, 1.18/2.84 in CCl4, 0.99/2.70 in toluene; tBu2C=C=O (14)
1.21 in diglyme or THF, 1.22 in CCl4; tBu2C(Cl)–CH=O (19) 1.20/
9.87 in [D6]DMSO, 1.23/9.82 in diglyme, 1.24/9.77 in toluene or
CCl4 (consistent with ref.[10]); tBuCH(OK)–CO–tBu (34) 0.89/1.16/
4.23–4.37 in THF; tBuCH(OLi)–CO–tBu 0.87/1.15/4.28–4.36 (br)
in THF; tBuCH(OH)–CO–tBu (38) 0.94/1.15/4.02 in THF or CCl4,
0.98/1.19/4.21 in CDCl3; tBu2CH–CO2Li (39) 1.17/hidden in tolu-
ene; tBu2CH–CO2Na 1.08/1.99 in diglyme; tBu2CH–CO2K 1.04/
and provided an opportunity to study the rarely observed
ET reaction from tBuMgCl (Scheme 8): the enolate
tBu2C=CH–OMgCl (45) was produced at room temp. via
the radical 44 within 35 min, while alternative mechanistic
1
possibilities were excluded through H NMR spectroscopy
in situ. The replacement of X = H by non-hydrogen atoms
X in tBu2CH–CO–X can impede carbonyl addition reac-
tions still further: the esters 36 and 41 and the anhydride
(tBu2CH–CO)2O (40) were shown to resist alkaline hydroly-
sis partially (36) or completely at 95 °C, and tBu2CH–CO– 2.05 in DMSO; (tBu2CH–CO)2O (40) 1.13/2.14 in diglyme without
6658
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664

Syntheses of Di-tert-butylketene and Related gem-Di-tert-butyl Compounds
or with DMSO, 1.10/hidden in toluene, 1.13/2.07 in CCl4. The ex-
periments with volatile compounds (but without the evolution of
significant amounts of gases) were performed in tightly closed ves-
sels at temperatures well below the b.p. of the solvent.
2-tert-Butyl-3,3-dimethylbutanoyl Chloride (13): The acid 11
(276 mg, 1.60 mmol) in dry CCl4 (2.0 mL) was heated with SOCl2
(0.180 mL, 2.48 mmol) under a dry double-walled reflux condenser
(fitted with an N2 bubbler) at 55 °C for up to 43 h. Residual SOCl2
and CCl4 were removed in a rotary evaporator under ca. 50 mbar
and 60 °C (ref.[4] b.p. 83–86 °C/12 Torr) for 3 min, leaving pure 13
2-tert-Butyl-3,3-dimethylbutanal (8) from 46: A solution of 2-tert-
butyl-3,3-dimethyl-1-(trimethylsiloxy)but-1-ene
(46,
229 mg,
1
(310 mg, 101%). H NMR (CDCl3): δ = 1.18 (s, 18 H), 2.90 (s, 2-
1.00 mmol) in methanol (0.4 mL) was stirred with aqueous HCl
(2 m, 0.1 mL) under argon for 40 min and was then poured into
distilled water (5 mL) and extracted with Et2O (3ϫ 8 mL). The
combined ethereal layers were washed with water until neutral,
dried with Na2SO4 under argon, filtered, and concentrated under
reduced pressure. The complete removal of Et2O caused some loss
of the slightly yellow liquid, which consisted entirely of the alde-
hyde 8 (crude yield 119 mg, 76%). 1H NMR (CCl4): δ = 1.08 (s, 18
H), 1.70 and 9.67 (AB system with 3J = 6.0 Hz) ppm. 1H NMR
(CDCl3): δ = 1.10 (s, 18 H), 1.75 and 9.77 (AB system with 3J =
6.3 Hz) ppm, consistent with ref.[59] 13C NMR (CDCl3): δ = 30.57
H) ppm.[35] 13C NMR (CCl4):
δ = 30.52 (2ϫCMe3), 36.35
(2ϫCMe3), 77.67 (C-2), 173.83 (C=O) ppm.
2-tert-Butyl-3,3-dimethylbut-1-en-1-one (14)
(a) From 19: A solution of LDA (16.5 mmol) in THF (10 mL) and
hexane (6.6 mL) was prepared in a Schlenk flask (50 mL) with
magnetic stirring at 3 °C under argon and then cooled to –70 °C.
The α-chloro aldehyde 19 (3.00 g, 15.7 mmol) in dry THF (5 mL)
was added dropwise over 20 min. After another 15 min, the mixture
was warmed up in an ice bath and acidified with HCl (2 m, 10 mL)
containing a little ice, then diluted with distilled water and ex-
tracted with Et2O (3ϫ 25 mL). The combined Et2O phases were
washed until neutral, dried with K2CO3 or Na2SO4, and carefully
concentrated to leave an orange-colored liquid (2.15 g). This almost
pure ketene (14) was distilled at room temp. und 0.03 mbar into a
trap cooled to –70 °C to give pure 14 (1.67 g, 69%) as a yellow
1
(qoct, J = 125.4 Hz, 2ϫCMe3), 34.14 (apparent sept, J = 3.8 Hz,
2ϫCMe3), 68.78 (ddm, 1J = 125.5, 2J = 22.5, 3J = 3.5 Hz, C-2),
206.61 (dd, 1J = 167.1, 2J = 8.0 Hz, C=O) ppm. IR (film): ν =
˜
2959, 2724 (w, OC–H), 1721 (C=O), 1478, 1370 cm–1.
2-tert-Butyl-3,3-dimethylbutan-1-ol (10): The α-chloro aldehyde 19
(314 mg, 1.65 mmol) and NaBH4 (190 mg, 5.02 mmol) in dry di-
glyme (3.0 mL) were heated at 150 °C for 3.5 h (copious precipi-
tate). After 2 d at room temperature (room temp.), the mixture was
dissolved in Et2O/H2O (slow evolution of H2 for 1 h). The Et2O
layer was washed first with water and then with HCl (2 m), shaken
with water (5ϫ for complete removal of diglyme), dried with
Na2SO4, and concentrated under 60 mbar at room temp. for 10 s
to furnish the highly contaminated alcohol 10. The only distillable
components of this material were 10 (49 mg, 19%) and a trace of
the aldehyde 8; b.p. of 10: 127–132 °C (bath temp.)/4 mbar. 1H
NMR (CDCl3): δ = 1.06 (s, 18 H), ca. 1.08 (hidden, 2-H), 3.73 (d,
1
liquid. H NMR (CDCl3): δ = 1.21 ppm.[35] 13C NMR (CDCl3): δ
= 31.3 (2ϫCMe3), 32.3 (2ϫCMe3), 52.2 (C-2), 203.9 (C=O) ppm,
consistent with ref.[10]. IR (film): ν = 2086, 1469, 1368, 1227 cm–1.
˜
(b) From 18: The chlorooxirane 18 (2.45 g, 12.8 mmol) in dry tolu-
ene (20 mL) was heated at 110 °C for 45 min and then cooled but
not worked up. Instead, the resulting mixture (containing 19) was
pipetted into a cooled solution (–70 °C) of LDA (25.7 mmol) in
THF (10 mL) and hexane (20 mL). After another 15 min, this yel-
low mixture was warmed up in an ice bath, acidified with ice-cold
HCl (2 m, 45 mL), and extracted with Et2O (3ϫ 15 mL). The com-
bined Et2O layers were washed neutral, dried with K2CO3, and
concentrated under 25 mbar to provide an orange-colored liquid
(925 mg) containing 14 (886 mg, 45%) and toluene only. This mate-
rial was used without purification; analogous specimens were dis-
tilled at 76–81 °C/80 mbar or at 30–35 °C/40 Torr (ref.[4] b.p. 74–
76 °C/47 Torr). It was difficult to distill the last portions of 14 out
of the residue.
1
3J = 3.5 Hz, CH2O) ppm. H NMR (400 MHz, [D6]benzene): δ =
1.01 (t, 3J = 3.5 Hz, 2-H), 1.03 (s, 18 H), 3.54 (d, 3J = 3.5 Hz,
CH2O) ppm. 13C NMR (CDCl3): δ = 30.97 (2ϫCMe3), 35.57
(2ϫCMe3), 60.51 (C-2), 63.68 (CH2O) ppm, in reasonable accord
with ref.[60] 13C NMR (100.6 MHz, [D6]benzene): δ = 31.27, 35.62,
59.94, 63.14 ppm.
2-tert-Butyl-2-chloro-3,3-dimethylbutanoyl Bromide (15): An NMR
tube charged with the α-chloro aldehyde 19 (50 mg, 0.26 mmol)
and PBr5 (1.31 mmol) in CCl4 (0.5 mL) was heated at 70 °C for
14 h for total consumption of 19. The mixture was dissolved in
Et2O and extracted with NaOH (2 m). The Et2O layer was washed
until neutral, dried with Na2SO4, and concentrated in vacuo. 1H
NMR (CDCl3): δ = 1.31 (s) ppm (ref.[10] δ = 1.05 ppm in CCl4).
13C NMR (CDCl3): δ = 30.12 (qsept, 1J = 127.0, 3J = 4.7 Hz,
2ϫCMe3), 45.15 (m, 2J = 3.7 Hz, 2ϫCMe3), 99.12 (m, C-2),
2-tert-Butyl-3,3-dimethylbutanoic Acid (11): A solution of the chloro-
oxirane 18 (1.91 g, 10.0 mmol) in dry toluene (18 mL) was heated
at 110 °C for 45 min and was then cooled and added dropwise at
–70 °C to LDA (20.3 mmol) in dry THF (10 mL) and hexane (ca.
10 mL). After 20 min at –70 °C, the mixture (containing the crude
ketene 14) was stirred with concentrated HCl (15 mL) at room
temp. for 2 h and then diluted with distilled water and Et2O. The
aqueous layer was shaken with more Et2O (2ϫ 10 mL) and dis-
carded. The combined Et2O phases were extracted with NaOH
(2 m, 3ϫ 10 mL) and then discarded. These alkaline aqueous
phases were combined, acidified with conc. HCl, extracted with
Et2O (3ϫ 15 mL), and discarded. The final three Et2O extracts
were combined, washed with distilled water until neutral, and dried
with Na2SO4, and the solvents were evaporated to dryness, yielding
crude 11 (1.14 g, 66%) with m.p. 69–72 °C (ref.[4] 72–74 °C or 80.5–
174.75 (s, C=O) ppm, in rough accord with ref.[10] IR (film): ν =
˜
1780 cm–1 (C=O; ref.[10] 1765 cm–1 in CCl4).
2,2-Di-tert-butyl-3-chlorooxirane (18): A two-necked flask (500 mL)
was fitted with an internal thermometer, a magnetic stirring bar,
and a large pressure-equalizing addition funnel with an argon bub-
bler. The flask was loaded with dry THF (70 mL), di-tert-butyl
ketone (1, 17.06 g, 120.0 mmol),[3] and dry dichloromethane
(23.16 mL, 30.69 g, 361.4 mmol) and was then cooled with stirring
at –70 °C. A solution of LDA (241 mmol), prepared by addition of
nBuLi (103 mL in hexanes, 240 mmol) to diisopropylamine
1
81.5 °C). H NMR (CDCl3): δ = 1.13 (s, 18 H), 2.20 (s, 2-H), 11.5
1
(br, OH) ppm. [35] 13C NMR (CDCl3): δ = 30.7 (qoct, J = 125.5,
3J = 4.6 Hz, 2ϫCMe3), 34.5 (m, 2J = 4.0 Hz, 2ϫCMe3), 64.5 (dm,
2
1J = 128 Hz, C-2), 180.5 (d, J = 8.3 Hz, C=O) ppm. IR (KBr): ν
˜
= 3600–2500 (br, O–H), 2959, 1704 (C=O), 1476 (w), 1371, 1245, (241 mmol) in dry THF (70 mL) at 3 °C, was added (not vice
1177, 939 (w), 712 (w) cm–1. An analogous preparation starting
with the addition of a THF solution of the α-chloro aldehyde 19
(in place of 18) to LDA yielded 11 (80%) with m.p. 71.5–73 °C.
versa!) dropwise with stirring from the addition funnel at such a
rate that the internal temperature remained below –50 °C (60 min).
After another 15 min, the cooling bath may be replaced by an ice
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6659

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
bath, whereupon the orange reaction mixture will turn into a black
suspension. The mixture was quenched by pouring it into a separat-
ing funnel (1 L) containing ice-cold aqueous HCl (2 m, 300 mL).
After extraction with Et2O (4ϫ 100 mL), the combined Et2O ex-
tracts were washed until neutral and dried with Na2SO4. The Et2O
was removed in a rotary evaporator at room temp. under 40 mbar
to leave pure 18 (22.59 g, 99%) as a brown liquid. This rather vola-
tile material should not be distilled but should be stored at –18 °C
in order to avoid its conversion into 19 and 27. Purification is un-
necessary for the subsequent transformations but may be per-
formed by a trap-to-trap distillation at 30 °C (bath temp.)/
0.02 mbar into a receiving trap cooled to –70 °C, providing 18 as a
2-tert-Butyl-2,3-dimethylbut-3-enal (27): This compound was ob-
tained along with 19, b.p. 77–108 °C (bath temp.)/4 mbar (ref.[10]
60–65 °C/12 Torr). 1H NMR (CDCl3): δ = 1.03 (s, 2-CMe3), 1.16
2
(s, 2-CH3), 1.84 (dd, |4J| = 1.43 and 0.68 Hz, 3-CH3), 4.74 (dq, J
2
≈ 0.8, |4J| = 0.68 Hz, 4-H trans to 3-CH3), 5.16 (dq, J ≈ 0.8, |4J| =
1.43 Hz, 4-H cis to 3-CH3), 9.80 (s, CH=O) ppm, assigned through
comparison with 4J(trans) = –0.6 Hz and 4J(cis) = –1.5 Hz of 2,3,3-
trimethylbut-1-ene.[61] 13C NMR (CDCl3): δ = 15.3 (2-CH3), 23.4
(3-CH3), 26.8 (2-CMe3), 35.8 (2-CMe3), 58.3 (C-2), 116.4 (CH2-4),
144.4 (C-3), 204.6 (C-1) ppm, assigned through HSQC of the pro-
ton-bearing 13C nuclei and comparisons of CMe3 and C-2 with
those[21] of 30. FT-IR (film on diamond, ATR): ν = 2959, 1722,
˜
1
viscous, colorless liquid. H NMR (CCl4 or CDCl3): δ = 1.08 and
1368 cm–1.
1.28 (2ϫs, 2 ϫCMe3), 5.22 (s, 3-H) ppm. 13C NMR (CCl4/[D12]-
cyclohexane, 10:1): δ = 28.4 and 30.2 (2ϫqsept, 1J = 126, 3J =
4.9 Hz, 2ϫCMe3), 36.7 (m, 2J = 3.8 Hz, CMe3 cis to Cl?), 37.4 (br
3-tert-Butyl-4,4-dimethylpent-2-enoic Acid (28): The acid fraction of
the photolysis products of the oxetanone 25 was filtered through
silica gel (60 Å, 171 mg) with light PE and then crystallized from
pentane at –18 °C to give colorless blocklets (21 mg, ~1%) with
m.p. 98–99 °C. 1H NMR (CDCl3): δ = 1.25 and 1.38 (2 ϫ s,
2ϫCMe3), 5.84 (s, 2-H), 10.8 (br, CO2H) ppm. 13C NMR (CDCl3):
δ = 31.63 and 32.34 (2 ϫ CMe3), 38.60 (br, 1 ϫ CMe3), 39.45
(1ϫCMe3), 115.57 (C-2), 166.76 (C-3), 174.78 (br, CO2H) ppm.
m, 2J = 3.7 Hz, CMe3 trans to Cl?), 70.1 (m, apparent J ≈ 4 Hz,
1
C-2), 72.7 (d, J = 215.2 Hz, C-3) ppm. IR (film): ν = 2971, 2930,
˜
1490 (s), 1394 (s), 1370 (s), 1343, 982, 951, 841, 789, 767 cm–1.
C10H19ClO (190.7): calcd. C 62.98, H 10.04; found C 63.52, H
10.02.
2-tert-Butyl-2-chloro-3,3-dimethylbutanal (19): The chlorooxirane
18 (9.536 g, 50.00 mmol) in dry toluene (75 mL) was heated under
a double-walled reflux condenser with drying tube at 110 °C for
45 min. Concentration in a rotary evaporator down to a final pres-
sure of 8 mbar at 60 °C afforded partially solidifying, brownish 19
(8.90 g, 93%) containing a trace of toluene. A sample distilling at
88–92 °C/15 Torr (ref.[10] b.p. 85 °C/12 Torr) with vigorous foaming
furnished waxy, block-shaped crystals with m.p. 81–84 °C (ref.[10]
83–85 °C). Because of its volatility, 19 should not be stored in an
FT-IR (diamond, ATR): ν = 3200–2400 (br, O–H), 1688 (s), 1615,
˜
1254 (s) cm–1. C11H20O2 (184.28): calcd. C 71.70, H 10.94; found
C 71.44, H 11.06.
4-tert-Butyl-2-[(Z)-2Ј,2Ј-dimethyl-1Ј-propylidene]-4,5,5-trimethyl-
tetrahydrofuran-3-one (29): The oxetanone 25 (700 mg, 2.27 mmol)
was heated overnight at 80 °C in a mixture of glacial acetic acid
(14 mL) and concentrated HCl (1.4 mL). The resulting mixture was
poured into distilled water (50 mL) and extracted with Et2O (4ϫ
15 mL). The combined extracts were shaken with NaOH (2 m, 3ϫ),
washed with water until neutral, dried with Na2SO4, and concen-
trated to yield 29 as a brownish liquid (469 mg, 82%) that solidified
slowly (m.p. 31–37 °C). Colorless platelets of analytically pure 29
were obtained from a pentane solution at –70 °C. M.p. 47–48 °C,
b.p. 150–170 °C (bath temp.)/12 mbar. 1H NMR (CDCl3): δ = 1.055
(s, 4-CH3), 1.060 (s, 4-CMe3), 1.143 (s, 1Ј-CMe3), 1.250 (s, 5-CH3
trans to 4-CMe3), 1.512 (s, 5-CH3 cis to 4-CMe3), 5.442 (s, 1Ј-
H) ppm, assigned through the NOESY correlations 1Ј-H ↔ 1Ј-
CMe3 and 4-CMe3 ↔ cis-5-CH3 ↔ trans-5-CH3 ↔ 4-CH3. 13C
NMR (CDCl3): δ = 16.1 (sharp q, 1J = 128.6 Hz, 4-CH3), 25.7 (qq,
1J = 127.0, 3J = 4.2 Hz, 5-CH3 cis to 4-CMe3), 27.1 (qq, 1J = 126.7,
1
evacuated desiccator. H NMR (CDCl3): δ = 1.25 (s, 18 H), 9.83
(s, CHO) ppm.[35] 13C NMR (CDCl3): δ = 29.67 (2ϫCMe3), 42.32
(2ϫCMe3), 93.20 (C-Cl), 200.71 (CH=O) ppm, in rough accord
with ref.[10] IR (film): ν = 1732 cm–1. A sample of 19 in [D ]DMSO
˜
6
solution remained unchanged over at least 8 h at 148 °C. A less
pure sample (obtained from 47.2 mmol of 18) was adsorbed on a
column of silica gel (90 g, 60 Å, ICN 63–200) and eluted with low-
boiling petroleum ether (PE) containing Et2O (2%). After a forerun
of 150 mL, the next 250 mL of eluent furnished pure 19 (6.01 g,
67%), followed by the elimination product 27.
4,4-Di-tert-butyl-2-(1-tert-butyl-2,2-dimethyl-1-propylidene)oxetan-
3-one (25): Di-tert-butyl ketone (1, 2.00 g, 14.1 mmol) and CH2Cl2
(0.920 mL, 14.4 mmol) were dissolved in dry THF (8 mL) and co-
oled to –70 °C with stirring. A solution of LDA (56.2 mmol) in dry
1
3
3J = 4.2 Hz, 5-CH3 trans to 4-CMe3), 27.6 (br qm, J = 126, J ≈
4.8 Hz, 4-CMe3), 29.5 (sharp qdm, 1J = 126.3, 3J = 3.4, 3J =
4.8 Hz, 1Ј-CMe3), 31.21 (dm, 2J = 1.6, 2J = 4.0 Hz, 1Ј-CMe3), 35.3
THF (16 mL) and hexanes (ca. 15 mL) was added dropwise over (m, apparent J = 3.7 Hz, 4-CMe3), 57.2 (septm, 3J = 2.4 Hz to both
15 min. After further stirring without cooling until the internal
temperature reached 0 °C (Ͻ4 h), the mixture had become orange-
colored and was poured into aqueous HCl (2 m, 80 mL) and then
extracted with Et2O (3 ϫ 25 mL). The combined extracts were
washed with distilled water until neutral, dried with Na2SO4, and
concentrated in vacuo, providing 25 as an almost pure yellow solid
(2.32 g) that was recrystallized from pentane at –20 °C to afford
colorless crystals of pure 25 (yield 89 %) with m.p. 45.5–47 °C
(ref.[18] 46–47 °C). 1H NMR (CDCl3): δ = 1.14 (s, 2ϫ4-CMe3), 1.30
5-CH3 groups, C-4), 86.9 (m, apparent J = 4.1 Hz, C-5), 116.7 (dm,
3
2
1J = 156.6, J = 4.5 Hz, C-1Ј), 146.2 (sharp d, J = 4.4 Hz, C-2),
3
3
205.7 (dq, J = 3.3, J = 3.8 Hz, C-3) ppm, assigned in the follow-
ing manner. The olefinic carbon nuclei C-1Ј and C-2 were recog-
nized through their coupling patterns as given above. The only CH3
1
group showing a sharp J quartet (no 3J coupling) must be 4-CH3.
The other methyl groups were unambiguously assigned through se-
lective {1H} decoupling experiments, which also provided the above
3
2J and J values as follows: {1Ј-H} Ǟ 1Ј-C(CH3)3 simplified to a
3
and 1.40 (2ϫs, 2 ϫ1Ј-CMe3) ppm. 13C NMR (CDCl3): δ = 28.3 quartet (1J) of septets with J ≈ 4.8 Hz, 1Ј-CMe3 as a sharpened m
(qsept, 1J = 125.9, 3J = 4.9 Hz, 2 ϫ 4-CMe3), 31.0 and 32.5
with J = 4.0 Hz, C-2 as a sharp s, and C-3 as a sharp q with J =
2
3
1
3
(2ϫqsept, J = 126.2, J = 4.9 Hz, 2ϫ1Ј-CMe3), 37.22 and 38.44
3.8 Hz; {cis-5-CH3} Ǟ C-5 simplified, trans-5-CH3 as a sharp 1J q
(2ϫm, 2J = 3.7 Hz, 2ϫ1Ј-CMe3), 38.37 (m, 2J = 3.8 Hz, 2ϫ4- (3J removed); {trans-5-CH3} Ǟ C-5 simplified, cis-5-CH3 as a
3
3
1
CMe3), 109.1 (m, J ≈ 3.7 Hz, C-4), 136.5 (m, J ≈ 3.6 Hz, olefinic
C-1Ј), 161.2 (s, C-2), 193.2 (s, C-3) ppm. IR (KBr): ν = 1780 (C=O),
phase-distorted yet sharp J q (3J removed); {1Ј-C(CH3)3} Ǟ C-1Ј
1
2
simplified to a sharp J d, 1Ј-CMe3 to a d with J = 1.6 Hz, and
˜
1562 (w), 1481, 1390, 1363, 1224, 1028, 963, 934 cm–1. This ketene 1Ј-C(CH3)3 to a d with J = 3.4 Hz; {4-C(CH3)3} Ǟ 4-C(CH3)3 as
3
“dimer” 25 may also be obtained from the chlorooxirane 18 with
LiTMP under similar conditions (yield 93% after recrystallization).
a sharp s, C-3 as a sharp d with 3J = 3.3 Hz (hence cis to 1Ј-H), C-
4 as a clear septet with 3J = 2.4 Hz, and C-5 simplified. IR (KBr): ν
˜
6660
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664

Syntheses of Di-tert-butylketene and Related gem-Di-tert-butyl Compounds
= 2959 (s), 1732 (s), 1655, 1478, 1376, 1340, 1271 (s), 1096, tion (97 mg) and distillation (140–150 °C bath temp./4 mbar) of the
862 cm–1. C16H28O2 (252.40): calcd. C 76.14, H 11.18; found C
76.35, H 11.49. The unequivocal assignments shown above allow
a straightforward interpretation of the NMR spectroscopic data
(Table S1 in the Supporting Information) at low temperatures in
terms of an impeded rotation of 4-tBu. The decoalesced signals
non-acidic product fraction. A corresponding run with the ketene
14 (employed in place of 13 and pyridine) at 150 °C required 21 h
for the complete consumption of 14 and partial (60%) conversion
of pivaloin (38) to the product ester 36, whereas traces of the ester
41 (from diglyme) appeared only after much longer periods of time.
(1H, 13C) of 4-tBu remained sharp down to –97 °C, whereas the 13
C
2,2,5,5-Tetramethyl-4-oxo-3-hexyl 2Ј-tert-Butyl-2Ј-deuterio-3Ј,3Ј-di-
methylbutanoate (37): Pulverized KOH (225 mg, 4.02 mmol) was
added to an NMR tube containing the α-chloro aldehyde 19
(94 mg, 0.49 mmol) in [D6]DMSO (0.80 mL). As described above
for 36, the black suspension produced the ketene 14 and a transient
amount of pivaloin (38) at 20 °C during 57 h. After 87 h at 60 °C,
the mixture was worked up as above to yield the acid [α-D]11
(14 mg, 16%) and the dideuteriated ester [3,2Ј-D2]36 (36 mg, 22%).
Isotope-modified NMR spectroscopic data[63] for [3,2Ј-D2]36 in
signals (100.6 MHz) of 4-CH3, cis-5-CH3, and trans-5-CH3 started
broadening (Table S1 in the Supporting Information), possibly be-
cause of a decelerating ring inversion.
2,2,5,5-Tetramethyl-4-oxo-3-hexyl 2Ј-tert-Butyl-3Ј,3Ј-dimethylbut-
anoate (36)
(a) From 19: The α-chloro aldehyde 19 (381 mg, 2.00 mmol) was
dissolved in dry DMSO (2.0 mL) and stirred magnetically with pul-
verized KOH (303 mg, 5.41 mmol) at room temp. in a tightly closed
round-bottomed flask (10 mL). After 6 h, 19 had been transformed
into a mixture (ca. 2:1) of ketene 14 and pivaloin (38). When 38
had disappeared after another 4 d at 20 °C (or after Ͻ4 d at 60 °C),
the contents were dissolved in Et2O/H2O. The Et2O layer was
washed until neutral, dried with Na2SO4 for ca. 1 h, and concen-
trated under moderately reduced pressure to furnish the ester 36
CD2Cl2: δH = 2.22 (11% of 2Ј-H), 5.10 (14% of 3-H) ppm; δC
=
2
34.12 (C-2, 82% D with Δ = –0.091 ppm), 35.19 (first C-3Ј, 86%
D with 2Δ = –0.089 ppm), 35.53 ppm (the broader second C-3Ј with
2Δ ≈ –0.089 ppm). Isotope-modified NMR spectroscopic data[63]
for [α-D]11 in CDCl3: δH = 2.20 (ca. 12% of 2-H) ppm; δC
=
=
1
63.97 ppm (C-2, ca. 87% D with 1Δ = –0.493 ppm and JC,D
19.3 Hz). This sample of ester [3,2Ј-D2]36 (0.11 mmol) was heated
in an NMR tube with KOH (106 mg, 1.89 mmol) in unlabeled
DMSO (0.7 mL) and [D12]cyclohexane (0.03 mL) at 95 °C for 8 h.
The workup procedure described above for 36 provided the mono-
deuteriated ester 37 (24 mg, 67%) and the acid [α-D]11 (5 mg, 26%,
formed through hydrolytic cleavage of 37). Isotope-modified NMR
spectroscopic data[63] for 37 in CD2Cl2: δH = 2.22 (16% of 2Ј-H),
5.12 (96% of 3-H) ppm; δC = 35.20 (C-3Ј, 81% D with 2Δ =
–0.090 ppm). Isotope-modified NMR spectroscopic data[63] for [α-
D]11 in CDCl3: δH = 2.21 (ca. 14% of 2-H) ppm.
along with residual ketene 14 (IR: ν = 2086 cm–1). The Et O extract
˜
2
of the acidified aqueous layer provided mainly the acid 11 (30 mg).
The pure ester 36 (171 mg, 26%) distilled at 137–145 °C (bath
temp.)/4 mbar, and a second distillation afforded the analytical
1
sample (133 mg). H NMR (CD2Cl2): δ = 1.01 (s, 3 CH3-1), 1.04
(br s, first 3 CH3-4Ј), 1.08 (sharp s, second 3 CH3-4Ј), 1.24 (sharp
s, 3 CH3-6), 2.22 (s, 2Ј-H), 5.11 (s, 3-H) ppm, assigned through
HMBC correlations (see below) consistent with the δH values of
1
pivaloin (38). H NMR (CDCl3): δ = 1.02, 1.06, 1.09, 1.26, 2.22,
5.12 ppm. 1H NMR (400 MHz, Cl2CD–CDCl2 at +80 °C): δ =
1.01, 1.05, 1.08, 1.23, 2.21, 5.13 ppm. 13C NMR (CD2Cl2): δ =
26.86 (3 CH3-1), 28.05 (3 CH3-6), 30.93 (all 6 CH3-4Ј), 34.22 (C-
2), 35.29 (sharp, first C-3Ј), 35.63 (br, second C-3Ј), 44.44 (C-5),
Lithium 2-tert-Butyl-3,3-dimethylbutanoate (39): Elemental lithium
(5.90 mg, 0.85 mmol) was dissolved in methanol (1.0 mL) in a small
round-bottomed flask (10 mL). The acid 11 (148 mg, 0.85 mmol)
64.3 (very br, C-2Ј), 78.03 (br, C-3), 174.53 (C-1Ј), 215.40 (C- was added, and the mixture was slowly concentrated in a rotary
4) ppm, assigned through HSQC (1JC,H) cross peaks in combina-
tion with the following HMBC correlations: for 3JC,H, 3-H Ǟ CH3-
evaporator to avoid splashing of the suddenly crystallizing product.
After 4 d under 4 mbar close to KOH and P4O10, the sample
weighed 155 mg (“102%”). 1H NMR (diglyme): δ = 1.10 (s,
1 Ǟ CH3-1 Ǟ C-3, 2Ј-H Ǟ all 6 CH3-4Ј Ǟ all 6 CH3-4Ј Ǟ C-2Ј,
2
and CH3-6 Ǟ CH3-6; for JC,H, CH3-1 Ǟ C-2, CH3-6 Ǟ C-5, 2Ј- 6ϫCH3), 2.06 (s, 2-H) ppm.[35]
H Ǟ the sharper C-3Ј Ǟ the sharper CH3-4Ј, and the broader CH3-
2-tert-Butyl-3,3-dimethylbutanoic Anhydride (40): A solution of the
4Ј Ǟ the broader C-3Ј. 13C NMR (CDCl3): δ = 26.73, 27.89, 30.76
+ 30.78 (3 + 3 CH3-4Ј), 33.96, 35.00, ca. 35.31 (very br), 44.20, ca.
64.0 (very br), ca. 77.6 (very br, hidden), 174.27, 215.32 ppm. The
three conspicuously broadened resonances of C-3, C-2Ј, and one
of the two C-3Ј indicate the freezing of some conformational mo-
tions (as also observed with 41); accordingly, these lines became
narrowed to normal linewidths at +80 °C. 13C NMR (100.6 MHz,
Cl2CD–CDCl2 at +80 °C): δ = 26.78, 27.86, 30.80 (all 6 CH3-4Ј),
33.91, 34.88, 35.15 (sharp), 44.14, 64.34 (sharp), 77.50 (sharp),
acyl chloride 13 (0.80 mmol), prepared as above from 11 (138 mg)
and dissolved in dry toluene (0.90 mL), was added to a round-
bottomed flask (10 mL) containing the dried lithium salt 39
(0.85 mmol, obtained above). The flask was stoppered tightly and
heated at 95 °C until the last traces of 13 had been consumed (2–
3.5 h). The mixture was dissolved in Et2O and distilled water and
was then extracted with NaOH (2 m, 2ϫ). The combined two
NaOH extracts were shaken with Et2O, which was combined with
the first Et2O layer, washed until neutral, dried with Na2SO4, con-
centrated to dryness, and dried further in vacuo to afford the al-
most pure anhydride 40 (176 mg, 67%). The acid 11 (36 mg, 13%)
was isolated as a byproduct from the acidified NaOH layer. Analyt-
ically pure anhydride 40 distilled at 146–153 °C (bath temp.)/
4 mbar, yielding colorless, bunchy fibers (109 mg, 42%) with m.p.
69–70.5 °C (ref.[33] 72–74 °C), mixed m.p. with the acid 11 43–
69 °C. 1H NMR (CDCl3 at 27 °C): δ = 1.15 (s, 36 H), 2.16 (s, 2ϫ2-
H) ppm.[35] 13C NMR (CDCl3 at 27 °C): δ = 30.66 (4ϫCMe3),
35.75 (br, 4ϫCMe3), 65.26 (very br, 2ϫC-2), 169.72 (2ϫC-
1) ppm; the two broadened resonances indicate the beginning of
some conformational freezing (compare 41). FT-IR (powder on
173.89, 214.81 ppm. FT-IR (film on diamond, ATR): ν = 2959 (s),
˜
2909, 2874, 1724 (s), 1705 (stronger), 1481, 1368, 1229, 1135 (s),
988 cm–1. C20H38O3 (326.52): calcd. C 73.57, H 11.73; found C
73.10, H 11.84.
(b) Independent Synthesis of 36 from Pivaloin (38): An NMR tube
(5 mm) was charged with pivaloin[62] (64 mg, 0.37 mmol), pyridine
(0.050 mL, 0.62 mmol), [D12]cyclohexane (0.030 mL), and a solu-
tion of the acyl chloride 13 (83 mg, 0.44 mmol) in dry diglyme
(0.65 mL). The 1H NMR spectra displayed the formation of ketene
14 (δ = 1.20 ppm, up to 0.12 equiv.) over 60 min at room temp.
Within 40 min at 150 °C, 13 and 38 disappeared almost completely
(and 14 completely), while the esters 36 and 41 were formed in a
87:13 ratio (further 2 h at 150 °C) that did not change upon isola-
diamond, ATR): ν = 3006, 2955 (s), 2912, 1797 (s), 1728 (w), 1366,
˜
1003 (s), 941 cm–1. FT-IR (film on diamond, ATR): ν = 1805, 1733
˜
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6661

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
(weaker), 1371, 1002, 915 cm–1. C20H38O3 (326.52): calcd. C 73.57,
H 11.73; found C 73.28, H 11.93. The anhydride 40 was formed
much more slowly from ketene 14 with the acid 11 in hot diglyme
(2 h at 150 °C) in the absence of bases.
NEt3 and THF. Almost pure 46 (2.18 g, 64%) was obtained
through fractional distillation at 91–113 °C/48 mbar. The analytical
sample was prepared through chromatography on a cold column
(diameter 15 mm, 25 g of silica gel, with pure low-boiling PE) with
95% recovery and a second distillation at 120–125 °C (bath temp.)/
3,6-Dioxaheptyl 2Ј-tert-Butyl-3Ј,3Ј-dimethylbutanoate (41): The
acyl chloride 13 (305 mg, 1.60 mmol), dry diglyme (1.0 mL), and
pyridine (0.160 mL, 2.00 mmol) were heated to 150 °C in a tightly
stoppered flask. The first trace of the product 41 was observed
after 5 h, along with residual 13 and substantial portions of the
ketene 14 and the anhydride 40 (δH = 1.08, 1.16, 1.20, and
1.14 ppm, respectively). After 86 h at 150 °C, the cooled mixture
was diluted with Et2O and HCl (2 m) and filtered to remove insolu-
ble resins. The Et2O layer was extracted with NaOH [2 m, for sepa-
ration from the pure acid 11 (23 mg, 8%)] and was then washed
with distilled water until neutral, dried with Na2SO4, and concen-
trated to give a mixture of 41 (206 mg, 47%) with the anhydride
40 (54 mg, 21%). A small sample of this mixture did not react with
NaOH (2 m, 10 equiv.) in [D8]dioxan at 85 °C (3 d) and 95 °C (6 h),
so 40 could not be separated from 41 through differential hydroly-
sis. The mixture was distilled at 132–151 °C (bath temp.)/6 mbar
and then adsorbed on silica gel (4 g) and eluted with PE/Et2O
(98:2) to yield pure 40, followed by contaminants. Further elution
1
58 mbar. H NMR (CDCl3): δ = 0.18 (s, OSiMe3), 1.15 (s, CMe3
1
anti to OSiMe3), 1.28 (s, syn-CMe3), 6.16 (s, 1-H) ppm. H NMR
([D6]benzene): δ = 0.10, 1.21, 1.51, 6.33, assigned through the 3JC,H
values for 1-H (see below) ppm. 13C NMR ([D6]benzene): δ = 0.00
1
3
1
3
(qsept, J = 118.6, J = 1.5 Hz, SiMe3), 32.54 (qm, J = 125.5, J
= 5.1 Hz, syn-CMe3), 32.74 (qm, 1J = 125.5, 3J = 5.0 Hz, anti-
CMe3), 36.55 (dm, J = 3.4, 2J = 3.7 Hz, anti-CMe3), 36.83 (dm,
3
3J = 6.5, J = 3.9 Hz, syn-CMe3), 134.94 (dm, J ≈ 8.2 Hz, C-2),
2
2
1
135.70 (sharp d, J = 175.4 Hz, C-1) ppm, assigned through selec-
tive {1H} decoupling experiments as follows: {anti-C(CH3)3} Ǟ
3
anti-C(CH3)3 decoupled and anti-CMe3 as a d with J = 3.4 Hz to
1-H; {syn-C(CH3)3} Ǟ syn-C(CH3)3 decoupled and syn-CMe3 as a
3
d with J = 6.5 Hz to 1-H; {anti- and syn-C(CH3)3} Ǟ C-2 as a d
2
with J ≈ 8.2 Hz. IR (film): ν = 2956, 2915, 1618, 1254, 1135 (s),
˜
879 (s), 844 cm–1. C13H28OSi (228.46): calcd. C 68.35, H 12.35;
found C 68.34, H 12.28.
3,6-Di-tert-butyl-2,2,7,7-tetramethyloct-4-ene-3,6-diol
(49):
A
with PE/Et2O (9:1, 80 mL) provided pure 41 as a colorless oil
(101 mg, 23%). H NMR (CDCl3): δ = 1.09 (s, 6ϫCH3), 2.23 (s, denser with drying tube, was charged with granulated elemental
2Ј-H), 3.38 (s, OCH3), 3.54 (m, CH2-5), 3.63 (m, CH2-4), 3.70 (m, lithium (278 mg, 40.0 mmol), dry THF (20 mL), and a magnetic
round-bottomed Schlenk flask (100 mL), fitted with a reflux con-
1
CH2-2), 4.21 (m, CH2-1) ppm, assigned through the NOESY corre-
stirring bar. An ice bath was employed to mitigate the moderately
exothermic reaction caused by the slow addition of chlorooxirane
lations C(CH3)3 ↔ 2Ј-H, and OCH3 ↔ CH2-5 ↔ CH2-4 ↔ CH2-
2 ↔ CH2-1. 13C NMR (CDCl3): δ = 30.72 (2ϫCMe3), 34.86 18 (1.907 g, 10.0 mmol) to the stirred mixture. After 2 d of contin-
(2ϫCMe3), 59.05 (OCH3), 62.19 (CH2-1), 64.44 (br, CH-2Ј), 69.24
(CH2-2), 70.34 (CH2-4), 71.94 (CH2-5), 174.54 (OCO) ppm, as-
signed through 1H/13C cross peaks in an HSQC (1JC,H) spectrum
ued stirring without cooling, the solution was poured into ice-cold
water (100 mL) through a plug of glass wool, which retained resid-
ual lithium metal (ca. 50%), and was rinsed with Et2O. The aque-
ous layer was extracted with more Et2O (3ϫ 50 mL) and then dis-
3
and in an HMBC spectrum that displayed the JC,H interactions
C(CH3)3 ↔ CH-2Ј, C(CH3)3 ↔ 2Ј-H, OCH3 ↔ CH2-5, OCH3 ↔ carded. The combined THF/Et2O layers were washed until neutral,
CH2-5, CH2-4 ↔ CH2-2, CH2-4 ↔ CH2-2, and CH2-1 ↔ OCO, in
dried with Na2SO4, and concentrated to furnish only the crude diol
49 (1.55 g, Ͻ99%). The colorless needles of pure 49 had a m.p. of
2
addition to the JC,H correlations 2Ј-H ↔ CMe3, CH2-5 ↔ CH2-
4, CH2–5 ↔ CH2-4, CH2-2 ↔ CH2-1, and CH2-2 ↔ CH2-1. The 151.5–152 °C (from pentane at –30 °C). 1H NMR (CDCl3): δ =
broad CH-2Ј resonance indicates the freezing of a conformational
process that was confirmed through low-temperature 13C NMR
spectroscopy. The first decoalescence at –15 °C gave rise to a signal
pair with ΔδC = 3.63 ppm in an intensity ratio of 13:87 that was
also shown by all other pairs of carbon nuclei below –45 °C (but
1.07 (s, 12ϫCH3, 13C satellites forming a d with 1JC,H = 125.2 Hz),
1.42 (s, 2ϫOH), 5.89 (s, olefinic 4- and 5-H, 13C satellites appear-
1
1
ing as the dd of the A part of an ABX system with JAX = JH,C
3
3
= 152.0 and JAB = JH,H = 15.7 Hz) ppm. 13C NMR (CDCl3): δ
1
3
2
= 29.08 (qsept, J = 125.2, J = 5.0 Hz, 12ϫCH3), 41.00 (m, J =
CH2-2 and the OCH3 group did not split). 13C NMR (CDCl3, 3.4 Hz, 4ϫCMe3), 82.10 (unresolved, C-3/-6), 130.39 (ddd, 1J =
100.6 MHz, –55 °C) for the 13/87 signal pairs: δ = 30.75/30.45
152, apparent 2J ≈ 4.4, 3J ≈ 2 Hz, C-4/-5) ppm, assigned through
(CMe3), 34.07/34.87 (CMe3), 62.54/61.83 (CH2-1), 66.70/63.07 selective {1H} decoupling experiments as follows: {C(CH3)3 and
(CH-2Ј), 69.99/70.12 (CH2-4), 71.61/71.48 (CH2-5), 174.92/174.70
(OCO). There were therefore only two significantly populated con-
formers, but their detailed structures remained unknown. FT-IR of
OH} Ǟ CMe3 as a s, C-3/-6 as a t with apparent J = 5.3 Hz, and
C-4/-5 as an apparent dd (X-part of ABX) with |1J + 2J| =
155.9 Hz. IR (KBr): ν = 3604 (sharp O–H), 2961, 2920, 1483, 1393,
˜
41 (film on diamond, ATR): ν = 2957, 2908, 2874, 1727 (s), 1370,
1367, 923 cm–1. C20H40O2 (312.54): calcd. C 76.86, H 12.90; found
C 77.02, H 12.82.
˜
1123 (s), 1054 cm–1. C15H30O4 (274.40): calcd. C 65.66, H 11.02;
found C 65.70, H 10.73.
2-tert-Butyl-2-chloro-3,3-dimethylbutan-1-ol (55): An NMR tube
(5 mm) was charged with the α-chloro aldehyde 19 (133 mg,
0.70 mmol), dry diglyme (0.7 mL), [D12]cyclohexane (0.030 mL),
2-tert-Butyl-3,3-dimethyl-1-(trimethylsilyloxy)but-1-ene (46):
A
mixture of the α-chloro aldehyde 19 (2.86 g, 15.0 mmol) and dry
THF (2.0 mL) was added slowly with magnetic stirring to cold
THF (25 mL) that contained ClSiMe3 (2.85 mL, 22.5 mmol) and
elemental lithium (312 mg, 45 mmol). Without further cooling, the
internal temperature rose to 33 °C within 60 min, whereupon LiCl
precipitated. After further stirring for 2.5 h, triethylamine
(4.15 mL, 30.0 mmol) was added at once. The suspension was fil-
tered from residual lithium through a glasswool plug that was
rinsed with Et2O (100 mL). This THF/Et2O solution was washed
with distilled water (not more than 3ϫ 25 mL), dried with Na2SO4,
and concentrated in a rotary evaporator down to 70 mbar, yielding
the intact enol ether 46 (4.00 g, no aldehyde 8) along with residual
1
TMS, and NaBH4 (49 mg, 1.30 mmol). H NMR spectra showed
the presence of dissolved NaBH4 (δH = –0.49 ppm) and revealed
the conversion of 19 (δH = 1.23 and 9.82) into the hydridoborate
54 (δH = 1.29 ppm) to be complete after 4.5 h at 20 °C, at which
point a trace of aldehyde 8 was already detectable. Without delay,
the contents were poured into a separating funnel that contained
ice and enough HCl (2 m) to destroy the residual hydrides (moder-
ate foaming by H2 formation!). The mixture was extracted with
Et2O (2ϫ 10 mL), and the combined extracts were washed with
HCl (2 m, 2ϫ), NaOH (2 m, 1ϫ), and distilled water until neutral.
After drying with Na2SO4 for not more than 1 h, the filtered Et2O
6662
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664

Syntheses of Di-tert-butylketene and Related gem-Di-tert-butyl Compounds
2
2
1
3
solution was concentrated in a rotary evaporator at Յ50 °C and
kept in a desiccator at 3 mbar for 30 min to furnish crude 55
(105 mg, Ͻ97%), which was stored at –18 °C. Most of the contami-
nants were removed by two successive filtrations through silica gel
(60 Å, 200 mg and then 435 mg) with low-boiling PE as the eluent
(50 mL in each filtration). The eluted material was dried at 3 mbar
for 20 min, furnishing 55 as a waxy, not completely pure product
(79 mg, 58%). 1H NMR (CDCl3): δ = 1.25 (s, 6ϫCH3), 2.01 (br s,
(dm, J = 5.4, J = 3.8 Hz, 2ϫCMe3), 54.6 (dm, J = 130.1, J =
4.2 Hz, C-2), 121.4 (d, 2J = 12, linewidth = 7 Hz, CN) ppm. IR
(film): ν = 2231 (w) cm–1.
˜
N-(2,2,4,4-Tetramethyl-3-pentyl)formamide (70): The aldehyde 8
(569 mg, 3.64 mmol) and hydroxylammonium chloride (330 mg,
4.75 mmol) in HCO2H/H2O (95:5, 3.20 mL)[53] were heated at re-
flux at 120 °C overnight (t1/2 ≈ 2 h).[44] The cooled solution was
diluted with iced water (45 mL) and extracted with Et2O (3ϫ
20 mL). The combined Et2O extracts were shaken with NaOH (2 m,
20 mL) and then with distilled water until neutral, dried with
MgSO4, and concentrated in vacuo. Distillation of the remaining
brown liquid (667 mg) at 14 Torr provided a forerun of heavily con-
taminated nitrile 69 (up to 140 °C bath temp., 185 mg), followed
by almost pure (Z)- and (E)-70 (74:26, 115 mg, 18%) at 140–180 °C
(bath temp.) as a slightly yellow, solidifying mixture (ref.[9] m.p. 89–
91 °C). Isomer (Z)-70: 1H NMR (CDCl3): δ = 1.04 (s, 6ϫCH3),
3.77 (d, J = 11.1 Hz, 3-H), 5.86 (unresolved, J Յ 11.1 Hz, NH),
8.28 (d, 3J = 2.0 Hz, CHO) ppm. 13C NMR (CDCl3): δ = 29.39
(6ϫCH3), 36.68 (2ϫCMe3), 62.02 (C-3), 160.86 (CHO) ppm. Iso-
mer (E)-70: H NMR (CDCl3): δ = 1.05 (s, 6ϫCH3), 2.76 (d, J =
10.9 Hz, 3-H), 6.50 (hardly resolved, quasi-dd, 3J ≈ 11.6 and
10.9 Hz, NH), 7.90 (d, 3J = 11.6 Hz, CHO) ppm. 13C NMR
(CDCl3): δ = 29.60 (6ϫCH3), 36.89 (2ϫCMe3), 68.90 (C-3),
165.37 (CHO) ppm.
1
OH), 3.93 (s, CH2O) ppm. 13C NMR (CDCl3): δ = 30.3 (qsept, J
= 126.6, 3J = 4.9 Hz, 6 CH3), 42.8 (oct, apparent J = 3.4 Hz, 2ϫC-
1
3), 65.3 (sharp t, J = 145.2 Hz, CH2OH), 91.4 (m, C-2) ppm. FT-
IR (film on diamond, ATR): ν = 3439 (br O–H), 2964 (s), 1395,
˜
1369 cm–1. Because of its ready transformation into 8, this chloro-
hydrin (55) could neither be recrystallized nor distilled for charac-
terization by elemental analyses or mass spectrometry.
2-Chloro-3,3-dimethylbutanal (63): Distilled at 50–60 °C (bath
temp.)/100 Torr (ref.[46] 38 °C/12 Torr). 1H NMR (CDCl3): δ = 1.09
3
3
3
(s, 3ϫCH3), 3.78 and 9.39 (AB system with J = 4.5 Hz) ppm.
1-Bromo-3,3-dimethylbut-1-ene (65 and 66): The reaction between
the aldehyde 8 (50 mg, 0.32 mmol) and 2,2,2-tribromo-2,2-dihy-
drobenzo-1,3,2-dioxaphosphole (64, 1.60 mmol)[48] in 1,2-dichloro-
ethane (0.5 mL) was complete within Ͻ5 min at room temp., as
shown by in situ 1H NMR spectroscopy, which also revealed the
formation of tBuBr (δH = 1.78 ppm) and exclusively the (Z) isomer
1
3
1
Supporting Information (see footnote on the first page of this arti-
cle): Low-temperature NMR (1H, 13C) chemical shifts δ and shift
non-equivalences of the 3-furanone derivative 29 (Table S1).
65. H NMR (ClCH2CH2Cl): δ = 1.24 (s, 3ϫCH3), 6.02 and 6.16
(AB system with 3J = 8 Hz, 1-/2-H) ppm.[49] After 23 d at room
temp., 65 had been transformed entirely into the (E) isomer 66. 1H
NMR (ClCH2CH2Cl): δ = 1.03 (s, 3ϫCH3), 6.00 and 6.25 (AB
3
system with J = 14 Hz, 1-/2-H) ppm.[49]
Acknowledgments
2-tert-Butyl-3,3-dimethylbutanal Oxime (67)
Funding of parts of this work by the Deutsche Forschungsgemein-
schaft is gratefully acknowledged. We thank Dr. David S. Stephen-
son for his assistance in some of the NMR spectroscopic investi-
gations.
(a) From 46: A solution of hydroxylammonium chloride (417 mg,
6.00 mmol) and sodium acetate (492 mg, 6.00 mmol) in ethanol
(4.83 mL) and distilled water (1.57 mL) was degassed by ultrason-
ification under argon gas for 15 min. After addition of the Si enol
ether 46 (457 mg, 2.00 mmol), the mixture was heated at 60 °C un-
der argon for 2 h and was then diluted with water (10 mL) and
shaken with Et2O (3ϫ 10 mL). The combined Et2O layers were
washed with a little water, dried with Na2SO4, and concentrated to
afford the almost pure oxime 67 (297 mg, 87%) as a colorless liquid
(ref.[6] m.p. 30 °C). 1H NMR (CCl4): δ = 1.05 (s, 6ϫCH3), 1.84
[1] T. T. Tidwell, Acc. Chem. Res. 1990, 23, 273–279.
[2] T. T. Tidwell in Ketenes II, 2nd ed., Wiley, Hoboken, New Jer-
sey, 2006.
[3] The imine 3, prepared from pivalonitrile with tert-butyl chlo-
ride and elemental lithium, is a convenient source of pure
ketone 1: R. Knorr, A. Donhärl, K.-O. Hennig, Liebigs Ann.
1996, 155–157.
3
and 7.33 (AB system with J = 10.1 Hz, 2- and 1-H, respectively),
1
8.15 (br, OH) ppm. H NMR (CDCl3): δ = 1.03, 1.88 and 7.46 (3J
[4] M. S. Newman, A. Arkell, T. Fukunaga, J. Am. Chem. Soc.
1960, 82, 2498–2501.
1
= 10.3 Hz) ppm. 13C NMR (CDCl3): δ = 30.75 (qoct, J = 125.1,
3J = 4.5 Hz, 6ϫCH3), 34.99 (m, 2J = 3.8 Hz, 2ϫCMe3), 57.45
[5] J.-E. Dubois, B. L. Zhang, C. Lion, Tetrahedron 1981, 37,
1
(dm, 1J = 127 Hz, 3J = 3.8 Hz, C-2), 154.11 [dd, J = 161 as ex-
4189–4194 (page 4191).
2
[6] C. Grundmann, S. K. Datta, J. Org. Chem. 1969, 34, 2016–
2018.
[7] G. A. Olah, A. Wu, O. Farooq, Synthesis 1989, 568.
[8] G. A. Olah, A. Wu, O. Farooq, Synthesis 1989, 566–567.
[9] T. G. Back, D. H. R. Barton, M. R. Britten-Kelly, F. S. Guziec,
J. Chem. Soc. Perkin Trans. 1 1976, 2079–2089; on pp. 2081
and 2082.
pected for the (E) conformation, J = 8.0 Hz, C-1] ppm.
(b) From 8: The protocol of Grundmann and Datta,[6] improved by
working under argon in an EtOH/H2O (1:1) mixture, furnished 67
(yield 68%) in a quality that required no distillation.
2-tert-Butyl-3,3-dimethylbutanenitrile (69): The addition of SOCl2
(0.64 mL, 8.82 mmol) to the undistilled oxime 67 (1.37 g,
8.00 mmol) in dry benzene (5 mL) at room temp. caused an imme-
diate evolution of gases, which ceased after warming to 60 °C for
10–30 min. The cooled solution was diluted with distilled water
(5 mL) and extracted with Et2O (3ϫ 10 mL). The combined Et2O
extracts were washed with distilled water (3ϫ 5 mL), dried with
Na2SO4, and concentrated in vacuo to leave a yellow liquid (1.06 g)
from which the nitrile 69 distilled at 80–85 °C/15 Torr (ref.[4] 76–
78 °C/10 Torr) as a colorless liquid (911 mg, 74%). 1H NMR
(CDCl3): δ = 1.19 (s, 6ϫCH3), 2.30 (s, 2-H) ppm. 13C NMR
(CDCl3): δ = 30.4 (qoct, 1J = 125.9, 3J = 4.5 Hz, 6ϫCH3), 35.4
[10] P. Hofmann, L. A. Perez Moya, I. Kain, Synthesis 1986, 43–
44.
[11] G. Köbrich, H. R. Merkle, Chem. Ber. 1966, 99, 1782–1792.
However, compare footnote 4 of ref.[12]
[12] A. Müller, M. Marsch, K. Harms, J. C. W. Lohrenz, G. Boche,
Angew. Chem. 1996, 108, 1639–1640; Angew. Chem. Int. Ed.
Engl. 1996, 35, 1518–1520, footnote 4.
[13] a) H. Taguchi, H. Yamamoto, H. Nozaki, J. Am. Chem. Soc.
1974, 96, 3010–3011; b) H. Taguchi, H. Yamamoto, H. Nozaki,
Bull. Chem. Soc. Jpn. 1977, 50, 1588–1591.
[14] a) We did not try to add nBuLi[10] to the mixture of 1 and
CH2Cl2 in THF, although this might also have worked, accord-
Eur. J. Org. Chem. 2010, 6651–6664
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6663

R. Knorr, K.-O. Hennig, B. Schubert, P. Böhrer
FULL PAPER
ing to: R. W. Hoffmann, H.-C. Stiasny, J. Krüger, Tetrahedron
1996, 52, 7421–7434; b) we were unable to couple ketone 1 to
the less stable carbenoids LiCHBr2 and LiCHI2.
[15] G. Köbrich, J. Grosser, W. Werner, Chem. Ber. 1973, 106, 2610–
2619.
[16] R. Knorr, C. Pires, C. Behringer, T. Menke, J. Freudenreich,
E. C. Rossmann, P. Böhrer, J. Am. Chem. Soc. 2006, 128,
14845–14853, compound 15 therein.
[35] For pertinent 1H NMR chemical shifts (δ) in other solvents,
see the collection at the beginning of the Exp. Sect.
[36] N. De Kimpe, L. De Buyck, R. Verhé, N. Schamp, Tetrahedron
1984, 40, 3291–3294.
[37] J. R. Shanklin, C. R. Johnson, J. Ollinger, R. M. Coates, J. Am.
Chem. Soc. 1973, 95, 3429–3431.
[38] J. Moskal, A. M. van Leusen, Recl. Trav. Chim. Pays-Bas 1987,
106, 137–141.
[17] When the deprotonation of CH2Cl2 and the final quenching
with acetic acid were performed at –110 °C in a “Trapp mix-
ture” (G. Köbrich, H. Trapp, Chem. Ber. 1966, 99, 670–679),
most of the ketone 1 was recovered, again only along with 18.
We thus did not succeed in preparing 20 and its dehydration
product tBu2C=CCl2, which would have been a candidate for
the “carbenoid chain” mechanism described in ref.[16]
[18] J. K. Crandall, G. E. Salazar, R. J. Watkins, J. Am. Chem. Soc.
1987, 109, 4338–4341, yield Յ13%.
[39] The “TOSMIC” reagent does not react with ketone 1 accord-
ing to: O. H. Oldenziel, D. van Leusen, A. M. van Leusen, J.
Org. Chem. 1977, 42, 3114–3118.
[40] V. Malatesta, D. Forrest, K. U. Ingold, J. Phys. Chem. 1978,
82, 2370–2373 (p. 2372).
[41] P. Lohse, H. Loner, P. Acklin, F. Sternfeld, A. Pfaltz, Tetrahe-
dron Lett. 1991, 32, 615–618, and refs. cited therein.
[42] J. K. Crandall, M. Apparu, Org. React. 1983, 29, 345–443.
[43] L. M. Baigrie, D. Lenoir, H. R. Seikaly, T. T. Tidwell, J. Org.
Chem. 1985, 50, 2105–2109.
[19] P. J. Wagner, C. A. Stout, S. Searles, G. S. Hammond, J. Am.
Chem. Soc. 1966, 88, 1242–1244.
1
[44] Values of the first t1/2 refer to the usually employed H NMR
[20] J. P. Wasacz, M. M. Joullié, U. Mende, I. Fuss, G. W. Griffin,
J. Org. Chem. 1976, 41, 572–574.
[21] J. E. Anderson, C. W. Doecke, H. Pearson, D. I. Rawson, J.
Chem. Soc. Perkin Trans. 2 1978, 974–978, compound 3d
therein.
concentrations (roughly 0.3 m) of the reacting compounds.
[45] V. J. Shiner, G. F. Meier, J. Org. Chem. 1966, 31, 137–142.
[46] Generated by the General Procedure (“Method B1”) reported
by: U. von Roman, J. Ruhdorfer, R. Knorr, Synthesis 1993,
985–992.
[22] X. Creary, Chem. Rev. 1991, 91, 1625–1678.
[23] R. N. McDonald, R. C. Cousins, J. Org. Chem. 1980, 45, 2976–
2984.
[24] The conversion of 19 to 27 on distillation or on long-term stor-
age might be initiated by an acid-catalyzed[22] C–Cl bond heter-
olysis.
[47] P. Duhamel, L. Duhamel, J. Gralak, Bull. Soc. Chim. Fr. 1970,
3641–3646.
[48] U. von Roman, R. Knorr, C. Behringer, J. Ruhdorfer, J. Prakt.
Chem./Chem.-Ztg. 1994, 336, 260–262, and cited lit.
[49] In accord with: R. B. Miller, G. McGarvey, J. Org. Chem. 1978,
43, 4424–4431.
[25] J. Gralak, J. Y. Valnot, Org. Prep. Proced. Int. 1979, 11, 107–
[50] a) G. Lohaus, Chem. Ber. 1967, 100, 2719–2729; b) G. Lohaus,
110; Chem. Abstr. 1979, 91, 140639u.
Org. Synth., Coll. Vol. VI 1988, 304–306.
[26] The rearrangement rates of other, less congested chlorooxir-
anes in toluene had to be measured at 200 °C by: S. A. Kline,
J. J. Solomon, B. L. Van Duuren, J. Org. Chem. 1978, 43, 3596–
3600.
[51] a) C. Fizet, J. Streith, Tetrahedron Lett. 1974, 15, 3187–3188;
b) J. Streith, C. Fizet, H. Fritz, Helv. Chim. Acta 1976, 59,
2786–2792.
[52] J. Liebscher, H. Hartmann, Z. Chem. 1975, 15, 302.
[53] G. A. Olah, T. Keumi, Synthesis 1979, 112–113.
[54] R. E. Gawley, Org. React. 1988, 35, 1–420 (pp. 9 and 26).
[55] A. Mannschreck, Tetrahedron Lett. 1965, 6, 1341–1347.
[56] H. Naganishi, J. D. Roberts, Org. Magn. Reson. 1981, 15, 7–
12.
[27] A similar elimination mode in an acidic (!) milieu was pos-
tulated by: H. Ogawa, Y. Takegutu, T. Imoto, Y. Taniguchi, H.
Kato, Tetrahedron Lett. 1979, 20, 3457–3460. According to
p. 188 of ref.,[2] “such an elimination appears unusual”.
[28] G. Böhrer, R. Knorr, P. Böhrer, B. Schubert, Liebigs Ann./Re-
cueil 1997, 193–202 (p. 197).
[57] L. A. LaPlanche, M. T. Rogers, J. Am. Chem. Soc. 1964, 86,
[29] K. Sato, T. Sekiguchi, S. Akai, Tetrahedron Lett. 2001, 42,
337–341.
3625–3628.
[58] G. E. Parris, E. C. Ashby, J. Am. Chem. Soc. 1971, 93, 1206–
1212 (Figure 5 therein).
[30] C. L. Stevens, E. Farkas, B. Gillis, J. Am. Chem. Soc. 1954, 76,
2695–2698.
[59] A. G. Davies, R. Sutcliffe, J. Chem. Soc. Perkin Trans. 2 1980,
[31] C. L. Stevens, B. T. Gillis, J. Am. Chem. Soc. 1957, 79, 3448–
819–824.
3451.
[60] S. H. Brown, R. H. Crabtree, J. Am. Chem. Soc. 1989, 111,
2946–2953 (p. 2952).
[32] D. Y. Curtin, A. Bradley, J. Am. Chem. Soc. 1954, 76, 5777–
5779.
[61] F. H. A. Rummens, J. S. Lomas, B. Tiffon, C. Coupry, N. Lum-
broso-Bader, Org. Magn. Reson. 1982, 19, 35–38.
[62] For a convenient synthesis of 38, see ref.[28]
[63] Deuterium-induced NMR chemical shifts over a distance of n
[33] Compound 40 was mentioned as a side-product by: E. Schau-
mann, Chem. Ber. 1982, 115, 2755–2765 (p. 2764).
[34] Compound 40 is tentatively shown in the (sc,ac) conformation
as determined for gaseous acetic anhydride by: G. Wu, C.
Van Alsenoy, H. J. Geise, E. Sluyts, B. J. Van der Veken, I. F.
Shishkov, L. V. Khristenko, J. Phys. Chem. A 2000, 104, 1576–
1587.
n
chemical bonds are defined as Δ = δ(deuteriated) – δ(undeut-
eriated).
Received: June 21, 2010
Published Online: October 19, 2010
6664
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6651–6664