Regioselective Dialkylations of N-(tert-Butyl)iminocyclopentane via Deprotonating One-Pot Procedures

Article abstract of DOI:10.1002/hlca.201800037

The title compound (1) was chosen as a model for the α/α′-regioselectivity of deprotonation and subsequent alkylation adjacent to the C=N bond. With the bulky base lithium N,N-diisopropylamide (LDA) as a catalyst, the one-pot deprotonation steps can be performed through titration with methyllithium, using gas-volumetric observation of the liberated methane. In the first step with ensuing methylation by iodomethane, the primary product is born at ?40?°C in its metastable (Z) configuration (kinetic control) and may be either isolated or converted in?situ at 30?°C into its thermodynamically favored (E)-isomer via cis to trans stereoinversion at the N-atom. Being slow enough on the laboratory time scale, this stereoinversion process can serve to control the regioselectivity of the second deprotonation/alkylation sequence as follows. The α,α′-products are formed from the intermediate (Z)-imine, whereas α,α-products result from the intermediate (E)-imine; in either case, syn deprotonation (cis to ^tBu at nitrogen) by LDA is apparently disfavored by the ^tBu group, so that anti deprotonation becomes obligatory. If a third one-pot deprotonation step is too slow with LDA, it may be performed with the stronger base butyllithium/HMPA which, however, reacts regio-unselectively. Regioselective one-pot, LDA-catalyzed deprotonation with alkylation by oxiranes (alone, or alternatingly with iodomethane) opens a short access to spiro-[2.4]heptan-4-ones.

Full text of DOI:10.1002/hlca.201800037

Article Type: Full Paper
Regioselective Dialkylations of N-(tert-Butyl)iminocyclopentane via
Deprotonating One-Pot Procedures
Rudolf Knorr,* Brigitte Neuner
Department Chemie, Ludwig-Maximilians-Universität München, Butenandtstrasse 5-13 (Haus F),
81377 München, Germany
* rhk@cup.uni-muenchen.de
Key words: Azaallyllithium; N-lithioenamines; sp2-nitrogen inversion; spiroacetals;
spiro[2.4]heptan-4-ones
Abstract. The title compound (1) was chosen as a model for the α/α´-regioselectivity of
deprotonation and subsequent alkylation adjacent to the C=N double bond. With the bulky base
lithium N,N-diisopropylamide (LDA) as a catalyst, the one-pot deprotonation steps can be performed
through titration with methyllithium, using gas-volumetric observation of the liberated methane. In
the first step with ensuing methylation by iodomethane, the primary product is born at –40 °C in its
metastable Z configuration (kinetic control) and may be either isolated or converted in situ at 30 °C
into its thermodynamically favored (E)-isomer via cis to trans stereoinversion at the nitrogen atom.
Being slow enough on the laboratory time scale, this stereoinversion process can serve to control
the regioselectivity of the second deprotonation/alkylation sequence as follows. The α,α´-products
are formed from the intermediate (Z)-imine, whereas α,α-products result from the intermediate (E)-
imine; in either case, syn deprotonation (cis to tBu at nitrogen) by LDA is apparently disfavored by
the tBu group, so that anti deprotonation becomes obligatory. If a third one-pot deprotonation
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/hlca.201800037
This article is protected by copyright. All rights reserved.

step is too slow with LDA, it may be performed with the stronger base n-butyllithium/HMPA which,
however, reacts regio-unselectively. Regioselective one-pot, LDA-catalyzed deprotonation with
alkylation by oxiranes (alone, or alternatingly with iodomethane) opens a short access to
spiro[2.4]heptan-4-ones.
1. Introduction
Regioselectively controlled deprotonation of ketimines (R3N=CR1R2) was developed through
meticulous mechanistic elucidations.[1 - 3] It may serve as a tool for the introduction of substituents
whose accurate (α- or α´) positioning would be difficult with the “parent” ketones O=CR1R2
(especially with monosubstituted cyclopentanones). Scheme 1 displays our presently employed
ketimine model system and the products of its anticipated regioselective trapping by an alkylating
agent AlkX: It illustrates how the ketimine configuration (Z or E) should determine the
regioselectivity of removing a proton from the anti position, namely, trans to the tert-butyl (“tBu”)
group at nitrogen. Such deprotonations are accompanied by a rapid rearrangement which directs
the entailing alkylation to give primary products with a syn relationship of the entering agent (“Alk”)
and the N–tBu group. This established1] mechanism explained the surprising formation of a (Z)-2,2-
product (Scheme 1) from an (E)-ketimine through that most advantageous rearrangement of an
intermediate 1-azaallyllithium compound (not shown here). The same mechanistic principles apply
to the monoalkylation of the title compound 1 with the initial formation of the (Z)-ketimine (with R =
alkyl in Scheme 1) and its second alkylation that generates the 2,5-product.
This article is protected by copyright. All rights reserved.

(H3C)3C
C(CH3)3
N
N
2
5
2
5
R
R
syn
syn
anti
anti
(E)-ketimine
(Z)-ketimine
+
-H
+
-H
o
o
-
-
-X
-X
+AlkX
+AlkX
C(CH3)3
(H3C)3C
N
N
R
2
2
5
R
Alk
5
Alk
(Z)-2,2-product
2,5-product
Scheme 1. Anti deprotonation and alkylation of the Z,E ketimine diastereomers.
The possibility to use the =N–tBu orientation as a control instrument depends on the direction
and velocity of the Z,E re-orientations. These Z,E stereoinversion processes were too fast for
controlling purposes in the case of an N-aryl model[4] that produced comparable amounts of the 2,5-
and 2,2-products.[5] The earlier studied N-alkylimine model systems were reported[1,2] to depend on
further factors; but they suggested to us that an N–tBu moiety might favor the anti over the
alternative syn deprotonation sufficiently to control the regioselectivity. Such an important
simplification in combination with the suitably slow Z,E interconversion convinced us to investigate
the =N–tBu model depicted in Scheme 1.
This article is protected by copyright. All rights reserved.

2. Results and discussion
2.1. Preparation and dimethylation of the ketimine 1.
The known[6] title compound N-(cyclopentylidene)-tert-butylamine (1) was prepared from
cyclopentanone and tert-butylamine (3 equiv) with TiCl4 (0.52 equiv),[7] using some modifications
that increased the yield of 1 slightly. Its NMR spectra were taken in CCl4 or [D6]benzene as the
solvent because its solutions in commercial CDCl3 deteriorated quickly. The bulky base lithium N,N-
diisopropylamide (LDA) had usually been employed to abstract a proton from the azaallyl-type (α)
position of ketimins (such as the 2-position of 1 in Scheme 2). With 1 and some of its intermediates
as the CH acids, we found it convenient to generate LDA in situ from N,N-diisopropylamine (0.14
equiv) and methyllithium as a base whose consumption could be measured simply by means of the
gas volume of the emanating methane. According to the well-stablished[1,2] deprotonation
mechanism, the azaallyllithium (an N-lithioenamine) intermediate 2 should be generated through
proton transfer from the anti position (H2C(2) in 1) to LDA; the by-product N,N-diisopropylamine will
then be reconverted to LDA by additional H3CLi. An accompanying fast rearrangement of 2 should
form the more stable[3,8] syn azaallyllithium 3. Thus, we performed this LDA-catalyzed
deprotonation of 1 in THF at –40 °C (General Procedure GP1) through titration with H3CLi in Et2O;
subsequent trapping of 3 at –70 °C with iodomethane (MeI, 1 equiv) generated the metastable
ketimine (Z)-4. This key compound controls the regioselectivity: Its anti position (H2C(5)) remains
available for a second deprotonation as long as (Z)-4 does not isomerize into the more stable (E)-4.
Therefore, the obtained solution of (Z)-4 and LDA was allowed to warm up to –40 °C only (GP2a), at
which temperature it was promptly titrated with a second portion of H3CLi (up to 1.15 equiv). Again,
the immediately generated azaallyllithium 5 (anti) rearranged quickly to 6 (syn). When the evolution
of CH4 ceased after 12 min, 6 was trapped with MeI (1.2 equiv) at –40 °C to give the ketimine 7K.
Like all other ketimines obtained in this work, 7K could be dissolved in cold aqueous HCl (2 M)
without being hydrolyzed, so that it could easily be separated from all nonbasic contaminants. Upon
alkalization (GP2c) of the HCl solution, 7K returned as the tautomeric enamine 7A without any other
isomeric form. Since small amounts of other basic contaminants were still present, this material was
subjected to acid-catalyzed, de-aminating hydrolysis by boiling aqueous H2SO4 (2 N, GP3), whereby it
furnished a distilled mixture (35% yield) of 2,5-dimethyl- (8) and 2-methylcyclopentanone (9) in a
molar ratio of 95:5. The presence of 9 disclosed that the conditions of the second methylation were
insufficient to consume 4 completely; however, this dimethylation of 1 was so highly regioselective
that no 2,2-product could be detected.
This article is protected by copyright. All rights reserved.

tBu
5
tBu
5
Li
5
tBu
2
Li
2
+MeLi
-CH4
N
N
N
o
2
2 (anti)
3 (syn)
+MeI
-LiI
tBu
1
tBu
Li
Li
tBu
< -40 °C
N
N
N
o
2
Me
+MeLi
-CH4
Me
Me
5
(Z)-4
H3O
6 (syn)
-LiI
5 (anti)
O
2
+
Me
+MeI
tBu
T
9
H
tBu
N
O
+
N
H3O
T
5
2
2
5
2
Me
5
Me
Me
Me
Me
Me
8
7K
7A
Scheme 2. Intermediates and products of LDA-mediated methylation on the 2,5-route; Me =
methyl, tBu = tert-butyl, T = bath temp. 130 °C.
The alternative “2,2-route” was opened when we warmed the THF/Et2O solution of the
metastable 2-methyl product (Z)-4 (Scheme 3) quickly from –40 °C to +30 °C and stirred it for 2
hours (GP2b). During this period of time,[9] (Z)-4 was transformed into the preponderant
diastereomer (E)-4. Although (E)-4 offered a normally less reactive tert-CH group at its anti position
C(2), the subsequent second titration with H3CLi led to a comfortably fast CH4 evolution with
formation of the azaallyllithium 10 (anti) that rearranged quickly to the thermodynamically
favored[3,8] isomer 11 (syn). Trapping of 11 with MeI (1.2 equiv) at –40 °C and subsequent stirring at
room temperature (rt) for 90 min generated the 2,2-dimethyl ketimine (E)-12 via its less stable
diastereomer (Z)-12. After separation from the nonbasic contaminants through extraction into cold
aqueous HCl (2 M, GP2c) and subsequent alkalization, the obtained sample of (E)-12 contained
minor basic side-products, so that the presence of an enamine or of a small 2,5-dimethyl fraction
This article is protected by copyright. All rights reserved.

could not yet be ascertained. In order to arrive at a reliable analysis, the basic fraction was
subjected to the acid-catalyzed, de-aminating hydrolysis (GP3), which afforded a distilled mixture
(45% yield) of 2,2-dimethyl- (13), 2,5-dimethyl- (8), and 2-methylcyclopentanone (9) in the molar
ratio of 92:3:3. Hence, the formation of 9 disclosed that the conditions of the second deprotonation
were again insufficient to consume the monomethyl ketimine 4 completely. The 13/8 = 92:3
regioselectivity seemed to point into the direction of the 87:13 ratio that may be expected[10] for the
(E)-4/(Z)-4 equilibrium (upper part of Scheme 3). Alternatively, leakage from the 2,2- to the 2,5-
route might have occurred by a slow protonation/deprotonation process (bottom part of Scheme 3)
that transformed 11 (syn) into 6 (syn) and may be catalyzed[11] by traces of a CH-acidic ketimine. We
actually discovered such a leakage under considerably harsher conditions : A slow conversion of 11
into 6 was detected after 3 hours in THF at 60 °C (but not yet after 64 hours at rt) through
methylation and hydrolysis as above, which furnished the 2,2-/2,5-products (ketones 13 and 8) as a
53:37 mixture instead of the above 92:3 ratio. The converse protonation/deprotonation process
under the same conditions with heating of 6 in the initial absence of 11 suggested that the
equilibrium ratio of 11/6 was ca. 6:87. These observations established an 11 → 6 leakage that
appeared to be significantly slower than the corresponding leakage process described[1] in the
literature. As a consequence, the 11 → 6 leakage can be excluded under the milder conditions of
our GP2b, which leaves the hypothesis of an LDA-repellent N–tBu group that impeded syn
deprotonation by LDA but admitted the observed anti deprotonations of both (Z)-4 and the (E,Z)-4
equilibrium mixture.
This article is protected by copyright. All rights reserved.

O
tBu
+
H
tBu
H3O
2,5-route
N
N
2
Me
2
T
Me
2
5
Me
Me
5
9
(Z)-4
7A
T
+
2,2-route
+30 °C
H3O
+
T
2
H3O
tBu
Li
tBu
0 °C
O
N
N
5
Me
Me
2
Me
Me
5
+MeLi
-CH4
8
10 (anti)
(E)-4
o
tBu
tBu
N
tBu
5
Li
+MeI
N
N
-LiI
-40 °C
2
2
Me
Me
Me
Me
Me
5
(E)-12
(Z)-12
11 (syn)
very slow
+
T
H3O
tBu
Li
2,5-route
O
N
O
2
2
5
Me
Me
Me
Me
Me
5
13
8
6 (syn)
Scheme 3. Intermediates and products of LDA-mediated methylation on the 2,2-route (bold
arrows); Me = methyl, tBu = tert-butyl, T = bath temp. 130 °C.
The second C(2) alkylation was regiospecific also with 2-iodopropane as one of the sterically
handicapped electrophiles: With the “LDA-repellent N–tBu” group adjusted for the 2,2-route, (E)-4
generated a 89:11 mixture of the ketimine 14K and its enamine 14A (75% yield of all basic products).
This article is protected by copyright. All rights reserved.

The presence of other basic contaminants thwarted our attempts to establish the absence of a 2,5-
isomer of 14; however, the acid-catalyzed, de-aminating hydrolysis (GP3) afforded a 55% yield of the
known[12,13] 2-isopropyl-2-methylcyclopentanone (15) without the 2,5-isomer. This confirmed the
2,2-regiospecific process as far as the missing 2,5-ketimine would have been expected to be
hydrolyzed faster than 14K.
tBu
tBu
+Me2CHI
+MeLi
N
N
10
(anti)
11
(syn)
2
2
Me
-CH4
0 °C
Me
5
5
-LiI
-40 °C
CHMe2
(Z)-14K
(E)-4
tBu
tBu
H
+
H3O
T
N
O
N
2
2
5
Me
Me
Me
CHMe2
2
5
5
CHMe2
CHMe2
(E)-14K
15
14A
Scheme 4. LDA-mediated, regiospecific isopropylation of (E)-4; Me = methyl, tBu = tert-butyl, T =
bath temp. 130 °C.
2.2. Convergent trimethylation.
As a short outlook beyond anti regioselectivity, the formation of tri- and tetramethylated
ketimines should be possible from both 2,5- (7A) and 2,2-dimethyl (12) precursors: (E)-12 (Scheme
5) must be deprotonated at C(5) either by the slower syn attack or via regeneration of the less
favored equilibrium component (Z)-12 and an ensuing anti deprotonation; in either case, (E)-12 is
expected to generate the 2,2,5-trimethyl ketimine 16 rather slowly. Actually, (E)-12 was prepared
through the GP2b process and treated in situ with additional N,N-diisopropylamine (0.60 equiv) and
H3CLi (1.2 equiv) at 42 °C for 2 hours, then quenched with MeI (1.2 equiv) at 0 °C. One repetition of
the latter deprotonation/methylation sequence (in the vain hope for tetramethylation) and
subsequent work-up (GP2c) afforded a basic product fraction (91% yield) that contained 16 and
unconsumed (E)-12 in a 48:52 ratio. This evidence for such a slow deprotonation of (E)-12 by LDA
had been obtained through acid-catalyzed, de-aminating hydrolysis (GP3, 1 hour) which left 53% of
ketimine 16 intact; the hydrolyzed and distilled portion (27% yield) contained a 77:23 mixture of 2,2-
This article is protected by copyright. All rights reserved.

dimethyl (13, from residual 12) and 2,2,5-trimethylcyclopentanone (17). Thus, both the
deprotonation of (E)-12 and the hydrolysis of 16 were slow processes.
+
+
+MeI
-LiI
H
tBu
+MeI -H
tBu
5
-H
tBu
N
N
N
-LiI
2
2
2
5
5
Me
Me
Me
Me
Me
Me
Me
7A
(E)-12
16
+
-H
+
+MeI
H3O
T
tBu
+
H3O
T
O
N
O
5
2
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
19
17
18
Scheme 5. 2,2,5-Trimethyl- (17) and (almost no) 2,2,5,5-tetramethylcyclopentanone (19) were
formed from enamine 7A or ketimine 12; Me = methyl, tBu = tert-butyl, T = bath temp. 130 °C.
Under the above conditions of a twofold LDA-mediated deprotonation/methylation sequence, the
2,5-dimethyl enamine 7A (from GP2a) furnished a basic fraction (65% yield) that contained mainly
the 2,2,5-trimethyl ketimine 16. Acid-catalyzed, de-aminating hydrolysis (GP3, 1 hour) left again
47% of 16 intact and afforded a distilled 93:7 mixture (9 % yield) of 2,2,5-trimethyl- (17) and 2,2-
dimethylcylopentanone (13, from residual 12), the latter as a reminiscence of the monomethyl
ketimine 4 that had survived the GP2a treatment (Scheme 3) and generated now the sluggish 2,2-
ketimine 12 at 42 °C. Tetramethylated compounds (18 or 19) were not observed in these runs with
LDA and MeI.
For a more efficient third deprotonation, we used the stronger base n-butyllithium (n-BuLi) in
combination with hexamethylphosphoramide (HMPA) which had been reported14] to prefer syn
deprotonation. The formerly sluggish 2,2-dimethyl ketimine (E)-12, prepared again in THF/Et2O with
LDA/H3CLi (GP2b), was now consumed completely by the twofold in situ treatment with the
sequence of deprotonation (2 hours at rt) and methylation. The basic fraction (96% yield) from GP2c
carried the 2,2,5-trimethyl- (16) and 2,2,5,5-tetramethyl ketimine (18) in a 99:1 ratio as determined
This article is protected by copyright. All rights reserved.

through acid-catalyzed, de-aminating hydrolysis (GP3, 1 hour) that afforded a mixture (8% yield) of
2,2,5-trimethyl- (17) and 2,2,5,5-tetramethylcyclopentanone (19) beside residual 16 (43%). Under
the same conditions, the enamine 7A was subjected to the twofold in situ treatment with n-BuLi (1.2
equiv) and HMPA (1 equiv) and MeI (1.2 equiv). The resultant basic fraction (yield 99%) was
hydrolyzed (GP3, 1 hour, yield 11%) to give 17 and 19 (99:1) again together with unconsumed 16
(52%). In short, trimethylation of ketimine 1 was conveniently performed using n-BuLi/HMPA for the
third deprotonation after the GP2a or GP2b process, whereas tetramethylation was almost
impossible because of the bulky N-tBu group.
The non-optimized one-pot procedures described in this Section exemplified how the ketimine 1
can successfully be mono- or trialkylated or regioselectively dialkylated.
2.3. Alkylation by oxiranes as the electrophiles
Selection of the 2,5-route can also be achieved with ethylene oxide (20a in Scheme 6) under the
conditions of obligatory anti deprotonation by LDA: In the spirit of Scheme 2, the first alkylation
(GP1) of ketimine 1 is expected to generate only (Z)-21a. At sufficiently low temperatures (GP2a)
that do not yet admit a Z to E stereoinversion, (Z)-21a will be deprotonated at its anti position (C(5))
to generate the azaallylithium 22a that undergoes the second alkylation to give the primary 2,5-
dialkylated product 23a. As ethylene oxide (20a) is unreactive toward both LDA and H3CLi under the
reaction conditions, the ketimine 1 can (but must not) be deprotonated in the presence of 20a (3.0
equiv) by LDA (0.14 equiv) with H3CLi (2.1 equiv). This is, of course, the most simple way to
deprotonate (Z)-21a as soon as it is born. Aqueous work-up disclosed that the ketimine product
expected from 23a had been cyclized with formation of the N,O-acetal[15] 24a: The IR spectrum of
crystalline 24a displayed neither C=O nor C=N vibrations, and the 13C (no C=O, no C=N functions)
1
and H NMR spectra revealed that 24a was a pair of diastereomers (ca. 1:1) whose resonances could
not be distinguished and assigned. The 2,5-constitution of 24a was established through acid-
catalyzed, de-aminating hydrolysis (GP3): The resultant, bridged spiro[4.4]acetal[16] 25a can be
formed from both diastereomers of 24a due to the known[17] configutational lability of spiroacetals
in strongly acidic solutions. The 2,5-stereoselective formation of 24a was further established with
the following derivatives. Both 24a and 25a were ring-opened in boiling conc. HCl (GP4) that
furnished the known[18] 2,5-bis(2´-chloroethyl)cyclopentanone (26) as a 4:1 trans/cis mixture. The
KOEt-mediated HCl elimination (GP5) from 26 completed the present, reasonably short
approach[19,20] to the known[21] dispiro ketone 27.
This article is protected by copyright. All rights reserved.

tBu
5
tBu
R
tBu
5
Li
tBu
2
Li
2
N
N
+MeLi
-CH4
N
N
o
OLi
2
O
R
2
5
5
R
20a,b
R
1
2 (anti)
3 (syn)
(Z)-21a,b
+MeLi
-CH4
O
R = CH3:
+60 °C?
R
tBu
Li
tBu
R
N
N
R
R
20a,b
R
LiO
OLi
OLi
5
2
R
R
R
(E)-21b
22a,b
23a,b
(H2O)
20-25 R
H
tBu
R
a
b
H
CH3
R
N O
HO
R
R
R = H: T
24a,b
conc. HCl
Cl
Cl
+
O
- t-BuNH2
H3O
5
2
1
2
8
O
7
R
R
R
R
O
O
6
7
2
4
3
9
R = H: T
conc. HCl
26
9
KOEt
3
6
1
4
5
8
25a,b
27
Scheme 6. LDA-mediated 2,5-dialkylation of 1 with two oxiranes (20a,b); Me = methyl, tBu = tert-
butyl, T = bath temp. 130 °C.
The above-mentioned unreactivity of ethylene oxide (20a) against LDA and H3CLi offered an
opportunity to study a dialkylation reaction of 1 under conditions of (Z)-21a → (E)-21a
stereoinversion (Scheme 7). The question was if the primary product (Z)-21a might undergo in-situ
deprotonation (the 2,5-route) less readily than stereoinversionwith subsequent generation of the
2,2-product (Z)-28. Using the same reactant ratios as above, H3CLi (2.1 equiv) was added dropwise
at 41 °C (rather than at –40 °C) to the rapidly stirred THF solution of ethylene oxide (20a, 3 equiv)
and N,N-diisopropylamine (0.14 equiv). After work-up (GP2c), the basic fraction was treated with
This article is protected by copyright. All rights reserved.

boiling conc. HCl (GP4) and furnished a mixture (ca. 1:1) of the 2,5-product 26 (from 24a in Schemes
6 and 7) and the 2,2-acetal 29 (from 28). In contrast to the bridged spiro[4.4]acetal 25a, the 2,2-
product dioxapropellane[3.3.3] 29 (a tricyclo[3.3.3.01,5]undecane) was obviously stable against hot
conc. HCl. The resultant 26/29 mixture established that (Z)- and (E)-21a had performed anti
deprotonations with subsequent alkylations by 20a to generate 24a and 28 in parallel. The results
convinced us to confine our further attempts to the 2,5-route.
This article is protected by copyright. All rights reserved.

tBu
tBu
N
+41 °C
N
OLi
OLi
2
2
5
5
(Z)-21a
(E)-21a
+MeLi -CH4
+MeLi -CH4
22a
O
O
20a
20a
tBu
N
tBu
N
OLi
LiO
OLi
2
2
5
5
+
OLi
23a
(Z)-28
(H2O)
(H2O)
24a
conc. HCl, T
conc. HCl, T
8
2
O
O
1
7
3
Cl
Cl
O
9
6
5
2
+
5
4
11
10
26
29
Scheme 7. LDA-mediated, parallel 2,5- and 2,2-dialkylation by ethylene oxide (20a) at 41 °C; Me =
methyl, tBu = tert-butyl, T = bath temp. 130 °C.
This article is protected by copyright. All rights reserved.

Returning to Scheme 6, we report that racemic trans-2,3-dimethyloxirane (20b) is significantly less
reactive than 20a yet (under suitably changed conditions) able to run through the 2,5-dialkylation
cascade of ketimine 1 downwards to the bridged spiro[4.4]acetal 25b. The initial
deprotonation/alkylation sequence of 1 with LDA/H3CLi and 20b afforded (Z)-21b smoothly; but this
alkoxide could not be forced in situ to evolve CH4 on treatment with LDA/H3CLi at temperatures up
to 60 °C, presumably because (Z)-21b had already isomerized to (E)-21b with an overcrowded anti
position (C(2)). Therefore, we started the one-pot procedure from 1 in THF with HMPA/H3CLi (syn
deprotonator)[14] in the absence of LDA and observed that CH4 (1 equiv) was liberated already at –
42 °C. In order to ascertain a complete first alkylation, the oxirane 20b was added in excess (3.1
equiv). As a consequence, the subsequent second titration with H3CLi at 0 °C evolved CH4 at first
quickly through deprotonation[22] of unconsumed 20b (2 equiv) by H3CLi, followed by the desired
deprotonation of 21b (1 equiv). The generated azaallyllithium 22b was trapped with additional 20b
(3.1 equiv) at 56 °C, which led to slow fading of the yellow color in the course of 2 hours. The acidic
work-up (GP2c) afforded an exuberant basic product fraction. Although the expected primary
product 23b had changed into a complex mixture of cyclized N,O-acetals (such as 24b) whose
composition could not be deduced from the NMR spectra, the predominant 2,5-dialkylation was
established through an attempted choride insertion with conc. HCl (GP4) that led only to de-
amination: The emerging bridged 2,5-spiroacetal 25b (36% yield with respect to 1) was obviously
resistant to ring opening by boiling conc. HCl and remained so even during 85 hours under these GP4
conditions. This kind of a Thorpe-Ingold effect can be ascribed to the four methyl groups since the
above methyl-free, bridged spiro acetal 25a had no problem to generate its dichloro ketone 26. The
NMR spectra of 25b were compatible with a single, C2-symmetric diastereomer; in particular, the
3
three-bond interproton coupling constant of J = 6.6 Hz of the symmetry-equivalent 2-/3-H and 6-/7-
H pairs were compatible with a cis relationship of these protons as expected for an SN2-type attack
of the azaallyllithiums (first 3, then 22b) on the trans-2,3-dimethyloxirane (20b).
2.4. Oxiranes as electrophiles combined with iodomethane
The non-enolizable, almost rigid monocyclopropyl ketone 34a (Scheme 8) was easily accessible on
the 2,2-route through the following one-pot procedure and subsequent manipulations. The
ketinime (E)-12 was prepared (Scheme 3) from ketinime 1 along the lines of GP2b and then
deprotonated in situ to give the azaallyllithium 30. In view of the above experience with
trimethylation (Section 2.2), this syn deprotonation was performed with HMPA/n-BuLi at rt for 2
This article is protected by copyright. All rights reserved.

hours, whereupon ethylene oxide (20a) was added and the mixture stirred for 60 min in a bath of 30
°C. Acidic work-up (GP2c) furnished a basic fraction containing mainly the N,O-acetal 32a that was
formed through cyclization of the primary product 31a. The full set of the strongest 13C NMR signals
could be assigned to a single, preponderant diastereomer of 32a, while the detection of minor
isomers was thwarted by the accompanying impurities. Neither IR nor the 13C NMR spectra of the
product mixture exhibited C=N absorptions; but de-aminating ring cleavage in boiling conc. HCl
(GP4) furnished the expected γ-chloro ketone 33a (69%) as a constitutional confirmation for 32a.
The ensuing cyclization (GP5) and distillation furnished the 5,5-dimethyl spiroketone 34a (59%). An
alternative method of preparing 34a through methylation (74%) will be commended further below.
tBu
Li
tBu
tBu
O
R
N
N
N
R
OH
2
Me
Me
2
Me
Me
+n-BuLi
Me
Me
20a,b
5
R
R
HMPA
31a,b
(E)-12
30 (syn)
20, 31-34 R
a
b
H
CH3
H
tBu
O
4
H
2
KOEt
O
3
conc.
R
R
R
N O
H
Me
Me
2
5
Me
Me
R
8
1
Me
Me
3
HCl
T
5
Cl
5
6
7
R
R
34a,b
33a,b
32a,b
Scheme 8. One-pot, n-BuLi/HMPA-mediated, third alkylation approach toward the non-enolizable
spiro[2.4]heptan-4-ones 34a,b; Me = methyl, tBu = tert-butyl, T = bath temp. 130 °C.
The more encumbered trans-2,3-dimethyloxirane (20b) was significantly less than 20a suited for
alkylation of the intermediate azaallyllithium 30 (Scheme 8). Under the same conditions as above,
the one-pot procedure of generating (E)-12 and its deprotonation by HMPA/n-BuLi with subsequent
alkylation by 20b (2.5 equiv, 55 °C for 2 hours) generated an unresolved mixture of basic products
including unconsumed (E)-12 as the major component. For a possible explanation, we propose that
20b and the strongly basic azaallyllithium 30 may have performed the planned first SN2 alkylation to
give 31b more slowly than a proton transfer[22] from 20b to 30, so that (E)-12 was regenerated.
After a series of further deprotonation/alkylation steps, the subsequent de-aminating ring cleavage
(GP4) in boiling conc. HCl (second SN2 step) and final cylization of 33b (GP5, third SN2 step) produced
a crude ketone mixture that required separation by distillations and column chromatography until a
small sample of almost pure 34b was obtained. These three SN2 steps from trans-20b to 34b predict
This article is protected by copyright. All rights reserved.

a cis relationship between the 1-/2-CH3 groups and also between the 1-/2-H atoms, which was
1
established by means of the H NMR spectrum of 34b (Figure 1) as follows. The startling shapes of
the multiplets at δ = 1.46 and 0.95 ppm (Figure 1a) were simulated as an A3XX´A´3 spectral system
3
(Figure 1b) for the CH3–CH–CH–CH3 part of 34b, using the interproton coupling constants J(AX) =
4
4
5
3
3J(A´X´) = 6.5 Hz, J(AX´) = J(A´X) = 0.25 Hz, J(AA´) = 0 Hz, and J(XX´) = 9.1 Hz. The high magnitude
3
of the latter 1-H/2-H coupling J(XX´) established the expected cis relationship. NOESY cross peaks
disclosed that the 1-/2-CH3 protons resided in a cis position to the H2C(7) protons; hence, 34b had
been formed 20b (trans) stereospecifically as a single diastereomer.[23] Its very low preparative yield
exemplified a limitation of our one-pot technique. On the other hand, 34b might perhaps be
accessible from trans-2,3-dimethyloxirane (20b) and the less basic enolate of 2,2-
dimethylcyclopentanone (13).
O
H
H
2
Me
Me
CH
1
3
5
6
CH3
a
7
34b
0.95
1.46
5-CH3
b
1-/2-H
1-/2-CH3
1-/2-CH3
1-/2-H
6-H
7-H
This article is protected by copyright. All rights reserved.

1
Figure 1. Total H NMR (400 MHz) spectrum of 34b: Multiplets at 0.95 and 1.46 ppm were
expanded in the insert (a) for comparison with their negative spectral simulation (b) that established
the 1,2-cis configuration.
All of the preceding examples described regioselective second alkylation steps of 1, but only one
of them presented regioselectivity with two different iodoalkanes (Scheme 4) on the 2,2-route. We
will now show in Scheme 9 that regioselectivity can be achieved with two different electrophiles also
on the one-pot 2,5-route as far as anti deprotonation by LDA remains the obligatory pathway. The
results will then enable us to examine a third alkylation step for its regioselectivity.
H
N O
1
tBu
Me
tBu
2
tBu
2
O
N
N
3
+MeLi
-CH4
Me
OLi
5
2
Me
(H2O)
4
20a
5
8
6
7
35
(Z)-4
36
conc. HCl, T
+n-BuLi/HMPA
- n-BuH
0 °C
O
5
2
Me
Cl
then
+MeI
37
H
tBu
H
tBu
N O
N O
KOEt
+
Me
Me
Me
O
Me
4
Me
32a
5
3
38
39
conc. HCl, T
conc. HCl, T
+LDA
+MeI
O
4
O
O
Cl
Me
+
2
Me
Me
2
5
2
Me
Me
5
Me
1
3
Cl
5
6
7
KOEt
33a
40
34a
This article is protected by copyright. All rights reserved.

Schem 9. LDA-mediated one-pot dialkylation of ketimine 1 via (Z)-4 and 35 affords 37 and 39
regiospecifically, whereas the third deprotonation/alkylation of 35 by n-BuLi/HMPA is unselective;
Me = methyl, tBu = tert-butyl, T = bath temp. 130 °C.
As we know from Section 2.1, the LDA-mediated anti deprotonation of (Z)-4 in Scheme 9 must
take place before N-inversion converts (Z)-4 into (E)-4. The temperature (–23 °C) chosen for
deprotonating (Z)-4 with LDA/H3CLi was sufficiently low, as shown by the subsequent alkylation with
ethylene oxide (20a) which occurred exclusively at the 5-position and formed first the alkoxide 35
and therefrom the two diastereomeric N,O-acetals 36 (ca. 1:1). De-aminating ring cleavage of 36 in
boiling conc. HCl (GP4) furnished the two possible dastereomers (3:7) of the γ-chloro ketone 37 as
the almost only products. The entailing cyclization (GP5) provided the known[24] monomethyl
spiroketone 39. Subsequent methylation of 39 to give 34a can be recommended as an emergency
tool if 39 showed up as a contamination of 34a after an incomplete second methylation during the
GP2b.
The above regiospecific formation of the alkoxide 35 offered an opportunity to examine a third
deprotonation in the one-pot system for its regioselectivity. Having noted that this in-situ operation
1
required HMPA/n-BuLi as a base, we learned now by H NMR that n-BuLi with HMPA was not
consumed (no proton transfer from C(5) or C(2)) until the temperature was raised to 0 °C. The two
emerging azaallyllithiums (not shown) were trapped with iodomethane, whereupon protolytic work-
up generated the isomeric N,O-acetals 32a and 38 (9:10) but not more than traces of enamines or
products that carried C=O or C=N double bonds. Although only a single major diastereomer of 38
could be recognized, the subsequent constitutional confirmation through de-aminating ring cleavage
in boiling conc. HCl (GP4) afforded two diastereomers of 40 in the mixture of the γ-chloro ketones
33a and 40 (43:57). This ratio quantified the regio-unselectivity for the methylation of alkoxide 35;
possible reasons for such a behavior might be an N-inversion of 35 at 0 °C or an intrinsically
indiscriminating attitude of the unavoidable base system n-BuLi/HMPA. Purification of the two γ-
chloroketones 40 was easy because 40 survived the cyclization (GP5) process in a KOH/EtOH solution
without substitution or any otherwise imaginable cyclizations, while the accompanying 33a was
cyclized with formation of the more volatile spiro ketone 34a that was separated through distillation
from the higher-boiling γ-chloroketones 40.
This article is protected by copyright. All rights reserved.

3. Conclusions
(i) Our one-pot procedures were devised on the basis of conditions described[1,2] for 2-
iminobutanes. As a comfortrable improvement, the observation of gas (CH4) evolution from H3CLi
enabled us to perform the LDA-catalyzed deprotonations at or above –40 °C completely within
much shorter[2] periods of time. An especially high deprotonation propensity of our
iminocyclopentanes may have been helpful to prevent the imminent E/Z equilibration.
(ii) The following mechanistic peculiarities[1 - 3,8] of ketimine anti deprotonation necessitated the
application of one-pot methods: The emerging anti azaallyllithium species changed quickly into its
syn configuration; in order to enter the 2,5-route, this syn species should be alkylated at or below
ca. –70 °C since it will generate a metastable Z product. With warming to a maximum of –40 °C, the
second deprotonation step had to be performed before the Z to E stereoinversion would take place.
Thus, this kind of kinetic control relied on a prompt treatment of the (Z)-product with a trapping
agent. At 30 °C, on the other hand, the Z product equilibrated with the thermodynamically favored E
isomer whose anti deprotonation and methylation furnished the 2,2-dimethylated product with
good (95%) regioselectivity.
(iii) In contrast, the third deprotonation step was unselective with the two bases used in this
work: Deprotonation by LDA was slower than N-stereoinversion, while n-BuLi/HMPA reacted fast
but admitted syn deprotonation.
(iv) Anti deprotonation was obligatory with the LDA-repellent N-tBu group that impeded syn
deprotonations. Many other bulky N-alkyl (but not[5] N-aryl) groups at nitrogen might also be
suitable, especially the chiral ones.[25,26] Our observations and conclusions were consistent with the
results of very instructive kinetic and mechanistic analyses of the regioselective deprotonation and
methylation of N-alkyliminocyclohexanes.[27]
(v) Trapping of the intermediate azaallyllithiums was fast with iodomethane or ethylene oxide but
very slow with trans-2,3-dimethyloxirane that reacted stereospecifically. The resultant basic
products were always stable in cold, dilute aqueous solution. The NMR spectra of these one-pot
products could be interpreted only partially because of cases with complicating asymmetry effects
and the presence of basic side-products (sometimes together with enamine isomers). Therefore,
reliable product assignments could be achieved only through de-aminating (often uncomfortably
This article is protected by copyright. All rights reserved.

slow) hydrolytic cleavage that afforded methylated cyclopentanones, spiroacetals, a tricyclic acetal
(dioxapropellane), or 2-choroethylated cyclopentanones. In order to achieve higher total conversion
with volatile and sensitive cleavage products, it may be profitable to perform this de-amination in a
distillation apparatus for up to 10 hours. HCl elimination from the γ-chloroketones provided both
known and new spiro[2.4]heptan-4-ones via a reasonably short route.
4. Experimental
4.1. General Remarks. All NMR chemical shifts δ were referenced to internal Me4Si.
Abbreviations for NMR spin coupling multiplicities were as follows: d = doublet, m = multiplet, q =
quartet, quat = quaternary, qi = quintet, s = singlet, sept = septet, sext = sextet, t = triplet.
4.2. General Procedure 1 (GP1): Monoalkylation of 1. The ketimine 1 (2.78 g, 20.0 mmol) and
dry N,N-diisopropylamine (0.393 mL, 2.8 mmol) were dissolved in anhydrous THF (15 mL) and stirred
at –40 °C under argon gas cover. After the dropwise addition of methyllithium (20.0 mmol) in Et2O
(13.8 mL) with brisk evolution of CH4 gas during 15 min, the deeply yellow solution was cooled at –
70 °C and treated with iodomethane (MeI, 1.25 mL, 20.0 mmol), which caused an immediate
precipitation of LiI from the weakly yellow solution. The resulting metastable product (Z)-4 was not
characterized but was warmed up for the second deprotonation/alkylation step either to –40 °C
(2,5-route, GP2a) or to +30 °C (2,2-route, GP2b).
4.3. General Procedure 2 (GP2): Second alkylation of 1. The freshly prepared (GP1) solution of
(Z)-4 (nominally 20 mmol) and N,N-diisopropylamine (2.8 mmol) in THF/Et2O (39 mL) was used for
the two regioselective second alkylations as follows.
4.3.1. GP2a (2,5-route). The stirred (Z)-4 solution was warmed up from –70 °C to –40 °C and
without delay titrated with H3CLi (up to 23 mmol) in Et2O (ca. 14 mL) during 15 min. The amount of
H3CLi should suffice to consume (Z)-4 completely. The evolution of CH4 ceased after 12 min,
whereupon this orange-colored solution of the azaallyllithium 6 (syn) was quenched by the quick
addition of the electrophile MeI (1.50 mL, 24.1 mmol at –40 °C) or of an oxirane (at –60 °C). After
slow warming to rt, the solution of the dialkylated derivative was either subjected to further
deprotonation or worked up as described below in GP2c.
4.3.2. GP2b (2,2-route). The above (Z)-4 solution, as obtained in the GP1, was warmed up quickly
and stirred at 30 °C for 2 hours so as to equilibrate (Z)-4 and (E)-4. The mixture was cooled in an ice-
bath and titrated with H3CLi as long as CH4 gas was liberated (1.12 equiv within 55 min). This
This article is protected by copyright. All rights reserved.

solution of the created azaallyllithium 11 (syn) was cooled and quenched by the quick addition of the
electrophile, either MeI (1.2 equiv at –40 °C) or an oxirane (at ≤ –70 °C). After 90 min at rt, the
resultant solution of the 2,2-dialkyl derivative was either subjected to further deprotonation or
worked up as follows.
4.3.3. GP2c (work-up method). The above solution from GP2a or GP2b (up to 20 mmol of
products) was poured onto a mixture of aqueous hydrochlorid acid (2 M, 30 mL) and crushed ice (30
g). The aqueous layer (if sufficiently acidic) should contain the protonated basic products and was
shaken with Et2O (to be discarded), then alkalized with aqueous NaOH (2 M), and extracted with
Et2O (3 ×). These Et2O extracts were combined, washed with distilled water until neutral, dried over
Na2SO4, filtered, and concentrated to yield the basic products and accompanying contaminants.
4.4 General Procedure 3 (GP3): De-aminating hydrolysis of the basic products. A sample of the
crude mixture of the prepurified basic products (GP2) was dissolved in the 20-fold volume of
aqueous H2SO4 (2 N) and heated to reflux (double-walled reflux condensor) at 130 °C (bath temp.).
After cooling to rt, the condensor was rinsed with Et2O and distilled water, and the combined
aqueous phases were extracted with Et2O (3 ×). The combined (four) organic phases were washed
with distilled water until neutral, dried over Na2SO4, filtered, and distilled to separate the volatile
ketones from other non-basic compounds. The above acidic aqueous layer was stirred in an ice-bath
during alkalization with aqueous NaOH (“30%”), then extracted with Et2O (3 ×) and discarded. These
latter Et2O extracts were combined, washed with distilled water until neutral, dried over Na2SO4,
filtered, and concentrated to yield the unconsumed basic products and accompanying contaminants.
4.5. General Procedure 4 (GP4): Reactions in boiling conc. HCl. The prepurified (GP2) basic
product fraction was mixed with concentrated hydrochloric acid (2.5 mL per mmol of the expected
products) in a round-bottomed flask that carried a double-walled reflux condensor to prevent the
volatile compounds from escaping by way of a steam-distillation. The stirred mixture was heated to
reflux (bath temp. 130 °C) for at least 1 hour. Any emanating gaseous HCl was diverted toward a
neutralizing absorbent. After cooling to rt, the condensor was rinsed with Et2O and distilled water (2
×) into the product mixture which was then poured out of the flask with more water and Et2O. The
aqueous layer was shaken with Et2O (3 ×), and the combined Et2O phases were washed with distilled
water until neutral, dried over Na2SO4, filtered, and concentrated at below 40 °C down to a pressure
of not less than 70 Torr because some γ-chloro ketones were quite volatile. The crude product may
be used for further transformations.
This article is protected by copyright. All rights reserved.

Note: In case of an intended purification, it may be important to remove THF completely from the
initially employed material because the cleavage of THF by conc. HCl would generate 1,4-
dichlorobutane that may distil together with the γ-chloro ketones.
4.6. General Procedure 5 (GP5): Preparing the spiro[2.4] ketones through HCl elimination. The
following doses were calculated per y molar equiv of each chloroethyl function of the employed γ-
chloro ketone. Solid KOH (0.29y g, 5.17y mmol) was stirred in a round-bottomed flask under inert
gas cover and mixed with ethanol (99%, 1.1y mL). The mixture became hot during the complete
dissolution of KOH. After cooling in an ice-bath, a solution of the γ-chloro ketone in ethanol (99%,
0.8y mL) was added dropwise within y min and began to precipitate KCl immediately. After further
stirring in the closed vessel at rt for 2 hours, the mixture was diluted with distilled water until all KCl
was dissolved, whereupon this solution was shaken with pentane (4 ×). The combined organic
phases were washed with aqueous Na2CO3 (4 ×), dried over Na2SO4, filtered, concentrated at 40 °C
down to a pressure of not less than 70 Torr, and finally distilled.
4.7. N-(Cyclopentylidene)-tert-butylamine (1). A round-bottomed, three-necked flask (4 L) was
fitted with a strong mechanical stirrer, a two-walled reflux condensor, and a pressure-equalizing
dropping funnel (250 mL). The flask was charged with cyclopentanone (112 mL, 1.26 mmol) and
anhydrous Et2O (900 mL), thereafter with tert-butylamine (440 mL, 4.21 mol) under argon gas cover.
With strong stirring of this mixture at –78 °C, a solution of TiCl4 (72.5 mL, 661 mmol) in dry pentane
(600 mL) was added dropwise during 1 hour and began to produce a yellow precipitate immediately.
The mixture was warmed up and refluxed for 3 hours, then cooled to rt and filtered. The filter cake
was washed with Et2O until the filtrate became colorless. The combined filtrates were concentrated
to give a dark-red oil (134.4 g) that was distilled in a dry Vigreux column. The slightly yellow distillate
(88.8 g, 51%) was an oil with b.p. 56 – 62 °C/12 Torr (Ref. [6]: 69 – 70 °C/20 Torr) and consisted of
ketimine 1 and its enamine[28] 1A (82:18).
1H NMR of 1 ([D6]benzene, 400 MHz) δ 1.27 (s, 9H, tBu), 1.36 (qi, 3J = 6.9 Hz, 2H, H2C(3)), 1.50 (qi,
3J = 6.9 Hz, 2H, H2C(4)), 1.94 (t, 3J = 7 Hz, 2H, H2C(2)), 2.27 (t, 3J = 7.3 Hz, 2H, H2C(5)) ppm;
13C NMR of 1 ([D6]benzene, 100.6 MHz) δ 23.5 (tm, 1J = 130 Hz, H2C(3)), 26.5 (tm, 1J = 131 Hz,
H2C(4)), 30.3 (qm, 1J = 125 Hz, tBu), 30.9 (obscured t, H2C(2)), 38.8 (tm, 1J = 129 Hz, H2C(5)), 55.4 (>
sept, 2J = ca. 4 Hz, Me3C), 172.9 (broadened qi, 3J = ca. 5.6 Hz, C=N) ppm;
This article is protected by copyright. All rights reserved.

1H NMR of enamine 1A ([D6]benzene, 400 MHz) δ 1.17 (s, 9H, tBu), 1.78 (qi, 3J = 7.2 Hz, 2H,
H2C(4)), 2.17 (tq, 3J = 7.3 Hz, apparent J ≈ 2 Hz, 2H, H2C(3)), 2.52 (tq, 3J = 6.9 Hz, apparent J ≈ 2 Hz,
2H, H2C(5)), 2.78 (broad s, 1H, NH), 4.36 (pseudo-qi, apparent J = ca. 1.9 Hz, 1H, HC(2)), assigned
through comparison with 12;
13C NMR of enamine 1A ([D6]benzene, 100.6 MHz) δ 22.2 (obscured tm, 1J = ca. 130 Hz, H2C(4)),
29.1 (obscured qm, 1J = ca. 125 Hz, tBu), 32.5 (tm, 1J = 130 Hz, H2C(3)), 36.5 (broadened td, 1J = 130
Hz, 3J = 8 Hz to 2-H, H2C(5)), 50.4 (m, 2J = ca. 3.7 Hz, Me3C), 92.9 (broadened dm, 1J = 159 Hz, HC(2)),
143.9 (unresolved, C(1)) ppm, assigned as above.
4.8. N-(2-Methylcyclopentylidene)-tert-butylamine (4). Upon preparation (GP1) of (Z)-4, the
thermodynamically more stable diastereomer (E)-4 was generated in situ through stirring for 2 hours
at 30 °C. Both (Z)-4 and (E)-4 were not characterized but constitutionally confirmed through their
common hydrolysis (GP3) product 2-methylcyclopentanone (9).
4.9. N-(2,5-Dimethylcyclopenten-1-yl)-tert-butylamine (7A). The GP2a protocol was used to
generate the enamine 7A from 1 (20 mmol) and to isolate the basic fraction as a yellow oil (3.07 g).
This sample contained minor basic contaminants but no ketimine form and no 2,2-dimethyl isomer
(12).
NMR of enamine 7A (CCl4, 400 MHz) δ 0.97 (d, 3J = 6.8 Hz, 3H, 5-CH3), 1.13 (s, 9H, tBu), 1.49 (s, 3H,
2-CH3), rest obscured by the contaminants;
13C NMR of enamine 7A (CCl4, 100.6 MHz) δ 13.0 (q, 2-CH3), 18.9 (q, 5-CH3), 29.5 (t, H2C(4)), 30.8
(q, tBu), 33.3 (t, H2C(3)), 40.4 (d, HC(5)), 51.7 (quat, Me3C), 114.8 (quat, C(2)), 144.6 (quat, C(1)) ppm.
4.10. 2,5-Dimethylcyclopentanone (8) and 2-methylcyclopentanone (9). The constitution of 7a
and the presence of the monomethyl ketimine 4 (5%) in the above run were determined through de-
aminating hydrolysis (GP3, 4 hours) of a sample (2.97 g) in boiling H2SO4 (2 N, 44 mL) with final
distillation at 20 – 60 °C (bath temp.)/12 Torr: Colorless mixture (775 mg) of the ketones 8 and 9 in
the molar ratio of trans-8/cis-8/9 = 60:35:5.
This article is protected by copyright. All rights reserved.

1H NMR of 8 (CCl4, 400 MHz) δ 1.03 (d, 3J = 6.8 Hz, 2-/5-CH3 of cis-8), 1.07 (d, 3J = 7.1 Hz, 2-/5-CH3
of trans-8) ppm in accord with Refs. [29] and [30];
1H NMR of 9 (CCl4, 400 MHz) δ 1.04 (d, 3J = ca. 7 Hz, 2-CH3) ppm;
13C NMR of trans-8 (CCl4, 100.6 MHz) δ 14.7 (2-/5-CH3), 30.1 (H2C(3)/(4)), 43.1 (HC(2)/(5)), 218.7
(C=O) ppm;
13C NMR of cis-8 (CCl4, 100,6 MHz) δ 15.2 (2-/5-CH3), 28.8 (H2C(3)/(4)), 42.0 (HC(2)/(5)), 219.0
(C=O) ppm in accord with Ref. [30];
13C NMR of 9 (CCl4, 100.6 MHz) δ 14.2 (2-CH3), 20.6 (H2C(4)), 31.9 (H2C(3)), 36.8 (H2C(5)), 43.3
(HC(2)), 217.0 (C=O) ppm in accord with Ref. [31] in CDCl3;
4.11. N-(2,2-Dimethylcyclopentylidene)-tert-butylamine (12). The GP2b protocol was employed
to generate (E)-12 from 1 (15.0 mmol). The basic product fraction (2.14 g) contained mainly (E)-12
but not more than a trace of an enamine.
1H NMR (CCl4, 400 MHz) δ 0.95 (s, 6H, 2 × 2-CH3), 1.17 (s, 9H, tBu), 1.52 (t, 3J = 7 Hz, 2H, H2C(3)),
1.71 (qi, 3J = 7.2 Hz, 2H, H2C(4)), 2.27 (t, 3J = 7.4 Hz, 2H, H2C(5)) ppm;
13C NMR (CCl4, 100.6 MHz) δ 21.7 (t, H2C(4)), 27.0 (q, 2 × 2-CH3), 30.0 (q, tBu), 30.3 (t, H2C(3)), 38.9
(t, H2C(5)), 44.0 (quat, C(2)), 54.6 (quat, Me3C), 176.3 (quat, C=N) ppm;
4.12. 2,2-Dimethylcyclopentanone (13). The constitution of 12 was ascertained by de-aminating
hydrolysis (GP3, 4 hours) of a sample (2.00 g) in boiling H2SO4 (2 N, 30 mL) and subsequent
distillation at 20 – 52 °C (bath temp.)/12 Torr: The ketone mixture (753 mg) consisted of 2,2-
dimethyl- (13), trans- and cis-2,5-dimethyl- (8), and monomethylcyclopentanone (9) in the molar
ratio of 92:2:1:3.
1H NMR of 13 (CDCl3, 400 MHz) δ 1.04 (s, 6H, 2 × 2-CH3), 1.81 (t, 3J = 6.8 Hz, 2H, H2C(3)), 1.90 (qi, 3J
= 7.1 Hz, 2H, H2C(4)), 2.26 (t, 3J = 7.6 Hz, 2H, H2C(5)) ppm in accord with Ref. [32];
13C NMR of 13 (CDCl3, 100.6 MHz) δ 18.7 (t, H2C(4)), 23.8 (q, 2 × 2-CH3), 37.1 (t, H2C(5)), 38.5 (t,
H2C(3)), 44.9 (quat, C(2)), 223.6 (quat, C=O) ppm in accord with Refs. [29], [32], and [33];
This article is protected by copyright. All rights reserved.

4.13. N-(2-Isopropyl-2-methylcyclopentylidene)-tert-butylamine (14K) and the enamine N-(5-
isopropyl-5-methylcyclopenten-1-yl)-tert-butylamine (14A). Following the GP1 protocol, the
ketimine 1 (20.0 mmol) was methylated to give (Z)-4. After 2 hours at 30 °C, the generated product
(E)-4 was deprotonated (GP2b) by H3CLi at 0 °C with slow evolution of CH4 (540 mL, 21.6 mmol in 55
min). The emerging azaallyllithium (syn-11) was not methylated but instead treated at –40 °C with
2-iodopropane (3.00 mL, 30.0 mmol). After 45 min at rt, the work-up protocol of GP2c (2 M HCl, 60
mL, 60 g of ice) furnished a yellow oil (3.30 g, 75%) that contained the ketimine 14K and its enamine
14A (89:11 in CCl4) but no trace of the 2,5-isomer.
1H NMR of 14K (CCl4, 400 MHz) δ 0.72 and 0.85 (2 × d, 3J = 6.8 Hz, 2 × 3H, 2 diastereotopic
isopropyl CH3), 0.91 (s, 3H, 2-CH3), 1.18 (s, 9H, tBu), 1.30 and ca. 1.6 (2 × m, 1 + 1H, diastereotopic
H2C(3)), 1.62 and 1.73 (2 × m, 1 + 1H, diastereotopic H2C(4)), 1.89 (sept, 3J = 6.9 Hz, 1H, 2-H), 2.00
(dm, 2J = 16 Hz, 1H of H2C(5)), 2.45 (dm, 2J = 16 Hz, 1H of H2C(5)) ppm;
13C NMR of 14K (CCl4, 100.6 MHz) δ 17.59 and 17.92 (2 × q, 2 diastereotopic isopropyl CH3), 21.9
(t, H2C(4)), 24.7 (q, 2-CH3), 30.0 (q, tBu), 30.7 (t, H2C(3)), 32.2 (t, H2C(5)), 34.1 (d, 2-CH), 50.2 (quat,
C(2)), 54.6 (quat, Me3C), 176.6 (quat, C=N) ppm;
1H NMR of 14A (CCl4, 400 MHz) δ 0.80 and 0.86 (2 × d, 3J = 6.8 Hz, 2 × 3H, 2 diastereotopic
isopropyl CH3), 1.00 (s, 3H, 5-CH3), 1.23 (s, 9H, tBu), 4.10 (t, 2J = 2.3 Hz, 1H, 2-H) ppm, rest obscured;
13C NMR of 14A (CCl4, 100.6 MHz) δ 17.56 and 17.94 (2 × q, 2 diastereotopic isopropyl CH3), 25.0
(q, 5-CH3), 28.4 (t, CH2), 28.9 (q, tBu), 29.8 (t, CH2), 33.9 (d, 5-CH), 49.9 (quat, C(5)), 51.8 (quat,
Me3C), 91.8 (d, HC(2)), 147.8 (quat, C(1)) ppm;
IR (film) ν 2966, 2872, 1680 (s, C=N), 1645 (w, C=C), 1460, 1368, 1360, 1228, 1213 cm–1.
See 15 for the constitutional confirmation.
4.14. 2-Isopropyl-2-methylcyclopentanone (15). The acid-catalyzed, de-aminating hydrolysis
(GP3) of 14K/14A (3.15 g) with aqueous H2SO4 (2 N, 80 mL, 3.5 hours) afforded the known[12,13]
ketone 15 but left 1.43 g (45%) of 14K/14A intact. The pure sample of 15 (158 mg) distilled at 95 –
97 °C (bath temp.)/12 Torr (Ref. [12]: 68 – 70 °C/13 Torr).
1H NMR (CCl4, 400 MHz) δ 0.78 and 0.87 (2 × d, 3J = 6.8 Hz, 2 × 3H, 2 diastereotopic isopropyl CH3),
0.92 (s, 3H, 2-CH3), 1.82 (sept, 3J = 6.8 Hz, 1H, 2-CH), 1.56, ca. 1.9, 2.02, and 2.22 (6 diastereotopic H
of H2C(3)/(4)/(5)) ppm;
This article is protected by copyright. All rights reserved.

13C NMR (CCl4, 100.6 MHz) δ 17.2 and 18.0 (2 × q, 2 diastereotopic isopropyl CH3), 18.5 (t, H2C(4)),
20.5 (q, 2-CH3), 31.5 (t, H2C(3)), 32.2 (d, 2-CH), 37.9 (t, H2C(5)), 50.8 (quat, C(2)), 219.1 (quat, C=O)
ppm, assigned through comparison with 13; IR (film) ν 1736 cm–1;
MS (70 eV) m/z (%) 140.1 (M+, 26). 98.1 (100).
4.15. N-(2,2,5-Trimethylcyclopentylidene)-tert-butylamine (16). The enamine 7A was generated
(GP2a) from ketimine 1 (5.0 mmol) in THFG/Et2O (ca. 12 mL) and treated in situ at rt with anhydrous
HMPA (0.876 mL, 0.894 g, 5.0 mmol) and then n-BuLi (6.0 mmol) in hexane (2.5 mL). The orange-
colored solution got a yellow tint withim the next 20 min. After further stirring at rt for 100 min, the
mixture was cooled in ice for the addition of iodomethane (0.375 mL, 6.0 mmol). The treatment
with n-BuLi and then iodomethane was repeated in an unsuccessful attempt to achieve
tetramethylation. Work-up according to GP2c furnished a yellow oil (900 mg) as the basic fraction
that contained the crude ketimine 16 and only a trace of its enamine; the very small portion of
tetramethylation became visible only after hydrolysis (to be seen below).
1H NMR of 16 (CCl4, 400 MHz) δ 0.91 (s, 3H, 2-CH3), 1.06 (d, 3J = 7.3 Hz, 3H, 5-CH3), 1.07 (s, 3H, 2-
CH3), 1.21 (s, 9H, tBu), 2.83 (pseudo-qi, 3J = 7.3 Hz, 1H, 5-H) ppm, rest obscured by contaminants;
13C NMR of 16 (CCl4, 100.6 MHz) δ 19.3 (q, 5-CH3), 28.7 (q, 2-CH3), 29.6 (q, 2-CH3), 30.6 (t, H2C(3)),
31.0 (q, tBu), 35.6 (t, H2C(4)), 35.8 (d, HC(5)), 43.5 (quat, C(2)), 55.4 (quat, Me3C), 178.7 (quat, C=N)
ppm;
4.16. 2,2,5-Trimethylcyclopentanone (17) and the tetramethyl congener. Acid-catalyzed, de-
aminating hydrolysis (GP3) of a sample (810 mg) of the above crude product 16 in boiling aqueous
H2SO4 (2 N, 22 mL, 1 hour) left 416 mg (52%) of 16 intact and afforded a distilled mixture (64 mg) of
2,2,5-trimethyl- (17) and 2,2,5,5-tetramethylcyclopentanone (19) in a 97:3 ratio with b.p. 40 – 85 °C
(bath temp.)/12 Torr. The same ratio and similar results were obtained by the same treatment of
the ketimine 12 with HMPA/n-BuLi/MeI and entailing hydrolysis under similar conditions.
1H NMR of 17 (CDCl3, 400 MHz) δ 0.98 and 1.07 (2 × s, 3 + 3H, 2 × 2-CH3), 1.12 (d, 3J = 6.8 Hz, 3H,
5-CH3), 1.50 (m, 1H), 1.68 (m, 1H), 1.81 (m, 1H), 2.15 (m, 1H), 2.21 (m, 1H) ppm in accord with Refs.
[29] and [34];
This article is protected by copyright. All rights reserved.

13C NMR of 17 (CDCl3, 100.6 MHz) δ 15.3 (q, 5-CH3), 24.1 and 24.8 (2 × q, 2 × 2-CH3), 28.0 (t,
H2C(4)), 36.5 (t, H2C(3)), 43.3 (d, HC(5)), 44.9 (quat, C(2)), 225.0 (quat, C=O) ppm in accord with Refs.
[33] and [34];
1H NMR of 19 (CDCl3, 400 MHz) δ 1.05 (s) ppm; 13C NMR of 19 (CDCl3, 100.6 MHz) δ 24.9 (q,
2,2,5,5-CH3), 34.9 (t, H2C(3)/(4)), 45.3 (quat, C(2)/C(5)) ppm in accord with Ref. [33].
4.17. 2,5-Bis(2-hydroxyethylation) of ketimine 1 to give the N,O-acetal 1-tert-butylamino-8-(2-
hydroxyethyl)-2-oxabicyclo[3.3.0]octane (24a). A dry three-necked flask (1 L) was charged with the
ketimine 1 (27.82 g, 200 mmol), anhydrous THF (200 mL), dry N,N-diisopropylamine (3.93 mL, 27.9
mmol), and a strong magnetic stirring bar. This solution was cooled to –60 °C under argon gas cover
and treated with a stock solution (4.08 M, 147 mL) of ethylene oxide (600 mmol) in THF.
Methyllithium (420 mmol) in Et2O (280 mL) was added dropwise with rapid stirring at –47 °C during
3 hours. The stirred mixture was warmed up to rt (1 hour) and then protonated with distillled water
(100 mL). The aqueous layer was shaken with Et2O (3 ×), and all four organic phases were combined
and washed with aqueous Na2CO3 (2 N, 3 ×), then dried over Na2SO4, filtered, and evaporated to
leave an orange-colored, viscous oil (48.32 g) that solidified partially (ca. one half). Recrystalllization
from cyclohexane afforded small, colorless blocks of 24a with m.p. 67.2 – 68.1 °C. A solution of
these crystals in CH2Cl2 or in CCl4/CDCl3 (2:1) contained two diastereomeric N,O-acetals (ca. 1:1, cis
and trans?) whose NMR spectra could not be disentangled but displayed no C=O or C=N resonances.
IR (KBr) ν 3394 (s, narrow, O–H), 3347 (w, very sharp N–H), 2956, 2920 (w), 2881, 1520 (w), 1480
(w), 1456 (w), 1362, 1352, 1231 (s), 1050 (s), 1013 (s) cm–1, no C=O and no C=N vibrations. Anal.
calcd for C13H25NO2 (227.35): C, 68.68; H, 11.08; N, 6.16. Found: C, 68.91; H, 10.87; N, 6.07.
4.18. The N,O-acetal 24b formed by dialkylation with trans-2,3-dimethyloxirane (20b). The
ketimine 1 (2.10 g, 15.1 mmol), dry HMPA (2.01 mL, 2.05 g, 11.5 mmol), anhydrous THF (15 ml), and
a magnetic stirring bar were placed in a three-necked, round-bottomed flask (100 mL) that was
fitted with a dry reflux condensor that was connected to a gas burette. The rapidly stirred solution
was cooled to –42 °C during the dropwise addition of H3CLi (ca. 15 mmol) in Et2O (ca. 10 mL), which
liberated CH4 (375 mL, 15.0 mmol) at first instantly and the last fifth within 10 min. The yellow
solution was cooled to –72 °C and treated with trans-2,3-dimethyloxirane (20b, 4.24 mL, 47.3
mmol). The colorless solution was warmed to 0 °C and titrated again with H3CLi: The generation of
CH4 continued up to rt until deprotonation of the monoalkylated intermediate 4 and of unconsumed
This article is protected by copyright. All rights reserved.

oxirane 20b (ca. 30 mmol)[22] was complete. The resultant yellow solution was treated with another
portion of 20b (47.3 mmol) and heated at a dry-ice-cooled condensor at 56 °C for 2 hours with
gradual decolorization. The acidic work-up protocol of GP2c furnished the basic product fraction
(3.52 g) as a viscous, yellow oil whose NMR spectra were too complicated for an instructive analysis;
therefore, the constitution of 24b was ascertained through de-aminatig ring cleavage as described
for 25b.
4.19. (3aα,5aβ)-Octahydrocyclopenta[1,2-b;1,5-b´]difuran (25a). This known[16] bridged
spiroacetal was obtained through acid-catalyzed, de-aminating hydrolysis (GP3, 2 hours) of 24a with
boiling aqueous H2SO4 (2 N). After work-up, it distilled as a colorless liquid with b.p. 143 – 148 °C
(bath temp.)/70 Torr (Ref. [16]: 155 – 160 °C (bath temp.)/70 Torr).
1H NMR (CCl4, 80 MHz): As in Ref. [16]; 13C NMR (CDCl3, 20.15 MHz) δ 31.14 (2 × CH2), 31.50 (2 ×
CH2), 47.7 (2 × CH), 69.2 (2 × OCH2), 128.8 ppm (1 × quat C) as in Ref. [16]. Anal. calcd for C9H14O2
(154.2): C, 70.10; H, 9.15. Found: C, 69.58; H, 9.24.
4.20. (3aα,5aβ)-2,3,6,7-Tetramethyloctahydrocyclopenta[1,2-b;1,5-b´]difuran (25b). A portion
(3.37 g) of the crude N,O-acetal 24b was de-aminated (GP4, 3.5 hours) in boiling conc. HCl which
yielded a non-basic oil (2.08 g) whose distillation afforded the bridged spiroacetal 25b (1.14 g, 36%).
Analytically pure 25b had b.p. 60 – 65 °C (bath temp.)/0.001 Torr and exhibited C2-symmetry
according to the 13C NMR data.
1H NMR (CDCl3, 400 MHz) δ 0.95 (d, 3J = ca. 7 Hz, 6H, 3-/6-CH3), 1.18 (d, 3J = 6.6 Hz, 6H, 2-/7-CH3),
1.45 (m, 2 × 1H, 4-/5-H), 1.87 (sext, 3J = ca. 6.6 Hz, 2H, 3-/6-H), 1.94 (m, 2 × 1H, second 4-/5-H), 2.06
(m, 2H. 3a-/5a-H), 4.11 (qi, 3J = ca. 6.6 Hz, 2H, 2-/7-H) ppm, assigned through DQCOSY, NOESY, and
HETCOR;
13C NMR (CDCl3, 100.6 MHz) δ 13.9 (q, 3-/6-CH3), 16.3 (q, 2-/7-CH3), 28.2 (t, H2C(4)/(5)), 41.1 (d,
HC(3)/(6)), 55.4 (d, HC(3a)/(5a)), 78.2 (d, HC(2)/(7)), 126.9 (quat, C(9)) ppm, assigned through DEPT
and HETCOR;
IR (film) ν 2963, 2935, 2872, 1460, 1379, 1116, 1097, 1063, 1052, 991, 960 cm–1.
Anal. calcd for C13H22O2 (210.3): C, 74.24; H, 10.54. Found: C, 74.39; H, 10.44.
This article is protected by copyright. All rights reserved.

4.21. 2,5-Bis(2-chloroethyl)cyclopentanone (26). The crude N,O-acetal 24a (48.32 g, ≤ 200
mmol) was heated (GP4) in boiling conc. HCl (500 mL) for 2.5 hours. After work-up according to GP4,
the brown oil (35.92 g, 86%) was distilled (2 ×) at 110 – 115 °C (bath temp.)/10–5 Torr to give 26 as a
colorless liquid[18] with an intense odor of pine and a trans/cis ratio of 4:1.
1H NMR of trans and cis 26 (CCl4, 400 MHz) δ 1.47, 1.67, 1.72, 2.09, 2.19, 2.26, and 2.42 (7 × m,
10H), 3.58 and 3.68 (2 × m, 4H, 2 × CH2Cl) ppm;
13C NMR of trans 26 (CCl4, 100.6 MHz) δ 27.6 (tm, 1J = 131 Hz, 2-/5-CH2), 33.2 (broadened t, 1J =
129 Hz, H2C(3)/(4)), 42.6 (tm, 1J = 150 Hz, 2 × CH2Cl), 46.2 (dm, 1J = 124 Hz, HC(2)/(5)), 217.3 (quat,
C=O) ppm;
13C NMR of cis 26 (CCl4, 100.6 MHz) δ 26.6 (tm, 1J = 130 Hz, 2-/5-CH2), 33.3 (obscured t,
H2C(3)/(4)), 42.5 (obscured tm, 1J = ca. 150 Hz, 2 × CH2Cl), 45.0 (obscured dm, 1J = ca. 125 Hz,
HC(2)/(5)), 217.3 (quat, C=O) ppm;
IR (film) ν 2963, 2870, 1732 (s), 1451, 1432, 1160, 652 cm–1.
Anal. calcd for C9H14Cl2O (209.1): C, 51.69; H, 6.75. Found: C, 52.49; H, 6.95.
4.22. Dispiro[2.1.2.2]nonan-4-one (27). This known[20,21] dispirocyclopentanone was prepared
according to GP5 but under mild reflux as reported in Ref. [19].
1H NMR (CDCl3, 400 MHz) δ 0.88 and 1.16 (2 × pseudo-qi, apparent J = 3.3 Hz, 2 × 4H, 2 × 1-/2-/6-
/7-H), 2.10 (s, 4H, H2C(8)/(9)) ppm;
13C NMR (CDCl3, 100.6 MHz) δ 17.4 (sharp t, 1J = 164.0 Hz, H2C(1)/(2)/(6)/(7) of the cyclopropane
part), 29.8 (tm, 1J = 132.0 Hz, H2C(8)/(9)), 30.0 (quat, C(2)/(5)), 218.9 (quat, C=O) ppm;
4.23. 2,8-Dioxatricyclo[3.3.3.01,5]undecane (29). The ketimine 1 (0.974 g, 7.0 mmol), dry N,N-
diisopropylamine (0.138 mL, 0.98 mmol), and a stock solution (4.88 mL) of ethylene oxide (20a, 21.0
mmol) were dissolved in anhydrous THF (7.0 mL). The rapidly stirred solution was heated to 41 °C
under inert gas cover and titrated with H3CLi in Et2O, which liberated 355 mL (14.2 mmol) of
methane. The work-up protocol GP2c afforded a viscous mixture (1.07 g, ≤ 67%) of 24a and the
corresponding 2,2-isomer (protonated 28) that was treated with boiling conc. HCl (GP4) and
This article is protected by copyright. All rights reserved.

furnished the dioxapropellane 29 (208 mg) with b.p. 102 – 112 °C (bath temp.)/12 Torr as a colorless
liquid.
1H NMR (CCl4, 80 MHz) δ 1.66 (broadened quasi-s, 10H, 2 × CH and 4 × CH2), 3.71 (m, 4H, 2 ×
OCH2) ppm;
13C NMR (CDCl3, 20.15 MHz) δ 34.2, 35.8, and 37.7 (3 × t, H2C(9)/(10)/(11)), 39.4 (t, H2C(4)/(6)),
58.5 (quat, C(5)), 68.2 (t, OH2C(3)/(7)), 127.7 (quat, C(1)) ppm;
Anal. calcd for C9H14O2 (154.2): C, 70.10 ; H, 9.15. Found : C, 69.58 ; H, 9.24.
4.24. 1-tert-Butylamino-8,8-dimethyl-2-oxabicyclo[3.3.0]octane (32a). The protocol of GB2b was
followed for the twofold deprotonatio and methylation of ketimine 1 (2.78 g, 20.0 mmol) that
afforded (E)-12 and methane (500 and 532 mL) at –40 °C. The stirred solution was kept at –40 °C
during the addition of HMPA (1.48 mL, 8.43 mmol) and then of n-BuLi (24.0 mmol) in hexane (9.60
mL). After stirring at rt for 2 hours, the solution was cooled again to –40 °C, treated with a stock
solution (10.2 mL) of ethylene oxide (40.0 mmol) in THF, and kept in a bath of 30 °C for 60 min. The
red solution was worked up (GP2c) to give a faintly yellow oil (4.38g) as the basic product fraction
that contained mainly one diastereomer of the N,O-acetal 32a. The constitution of 32a was
confirmed below through conversion into 33a.
1H NMR (CCl4, 400 MHz) δ 0.90 (s, 3H, 1 × 8-CH3), 0.97 (s, 3H, 1 × 8-CH3), 1.12 (s, 9H, tBu), 1.48,
1.55, 2.06, 2.37, 2.53, and 3.65 (6 × m, 6 × ca. 1H), 3.98 (td, 3J = 8.5 Hz, 4J = 2 Hz, 1H) ppm, rest
obscured;
13C NMR (CCl4, 100.6 MHz) δ 22.7 (qm, 1J = 124 Hz, 8-CH3), 26.0 (qm, 1J = 124 Hz, second 8-CH3),
30.3 (obscured tm, 1J = ca. 122 Hz, CH2), 32.1 (qm, 1J = 125 Hz, tBu), 35.4 (tm, 1J = 128 Hz, CH2), 39.6
(tm, 1J = 128 Hz, CH2), 46.0 (broadened dm, 1J = 135 Hz, HC(5)), 48.3 (narrow m, C(8)), 49.9 (> sext,
apparent J = 3.8 Hz, Me3C), 67.5 (tm, 1J = 144.5 Hz, H2C(3)), 108.3 (quat, C(1)) ppm;
IR (CCl4) ν 2959, 2869, 1462, 1362, 1222 cm–1.
4.25. 5-(2-Chloroethyl)-2,2-dimethylcyclopentanone (33a). The above N,O-acetal 32a (4.29 g, ≤
19.6 mmol) was de-aminated by boiling conc. HCl (GP4, 2 hours), which afforded the crude γ-chloro
ketone 33a (2.78 g, almost 80%) as a slightly yellow oil. The analytically pure, colorless liquid (2.39 g,
69% yield with respect to 1) had b.p. 69 – 71 °C/0.001 Torr.
This article is protected by copyright. All rights reserved.