Studies on the carbenium-iminium ions derived from N-methylmorpholine-N- oxide (NMMO)

Article abstract of DOI:10.1016/j.tet.2003.11.015

Two carbenium-iminium ions, an exo-centered species 2 and a ring-centered form 3, are derived from the widely used oxidant and cellulose solvent N-methylmorpholine-N-oxide (1) by heterolytic degradation. 3 rearranges into 2 in the presence of water, in an endothermic, bimolecular reaction involving a highly organized transition state, which is the first example of a carbenium-iminium ion interconversion. The reaction mechanism was investigated by a combined approach consisting of trapping reactions, isotopic labeling, kinetic studies, and computations on the DFT level.

Full text of DOI:10.1016/j.tet.2003.11.015

Tetrahedron 60 (2004) 301–306
Studies on the carbenium-iminium ions derived from
N-methylmorpholine-N-oxide (NMMO)
*
Thomas Rosenau, Antje Potthast and Paul Kosma
Institute of Chemistry, University of Natural Resources and Applied Life Sciences, Christian-Doppler-Labor, A-1190 Vienna, Austria
Received 2 September 2003; revised 29 October 2003; accepted 7 November 2003
Abstract—Two carbenium-iminium ions, an exo-centered species 2 and a ring-centered form 3, are derived from the widely used oxidant
and cellulose solvent N-methylmorpholine-N-oxide (1) by heterolytic degradation. 3 rearranges into 2 in the presence of water, in an
endothermic, bimolecular reaction involving a highly organized transition state, which is the first example of a carbenium-iminium ion
interconversion. The reaction mechanism was investigated by a combined approach consisting of trapping reactions, isotopic labeling,
kinetic studies, and computations on the DFT level.
q 2003 Published by Elsevier Ltd.
1
. Introduction
species which are formed by cleavage of the N–O bond
with simultaneous transfer of one electron. In these
processes NMMO acted as a one-electron oxidant. In the
6
N-Methylmorpholine-N-oxide (NMMO, 1) is a frequently
1
applied oxidant in organic synthesis, which is mainly
applied in combination with catalytic amounts of transition
present work, we wish to communicate the peculiar
behavior of carbenium-iminium ions (C-I ions, Mannich
intermediates) derived from NMMO, the second major class
of NMMO-derived intermediates. These rather labile
species are generated by breaking the N–O bond of the
amine N-oxide with concomitant two-electron transfer. In
the case of NMMO, two of such species can be expected, the
exo-centered C-I ion 2 and the ring-centered C-I ion 3.
2
metal catalysts. It is also used in bulk quantities as a solvent
for cellulose in the industrial Lyocell process, which is a
new and environmentally benign approach to production of
3
man-made cellulosic fibers. NMMO dissolves cellulose
directly without chemical derivatization to give a dope
which is spun simply into water. This is in complete contrast
to viscose rayon production, which requires intensive use of
both derivatization and spinning bath chemicals.
NMMO is a relatively strong oxidant that tends toward
uncontrollable decomposition reactions. Both upon usage as
4
oxidant in organic synthesis, as well as upon industrial
5
utilization of NMMO as cellulose solvent there have been
reports on the instability of NMMO solutions and on the
occurrence of unpredictable thermal runaway reactions.
Due to these irregularities great caution must be exercised in
all processes involving NMMO. It is evident that the interest
in side reactions of NMMO and the NMMO-derived
chemical species involved therein originate from very
different areas of NMMO utilization.
C-I ions have been shown to be able to cause O-alkylation of
tertiary amine N-oxides. The alkylation step is immediately
followed by fragmentation in a concerted mechanism
regenerating the C-I ion. The reaction, being an auto-
7
catalytic and highly exothermic process, thus easily
In previous studies on side reactions of NMMO we have
reported on radicals derived from this amine N-oxide, i.e.,
becomes uncontrollable causing complete degradation of
the parent amine N-oxide. Also in the case of NMMO
reaction mixtures, C-I ion intermediates catalyze its
autocatalytic degradation, and are thus mainly responsible
for the observed instabilities and exothermicities in NMMO
Keywords: Carbenium-iminium ions; N-Methylmorpholine-N-oxide
(NMMO); Mannich intermediates; Trapping; Isotopic labeling; Cellulose
solvents.
4
reaction mixtures, which gave the impetus to the present
study.
*
Corresponding author. Tel.: þ43-1-36006-6071; fax: þ43-1-36006-6059;
e-mail address: thomas.rosenau@boku.ac.at
0040–4020/$ - see front matter q 2003 Published by Elsevier Ltd.
doi:10.1016/j.tet.2003.11.015

302
T. Rosenau et al. / Tetrahedron 60 (2004) 301–306
. Results and discussion Table 1. Different reaction conditions for the conversion of C-I ion 3 into
C-I ion 2 followed by trapping
2
Even though 3 might appear more stable at a first glance,
since its ring-centered positive charge seems ‘better
accommodated’ than the one in 2, it is the latter which
occurs predominantly in all different types of side reactions
starting from NMMO. In fact, intermediate 2 was produced
Starting material
(
Solvent
50 mL), treatment
(24 h) before addition of trap
Product ratio
2
3
2
3
2
3
2
3
2
3
2
3
2
3
2
3
3
3
3
Dry toluene, stirring at rt
Dry toluene, stirring at rt
Wet toluene (2% H
Wet toluene (2% H
Dry dioxane, stirring at rt
Dry dioxane, stirring at rt
Dioxane/water (3/1), stirring at rt
Dioxane/water (3/1), stirring at rt
Dry toluene, reflux
5 only
6 only
5 only
12/88
5 only
6 only
5 only
8/92
5 only
6 only
5 only
83/17
5 only
6 only
5 only
79/21
5 only
90/10
5 only
8
exclusively in NMMO reaction mixtures containing water,
whereas in non-aqueous solutions of NMMO also smaller
2
2
O), stirring at rt
O), stirring at rt
9
amounts of 3 were found besides 2. Thus, both the presence
of water and higher temperatures generally seemed to
promote the consumption of 3 in favor of 2.
To study the reactions of these two NMMO-derived
Mannich intermediates in more detail, the crystalline
chloride forms of 2 and 3 were used, which were prepared
according to Scheme 1. As reporter method, the trapping
with 2-acetonaphthone (4) in a Mannich reaction was used.
The trapping agent reacts with C-I ions in a fast and neat
reaction to give the corresponding Mannich bases 5 and 6,
respectively, which are readily extractable even from very
complex mixtures due to their lipophilic naphthalene
moiety. Another advantage is the usability of the trap also
at elevated temperatures (e.g., in NMMO melts), since the
resulting Mannich bases are relatively thermostable and do
not undergo elimination to the a,b-unsaturated ketone
below 130 8C.
Dry toluene, reflux
Wet toluene (2% H
Wet toluene (2% H
Dry dioxane, reflux
Dry dioxane, reflux
Dioxane/water (3/1), reflux
Dioxane/water (3/1), reflux
Dry DMF, stirring at 120 8C
DMF/water (9/1), stirring at 120 8C
Dry DMF, stirring at 120 8C
2
2
O), reflux
O), reflux
corresponding Mannich bases 5 and 6, respectively,
quantitatively. The same was true for the reaction of 2 in
aqueous dioxane, 3 afforded small amounts of 5 besides the
expected trapping product 6. Refluxing 2 in aqueous
dioxane for 24 h before carrying out the trapping reaction
had no effect as compared to the room temperature process:
only 5 was detected. In the case of 3, however, 5 and 6 were
formed in a 83/17 ratio. Thus, the major part of 3 must have
As shown in Table 1, trapping of both 2 and 3 in non-
aqueous organic solution at room temperature provided the
1
0
rearranged during the refluxing into 2. A similar refluxing
test carried out in dry dioxane or dry toluene gave no
indications of such a reaction: in the case of 2 as the starting
material only 5 was produced, and in the case of 3 only 6. In
refluxing wet toluene, in contrast, the rearrangement of 3
into 2 occurred. In general, replacing the refluxing
procedure by room-temperature stirring provided essen-
tially the same results in all solvents, but the 3 into 2
conversion proceeded much slower. The results of the
experiments are summarized in Table 1. It was evident that
higher temperatures and the presence of water favored 2 at
the expense of 3.
First, we assumed that 3 might undergo a [1,3]-sigmatropic
shift to give 2, meaning that 2 basically would be a more
stable tautomeric form of N-(alkylene)-morpholinium ions.
If this assumption was true, 3-d should have rearranged
3
into 2-d , which in turn should have produced 5a upon
3
trapping (see Scheme 2). However, this product was not
observed, but only 5b, the bisdeuterated compound, was
found. No incorporation of deuterium into the N-methylene
group of the morpholine ring—and thus no [1,3]-shift—had
occurred. Consequently, the proton entering the morpholine
ring must have come from the solvent. Indeed, reaction of 3
in dioxane/D O (or toluene/D O) provided 5c, which
2
2
proved the participation of water as a solvent component
in the reaction. Thus, a solvent proton is incorporated into
the morpholine ring, while a proton from the exo-methyl
group is lost into the solvent (Scheme 2).
Scheme 1. Synthesis of stable carbenium-iminium chlorides 2 and 3 and
their trapping with 2-acetonaphthone in a Mannich reaction.
Kinetic studies have shown the reaction to be of second

T. Rosenau et al. / Tetrahedron 60 (2004) 301–306
303
a molecule of water were involved in the rate-determining
step of the reaction, having a highly organized transition
state.
The stability of the two C-I ions was assessed by means of
2
1
DFT computations, which showed that 2 is by 2.9 kJ mol
1
4
more stable than 3. In the minimum geometries of both 2
2
and 3 the positive carbons are sp -hybrids with trigonal
planar geometry (Scheme 3). The four atoms of the
N-(methylene) group in 2 lie in one plane; the morpholine
ring is nevertheless able to adopt its typical chair
conformation. In contrast, the morpholine ring of C-I-ion
3
is forced into a distorted chair geometry by the trigonal
planar geometry of the carbenoid carbon. Thus, the spatial
arrangement of the three substituents at the carbenium
center can only in 2 be accommodated without additional
strain, the resulting twist of the morpholine chair rendering
3
energetically more unfavorable than 2 (Scheme 3).
However, this rather small energy difference could neither
account for the experimentally observed preference of 2, nor
explain the conversion of 3 into 2. Thus, in the next step a
1
5
DFT analysis of the system (3þH O) was performed,
2
which now produced a plausible picture that agreed with the
experimental data (Scheme 3).
According to Scheme 3 the conversion 3!2 follows a
reaction path that requires a rather high activation energy of
DH ¼173.8 kJ mol , which compares very favorably with
#
21
2
1
the experimental value of E ¼169.9 kJ mol . Going from
A
3
to the transition state, the positive charge must be
localized at the nitrogen, so that the double bond becomes
located and the resonance stabilization of the carbenium-
iminium ion is lost. In the transition state, water and the
CI-ion 3 arrange in a way that places the reaction centers
into a favorable six-membered chair-geometry. One water
Scheme 2. Use of isotopically labeled material to clarify the reaction
mechanism of the C-I ion conversion.
˚
hydrogen is placed into a distance of 1.26 A to the ring-
order, i.e., first order with respect to both water and C-I ion
. Thus, the reaction was assumed to be a bimolecular
carbon, which comes close to the length of a C–H single
˚
bond in the N–CH groups of N-alkylamines with 1.1 A.
2
3
process according to rate law d[2]/dt¼k[3][H O]. The
The ring carbon reaction center, which is nearly trigonal
planar in 3, already adopts a tetragonal environment in the
transition state, which removes the strain from the six-
membered ring system. The C–C–N angle (109.838) almost
2
kinetic rate constant was determined to be
. Evaluation of the temperature
k¼1.40£10 L mol21
2
4
s
dependence of k according to the Eyring equation
21 11
1
2
produced the activation parameters of the reaction: a rather
#
21
large activation enthalpy DH ¼167.4^7.2 kJ mol , the
#
Arrhenius activation energy thus being E ¼DH þ
A
2
1
RT¼169.9^7.2 kJ mol at 298 K, and a strongly negative
#
21
activation entropy DS ¼2142^13.8 J(mol K) indicating
a high degree of order in the transition state with limited
mobility of the two coreactants as compared to the starting
material.
The transmutation of 3 into 2 was prevented by addition of
2
either acids or alkali. In the case of OH the C-I ion was
evidently quenched to the corresponding N-hydroxyalkyl
compound, so that neither rearrangement nor reaction with
3
1
the trapping agent occurred. The presence of strong acids
1 equiv. rel. to 3) suppressed the transformation of 3 into 2
(
almost completely, whereas simple trapping proceeded
faster than in the neutral case due to acid catalysis effect.
This led us to the assumption that the basic oxygen functions
in water were required for the C-I ion conversion, which
was no longer present due to protonation in acidic media. In
summary, the kinetic data indicated that a molecule of 3 and
Scheme 3. Schematic representation of the computed reactant and
transition state geometries as well as reaction energetics for the C-I ion
21
conversion of 3 into 2. Values are given in kJ mol . The trigonal planar
environments of the carbenoid carbons are shaded in gray. The participating
water molecule is circled by a gray line.

304
T. Rosenau et al. / Tetrahedron 60 (2004) 301–306
3
equals the normal value of a tetrahedral angle in carbon sp
hybrids. Compared to the N–CH3 group in N-methyl-
3. Experimental
˚
morpholine having a C–H length of 1.097 A, the activated
˚
C–H bond in the exo-CH group is stretched to 1.61 A,
3.1. General
3
which indicates advanced bond cleavage. The exo-carbon
2
almost obtained the trigonal planar geometry of an sp -
All chemicals were commercially available. Thin layer
chromatography (TLC) was performed on silica gel 60
plates (5£10 cm, 0.25 mm) with fluorescence detection
under UV light at 254 nm. Column chromatography was
performed on silica gel G60 (40–63 mm). Melting points,
determined on a Kofler-type micro hot stage with Reichert-
hybrid, with the H–C–H plane standing roughly perpen-
dicular on the C–N–C plane of the morpholine ring. The
distances of the three protons bound to the water oxygen are
˚
nearly equal (0.99–1.08 A), close to that in a hydronium ion
þ
state is rather close to that of the products, 2 and H O,
˚
1
H O (0.98 A). Thus, the overall geometry of the transition
3
Biovar microscope, are uncorrected. H NMR spectra were
1
3
recorded at 300.13 MHz, C NMR spectra at 75.47 MHz in
CDCl as the solvent and TMS as the internal standard. Data
2
indicating a ‘late’ transition state on the reaction coordinate.
The water involved has consequently the effect of a catalyst,
which per definitionem is regenerated after the reaction,
although the conversion of 3 into 2 is accompanied by a
proton exchange in that water molecule: one proton of the
3
1
3
are given in ppm. C peaks were assigned by means of
APT, HMQC and HMBC spectra; ‘d.i.’ denotes peaks with
double intensity.
exo-CH group is taken up, while a proton is donated to
3
Computations, as implemented through Spartan Pro 02 by
Wavefunction, Inc., Irvine, CA, USA, were carried out on
geometries pre-optimized by the semi-empirical PM3
method. For full geometry optimization the widely
employed B3LYP hybrid method, which includes a mixture
of HF and DFT exchange terms and the gradient-corrected
the ring carbon. Finally, going from the transition state to
the product 2 requires rotation of the exo-methylene group,
so that the H–C–H plane and the C–N–C plane fall
together, re-establishing carbenium-iminium resonance
stabilization.
1
6
correlation functional of Lee, Yang and Parr parametrized
1
by Becke, was used, along with the double-zeta split
7
The absence of the reverse reaction 2!3 can be explained
by the stability of intermediate N-hydroxymethyl-morpho-
line (HMM), which is formed by addition of a hydroxyl ion
to 2 (or addition of water with subsequent loss of a proton).
HMM subsequently gives morpholine and formaldehyde,
from which regeneration of 2 under neutral conditions is
very slow. In contrast, 3-hydroxy-N-methylmorpholine,
formed by formal addition of a hydroxyl ion into 3-position
of 3, is less stable, so that 3 can readily be regenerated by the
reversed process. Moreover, formation of 2-(amino-
methyl)ethoxy-acetaldehyde—the open-chain form of the
cyclic semiaminal 2-hydroxy-N-methylmorpholine—is
negligible and does not consume 3, as does the formation of
morpholine and HCHO from HMM in the case of C-I ion 2.
1
8
valence basis sets 6-31þG*, which includes diffuse
functions. Transition states and minima were confirmed
by analysis of the calculated vibrational spectrum, and by
intrinsic reaction coordinate analysis. For all transition
states the number of imaginary frequencies was 1, for all
minimum geometries it was 0.
3.1.1. N-(Methylene)morpholinium chloride (2). Forma-
line (35%, 8.6 mL, 0.1 mol) was cooled to 210 8C (ice/
NaCl bath) and morpholine (17.4 g, 0.2 mol) was added
under stirring. Within about 30 min the mixture was allowed
to reach r.t. Anhydrous solid K CO was added until the
2
3
aqueous phase disappeared. The solids were crushed by a
sufficiently powerful magnetic stirrer, removed by filtration,
washed with Et O, and the combined phases were dried
In summary, the NMMO-derived ring-centered carbenium-
iminium ion 3 is rearranged into its counterpart, the
NMMO-derived exo-centered carbenium-iminium ion 2 in
the presence of water. The conversion is an endothermic,
bimolecular reaction involving 3 and water. The two
coreactants pass through a highly organized transition
state in which water, acting as a catalyst, is simultaneously
accepting and donating a proton. The generated 2 is
consumed by formation of the quite stable N-hydroxy-
methylmorpholine, morpholine and formaldehyde, so that
the reversed reaction 2!3 is largely prevented. These
findings agree completely with empirical data from NMMO
oxidation chemistry and from the Lyocell fiber-making
process. Especially under the rather drastic reaction
conditions of the latter—working in a melt of NMMO
monohydrate at temperatures around 100 8C—the C-I ion
conversion is likely to account for the observed nearly
complete absence of 3.
2
again over K CO . Et O was stripped off and the residue
2
3
2
was distilled under reduced pressure to give bis(4-
morpholino)methane, bp ¼78 8C. This product was dis-
3
5
solved in dry Et O in an inert atmosphere. Acetyl chloride
2
(7.9 g, 7.1 mL, 0.1 mol) was added at 0 8C under stirring.
The resulting precipitate was collected under exclusion of
moisture, washed with Et O, and dried in vacuo to give
2
N-(methylene)morpholinium chloride (2) as white crystals
1
(8.05 g, 80.4%). H NMR (CDCl ): d 3.58 (t, 4H), 4.08 (t,
3
1
3
4H), 8.53 (s, b, 2H). C NMR: d 56.2 (d.i.), 65.5 (d.i.),
167.7. Anal. calcd for C H NOCl (135.59): C 44.29, H
5
10
7.43, N 10.33, Cl 26.15. Found: C 44.11, H 7.62, N 10.05, Cl
25.89. Computated thermodynamic data for the cation
2
1
(without anion): DH¼100.5 kcal mol
,
DS¼78.05
cal (mol K)2
vibrational energy 96.6 kcal mol
1
,
DG ¼77.2 kcal mol
21
,
zero-point
2
98
2
1
.
3.1.2. 4-Methyl-3,6-dihydro-2H-[1,4]oxazinium chloride
The described process is the first example of a direct
conversion between two Mannich intermediates. Studies on
the general synthetic applicability of the reaction with
regard to a N-demethylation involving a carbenium-
iminium ion conversion step are underway.
(3). 3-Hydroxy-4-methylmorpholine (11.72 g, 0.1 mol) was
cooled to 210 8C (ice/NaCl bath) and morpholine (17.4 g,
0.2 mol) was added under stirring. Within about 30 min the
mixture was allowed to reach rt. Anhydrous solid K CO
2
3
was added until the aqueous phase disappeared. The solids

T. Rosenau et al. / Tetrahedron 60 (2004) 301–306
305
were removed by filtration, washed with Et O, and the
2
3.2.2. 3,3-Dideutero-3-(3-deutero-morpholin-4-yl)-pro-
1
combined phases were dried again over K CO . Et O was
2
pio-naphthone (5a). H NMR (CDCl ): d 2.54 (m, 1H,
3
3
2
3
stripped off and the residue was distilled under reduced
pressure to give bis(4-morpholino)methane, bp ¼112 8C.
CHD), 2.68 (t, 2H, J¼5.2 Hz, N–CH –CH –O), 2.98 (s,
2
2
3
b, 1H), 3.12 (s, 2H, CH –CD –CO), 3.76 (t, 2H, J¼5.2 Hz
N–CH –CH –O), 3.78 (dd, 2H, N–CHD–CH –O), 7.58–
2 2 2
7.65 (m, 2H, CH), 7.86–8.05 (m, 4H, CH), 8.45 (s, 1H,
Ar
CH). C NMR: d 30.1, 51.0 (quint), 51.2 (t), 57.7, 63.9,
2
0
2
2
The product was dissolved in dry Et O in an inert
2
Ar
Ar
atmosphere. Acetyl chloride (7.9 g, 7.1 mL, 0.1 mol) was
added at 0 8C under stirring. The resulting precipitate was
13
collected under exclusion of moisture, washed with Et O,
2
64.7, 124.4, 128.0, 128.8, 129.5, 130.1, 130.8, 131.5, 133.6,
134.2, 136.6, 197.4. Anal. calcd for C H NO D (272.36):
C 74.96, H 7.03, N 5.14. Found: C 75.08, H 6.98, N 5.20.
and dried in vacuo to give 4-methyl-3,6-dihydro-2H-
[
1
7
16
2 3
1,4]oxazinium chloride (3) as white crystals (6.25 g,
1
6
2.4%). H NMR (CDCl ): d 3.42 (s, 3H), 3.68 (d, b, 2H,
3
3
J¼6.1 Hz), 4.19 (m, 2H), 4.62 (m, 2H), 9.15 (t, b,
3.2.3. 3,3-Dideutero-3-(morpholin-4-yl)propio-
1
3
13
J¼6.1 Hz). C NMR: d 46.9, 55.1, 62.0, 72.4, 147.7.
naphthone (5b). H NMR (CDCl ): d 2.24 (s, b, 1H),
3
3
Anal. calcd for C H NOCl (135.59): C 44.29, H 7.43, N
5
2.65 (t, 4H, J¼5.0 Hz, N–CH –CH –O), 3.36 (s, 2H,
10
2
2
3
1
Computed thermodynamic data from the cation (without
0.33, Cl 26.15. Found: C 44.41, H 7.25, N 10.17, Cl 26.03.
CH –CD –CO), 3.74 (t, 4H, J¼5.0 Hz N–CH –CH –O),
2
2
2
Ar
2
Ar
7.58–7.65 (m, 2H, CH), 7.82–8.06 (m, 4H, CH), 8.44
2
1
21
Ar
13
anion): DH¼100.4 kcal mol
,
DS¼80.1 cal (mol K)
,
(s, 1H, CH). C NMR: d 30.2, 51.2 (quint), 52.9 (d.i.),
64.6 (d.i.), 124.4, 128.1, 128.8, 129.6, 130.0, 130.7, 131.6,
133.4, 134.3, 136.6, 198.0. Anal. calcd for C H NO D
2 2
2
1
DG298¼76.5 kcal mol , zero-point vibrational energy
6.0 kcal mol2
1
.
9
1
7
17
(271.36): C 75.24, H 7.06, N 5.16. Found: C 75.44, H 7.12,
N 5.34.
3.2. General experimental procedure for trapping of
carbenium-iminium ions
3
(5c). H NMR (CDCl ): d 2.54 (m, 1H, CHD), 2.67 (t,
.2.4. 3-(3-Deutero-morpholin-4-yl)propionaphthone
1
A solution of the carbenium-iminium chloride 2 or 3
(
3
3
1 mmol) was dissolved in the respective solvent (20 mL),
2H, J¼5.1 Hz, N–CH –CH –O), 2.84 (s, b, 1H), 2.91 (t,
2
2
3
3
for the solvents used see Table 1. The mixture was stirred
for 24 h either under reflux or at room temperature (see
Table 1). At room temperature, a solution of 2-aceto-
naphthone (0.51 g, 7 mmol) in the same solvent (20 mL)
and H SO (conc., 0.1 mL) was added at once. The
2H, J¼7.0 Hz, CH –CH –CO), 3.44 (t, 2H, J¼7.0 Hz,
2 2 2 2
2
2
3
CH –CH –CO), 3.76 (t, 2H, J¼5.1 Hz, N–CH –CH –
O), 3.78 (dd, 2H, N–CHD–CH –O), 7.58–7.65 (m, 2H,
2
Ar
Ar
Ar
13
CH), 7.86–8.08 (m, 4H, CH), 8.45 (s, 1H, CH).
C
NMR: d 30.1, 51.4, 51.5 (t), 57.7, 64.0, 64.7, 124.3, 128.0,
128.7, 129.5, 130.1, 130.7, 131.7, 133.6, 134.2, 136.8,
198.0. Anal. calcd for C H NO D (270.35): C 75.52, H
2
4
consumption of C-I ions was complete after less than
0 min (DC control). After additional stirring for 30 min,
1
1
7
18
2
water (100 mL) and chloroform (100 mL) were added and
the phases were separated. The organic phase was washed
twice with water and dried over Na SO . The solvents were
carefully removed in vacuo. For determination of the
product composition, the remaining waxy solid was
re-dissolved in n-hexane (15 mL). Ethereal HCl (3 mL,
7.08, N 5.18. Found: C 75.35, H 7.18, N 5.29.
3.2.5. 2-(4-Methyl-morpholin-3-yl)-1-naphthalen-2-yl-
1
2
4
ethanone (6). H NMR (CDCl ): d 2.42 (s, 3H, CH ),
3
3
2.67 (m, 2H, N–CH ), 3.10 (m, 1H, N–CH), 3.18–3.38 (m,
2
2H, CH –CO), 3.68 (m, 2H, N–CH –CH –O), 3.89 (m,
2
2
2
2
M) was added, the mixture was left standing for 30 min,
2H, N–CH–CH –O), 4.12 (s, b, 1H), 7.54–7.62 (m, 2H,
2
Ar
Ar
Ar
13
and the resulting white, crystalline precipitate was removed
by filtration and dried in vacuo. An aliquot of the solid was
dissolved in CDCl and directly analyzed by NMR using the
CH), 7.84–8.02 (m, 4H, CH), 8.46 (s, 1H, CH).
C
NMR: d 39.7, 49.1, 53.8, 61.2, 65.3, 71.4, 124.6, 128.0,
128.8, 129.3, 130.3, 131.0, 131.4, 133.0, 134.4, 136.6,
198.7. Anal. calcd for C H NO (269.35): C 75.81, H
3
integrals of the b-methylene group in naphthone 5
1
7
19
2
(
(
,2.95 ppm) and the b-methine proton in naphthone 6
,3.10 ppm) for determination of the product ratio. For
7.11, N 5.20. Found: C 76.01, H 7.20, N 5.13.
purification of the products, the above crude solid
obtained after drying and solvent evaporation was re-
dissolved in 5 mL of toluene and chromatographed on silica
gel (toluene/ethyl acetate, v/v¼9/1). For better storage, the
obtained waxy morpholinonaphthones were converted into
their crystalline hydrochlorides by treatment with ethereal
HCl.
Acknowledgements
The financial support by the Austrian Fonds zur F o¨ rderung
der wissenschaftlichen Forschung, project P-14687
Chemistry of amine N-oxides) is gratefully acknowledged.
The authors would like to thank Dr. Andreas Hofinger,
Institute of Chemistry at the University of Natural
Resources and Applied Life Sciences Vienna, for recording
the NMR spectra.
(
1
3
(
.2.1. 3-(Morpholin-4-yl)propionaphthone (5). H NMR
3
CDCl ): d 2.18 (s, b, 1H), 2.60 (t, 4H, J¼5.0 Hz, N–CH –
3
2
3
2 2 2
CH –O), 2.93 (t, 2H, J¼7.0 Hz, CH –CH –CO), 3.46 (t,
Ar
3
3
2
N–CH –CH –O), 7.58–7.66 (m, 2H, CH), 7.86–8.08
H, J¼7.0 Hz, CH –CH –CO), 3.78 (t, 4H, J¼5.0 Hz
2
2
2 2
Ar
Ar
13
(
m, 4H, CH), 8.45 (s, 1H, CH). C NMR: d 33.7, 53.1,
References and notes
5
1
3.2 (d.i.), 64.7 (d.i.), 124.4, 128.2, 128.8, 129.6, 130.1,
30.8, 131.6, 133.5, 134.2, 136.8, 198.0. Anal. calcd for
1. Albini, A. Synthesis 1993, 263–277.
2. For a review see: Schr o¨ der, M. Chem. Rev. 1980, 80, 187–213.
3. (a) Chanzy, H. J. Polym. Sci., Polym. Phys. Ed. 1980, 18,
C H NO (269.35): C 75.81, H 7.11, N 5.20. Found: C
1
7
19
2
7
5.92, H 7.29, N 5.22.

306
T. Rosenau et al. / Tetrahedron 60 (2004) 301–306
1137–1144. (b) Chanzy, H.; Nawrot, S.; Peguy, A.; Smith, P.
J. Polym. Sci. 1982, 20, 1909–1924.
the reaction time, and the reaction became pseudo-first order
0
with regard to the starting material 3, i.e., d½2Š=dt ¼ k ½3Š; with
0
4
. (a) Black, D. S. C.; Deacon, G. B.; Thomas, N. C. Inorg. Chim.
Acta 1983, 65, L75–L76. (b) Ringel, C. Z. Chem. 1969, 9, 188.
. (a) Brandner, A.; Zengel, H. G. German Patent DE-OS 3,034,
k ¼ k½H2OŠ.
12. Graphical representation of lnðk=TÞ versus 1=T produced a
#
5
straight line y ¼ ax þ b with the slope a ¼ 2DH =R in
2
1
#
21
685, 1980. (b) Brandner, A.; Zengel, G. H. Chem. Abstr. 1982,
J mol und b ¼ DS =R þ 23:76 in J(mol K)
.
97 7727d Chem. Abstr. (c) Firgo, H.; Eibl, M.; Kalt, W.;
Meister, G. Lenz. Ber. 1994, 74, 80–90.
. Rosenau, T.; Potthast, A.; Sixta, H.; Kosma, P. Tetrahedron
13. Trapping after reacidification proved that indeed not only the
trapping reaction, but also the rearrangement itself was
prevented.
6
7
8
9
2
002, 58(15), 3073–3078.
14. For procedure and for the calculated thermodynamic values
see Section 3.
. Rosenau, T.; Potthast, A.; Kosma, P.; Chen, C. L.; Gratzl, J. S.
J. Org. Chem. 1999, 64, 2166–2167.
15. This comprised localization of transition states and geometry
minima, supported by vibrational frequency analysis and
intrinsic reaction coordinate analysis.
. This applies for instance to reaction mixtures containing
NMMO monohydrate, or NMMO-cellulose spinning dopes.
. Rosenau, T.; Potthast, A.; Sixta, H.; Kosma, P. Progr. Polym.
Sci. 2001, 26(9), 1763–1837.
16. (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. 1988, B37, 785.
(b) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys.
Lett. 1989, 157, 200.
1
0. It was checked that 6 was stable under the prevailing
conditions, and did not transform into 5. Thus, formation of
17. (a) Becke, A. Phys. Rev. 1988, A38, 3098. (b) Becke, A. J.
Chem. Phys. 1993, 98, 5648.
5
transformation of the trapping products.
must be due to the conversion of 3 into 2 rather than a
18. (a) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28,
213. (b) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley,
J. S.; Gordon, M. S.; DeFrees, D. J.; Pople, J. A. J. Chem.
Phys. 1982, 77, 3654.
1
1. The overall reaction order was determined using equivalent
starting concentrations. Applying a large excess (80 equiv.) of
water, the water concentration could be regarded constant over