Nitronyl and imino nitroxides: Improvement of Ullman's procedure and report on a new efficient synthetic route

Article abstract of DOI:10.1002/1521-3765(20010504)7:9<2007::AID-CHEM2007>3.0.CO;2-7

The synthesis of nitronyl and imino nitroxides has been reexamined with the aim of both increasing yields and of offering opportunities for new structures. The conditions for the formation of 2,3-bis(hydroxyamino)-2,3-dimethylbutane, the key intermediate of Ullman's route, have been carefully studied, and a new procedure is proposed, which affords the free base in a very pure form and up to 60% yield. Full characterization of this intermediate including an X-ray crystal structure is presented. An alternative synthetic route through 2,3-diamino-2,3-dimethylbutane and the corresponding imidazolidines which bypasses the delicate synthesis of the bis(hydroxyamino) compound is described. It is shown that 3-chloroperbenzoic acid is an effective oxidant for the transformation of adequately substituted imidazolidines into nitronyl nitroxides, which are obtained in high yield. An illustration of the potentialities of this new route, a new nitronyl nitroxide with two ethyl substituents in positions 4 and 5 of the imidazoline ring, is reported. The scope and limitations of the two routes are discussed.

Full text of DOI:10.1002/1521-3765(20010504)7:9<2007::AID-CHEM2007>3.0.CO;2-7

FULL PAPER
Nitronyl and Imino Nitroxides: Improvement of Ullmanꢀs Procedure
and Report on a New Efficient Synthetic Route
Catherine Hirel,[a] Kira E. Vostrikova,[a] Jacques PeÂcaut,[a] Victor I. Ovcharenko,[b]
and Paul Rey*[a]
Abstract: The synthesis of nitronyl and
imino nitroxides has been reexamined
with the aim of both increasing yields
and of offering opportunities for new
structures. The conditions for the for-
mation of 2,3-bis(hydroxyamino)-2,3-di-
methylbutane, the key intermediate of
Ullmanꢀs route, have been carefully
studied, and a new procedure is pro-
posed, which affords the free base in a
very pure form and up to 60% yield.
Full characterization of this intermedi-
ate including an X-ray crystal structure
is presented. An alternative synthetic
route through 2,3-diamino-2,3-dimethyl-
butane and the corresponding imida-
zolidines which bypasses the delicate
synthesis of the bis(hydroxyamino) com-
pound is described. It is shown that
3-chloroperbenzoic acid is an effective
oxidant for the transformation of ade-
quately substituted imidazolidines into
nitronyl nitroxides, which are obtained
in high yield. An illustration of the
potentialities of this new route, a new
nitronyl nitroxide with two ethyl sub-
stituents in positions 4 and 5 of the
imidazoline ring, is reported. The scope
and limitations of the two routes are
discussed.
Keywords: amines ´ heterocycles ´
imidazolidines
nitroxides
´
imidazolines
´
and oxidation stages proceed with acceptable yields and are
reproducible, preparation of the bis(hydroxyamino) inter-
mediate is poorly reproducible and often leads to frustrating
results. Its formation in moderate yield and questionable
purity was published over thirty years ago;[10] it is obtained
using the classical reduction process with zinc in ammonium
chloride buffered solution of the dinitro analogue, 1.[11] This
synthetic procedure has recently received attention, and the
understanding of important features of the redox process has
led to a new procedure, which affords the sulfate salt of 2 in a
fairly good yield; the synthesis of the pure free base however,
is still a challenge.[12] Thus, the overall yield of nitroxide from 1
is often less than 50% and mainly reflects the difficulties in
preparing the bis(hydroxyamino) compound. However, since
the dinitro precursor is readily available in most cases, yield
and purity of the bis(hydroxyamino) (2) intermediate are not
a matter of great concern, and all nitronyl nitroxides reported
so far were prepared using this procedure.
However, problems associated with the preparation of this
key intermediate are probably the cause of the limited
number of developments reported so far that concern
structural variations in nitronyl nitroxides. Indeed, while any
aldehyde may give a nitronyl nitroxide, uncertainty about
purity of the bis(hydroxyamine) may lead to delay or
abandonment of attempts to synthesize specific nitroxides
that require aldehydes, which are difficult to prepare.
Accordingly, we performed a thorough study of the formation
of compound 2 and produced a new workup, which affords the
Introduction
Nitronyl and imino nitroxide free radicals were described in
the seventies by Ullman in his pioneering work,[1±3] and since
then they have not attracted much attention until the last
decade of the century. This recent renewal of interest mainly
stems from the use of these paramagnetic species as building
blocks for designing molecular magnetic materials.[4] Success-
ful achievements in this field, such as purely organic
ferromagnetically ordered solids[5, 6] and metal-organic ex-
change-coupled complexes that exhibit versatile magnetic
properties,[7, 8] have triggered the synthesis of hundreds of
these free radicals.
The synthesis of nitronyl and imino nitroxides relies almost
exclusively[9] on the condensation of 2,3-bis(hydroxyamino)-
2,3-dimethylbutane (2) with an aldehyde and oxidation of the
condensation product (Scheme 1).[2] While the condensation
Â
[a] Dr. P. Rey, C. Hirel, Dr. K. E. Vostrikova, Dr. J. Pecaut
Â
CEA, Departement de Recherche
Á
Â
Fondamentale sur la Matiere Condensee
Service de Chimie Inorganique et Biologique
Laboratoire de Chimie de Coordination (UMRCNRSNo. 5046)
17 rue des Martyrs, 38054 Grenoble cedex 09 (France)
Fax : (‡33)476885090
E-mail: rey@drfmc.ceng.cea.fr
[b] Prof. V. I. Ovcharenko
Laboratory of Polyspin Systems
International Tomography Center, Institutskaya 3A
630090 Novosibirsk (Russia)
Chem. Eur. J. 2001, 7, No. 9
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2007 $ 17.50+.50/0
2007

P. Rey et al.
FULL PAPER
enantiotopic methyl groups in positions 4 and 5 (Scheme 1,
compound 7) of the imidazoline ring that would be involved
in the preparation of all series of chiral nitronyl nitroxides
cannot be envisioned because it would require a tedious study
for establishing a reliable synthetic procedure for the corre-
sponding bis(hydroxyamino) intermediate.
Considering that flexible syntheses of new nitronyl ni-
troxides are central to continued progress in the field of
molecular magnetism, we decided to explore other synthetic
routes toward these important open-shell intermediates.
There are two general methods for obtaining nitroxide free
radicals: the first, related to Ullmanꢀs procedure, involves
oxidation of hydroxyamines by mild oxidizing reagents,[13±15]
the second converts amines directly to free radicals by use of
strong oxidants.[13, 14, 16±18] Conversion of adequately substitut-
ed imidazolines into nitronyl nitroxides by sodium tungstate
catalyzed hydrogen peroxide oxidation in a rather low yield
has been reported which is a possibility for a new synthetic
route.[9]
In addition to a more efficient synthesis of the bis(hydroxy-
amino) intermediate, we describe in this paper an alternative
new synthetic route through 2,3-diamino-2,3-dimethylbutane,
3, which bypasses the usual bis(hydroxyamino) precursor.
Potentialities of this new synthetic route are illustrated by the
synthesis of already known free radicals in high yield and by
the straightforward preparation of a newly substituted
nitronyl nitroxide (Scheme 2). The scope and limitations of
this new synthetic route are discussed and compared with
those of Ullmanꢀs procedure.
Scheme 1. The bis(hydroxyamino) and diamino routes to nitronyl and
imino nitroxides. a) Zn/NH4Cl, THF/H2O; b) Sn, HCl, reflux; c) and
d) RCHO, methanol, CHCl3 or diethyl ether; e) NaIO4 or MnO2, CH2Cl2/
H2O; f) m-chloroperbenzoic acid (2 equivalents)/CH2Cl2; g) Na2WO4/
H2O2; h) see ref. [9]; i) NaNO2/HCl, H2O/CH2Cl2. R ˆ phenyl (a),
p-nitrophenyl (b), 3-pyridyl (c), H (d), methyl (e), and tert-butyl (f).
pure free base 2 in a highly pure form and a yield of up to
60%; we report its full characterization including a X-ray
diffraction study.
Although tetramethylated nitronyl nitroxides (Scheme 1, 7)
are thus satisfactorily obtained for most purposes, the Ullman
route suffers from major drawbacks: i) the delicate synthesis
of the bis(hydroxyamino) compound could hardly be trans-
posed to differently substituted dinitro precursors; ii) ade-
quately substituted vic-diamines are precursors of nitronyl
nitroxides, which, according to this synthetic scheme, would
have to be oxidized to dinitro compounds and then reduced to
bis(hydroxyamino) compounds with increasing uncertainties
about the yield. Accordingly, there are synthetic problems
that would be solved only with great difficulty following
Ullmanꢀs synthetic scheme. For example, the yield is not high
enough to trigger the synthesis of perdeuterated nitronyl
nitroxides, which would have utility in neutron diffraction
studies. More importantly, structural variations that involve
Scheme 2. Illustration of the synthetic potentialities of the diamino route:
meso-4,5-diethyl-4,5-dimethyl-2-phenyl-4,5-dihydro-1H-imidazolyl-3-ox-
ide-1-oxy. a) benzaldehyde/diethyl ether; b) m-chloroperbenzoic acid/
CH2Cl2, NaIO4.
Results and Discussion
The bis(hydroxyamino) route: As stated before, the only
delicate step of Ullmanꢀs route is the synthesis of compound 2.
Its preparation was performed more than one hundred times
in our laboratories using the Lamchem ± Mittag procedure,[10]
which was progressively modified. These studies ascertained
the redox mechanism, established the stability of the bis(hy-
droxyamino) derivative when pure, and explained the for-
2008
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2008 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 9

Nitronyl and Imino Nitroxides
2007±2014
mation of side products such as 2-methyl-substituted nitronyl
nitroxide.[12] Concerning this point, it is worth mentioning that
alcohols should not be used as solvent. They are oxidized by
the bis(hydroxyamine) into aldehydes, which condense in situ
and are air oxidized into nitroxides. Since this process is
probably catalyzed by impurities and is unpredictable,[19] use
of nonpurified methylene chloride or chloroform, which are
usually stabilized with 0.2 ± 0.4% of ethanol may also give the
undesirable free radical, when compound 2 is extracted in the
presence of a base. The pink color (low concentration of
nitroxide) disappears during extraction, but decomposition
products contaminate the bis(hydroxyamine), which is far less
stable. These observations open questions about catalytic
redox reactions that involve hydroxyamines, which are out of
the scope of this study.
Use of THF instead of ethanol in the modified Lamchem ±
Mittag procedure (Procedure A) not only avoids formation of
undesired nitroxides but also has a strong influence on the
course of the reduction. While in ethanol/water (1:1) ammo-
nium chloride is soluble and the dinitro compound sparingly
soluble, the reverse situation is observed when THF/water
(2:1) is used. Actually, the kinetics of the reduction are
strongly dependent on the concentration of NH4Cl in the
solution. The course of the reaction is thus strongly affected
since in the first case there is a rather weak concentration of
dinitro compound and a large concentration of NH4Cl, while
in the second case the concentration of dinitro compound is at
a maximum, and that of NH4Cl is monitored by the quantity of
water. Formation of side products is thus minimized; in
particular, provided that the temperature is kept below 128C,
formation of the diamine is almost negligible using a 2:1 ratio
of THF/water.
Another improvement to the Lamchem ± Mittag procedure
comes from using eight equivalents of NH4Cl. This quantity
brings just the right number of anions required for four atom
grams of Zn. Examination of the literature shows that the
quantity of NH4Cl used (2 ± 4 equivalents) does not rely on
any experimental finding except that it is in line with the
incorrect belief that NH4Cl had only the role of buffering the
solution. Actually, use of eight equivalents does result in a
well-defined Zn complex, [Zn(NH3)2Cl2].[20]
However, although the synthesis does afford the pure free
base 2, the yield is spread over a large range, and the synthesis
is not well reproducible. This is likely to be the consequence of
adding Zn portionwise, which is a hardly reproducible process
and this probably results in a great deal of local heating of the
solution. This drawback is overcome in Procedure B, in which
the dropwise addition of NH4Cl in solution allows one to
monitor the temperature and the reaction rate by means of a
constant and weak concentration of the salt. One observes
that the role of NH4Cl is dramatic: compound 2 appeared at
the same time as a few drops of the salt were added.[21]
Therefore, since these parameters are well controlled, repro-
ducibility is enhanced, and one obtains a reproducible yield,
which was not possible using Procedure A.
analyzing the solution content that corresponds to
[Zn(2)2Cl2]. Mass spectrometry analysis of the recovered
solid exhibits the expected mass and the characteristic
isotopic pattern of Zn complexes, which, with the results of
elemental analyses, demonstrate unambiguously that com-
pound 2 is filtered as its ZnII complex.
Accordingly, the reaction should be written as follows
[Eq. (1)].
2C6H12(NO2)2 ‡ 8Zn ‡ 16NH4Cl
! 7[Zn(NH3)2Cl2] ‡ [C6H12(NHOH)2]2ZnCl2 ‡ 4H2O ‡ 2NH3 (1)
Hydroxyamino derivatives are known to be unstable and
particularly subject to oxidation even on exposure to air. This
seemed also true for compound 2, which, when prepared
according to the Lamchen ± Mittag procedure, often decom-
posed at room temperature in a few days. At lower temper-
ature (4 ± 68C) decomposition occurs slowly with the produc-
tion of acetoxime, which sublimes in the flask. Therefore, it
clearly appeared that the Lamchen ± Mittag workup, mainly
the acidification of the crude reaction filtrate before evapo-
ration, was based on the incorrect belief that the desired
bis(hydroxyamino) derivative is volatile. This fact gave
support to the new workup, which avoids acidification of the
crude product and increases the yield.
Although this procedure affords 2,3-bis(hydroxyamino)-
2,3-dimethylbutane in pure form and acceptable yield, one
should be aware that the use of Zn from another source
(different particle size distribution) may result in a different
yield and purity. As well, in Procedure B, the use of less or
more water than usual may also have the same consequence.
Even when pure, in solution the compound is rather
unstable at room temperature and transforms slowly into
acetoxime and other nonidentified products.[12] In contrast, in
the solid state it seems to be indefinitely stable (up to two
years at room temperature without change). Full character-
ization by X-ray diffraction was performed without precau-
tion, and the data did not show any sign of crystal destruction
during data collection at room temperature. The structure
(Figure 1) is unexceptional. The molecule adopts a staggered
conformation with a NCCN dihedral angle of 71.6(4)8. One of
the hydroxyamino groups is statistically disordered on two
positions with respective occupancies of 0.1 and 0.9, which
Finally, we addressed the problem of the true nature of the
solution of 2 in THF/H2O. One notes that the precipitate of
[Zn(NH3)2Cl2] is recovered in only 85 ± 88% yield, and that 1ꢂ8
of the added Zn metal should be in solution. This is verified by
Figure 1. View of the molecular structure of 2. Relevant bond lengths [Š]
and angles [8] are as follows: O(1) N(1) ˆ 1.451(1), O(2) N(2) ˆ 1.507(9),
C(1) C(4) ˆ 1.568(1); N(1) C(1) C(4) N(2) ˆ 71.6(4).
Chem. Eur. J. 2001, 7, No. 9
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2009 $ 17.50+.50/0
2009

P. Rey et al.
FULL PAPER
each share the hydrogen atom (not disordered) of their
hydroxyl group with a neighboring molecule. Therefore, the
molecules are arranged into dimers as shown in Figure 1. In
addition, the second hydroxyamino group is also engaged in
hydrogen bonding with another dimer so that the compound
should be viewed as zig-zag chains that run along the c axis.
All distances and angles are within the expected range and
need no further comments.
(see Experimental Section). This extension of the new route
has been investigated for both an aromatic (5a) and an alkyl
(5e)[29] dimethyl acetal with equal success. These heterocycles
are oils or solids that have a low melting point; there are a few
exceptions like 5b, which crystallizes from methanol and
whose structure was determined by X-ray diffraction, which
gave us the opportunity of fully characterizing these new
compounds.[30]
The production of compound 2 in pure form is important.
Indeed, the presence of impurities has dramatic effects on the
following steps of the synthesis of nitronyl nitroxides and, in
particular, it partly accounts for the presence of imino
nitroxides in variable amount in almost all nitronyl nitroxide
preparations. Dehydration of the 1,3-dihydroxy-imidazolidine
compound 3 (Scheme 1) seems to be promoted not only by
sodium periodate during the oxidation stage and SiO2 if the
oxidation is not complete, but also by the presence of
impurities during the condensation process.
In conclusion, since the dinitro precursor 1 is a readily
available product, and that this new workup brings pure
intermediate 2 in acceptable yield, the synthesis of nitronyl
nitroxides proceeds with higher yields. In particular, this new
method opens up new possibilities when the aldehyde is a rare
compound.
While hydroxyamines are oxidized into nitroxides by several
mild oxidants such as PbO2, NaIO4, or Ag2O,[13±15] conversion
of amines into nitroxides requires more powerful oxidizing
agents.[16±18] As mentioned before, a few reports describe
attempts for synthesizing tetramethylated imidazolines (6),
which were converted to nitroxides through phosphotungstic
acid or NaWO4 catalyzed oxidation with hydrogen peroxide in
a rather low yield.[9] In our laboratory, these oxidizing systems
converted imidazolidines (5) into imino nitroxides (8) with a
similar low yield. Nevertheless, we have found that 3-chloro-
perbenzoic acid, introduced in the seventies for converting
substituted amines into the corresponding aliphatic ni-
troxides,[17, 18] was effective also for efficiently transforming
imidazolidines into nitronyl nitroxides. This reaction is carried
out at low temperature (6 ± 88C) in the presence of a base
(NaHCO3), which removes m-chlorobenzoic acid from the
organic phase. However, we have observed that in this basic
medium compound 7d (R ˆ H) is not stable. As described
previously by Ullman,[31, 32] NaHCO3 is a strong enough base
to convert the nitronyl nitroxide into the corresponding
carbanion, which disproportionates into diamagnetic com-
pounds. This compound is better prepared without the use of a
base and purified by chromatography on alumina.
Formally, oxidation of imidazolidines 5 into nitronyl ni-
troxides 7 requires 3.5 equivalents of peracid and should
involve several steps. Susceptibility of these heterocycles to
oxidizing agents is well-documented; they have even been
used as reducing agents for activated double bonds.[33]
Consistently, one observes the presence of a weak [M 2]
peak in the mass spectra of all compounds 5a ± f, which
corresponds to imidazolines 6a ± f. However, peracid oxida-
tion of amines is known to proceed by nucleophilic attack of
the amine on a hydroxy group of the peracid, which would
afford a mono- or dihydroxyimidazolidine as an intermedi-
ate.[34] In order to ascertain the oxidation pathway (5 ! 6 ! 7
or 5 ! 4 ! 7), we have reacted imidazolidines 5a,b (aromatic-
substituted) with two equivalents of 3-chloroperbenzoic acid
at low temperature (68C). One does not observe the
appearance of the characteristic color of nitroxides nor the
formation of 6a,b but the production of the corresponding
1,3-dihydroxyimidazolidine 4a,b. It is noteworthy that under
the same conditions aliphatic-substituted imidazolidines
(5d ± f) are more rapidly converted into free radicals as are
aliphatic amines; in this case, the dihydroxy compounds could
not be characterized. Finally, neither PbO2 nor NaIO4 were
able to perform these transformations.
The diamino route: In contrast to the bis(hydroxyamino)
derivative 2, the diamino compound 3 is easily prepared.
Corresponding to the full reduction of the dinitro precursor, it
is obtained pure and in high yield by using drastic reducing
conditions.[11, 22] Condensation with aldehydes is quantitative,
and the resulting imidazolidines are easily oxidized by
3-chloroperbenzoic acid. We have investigated this new route
for a few representatives of aromatic- (7a ± c) and aliphatic-
(7d ± f) substituted nitronyl nitroxides.
Although the condensation of aldehydes with a-diamines is
a well-known preparation of imidazolidines (5),[23±26] surpris-
ingly, no investigations have been reported that involve
tetramethyl-substituted fragments suitable for the prepara-
tion of stable nitronyl nitroxides. Despite the steric crowding
of the gem-dimethyl groups, the condensation is fast in diethyl
ether or chloroform; it is likely that steric hindrance is
overcompensated for by the basicity of the amino groups,
which is magnified by the presence of these methyl groups.
However, electronic or steric effects of the substituent are
reflected in the formation of 5b (p-nitrophenyl) and 5 f (tert-
butyl), which required boiling in chloroform to reach com-
pletion. These effects are also probably the cause of the low
yields reported by Ullman (bis(hydroxyamino) route)[1] for
the corresponding two nitroxides 7b and 7 f (24 and 47%,
respectively, from the corresponding aldehydes), and one can
note the efficiency of the new route, which produces these
radicals in 66 and 64% yield from 1.
Being aminals, imidazolidines (cyclic form) are also ex-
pected to exist as amino-imines (open form).[27] At room
temperature however, 1H and 13C NMR spectra performed in
D2O in the pH range 1 ± 10 are consistent only with the cyclic
form. Moreover, imidazolidines that are described as being
hydrolyzed in acidic medium[28] are formed in good yield when
acetals are used instead of aldehydes and hydrolyzed in situ
As observed for compound 2, when pure in the solid state
these bis(hydroxyamines) are stable enough to be crystallized
in methanol and analyzed by X-ray diffraction at room
temperature. The structure of 4b (p-nitrophenyl) is displayed
in Figure 2. The imidazolidine ring is highly distorted (atoms
2010
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2010 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 9

Nitronyl and Imino Nitroxides
2007±2014
severe limitation to this new route. Indeed, many functional
groups are sensitive to this reagent and would probably be
oxidized during the synthesis. This problem has not been
thoroughly investigated.
Examination of Table 1 shows that this new synthetic route,
which proceeds through well-established pathways and af-
fords nitronyl nitroxides in good yield. Moreover, it can be
extended in a straightforward way to other systems. As an
example, we performed the synthesis of nitroxide 9 (Scheme 2)
derived from meso-3,4-dimethyl-3,4-dinitrohexane, which
illustrates the synthetic potentialities of this new route.[35]
The structural (Figure 3) and magnetic properties (Curie
law) of this new nitroxide are unexceptional and will be
discussed in a forthcoming paper with those of chiral nitroxides
based on the closely related (R,R)- or (S,S)-precursors.
Figure 2. View of the molecular structure of 4b. Relevant bond lengths [Š]
and angles [8] are as follows: O(1) N(1) ˆ 1.440(2), O(2) N(2) ˆ 1.454(2),
N(1) C(1) ˆ 1.478(3), N(2) C(2) ˆ 1.454(3).
are distributed in
a
0.27,
Table 1. NMR (CDCl3, d) characteristics of imidazolidines and yield of corresponding nitronyl nitroxides based
on 1.
‡ 0.30 Š range from the mean
plane), and the phenyl ring is
almost orthogonal to the mean
plane (868). The molecules are
arranged in head-to-tail dimers
through hydrogen bonds be-
tween one hydroxyl group of
one molecule and the nitro
group of the second molecule;
the dimers exhibit double hy-
drogen bonding between the
second NOH group of each
molecule (six-membered ring)
that leads to sheets parallel to
the ab plane. As expected, the
Compound
m.p. [8C]
1H NMR
13C NMR
yield
s 1.13, 1.04 (a,b)
s 2.38 (e), s 5.12 (d)
m 7.2 ± 7.6 (g,h,i)
24.1, 24.9 (a,b)
63.2, 73.6 (d)
126.8 (i), 127.7 (h)
81
38 ± 40
128.3 (g), 144.5 (f)
138 ± 139
s 1.12, 1.23 (a,b)
s 2.10 (e), s 5.28 (d)
q 7.77 ± 7.82, q 8.20 ± 8.26 (g,h)
24.1, 25.8 (a,b)
63.6 (c), 73.1 (d)
124.0, 128.0 (f,g,h,i)
73
77
intermolecular
arrangement
and the molecular structure as
well are very different from the
imidazolidine 5b and nitroxide
analogues.[5]
<
12
s 1.12, 1.21 (a,b)
24.1, 25.9 (a,b)
s 2.37 (e), s 5.22 (d)
63.4 (c), 71.8 (d)
m 7.9 ± 8.9 (f,g,h,i,j)
149.3, 149.4 (f,g,h,i,j)
123.6, 134.5, 140.1
Surprisingly, conversion of
the
bis(hydroxy)-imidazoli-
H
98 ± 100
s 1.06, 1.07 (a,b)
s 2.09 (e), s 3.9 (H)
24.3 (a,b), 61.0 (c)
61.9 (d)
71
82
dines (4) into nitroxides by m-
chloroperbenzoic acid is ex-
ceedingly slow and almost in-
effective for aromatic-substitut-
ed representatives in the ab-
sence of air or heating. There
are two reasons for this behav-
ior: i) m-chloroperbenzoic acid
is a better oxidizing agent for
CH3
s 1.06, 1.07 (a,b)
d 1.24 (CH3), s 1.9 (e)
q 4.17 (d)
24.3, 24.6 (a,b)
26.1 (CH3), 63.1 (c)
67.6 (d)
oil
s 0.93 (tert-butyl)
s 1.07, 1.13 (a,b)
s 2.02 (e), s 3.83 (d)
24.6 (tert-butyl), 26.4, 26.6
(a,b), 62.2 (c), 80.1 (d)
65
strong nucleophilic substrates like amines;[17, 18] this is in line
with the fairly rapid oxidation of alkyl-substituted species;
ii) in our oxidizing procedure, hydroxylated intermediates
such as 4 are likely to be very soluble in the aqueous phase, in
which m-chloroperbenzoic acid is almost absent. This draw-
back is easily overcome by using sodium periodate, which
insures completion of the last oxidation step (4 ! 7).
Conclusion
During the last ten years Ullmanꢀs route towards nitronyl
nitroxides has shown its broad scope through the diversity of
free radicals that have been prepared. As stated before, any
aldehyde may condense with compound 2 and give a nitronyl
nitroxide. The generality of the synthesis is illustrated by
several recent reports that show that protected aldehydes
(dialkyl acetals) are also suitable for the synthesis.[36] More-
over, formation of the corresponding nitroxides requires only
mild oxidants that are expected to preserve the integrity of the
Finally, it should be noted that oxidation of 5c (R ˆ 3-
pyridyl) proceeds with the formation of a small amount (8%)
of pyridyl-N-oxide-substituted nitronyl nitroxide, which
shows that the use of m-chloroperbenzoic acid might be a
Chem. Eur. J. 2001, 7, No. 9
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2011 $ 17.50+.50/0
2011

P. Rey et al.
FULL PAPER
powder was continuously extracted in a Soxlet apparatus protected from air
for 16 hours with methylene chloride (400 mL). Slow cooling of the organic
phase to room temperature afforded 5.6 g of white crystals (m.p. 181 ±
1838C), which were suitable for a X-ray diffraction study.
Elemental analysis calcd (%) for C6H16N2O2: C 48.63, H 10.88, N 18.90, O
21.59; found C 48.76, H 11.13, N 18.87, O 21.31. The rest of the solution was
concentrated in vacuo to 100 mL, and pentane (500 mL) was added; this
caused crystallization of a further 5.8 g crop (elemental analysis calcd (%):
C 48.63, H 10.88, N 18.90, O 21.59; found C 42.28, H 9.35, N 15.23, O 17.85),
which was recrystallized from THF (100 mL) at 48C to yield the pure
compound 2 (3.7 g, yield 63%).
1H NMR (200 MHz, 208C, D2O): d ˆ 1.25 (s); 13C NMR (200 MHz, 208C,
D2O): d ˆ 21.5 (CH3), 63.3 (CH3 C CH3). Yields of five different runs
were as follows: 43, 35, 56, 63, and 68%.
Procedure B: Compound 2,3-dimethyl-2,3-dinitrobutane (17.6 g, 0.1 mol)
was dissolved in a mixture of tetrahydrofuran (300 mL) and water (50 mL).
Zn powder (27 g) was added all at once to this solution cooled to 8 ± 108C in
an ice bath. A solution of NH4Cl (43 g, 0.8 mol) in H2O (150 mL) was added
dropwise to this slurry at such a rate (ꢀtwo hours) that the temperature of
the reaction did not exceed 128C. Then stirring was continued at 108C for
one hour, and the flask stored in a fridge (4 ± 68C) for 16 hours. As for
Procedure A, the slurry was filtered, and the precipitate carefully washed
with THF (4 Â 100 mL). The precipitate was then dried by three washings
with diethyl ether and carefully collected (59 g, Zn content: 22.6 g,
0.346 atomgrams). The solution was evaporated under vacuo until THF
ceased to distill off. Then the solution was protected from air, and sodium
carbonate (50 g) and sodium chloride (30 g) were added with cooling.
Continuous extraction with chloroform (400 mL) was performed over
18 hours. A white powder was obtained (9.4 g, 63%, m.p. 1828C).
Figure 3. View of the molecular structure of 9. Relevant bond lengths [Š]
and angles [8] are as follows: N(1) O(1) ˆ 1.287(1), N(2) O(2) ˆ 1.283(2),
N(1) C(1) ˆ 1.346(2), N(2) C(1) ˆ 1.344(2); phenyl-imidazoline ˆ 33.8(4).
majority of functional groups that one would wish to
introduce in the free radical. In this respect, this route is of
general use.
In contrast, the diamino route might be considered as of less
general use because the use of 3-chloroperbenzoic acid could
be a severe disadvantage for the exclusion of oxidant sensitive
groups such as amino groups, double bonds, and thioether in
the final free radical. However, it proceeds with higher yields
and it opens new perspectives for the design of new nitronyl
nitroxides because it bypasses the delicate synthesis of 2,3-
bis(hydroxyamino)-2,3-dimethylbutane and can be easily
extended to any adequately substituted a-diamino compound
as shown for 3,4-diamino-3,4-dimethylhexane.
Elemental analysis calcd (%) for C6H16N2O2: C 48.63, H 10.88, N 18.90, O
21.59; found 48.79, H 11.01, N 18.91, O 21.47.
Dichloro-bis[2,3-bis(hydroxyamino)-2,3-dimethylbutane]-zinc: The reduc-
tion was performed as described above except that the solution in THF/
water was evaporated under vacuum at room temperature to dryness. The
resulting solid was dissolved in hot CH2Cl2 and filtered, and the solution
evaporated again to a solid phase, which was dried several times with
pentane. A white solid was obtained (yield 16.9 g, 76%).
Elemental analysis calcd (%) for C12H32N4O4Cl2Zn: C 33.48, H 7.50, N
13.02, Cl 16.26, Zn 14.86; found C 33.60, H 7.81, N 12.91, Cl 15.97, Zn 15.03;
‡
MS: m/z (%): [Zn(2)2Cl] , 397; the isotopic pattern of Zn complexes was
successfully modeled.
Experimental Section
2,3-Diamino-2,3-dimethylbutane (3): This was obtained by
a slight
modification of reported procedures.[11, 22] Compound 2,3-dimethyl-2,3-
dinitrobutane 1 (17.6 g, 0.1 mole) was suspended in concentrated hydro-
chloric acid (37%, 150 mL). Then, granular Sn (100 g) was added by
portions, and the mixture refluxed for three hours. After cooling, the clear
solution was extracted with diethyl ether, and the aqueous phase made
strongly basic with sodium hydroxide pellets (60 g) and centrifuged to get
rid of a sticky precipitate. The solution was extracted with methylene
chloride (4 Â 50 mL), the solvent distilled off, and the residual oil dried with
pentane. Compound 2,3-diamino-2,3-dimethylbutane (9.4 g, 81%) was
obtained as a colorless solid with a low melting point.
1H NMR (200 MHz, 208C, CDCl3): d ˆ 1.07 (s, CH3), 1.26 (brs, NH2);
13C NMR (200 MHz, 208C, CDCl3): d ˆ 26.9 (CH3), 55.2 (CH3 C CH3).
Dry HCl was bubbled through the methylene chloride distillate, and an
additional crop of the chlorhydrate (1.9 g, 10%) was obtained (overall yield
91%).
Synthesis: Compound 2,3-dimethyl-2,3-dinitrobutane was prepared ac-
cording to the reported procedure and crystallized from methanol
(10 mLg 1).[11] Compound 3-chloroperbenzoic acid and Zn (<10 mm) were
purchased from Aldrich. The former was washed with a phosphate buffer
(pH 7.5), dissolved in CH2Cl2, the solution dried over sodium sulfate and
evaporated, and the resulting solid dried again by evaporation of pentane;
iodometric assay indicated 99% purity. The latter was activated by
successive washings with HCl (2%), H2O, ethanol, and diethyl ether.
Methylene chloride and chloroform were purified by filtration through
alumina (Activity 1). All other reagents were used as received.
2,3-Bis(hydroxyamino)-2,3-dimethylbutane
Procedure A: Compound 2,3-dimethyl-2,3-dinitrobutane (17.6 g, 0.1 mol)
was dissolved in tetrahydrofuran (THF, 300 mL). NH4Cl (43 g, 0.8 mole) in
water (150 mL) was added to this solution (precipitation of finely divided
NH4Cl occurred). The two-phase system was cooled in a ice bath (8 ± 128C),
and oxygen was excluded by Ar bubbling. Then, Zn powder (27 g,
4 atomgrams) was added by portions over 100 min (ꢀ3 g/10 min), while
the temperature was kept below 128C. Stirring was continued for 90 min,
and the flask placed for 15 hours (a night) in a refrigerator (4 ± 68C).[37]
Then, the mixture was filtered, the precipitate carefully washed with THF
until compound 2 was not detected in the washings[21] (4 Â 100 mL), and the
solution concentrated in vacuo until a waxy solid was obtained. Note that
this solid may be used directly in the synthesis of nitroxide free radicals,
particularly for condensation with commercially available and inexpensive
aldehydes. It contained ꢀ30 ± 50% of the free base.
2-(R)-4,4,5,5-Tetramethylimidazolidines (5)
From benzaldehyde: General procedure was used as for 5a. Benzaldehyde
(1.85 g, 0.017 mole) in diethyl ether (10 mL) was added dropwise to a ice
cooled solution of 2,3-diamino-2,3-dimethylbutane 3 (2 g, 0.017 mol) in
diethyl ether (20 mL). Condensation was complete in a few minutes. The
solution was dried over Na2SO4 and evaporated (5a: 3.5 g, 99%; 5b:[38]
92%; 5c: 95%; 5d: 89%; 5e: 78%; 5 f:[38] 95%). Single crystals of 5b were
obtained from solutions in methanol at 68C.
From benzaldehyde dimethylacetal: General procedure was used as for 5a.
Benzaldehyde dimethylacetal (2.6 g) was added to a solution of diamine 3
(2 g) in methanol (20 mL). Then, the pH of the solution was adjusted to 3 ±
4 by dropwise addition of sulfuric acid (1n). Stirring was continued for one
H2O (20 mL), sodium carbonate (30 g), sodium chloride (20 g), and
anhydrous sodium sulfate (20 g) were added to this solid. The resulting
2012
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2012 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 9

Nitronyl and Imino Nitroxides
2007±2014
hour at room temperature, the methanol evaporated, and the residue
dissolved with cooling in sodium hydroxide (6n, 50 mL). Extraction with
methylene chloride (4 Â 50 mL), drying, and evaporation gave a thick oil
(2.76 g, 78%), which crystallized in the refrigerator (5a: 78%; 5e: 67%).
during collection. Unique intensities with I > 10s(I) selected within all data
frames using the SAINT program were used to refine the cell parame-
ters.[41] The substantial redundancy in data allowed empirical absorption
corrections to be applied using multiple measurements of equivalent
reflections with the SADABS program. Space groups were derived from
systematic absences and were confirmed by the successful solution of the
structure determination. Complete information regarding crystal data and
data collection parameters is found in the CCDC data (see below).
Characteristics of imidazolidines 5a ± f are reported in Table 1. NMR
spectra showed that the imidazolidines did not need to be purified for use in
the following synthetic steps.
1,3-Dihydroxy-2-(R)-4,4,5,5-tetramethylimidazolines (4): Saturated NaH-
CO3 (30 mL) and, dropwise, 3-chloroperbenzoic acid (860 mg, 0.005 mole)
in methylene chloride (50 mL) were added to a ice cooled solution of
2-phenyl-4,4,5,5-tetramethylimidazolidine 5a (500 mg, 0.0025 mol) in
methylene chloride (50 mL). One did not observe any color change. The
mixture was stirred for half an hour at low temperature, filtered, and the
organic phase dried (Na2SO4) and evaporated under vacuum (406 mg,
70%, m.p. 224 ± 2268C).
Crystal structure determinations 2: System monoclinic, space group C2/c;
Tˆ 293 K, a ˆ 23.644(2) Š, b ˆ 6.134(1) Š, c ˆ 13.462(1) Š, b ˆ
121.868(2)8, Vˆ 1658.2(3) Š3, 1992 unique reflections, 165 parameters
refined, R(F) ˆ 0.052, Rw(F) ˆ 0.145.
Crystal structure determinations 4b: System monoclinic, space group P21/c;
Tˆ 293 K, a ˆ 6.604(1) Š, b ˆ 10.873(1) Š, c ˆ 20.030(3), b ˆ 92.220(4)8,
Vˆ 1437.2(4) Š3, 3473 unique reflections, 257 parameters refined, R(F) ˆ
0.063, Rw(F) ˆ 0.144, residual electron density 0.27, 0.30.
Crystal structure determinations 5b: System monoclinic, space group P21/n;
Tˆ 293 K, a ˆ 7.493(1) Š, b ˆ 8.038(6) Š, c ˆ 21.939(2) Š, b ˆ 91.943(2)8,
Vˆ 1320.6(1) Š3, 3203 unique reflections, (1020 with I > 2s(I)), 239 pa-
rameters refined, R(F) ˆ 0.040, Rw(F) ˆ 0.089; residual electron density
0.33, 0.37.
Elemental analysis calcd (%) for C13H20N2O2: C 66.07, H 8.53, N 11.85, O
13.54; found C 66.25, H 8.56, N 11.95, O 13.84; 1H NMR (200 MHz, 208C,
CDCl3): d ˆ 1.18 (s, 6H; CH3), 1.19 (s, 6H; CH3), 1.62 (brs, 2H; OH), 4.82
(s, 1H; CH), 7.36 ± 7.59 (m, 5H; phenyl).
Mass spectra showed the presence of traces of mono-hydroxylated
imidazolidines, which were not detected in NMR spectra. Crystallization
from methanol afforded colorless single crystals of 4b.
Crystal structure determinations 9: System monoclinic, space group P21/c,
Tˆ 293 K, a ˆ 10.060(1), b ˆ 8.761(1), c ˆ 16.554(3), b ˆ 102.116(3)8, Vˆ
1426.5(4) Š3, 3451 unique reflections, 256 parameters refined, R(F) ˆ 0.05,
Rw(F) ˆ 0.118, residual electron density 0.193, 0.242.
The data were processed by the SAINT data reduction software, and the
structures were solved by direct methods included in the SHELXTL 5.03
package.[42] All atoms were located on difference Fourier syntheses. Non-
hydrogen atoms were refined anisotropically on F 2 while hydrogen atoms,
located by calculations, were refined isotropically. Final results (R factors,
coefficients of the weighting scheme, and final residual electron densities)
are found in the CCDC data (see below).
Nitronyl nitroxides:
A solution of m-chloroperbenzoic acid (4.25 g,
0.025 mol) in methylene chloride (50 mL) was added dropwise to a mixture
of 2-phenyl-4,4,5,5-tetramethylimidazolidine 5a (2 g, ꢀ0.01 mol) in meth-
ylene chloride (100 mL) and saturated NaHCO3 (60 mL) in an ice bath.
The typical nitronyl nitroxide purple (or red for aliphatic-substituted) color
slowly developed. Stirring was continued for one hour, and a solution of
NaIO4 (3.1 g, ꢀ0.015 mol) in water (50 mL) was added dropwise.
Appearance of the nitronyl nitroxide was followed by TLC (SiO2, ethyl
acetate). Drying (Na2SO4), evaporation of the organic phase, and
chromatography (SiO2, ethyl acetate) led to pure nitronyl nitroxide
(1.8 g, 78%) and the corresponding imino nitroxide (278 mg, 12%). This
procedure did not work for preparing nitroxide 5d, for which oxidation
should be conducted in absence of NaHCO3 (see preceding section). All
free radicals 7a ± f were identical to authentic samples prepared according
to Ullmanꢀs procedure.[1±3]
Crystallographic data (excluding structure factors) have been deposited
with the Cambridge Crystallographic Data Centre as supplementary
publication nos. CCDC-149144, CCDC-149195, CCDC-149196, and
CCDC-152832 for compounds 2, 4 ± 5b, and 9, respectively. Copies of the
data can be obtained free of charge on application to CCDC, 12 Union
Road, Cambridge CB21EZ, UK (fax: (‡44)1223-336-033; e-mail: deposit
@ccdc.cam.ac.uk).
meso-4,5-Diethyl-4,5-dimethyl-2-phenyl-4,5-dihydro-1H-imidazolyl-3-ox-
ide-1-oxy: Benzaldehyde (750 mg, 7 mmol) was added to meso-3,4-
diamino-3,4-dimethylhexane (1 g, 7 mmol), obtained as described else-
where,[35] in diethyl ether (50 mL), and the solution stirred at room
temperature for 15 hours. Drying (Na2SO4) and evaporation of the solvent
afforded the corresponding imidazolidine (1.5 g, 92%) as a mixture of both
isomers. Since oxidation of both compounds resulted in the same nitroxide,
the crude product was dissolved in CH2Cl2 (100 mL) and saturated aqueous
NaHCO3 (50 mL). Then, a solution of m-chloroperbenzoic acid (3 g,
17 mmol) in CH2Cl2 (30 mL) was added dropwise to the cooled mixture
(68C), and after stirring for one hour a solution (H2O, 20 mL) of NaIO4
(2.25 g, 10 mmol) was added. Stirring was continued for one hour at 68C,
and the organic phase was dried (Na2SO4) and evaporated. The crude
product was purified by chromatography (SiO2, ethyl acetate). Nitroxide 9
(1.373 g, 83%, m.p. 70 ± 718C) was crystallized from petroleum ether to
afford single crystals suitable for a X-ray diffraction study.
Acknowledgements
Help from Dr. C. Lebrun for recording MS spectra is gratefully acknowl-
edged. This work was performed with support from the Centre National de
Á
la Recherche Scientifique, the Commissariat a lꢀEnergie Atomique, and the
European CommunityHCM program CHRXCT920080. C.H. acknowl-
edges financial support from the Region Rhone-Alpes, program Emer-
gence.
Elemental analysis calcd (%) for C15H21N2O2: C 68.94, H 8.10, N 10.72;
found C 68.83, H 7.96, N 10.89.
[1] J. H. Osiecki, E. F. Ullman, J. Am. Chem. Soc. 1968, 90, 1078.
[2] E. F. Ullman, J. H. Osiecki, D. G. B. Boocock, R. Darcy, J. Am. Chem.
Soc. 1972, 94, 7049.
[3] E. F. Ullman, L. Call, J. H. Osiecki, J. Org. Chem. 1970, 35, 3623.
[4] For recent reviews, see: V. I. Ovcharenko, R. Z. Sagdeev, Russ. Chem.
Rev. 1999, 68, 345; Magnetism: A Supramolecular Function (Ed.: O.
Kahn), Kluwer, Dordrecht, 1996; Mol. Cryst. Liq. Cryst. (Eds.: K.
Itoh, J. S. Miller, T. Takui), 1997, 305/306, 1; Mol. Cryst. Liq. Cryst.
(Ed.: O. Kahn), 1999, 334/335, 1.
Imino nitroxides: Na2WO4 (100 mg) and 5a (500 mg) were dissolved in
water/methanol (1:1, 10 mL). Then, H2O2 (1 mL, 30%) was added, and the
solution was stirred at room temperature for 12 hours. Extraction with
CH2Cl2 followed by chromatography (SiO2, ethyl acetate) afforded the
corresponding imino nitroxide 8a (223 mg, 41%). Imino nitroxides were
obtained quantitatively from the nitronyl nitroxides (7) using the reported
procedures.[39, 40]
Crystal structure determination: Selected crystals of 2, 4b, 5b, and 9 were
analyzed using a SiemensSMARTCCD area detector three-circle diffrac-
tometer (MoKa radiation, graphite monochromator, l ˆ 0.71073 Š). The
cell parameters were determined with intensities detected on three batches
of 15 frames with a 10 s exposure time for each. For three settings of F and
2V, 1200 narrow data frames were collected for successive increments of
0.38 in w. A full hemisphere of data was collected. At the end of collection,
the first 50 data frames were recollected in order to check eventual decay
[5] M. Kinoshita, Physica B 1995, 213 ± 214, 257, and references therein.
[6] R. Chiarelli, M. Novak, A. Rassat, J.-L. Tholance, Nature 1993, 363,
147.
[7] A. Caneschi, D. Gatteschi, R. Sessoli, P. Rey, Acc. Chem. Res. 1989, 22,
392.
[8] A. Caneschi, D. Gatteschi, P. Rey, Prog. Inorg. Chem. 1991, 39, 331.
[9] Formation of 2-phenyl imidazoline from 2,3-diamino-2,3-dimethylbu-
tane and phenyl thioamide followed by oxidation with NaWO4/H2O2
Chem. Eur. J. 2001, 7, No. 9
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2013 $ 17.50+.50/0
2013

P. Rey et al.
FULL PAPER
has been reported.H. G. Aurich, H. Czepluch, K. Hahn, Tetrahedron
Lett. 1977, 50, 4373.
[24] L. F. Fieser, M. Fieser, Reagents for Organic Synthesis, Vol. 1, Wiley,
New York, 1967, p. 187.
[10] M. Lamchen, T. W. Mittag, J. Chem. Soc. C 1966, 2300.
[11] R. Sayre, J. Am. Chem. Soc. 1955, 77, 6689.
[25] S. Witek, A. Bielawska, J. Bielawski, J. Heterocycl. Chem. 1980, 14,
1313.
[12] V. I. Ovcharenko, V. I. Fokin, P. Rey, Proc. VI Int. Conf. Mol. Mag.
Mol. Cryst. Liq. Cryst. 1999, 334, 109; V. I. Ovcharenko, V. I. Fokin,
G. V. Romanenko, I. V. Korobkov, P. Rey, Russ. Chem. Bull. 1999, 48,
1519.
[13] A. R. Forrester, J. M. Hay, R. H. Thomson, Organic Chemistry of
Stable Free Radicals, Academic Press, London, 1968, p. 180.
[14] L. B. Volodarsky, V. A. Reznikov, V. I. Ovcharenko, Synthetic Chem-
istry of Stable Nitroxides, CRC, Boca Raton, 1994, p. 23.
[15] J. F. W. Keana, Spin Labeling in Pharmacology (Ed.: J. L. Holtz-
mann), Academic Press, New York, 1984, p. 1.
[26] D. Cuvinot, P. Mangeney, A. Alexkis, J.-F. Normant, J. Org. Chem.
1989, 54, 2420.
[27] G. Parrinello, R. Mülhaupt, J. Org. Chem. 1990, 55, 1772.
[28] H.-W. Wanzlick, W. Löchel, Chem. Ber. 1953, 86, 1463.
[29] Compound 5e is better obtained from acetaldehyde dimethylacetal
because condensation with the unprotected acetaldehyde resulted in
some trimer, which is found as impurity in the imidazolidine.
[30] The structure is unexceptional. Structural features are reported in the
CCDC data.
[31] D. G. B. Boocock, R. Darcy, E. F. Ullman, J. Am. Chem. Soc. 1968, 90,
5945.
Á
[16] R. Briere, H. Lemaire, A. Rassat, Bull. Soc. Chim. Fr. 1965, 3273.
[17] G. Chapelet-Letourneux, H. Lemaire, A. Rassat, Bull. Soc. Chim. Fr.
1966, 3283.
[18] T. Toda, E. Mori, K. Maruyama, Bull. Chem. Soc. Jpn. 1972, 45, 1904.
[19] We noticed that use of crude 2,3-dimethyl-2,3-dinitrobutane, which
probably contains brominated derivatives as impurity, often leads to
the undesired methyl-substituted nitronyl nitroxide.
[32] D. G. B. Boocock, E. F. Ullman, J. Am. Chem. Soc. 1968, 90, 6873.
[33] H. Chikashita, S. Nishida, M. Miyazaki, Y. Morita, K. Itoh, Bull.
Chem. Soc. Jpn. 1987, 60, 737.
[34] K. M. Ibne-Rasa, J. O. Edwards, J. Am. Chem. Soc. 1962, 84, 763.
[35] M. F. Asaro, I. Nakayama, R. B. Wilson, Jr., J. Org. Chem. 1992, 57,
778.
[20] Elemental analysis calcd (%) for [Zn(NH3)2Cl2]: H 3.55, N 16.44, Cl
41.62, Zn 38.38; found H 3.51, N 15.67, Cl 39.95, Zn 38.31. This shows
that the precipitate contains a few percent of [Zn(NH3)(OH)Cl] as
already observed.[12]
[36] Y. Pei, O. Kahn, M. A. Aebersold, L. Ouahab, F. Le Berre, L. Pardi, J.-
L. Tholence, Adv. Mater. 1994, 6, 681.
[37] If it is put in a fridge, the precipitate is easier to filter but this does not
increase the yield.
[21] Detection of compound 2 was performed as follows: Two drops of
HCHO (37% in water) and a saturated aqueous solution of NaIO4
(1 mL) were added successively to the solution (1 mL). Development
of a pink color due to the presence of the corresponding nitronyl
nitroxide indicates that compound 2 is present. This test is not
quantitative because the nitroxide is not stable in this solution, and the
color fades rapidly.
[38] Reported yield for condensation in refluxing chloroform.
[39] F. Lanfranc de Panthou, Thesis dissertation, Grenoble (France), 1994.
[40] P. Rey, D. Luneau, Supramolecular Engineering of Synthetic Metallic
Materials (Eds.: J. Veciana, C. Rovira, D. B. Amabilino), NATO ASI
Series, Kluwer, 1999, 518, p. 145.
[41] SAINT Version 4.050, Copyright 1992 ± 1995, Siemens Analytical
X-ray Instruments.
[22] C.-M. Che, K.-Y. Wong, H.-W. Lam, K.-F. Chin, Z.-Y. Zhou, T. C. W.
Mak, J. Chem. Soc. Dalton Trans. 1993, 857.
[42] SHELXTL Version 5.030, Copyright 1994, Siemens Analytical X-ray
Instruments.
[23] K. Hofmann, The Chemistry of Heterocyclic Compounds (Ed.: A.
Weissberger), Interscience Publishers, New York, 1953, p. 213.
Received: September 6, 2000 [F2716]
2014
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0709-2014 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 9