Improved preparation of tosyldiazomethane

Article abstract of DOI:10.1016/S0040-4039(00)00763-2

Treatment of the carbamate 2a by i-amyl nitrite and TMSCl in presence of a reduced amount of pyridine, followed by decomposition of the resulting nitroso derivative 3a with activated alumina afforded in high yield the pure, crystalline, diazosulfone 1a, a

Full text of DOI:10.1016/S0040-4039(00)00763-2

Tetrahedron Letters 41 (2000) 5489±5493
Improved preparation of tosyldiazomethane
C. Plessis,a D. Uguen,a,* A. De Cianb and J. Fischerb
a
Á Â Á Â
Laboratoire de Synthese Organique, associe au CNRS, Ecole Europeenne de Chimie, Polymeres et Materiaux,
Universite Louis Pasteur 25, rue Becquerel, 67087 Strasbourg, France
Â
Â
 Â
Laboratoire de Cristallochimie etChimie Structurale, associe auCNRS, Universite Louis Pasteur, 67070Strasbourg, France
b
Received 20 April 2000; accepted 10 May 2000
Abstract
Treatment of the carbamate 2a by i-amyl nitrite and TMSCl in presence of a reduced amount of
pyridine, followed by decomposition of the resulting nitroso derivative 3a with activated alumina a€orded
in high yield the pure, crystalline, diazosulfone 1a, a process which could be extended to the preparation of
the parent diazosulfone 1b. # 2000 Elsevier Science Ltd. All rights reserved.
As part of a synthetic work, we had to prepare various tosylmethyl ethers and an obvious,
straightforward way for doing so was, as indicated, to react the corresponding alcohols with
tosyldiazomethane 1a in presence of an acid catalyst (Scheme 1).1 The preparation of the diazo-
sulfone 1a was thus examined.
Scheme 1.
According to the organic syntheses directions,2 the carbamate 2a, which was formed eciently
by stirring for a few days a solution of ethyl carbamate and sodium p-toluenesul®nate in 30%
formalin diluted with methanol and formic acid, was reacted with nitrosyl chloride in pyridine,
used both as a reactant and a solvent (Scheme 2). Dilution with water of the crude reaction
mixture then a€orded the N-nitroso derivative 3a, the treatment of which with basic alumina in
ether a€orded the desired diazo compound 1a.
* Corresponding author.
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00763-2

5490
Scheme 2. Reagents and conditions: according to Ref. 1: (1) sodium p-toluenesul®nate in formalin/formic acid; (2)
NOCl in pyridine; (3) basic alumina in ether/CH2Cl2
Some pure 1a could indeed be obtained this way. However, after repeated experiments on
various scales (ca. 0.1 to 10 g), we noticed that not only the yields but also the physical state and,
to a lesser extent, the purity3 of this product varied rather erratically. Hence, though in a few
cases the ®nal treatment of 3a with basic alumina a€orded as reported, though in lower yield, the
diazosulfone 1a as a yellow solid, in many instances an impure orange to reddish oil resisting
crystallisation was isolated in variable yields. With the aim to elucidate the origin of this
inconsistency, we re-examined each step of this otherwise valuable procedure.4
It could obviously be surmised that much of these diculties originated from the conditions of
the ®nal 3a±1a conversion. As pointed out in the relevant publication,2 the time required for
observing complete disappearance of the starting nitroso derivative 3a was highly dependent on
the speci®cationÐi.e. basic or neutralÐof the used alumina, the former being, with regards to the
kinetic of this process, the more ecient. The trouble, however, also noticed in this publication,
was the formation of red-coloured side-products at the surface of the alumina when the basic
form was used. Though no other product than 1a appeared (TLC) to be eluted during ensuing
®ltration and washing operations, trace amounts of these impurities could be present however,
thus preventing the crystallisation of the oily 1a thus obtained. The attempted use of neutral
alumina looked at ®rst bene®cial since only a faint pink colour developed in these conditions.
But, in agreement with previous observations,2 though the clean formation of 1a could be
substantiated (TLC, NMR), the reaction then proceeded only very slowly and repeated
treatments of the resulting 1a±3a mixture with alumina, detrimental to the yield owing to the
moderate stability of 1a, were necessary to observe a full conversion of 3a.
These problems could ®nally be solved by ®rst heating at ca. 300ꢀC commercial (Grade I) basic
alumina in an oven for 12 h, the dry powder thus obtained being then slurried in ether with
cooling (ice bath) and vigorous stirring before being treated by a solution of 3a in CH2Cl2, any
adventitious moisture being constantly excluded during all operations. After several hours of
stirring at rt, a simple elimination of the solids by ®ltration, followed by evaporation of the
solvents, gave reproducibly the pure (TLC, NMR) diazosulfone 1a as an oil crystallising almost
instantly in the cold (ca. ^30ꢀC).
Another crucial point appeared to be the quality of the nitroso compound used: the purer the
starting nitroso derivative, the higher the yield in crystalline diazosulfone 1a. Unfortunately, due
to the use of pyridine as solvent in the nitrosation step, the isolation of 3a proceeded unsatis-
factorily: whereas in a few cases 3a crystallised out readily by dilution of the reaction mixture
with water as recommended,2 in most experiments this dilution resulted only in the slow appear-
ance of a gummy solid having a strong smell of pyridine, and a classical extraction, followed by a
puri®cation by chromatography was necessary to obtain pure 3a in reduced yield.
In order to avoid the use of a large excess of pyridine, we became interested by a report on the
clean conversion of oximes into the corresponding ketones by treatment with a mixture of t-butyl
nitrite and TMSCl.5 Though not explicitly mentioned in this publication, it could be supposed

5491
that TMSCl induced ®rst the formation of the O-TMS derivative of nitrous acid, then possibly
decomposed into nitrosyl chloride as indicated (Scheme 3).
Scheme 3.
In the event, simply mixing 2a with an organic nitrite and TMSCl in an inert solvent and in
presence of a reduced amount of an amine base would have resulted in the formation of 3a
(Scheme 4). This proved to be the case.
Scheme 4. Reagents and conditions: (1) i-amyl nitrite (1.4 equiv.), pyridine (1.4 equiv.), TMSCl (2.8 equiv.), CH2Cl2;
rt, 1 day; (2) activated basic alumina, ether/CH2Cl2; overnight, then crystallisation from ether/hexane
Sequential addition of i-amyl nitrite, preferred over t-butyl nitrite for its higher stability, and
TMSCl to a CH2Cl2 solution of 2a and pyridine, each reagents being in slight excess, resulted
after one day of stirring at rt in the obtention of pure 3a in high yield (Table 1). Moreover,
submitting this compound to the treatment with activated alumina as above a€orded almost
quantitatively the diazosulfone 1a as an orange oil which crystallised on standing in the freezer
overnight (90% yield overall, from 2a).6 Interestingly, submitting the carbamate 2b, which was
formed by condensing acetaldehyde and p-toluenesul®nic with ethyl carbamate, to the same two-
step procedure gave similarly, the parent, crystalline, diazosulfone 1b (Table 1).7
Table 1
Preparation of the diazosulfones 1 from carbamates 2 by the i-amyl nitrite-TMSCl process
Slow recrystallisation of 1a was then attempted in an ether/hexane mixture, at low temperature,
as described (Fig. 1).2 To our delight, one of the crystals thus obtained (67.2%) proved suitable
for X-ray analysis, which was run by maintaining the temperature at ^100ꢀC throughout in order
to minimise the decomposition of 1a.8
As can be seen on the Chem-3D structure computed from the resulting crystal data, the length of
the carbon±nitrogen bond (ca. 1.294 A) is close to the value (ca. 1.32 A) obtained for the carbon±
nitrogen bond in diazomethane,9a which suggested at ®rst sight that the term tosyldiazomethane

5492
Figure 1.
used to designate 1a is justi®ed. However, the relatively short bond (1.7 A)9b of the sulfur atom
with the carbon atom bearing the diazo residue is indicative of some delocalisation of the negative
charge on the sulfonyl group, which is consistent with previous NMR9c and IR9d studies.
Additionally, besides being chiral, each molecule of tosyldiazomethane 1a is arranged in the
crystal unit so that each central nitrogen atom, which should theoretically share a positive charge,
is bonded to one of the two oxygen atoms of the neighbouring SO2 residue, a complexation which
is reminiscent of the metal±SO2 interaction displayed by anionised sulfones in the crystalline
state.9b
In conclusion, new conditions for converting eciently the carbamate 2 into the corresponding
N-nitroso derivatives 3 has been found and, although this has not yet been fully explored, it can
be predicted that the combination of an alkyl nitrite with TMSCl used presently should also be
useful for performing related nitrosylation reactions. Additionally, improved conditions for the
known alumina-induced decomposition of compounds 3 into the diazosulfones 1 have been
identi®ed. Both improvements contribute to making this method for preparing diazosulfones much
ecient, permitting inter alia to obtain a crystal of 1a suitable for an apparently unprecedented
X-ray analysis. Further study of this interesting class of diazo compounds is actively pursued, the
results of which will be published in due course.
References
1. Bruckner, R.; Peiseler, B. Tetrahedron Lett. 1988, 29, 5233±5236.
2. van Leusen, A. M.; Strating, J. Org. Synth. 1977, 57, 95±102. This procedure is based on the work of Strating and
van Leusen (van Leusen, A. M.; Strating, J. Rec. Trav. Chim. Pays-Bas 1965, 84, 151±164 and references cited
therein), who discovered this class of diazo compounds.
3. Most of the crude samples of 1a thus obtained appeared pure in NMR. In a few cases, however, a weak singlet at
1
6.8 ppm, possibly indicative of the presence of 1,2-bis-tosylethylene, was also perceptible in the H NMR spectra.
4. Sulfones 1 cannot be obtained from the corresponding p-tolyl alkylsulfones by using the diazo transfer method of
Regitz (see: Dieckmann, J. J. Org. Chem. 1965, 30, 2272±2275). Recently, a procedure based on the related
alumina-induced decomposition of 1-diazo-1-tosyl-acetophenone has been proposed (Korneev, S.; Richter, C.
Synthesis 1995, 1248±1250). However, owing to the ease with which the starting carbamate 2a is prepared, this
method presents no clear advantage over the van Leusen's process.
5. Lee, J. G.; Kwak, K. H.; Hwang, J. P. Tetrahedron Lett. 1990, 31, 6677±6680.

5493
6. All reactions were performed in ¯ame-dried ¯asks in a dry argon atmosphere and using well-dried solvents and
reagents, all vessels used being protected from daylight with a foil of aluminium during any operation. Protocol
for the nitrosylation of 2a: To a stirred solution of 2a (19 g; 738 mmol) in CH2Cl2 (100 ml) were added
successively, by means of a syringe: (i) pyridine (8.5 ml; 105 mmol); (ii) i-amyl nitrite (14 ml; 105 mmol); (iii)
TMSCl (26.7 ml; 210 mmol). The resulting clear solution was stirred for 24 h at rt (should a TLC indicate the
presence of residual 2a, slight excess (0.2 equiv.) of each reagent will be added in the same order, the stirring being
pursued until complete disappearance of the starting material). The reaction mixture was then poured into 10%
aqueous NaHCO3 (200 ml). After 30 min stirring, the aqueous layer was extracted with CH2Cl2 (2Â50 ml). The
pooled organic extracts were then washed successively with 1 N HCl (2Â150 ml), brine (2Â50 ml), 10% NaHCO3
(50 ml), then brine (20 ml), and dried (MgSO4). Rotoevaporation (without heating) of the solvents left a pale-
yellow powder (21 g) which was recrystallised in CH2Cl2/Et2O to give 2a (19.2 g; 67 mmol; 90.8%) as a bright-
yellow powder; mp 88±89ꢀC (lit.2 mp 87±89ꢀC); H NMR: 1.33 (t, J=7.12 Hz, 3H), 2.4 (s, 3H), 4.42 (q, J=7.12
1
Hz, 2H), 5.06 (s, 2H), 7.31 (m, 2H), 7.63 (m, 2H); 13C NMR: 14.1, 21.8, 58.5, 65.4, 128.5, 130.6, 135.4, 145.9,
152.4. The 3±1 conversions were performed according to Ref. 2 except that the ¯ask containing the basic alumina
(Merck; 3 g/mmol) was ®rst heated in an oven at 300ꢀC for 12 h, then immediately transferred in a desiccator
where it was allowed to cool to rt in vacuo. It was then connected to an argon line, and cooled (ice/methanol)
before the ether was added. After 12 h stirring at rt, on a 7.2 g scale (25 mmol), the diazosulfone 1a was obtained
as bright-yellow needles (3.3 g; 67.2%) after recrystallisation from a 2:1 ether:hexane mixture.
1
7. Selected data: 2b: H NMR: 1.04 (t, J=6.8 Hz, 3H), 1.56 (d, J=6.7 Hz, 3H), 2.39 (s, 3H), 3.86 (q, J=6.8 Hz),
4.85±5.1 (m, 1H), 5.57 (d, J=10.3 Hz, 1H (NH)), 7.29 (m, 2H), 7.75 (m, 2H); 13C NMR: 13.2, 14.5, 21.7, 61.6,
67.4, 129.5, 129.8, 133.4, 145.2, 154.8; 3b: 1H NMR: 1.45 (t, J=7.15 Hz, 3H), 1.67 (d, J=7.2 Hz, 3H), 2.43 (s, 3H),
4.53 (qd, J=7.15, 1.6 Hz, 2H), 5.75±5.9 (m, 1H), 7.34 (m, 2H), 7.68 (m, 2H); 13C NMR: 13.3, 14.5, 21.8, 61.7, 67.4,
1
129.4, 129.8, 133.3, 145.3, 154.9; 1b: 13C NMR: 8.95, 21.7, 59.3, 126.7, 130.1, 139.3, 144.4. H and 13C NMR at
200 and 50 MHz, respectively, in CDCl3.
8. Crystal data for 1a: C8H8N2O2S, M.W.=196.23, colourless, orthorhombic, a=7.791(3), b=20.389(6), c=5.653(2)
A, Z=4, dcalc=1.45 g cm^3, ꢀ=2.906 mm^1, space group P212121. Data were collected on a crystal of dimensions
0.35Â0.35Â0.12 mm3 using a Philips PW1100/16 automatic di€ractometer and graphite monochromated CuKꢁ
radiation (l=1.5418 A) at ^100ꢀC. 672 re¯ections measured (3ꢀ<2ꢂ<54ꢀ) with 603 having I>3 ꢃ(I). Empirical
absorption corrections, transmission factors between 0.28 and 1. The structure was determined using direct
methods. Hydrogen atoms were introduced in structure factor calculations as ®xed contributors with d(X-
H)=0.95 A and B(H)=1.3*Beqv(X) A2. Re®nements against l Fl using the OpenMoleN package on a DEC
Alpha workstation and anisotropic temperature factors for all non-hydrogen atoms. The absolute structure was
determined re®ning Flack's x parameter. Final results: R(F)=0.052, Rw(F)=0.068, GOF=1.657.
9. (a) Regitz, M.; Waas, G. Diazo Compounds. Properties and Synthesis; Academic Press: London, 1986; (b) To be
compared with the 1.66 A value obtained for the corresponding C±S bond in anionised benzyl phenylsulfone
(Reetz, M. T.; Hutte, S.; Goddard, R. Eur. J. Org. Chem. 1999, 2475±2478 and references cited therein); (c)
Cerioni, G.; Culeddu, N.; Saba, A. Magn. Reson. Chem. 1993, 31, 829±831; (d) Engberts, J. B. F. N.; Zwanenburg,
B. Tetrahedron 1968, 24, 1737±1754.