Reduction of alkyl and aryl azides with sodium thiophosphate in aqueous solutions

Article abstract of DOI:10.1016/j.tetlet.2011.03.083

A simple aqueous method for the conversion of alkyl and aryl azides into the corresponding amines using trisodium thiophosphate is presented. Thiophosphate is converted into phosphate during these formal reduction processes.

Full text of DOI:10.1016/j.tetlet.2011.03.083

Tetrahedron Letters 52 (2011) 2730–2732
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Reduction of alkyl and aryl azides with sodium thiophosphate
in aqueous solutions
⇑
Jennifer L. Norcliffe, Louis P. Conway, David R. W. Hodgson
Department of Chemistry, Durham University, Science Laboratories, South Road, Durham DH1 3LE, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
A simple aqueous method for the conversion of alkyl and aryl azides into the corresponding amines using
trisodium thiophosphate is presented. Thiophosphate is converted into phosphate during these formal
reduction processes.
Received 7 January 2011
Revised 1 March 2011
Accepted 18 March 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Aqueous
Azide
Reduction
Thiophosphate
The formation of primary amines is a cornerstone of organic
synthesis. Classical methods for the preparation of these systems
from alkyl halides revolve primarily around the use of phthalimide
(Gabriel) followed by deprotection, or the displacement of halide
ions with azide followed by reduction of the resulting organic
azide. The reduction of azides is commonly achieved through me-
tal-based hydrogenations, hydride-based reductions or the Stau-
dinger procedure using PPh3. Less common methods include the
use of H2S and hydrogen transfer methods from hydrazine. In addi-
tion, reductions that proceed in aqueous solvent systems have
been developed, where transition metal catalysts in combination
with hydride reagents1 or transition metal based single electron
transfer reducing agents2 are employed. While these methods are
reliable, they each present their own limitations in terms of cost,
convenience and disposal of effluents.
We began our investigations using benzyl azide (4a)10 as a con-
venient model. As the quality of commercially available sodium
thiophosphate was found to be variable, a quantity was synthes-
ised according to a literature procedure.11 The product was found
to be free of phosphate and anhydrous, according to 31P NMR
and thermogravimetric analyses. Initially, we explored the use of
a fully aqueous solvent system where we employed 1 equiv of
thiophosphate ion (Table 1, entry 1). The azide 4a was refluxed
with the thiophosphate for 3 h before being subjected to standard
work-up. Unfortunately, only 77% reduction to the amine was
observed, which we attributed to the limited solubility of the azide
substrate. To overcome the poor solubility, we adopted a mixed
water/iso-propanol solvent system and repeated the procedure
using 1 equiv of thiophosphate. After work-up, reduction had
improved to 93% (entry 2), however, complete conversion of the
thiophosphate into phosphate was observed by 31P NMR spectros-
copy. We surmised that the thiophosphate ion underwent a
Recently, in an attempt to form an N-phosphorylated 50-amino-
50-deoxy-nucleoside 23,4 from the azidonucleoside 1, we investi-
gated the use of thiophosphate ions (Scheme 1). Our procedure
was based on precedents set with reactions between thiocarboxy-
late systems and organic azides for the formation of carboxylic
amides and N-acyl sulfonamides.5–8 However, we found that reac-
tion mixtures containing thiophosphate ions effected a formal
reduction of the azide 1 to amine 3 rather than the formation of
the desired phosphoramidate 2.9
H
O
N
G
P
O
O
O
N
G
3
OH OH
2
Na SPO
3
3
O
H O
2
With this transformation in mind, we now present details of the
use of thiophosphate as a general reagent for the reduction of or-
ganic azides to their corresponding amines.
OH OH
reflux, 1 h
H N
2
G
O
1
OH OH
3
⇑
Corresponding author. Tel.: +44 191 33 42123.
E-mail address: d.r.w.hodgson@durham.ac.uk (D.R.W. Hodgson).
Scheme 1. Reaction of 50-amino-50-deoxyguanosine with thiophosphate.
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.03.083

J. L. Norcliffe et al. / Tetrahedron Letters 52 (2011) 2730–2732
2731
Table 1
Table 2
Preliminary investigations with benzyl azide (4a)
Preparation of amines from azides
Entry Solventa
Conversion
Conversion of
SPO33ꢀ equiv
of 4a into 4b (%)b
SPO33ꢀ into PO43ꢀ (%)c
Entry
1
Substrate
Product
t (h)
Yielda (%)
1
2
H2O
1
1
3
0
3
77
93
100
0
100
100
80
—
100
3
93
54
H2O/iPrOH
H2O/iPrOH
H2O/iPrOH
H2O/iPrOH
2
3
n-C8H17–N3
MeO2CCH2N3
n-C8H17–NH2
16
1
3
b
ꢀO2CCH2NH2
—
4d
5e
—
4c
3
56
43
H2O/iPrOH was used in 2:1 ratio.
a
Determined by 1H NMR spectroscopy of the crude product mixture after work-
b
up.
5d
48
Determined by 31P NMR spectroscopy of the crude reaction mixture before
c
work-up.
d
Reaction performed under the same conditions as entries 1–3, but without
Na3SPO3.
Reaction performed at 95 °C in an NMR tube under the same concentration
conditions as entry 3, but without benzyl azide (4a).
6
3
72
87
e
7
8
16
3
e
NC(CH2)3–N3
ꢀO2C(CH2)3–NH2
—
desulfurization process leading to the production of thiolate ions.
Thiolate ions may also act as a reducing agent, but can be lost from
the reaction in the form of H2S. To overcome this, we increased the
number of equivalents of thiophosphate ion being employed and
repeated the procedure, whereupon benzyl azide (4a) was cleanly
converted into benzylamine (4b) (entry 3). 31P NMR spectroscopy
performed on the crude reaction mixture after reflux showed that
more than 1 equiv of thiophosphate was consumed. To confirm
that thiophosphate was the source of the reduction process, benzyl
azide (4a) was heated in the solvent mixture without thiophos-
phate, and unreacted azide was recovered from the reaction mix-
ture (entry 4). In addition, thiophosphate was heated in the
absence of benzyl azide (4a), and decomposition to phosphate
was still observed (entry 5).
a
Yields were determined after work-up without further purification.
Reaction was performed in D2O at 95 °C in an NMR tube. Glycine was not
b
isolated from the reaction mixture, but its identity was confirmed by 1H and 13C
NMR spectroscopy.
c
The HCl salt of the substrate was used with 1 equiv of NaOH added to the
reaction mixture.
10 equiv of SPO33ꢀ were used, added over two portions, one at the outset of the
d
reaction, and one after 24 h reflux.
e
Reaction was performed in D2O at 90 °C in an NMR tube, producing a mixture of
c
-aminobutyric acid (GABA) and 2-pyrrolidone. This mixture was then hydrolysed
by the addition of NaOD. GABA was not isolated from the reaction mixture, but its
identity was confirmed by 1H and 13C NMR spectroscopy.
With an appreciation of the desulfurization process in hand, and
having achieved clean conversion of benzyl azide (4a) into 4b, we
explored the scope of the SPO33ꢀ-based reduction with a range of
azide substrates. The azide substrates were prepared using estab-
lished aqueous (Table 2, entries 3 and 6),12,13 DMSO-based proce-
dures (entries 1, 2, 7 and 8),10,14 or obtained commercially
(entries 4 and 5). Reductions were carried out with 3 equiv of
SPO33ꢀ, unless stated otherwise.
conversion to cyclohexyl amine, and was isolated in good yield
via standard work-up. After 3 h of heating, 1-azido-3-cyanopro-
pane (entry 8) was reduced to a mixture of
c-aminobutyric acid
(GABA) and 2-pyrrolidone. Addition of sodium deuteroxide and a
further 16 h of heating afforded GABA as the sole product, as con-
firmed by spiking with authentic GABA.
Mechanistically, we can consider three possible pathways for
the formal reduction process (Scheme 2). Initially, in our work with
nucleoside substrate 1, we aimed to form phosphoramidate 2, via a
stepwise, or concerted route, analogous to pathway A. Kinetic
In all cases the azide functionality was reduced to the amine.
With simple 1°- and 2°-alkyl systems (entries 1, 2, 6 and 7), the
products were isolated from the reaction mixture using standard
work-up procedures. For
a-azidoethyl acetate (entry 3), complete
hydrolysis of the methyl ester was also observed, which likely
arises because of the basic nature of trisodium thiophosphate solu-
tion in combination with the prolonged reflux. The glycine product
was not isolated, but its identity was confirmed by ‘spiking’ the
crude D2O-based reaction mixture with authentic glycine. The aro-
matic system, p-amino-azidobenzene (entry 4), showed complete
reduction in trial 1H NMR experiments in D2O, and, thus, we attri-
bute the relatively low isolated yield to poor extraction of the dia-
mino product on work-up. The 3°-azido adamantyl system showed
conversion into the amine (entry 5), however, the process occurred
very slowly even in the presence of elevated concentrations of
SPO33ꢀ and extended reflux. In the presence of 5 equiv after 24 h
of reflux, complete consumption of thiophosphate was observed,
however, only partial reduction of the azide was observed. An
additional 5 equiv of SPO33ꢀ were added and the mixture was re-
fluxed for an additional 24 h resulting in complete reduction of
the azide and an isolated yield of 43% of 1-aminoadamantane. 2-
Azido-1-phenylethanol (entry 6) was reduced to the corresponding
amine in a yield of 72%. Cyclohexyl azide (entry 7) showed clean
Scheme 2. Potential mechanisms for reduction by SPO33ꢀ
.

2732
J. L. Norcliffe et al. / Tetrahedron Letters 52 (2011) 2730–2732
is consistent with pathway C, where the thiolate ions build-up as
the reducing agent is required. In addition, the 1H spectra con-
tained signals corresponding to 4-azido-aniline and p-diaminoben-
zene alone, with no evidence of the formation of intermediates.
This is reinforced by the fact that 31P spectra show clean conver-
sion of SPO33ꢀ into phosphate ions. With these observations in
mind, we tentatively put forward a mechanistic proposal in line
with pathway C with a rate-limiting attack of thiolate ion followed
by fast decomposition of the resulting adduct. On this basis, SPO33ꢀ
represents a convenient, ‘caged’ form of thiolate ions which are
known reducing agents for organic azides.
In summary, we have developed a convenient, aqueous method
for the reduction of organic azides to amines. The by-product from
the reduction process is inorganic phosphate ions, which apart
from being easy to remove on work-up, represent an innocuous
effluent. Future studies will centre on the mechanism of the reduc-
tion process and exploring the reactions of esters of thiophosphoric
acid with azides.
Figure 1. 1H and 31P NMR kinetic experiments on the reduction of 4-azido-aniline
to p-diaminobenzene by SPO33ꢀ at 70 °C in D2O. Blue squares represent the
remaining number of equivalents of SPO33ꢀ determined by 31P NMR. Red circles
represent the number of equivalents of p-diaminobenzene determined by 1H NMR.
Acknowledgements
studies on phosphoramidates 84,15 suggest they would be short-
lived under the reflux conditions, and their breakdown would
result in the formation of amines 9 and phosphate ions.
We thank the Engineering and Physical Sciences Research
Council for a Vacation Bursary for J.L.N. and a Ph.D. studentship
for L.P.C.
While we cannot exclude pathway A conclusively, the intramo-
lecular capture of the nitrogen anion by the phosphoryl dianions 6,
or a concerted addition of the sulfur anion of SPO33ꢀ to azides 5 to
give cyclic intermediates 7, seems improbable on the basis of
charge repulsion. This contrasts with the formation of iminophos-
phorane intermediates during Staudinger reductions where a
nitrogen anion is captured by a cationic phosphorus centre, and
the formation of cyclic intermediates between azides and thiocarb-
oxylic acids en route to amides. However, protonation of one of the
phosphoryl oxygen anions could occur in the aqueous solvent,
reducing the intramolecular repulsion. Pathway B would also give
rise to a thiophosphoryl-azide adduct 6, which, on N-protonation,
could decompose via loss of the phosphoryl group, followed by for-
mal loss of N2S, to give amine 9. Pathway C would also produce
thiolate intermediate 10 through direct addition of SH– formed
from desulfurization of SPO33ꢀ and would likely proceed via the
same mechanism as H2S-based reductions of azides. In order to
gain some understanding of the potential mechanism, we per-
formed 1H and 31P NMR kinetic studies on the reduction of 4-azi-
do-aniline (Fig. 1).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2011.03.083.
References and notes
1. Fringuelli, F.; Pizzo, F.; Vaccaro, L. Synthesis 2000, 646–650.
2. Lee, J. G.; Choi, K. I.; Koh, H. Y.; Kim, Y.; Kang, Y.; Cho, Y. S. Synthesis 2001, 81–
84.
3. Williamson, D.; Cann, M. J.; Hodgson, D. R. W. Chem. Commun. 2007, 5096–
5098.
4. Williamson, D.; Hodgson, D. R. W. Org. Biomol. Chem. 2008, 6, 1056–1062.
5. Shangguan, N.; Katukojvala, S.; Greenburg, R.; Williams, L. J. J. Am. Chem. Soc.
2003, 125, 7754–7755.
6. Barlett, K. N.; Kolakowski, R. V.; Katukojvala, S.; Williams, L. J. Org. Lett. 2006, 8,
823–826.
7. Kolakowski, R. V.; Ning, S. G.; Williams, L. J. Tetrahedron Lett. 2006, 47, 1163–
1166.
8. Kolakowski, R. V.; Shangguan, N.; Sauers, R. R.; Williams, L. J. J. Am. Chem. Soc.
2006, 128, 5695–5702.
ˇ ´
9. Brear, P.; Freeman, G. R.; Shankey, M. C.; Trmcic, M.; Hodgson, D. R. W. Chem.
The desulfurization of SPO33ꢀ to afford phosphate ions and thio-
late appears to exhibit zero-order kinetics. This is likely attribut-
able to the pH dependence of desulfurization, where lower pH
leads to a greater rate of desulfurization. Measurements of the
pH of a solution of SPO33ꢀ confirm that the pH decreases from
ꢁ11 to ꢁ9 over the course of 3 h. The appearance of p-diaminoben-
zene displays a lag period that suggests a requirement for the
build-up of an intermediate that acts as the reducing agent. This
Commun. 2009, 4980–4981.
10. Alvarez, S. G.; Alvarez, M. T. Synthesis 1997, 413–414.
11. Yasuda, S. K.; Lambert, J. L. J. Am. Chem. Soc. 1954, 76, 5356.
12. Acquaah-Harrison, G.; Zhou, S.; Hines, J. V.; Bergmeier, S. C. J. Comb. Chem.
2010, 12, 491–496.
13. Ju, Y. H.; Kumar, D.; Varma, R. S. J. Org. Chem. 2006, 71, 6697–6700.
14. Yuan, H.; Silverman, R. B. Bioorg. Med. Chem. 2006, 14, 1331–1338.
15. Benkovic, S. J.; Sampson, E. J. J. Am. Chem. Soc. 1971, 93, 4009–4016.