Thermal Decomposition of 3′-Azidothymidine

Article abstract of DOI:10.1023/A:1026081917811

The kinetics of thermal decomposition of 3′-azidothymidine (AZT) in melt at 140-201°C were studied by the thermogravimetric and manometric methods. The process follows the first-order kinetics and is decribed by the Arrhenius equation: k = 10^14.7±0.5

Full text of DOI:10.1023/A:1026081917811

Russian Journal of Organic Chemistry, Vol. 39, No. 5, 2003, pp. 727 730. Translated from Zhurnal Organicheskoi Khimii, Vol. 39, No. 5, 2003,
pp. 774 777.
Original Russian Text Copyright
2003 by Nedel’ko, Korsunskii, Chukanov, Larikova, Korchevskaya, Malin, Shcherbinin, Ostrovskii.
Dedicated to Prof. G.I. Koldobskii
on the 70th Anniversary of His Birth
Thermal Decomposition of 3 -Azidothymidine
V. V. Nedel’ko1, B. L. Korsunskii1, N. V. Chukanov1, T. S. Larikova1,
E. V. Korchevskaya2, A. A. Malin2, M. B. Shcherbinin2, and V. A. Ostrovskii2
1
Institute of Chemical Physics Problems, Russian Academy of Sciences,
Chernogolovka, Moscow oblast, 142432 Russia
e-mail: vnedelko@icp.ac.ru
2
St. Petersburg State Institute of Technology, Moskovskii pr. 26, St. Petersburg, 198013 Russia
Received December 20, 2002
Abstract The kinetics of thermal decomposition of 3 -azidothymidine (AZT) in melt at 140 201 C were
studied by the thermogravimetric and manometric methods. The process follows the first-order kinetics and
is decribed by the Arrhenius equation: k = 1014.7 0.5 exp[ (38100 900)/RT] (1/s). The kinetic and activation
parameters are typical of thermal decomposition of aliphatic azides; they suggest that decomposition of the
azido group is the rate-determining stage. The formation of nitrogen (1 mol per mole of AZT), furan, methyl
alcohol, water, and carbon dioxide was confirmed by mass spectrometry. Thymine was found to be one of
the condensed products of thermal decomposition of 3 -azidothymidine.
3 -Azidothymidine (AZT) is an efficient inhibitor
of the HIV reverse transcriptase. Despite the fact that
some new inhibitors of the reverse transcriptase and
protease have appeared in the pharmaceutical market,
the significance of AZT as the most important anti-
HIV drug has not been reduced [1].
In the recent publications, considerable attention
is given to kinetic studies of processes leading to
formation of AZT [2, 3]. These studies are important
from the viewpoint of understanding the mechanisms
of some key reactions in nucleoside synthesis, as well
as for solving a number of technological problems.
Nevertheless, the available data cannot be regarded as
sufficient without information on thermal decomposi-
tion of products and intermediates of the above reac-
tions. Quantitative data on their thermal transforma-
tions may be useful for interpretation of biological
activity of AZT and its metabolism.
Thermal decomposition of nucleosides was studied in
[6 8]. Comparison of the activation parameters typical
of these processes makes it possible to substantiate
a probable mechanism of thermal decomposition of
3 -azidothymidine.
When a sample of AZT was heated in air at a rate
of 4 deg/min, no decomposition was observed up to
180 C (Fig. 1). Above that temperature, stepwise
decomposition occurred. The first stage was accom-
panied by a weight loss of 20% (up to 230 C). The
subsequent decomposition (230 260 C) was slow.
In the present work we studied the kinetics of
thermal decomposition of AZT and proposed possible
ways and mechanism of this process. A molecule of
AZT contains two fragments whose thermal decom-
position may initiate profound decomposition of
the entire molecule. These are the azido group and
nucleoside moiety. Thermal decomposition of alkyl
and aryl azides has been well documented [4, 5].
Fig. 1. Plot of the weight loss ( m) versus temperature
in the nonisothermal decomposition of AZT in air (sample
weight 10 mg, heating rate 4 deg/min).
1070-4280/03/3905-0727$25.00 2003 MAIK Nauka/Interperiodica

728
NEDEL’KO et al.
Figure 2 shows kinetic curves for thermal decom-
position of AZT. The process follows the first-order
kinetics, as indicated by the linear plots in the coor-
dinates ln(m
m) , where m is the total weight
loss in the first decomposition stage (20% of the
initial weight), and m is the weight loss by a time
moment . The rate constants k1 105 at 140, 158,
169, 183, 192, and 201 C are, respectively, 0.35, 2.9,
1
6.8, 28, 81, and 147 s . Manometric study of the first
decomposition stage in the temperature range from
160 to 201 C also revealed the first-order kinetics.
The rate constants k1 105 at 160, 170, 181, 191, and
1
201 C, are respectively, 3.6, 9.2, 21, 64, and 146 s .
Fig. 2. Kinetics of thermal decomposition of AZT in air
at (1) 140 C, (2) 158 C, (3) 169 C, (4) 183 C, (5) 192 C,
and (6) 201 C; thermogravimetry, sample weight 40 mg.
The rate constants for thermal decomposition of AZT,
determined by the thermogravimetric (k1) and mano-
metric methods (k1), are very similar. The above listed
values fit a common Arrhenius equation (Fig. 3):
1
k = 1014.7 0.5 exp[ (38100 900)/RT] (s );
r = 0.997, s = 0.063, n = 11.
In thermogravimetric experiments decomposition
of AZT occurs in an open reaction system, whereas
manometric experiments are carried out in a vessel
with a limited volume. Therefore, the similarity
between the rate constants obtained in the two series
of experiments suggests that there is no appreciable
effect of AZT decomposition products on the kinetics
of the process. On the other hand, comparison of the
thermogravimetric (in air) and manometric kinetic
data (in a vacuum) indicates once more that the reac-
tion under study is not accompanied by thermo-
oxidative processes.
By cooling the gaseous AZT thermolysis products
in the Bourdon gage from 25 C to 196 C (liquid
nitrogen) we found that the volume fraction of nitro-
gen among these products is 95.2%. According to the
mass spectrometric data, nitrogen is the major gaseous
product formed in the first stage of thermal decom-
position of AZT (190 C, 10 h); also, carbon dioxide
(no more than 3 vol. %), methanol, and furan were
detected: m/z (Irel, %): 68 (1.0) [CH CHOCH CH]+;
44 (3.9), 42 (0.3) [CO2]+; 40 (0.3), 39 (1.4), 32 (0.4),
31 (0.7), 29 (1.3), 28 (100) [N2]+, 18 (3.2) [H2O]+,
17 (1.2), 16 (0.5), 15 (0.3), 14 (5.2), 12 (0.2).
Fig. 3. Arrhenius plot for the first-order rate constant of
thermal decomposition of AZT according to (1) thermo-
gravimetric and (2) manometric measurements.
The rate of thermolysis increases on further heating,
presumably as a result of thermooxidative processes.
It seems to be important that the kinetic curve for
wieght loss of AZT almost does not change up to
250 C on replacement of air by argon (1 atm). This
means that the first decomposition stage is a purely
thermal process which is not accompanied by oxida-
tion. Under conditions of a dynamic vacuum (con-
tinuous evacuation of the reaction vessel) above
170 C, the contribution of sample evaporation
becomes appreciable. Taking the above stated into
account, the subsequent termogravimetric study of
AZT decomposition under isothermal conditions was
performed in air at temperatures above the melting
point.
The solid residue obtained after complete decom-
position of AZT in the temperature range 140 201 C
has a dark color. Its weight fraction is about 80%.
The yield of volatile products approaches 20%, i.e.,
it is almost twice as large as the theoretical weight
RUSSIAN JOURNAL OF ORGANIC CHEMISTRY Vol. 39 No. 5 2003

THERMAL DECOMPOSITION OF 3 -AZIDOTHYMIDINE
729
Scheme 1.
fraction of nitrogen (10.5%) provided that each azido
EXPERIMENTAL
group of AZT gives rise to one nitrogen molecule.
Thus the formation of nitrogen by decomposition of
the azido group is accompanied by degradation of the
molecular structure of AZT. Analysis of the IR spec-
trum of the solid residue obtained by decomposition
of AZT at 158 C (48 h) shows strengthening of hydro-
gen bonds involving hydroxy group (a shoulder at
Thermogravimetric experiments were performed
with the use of an ATV-14M automatic electronic
thermobalance which was designed and manufactured
at the Institute of Chemical Physics Problems, Rus-
sian Academy of Sciences [9]. This device allowed us
to perform kinetic measurements in a vacuum, in air,
and in an inert atmosphere (krypton, argon, 1 atm)
both under isothermal conditions and in the tempera-
ture-programming mode. A Bourdon tube pressure
gage was used in manometric measurements. The
residual pressure before decomposition did not exceed
0.1 torr. Gaseous and condensed thermolysis products
were identified by IR spectroscopy using a Specord
75IR spectrophotometer and mass spectrometry using
an MI-1201-V mass spectrometer (70 eV, 25 C).
Initial 3 -azidothymidine was synthesized by the
procedure described in [10] and was purified by
sublimation; it contained no less than 99.4% of the
main substance (HPLC; Perkin Elmer 200; UV de-
tector, 260 nm; 4.6 250 nm column packed with
Nucleosil C18; eluent acetonitrile 0.1 M potassium
dihydrogen phosphate).
1
3350 cm ). Treatment of the solid residue with water
and subsequent evaporation of the extract gave yel-
lowish crystals which became colorless on washing
with carbon tetrachloride. The colorless crystalline
product was identified as a pyrimidine base (thymine)
by TLC on silica gel (Silufol) using an authentic
sample. Their identity was also proved by the IR
spectra and melting test: no depression of the melting
point was observed on mixing with an authentic
sample (Aldrich, mp 316 C).
The activation parameters for thermal decomposi-
tion of AZT fall into a narrow range typical of decom-
position of aliphatic azides [4, 5]. This fact, as well
as the formation of nitrogen as the major gaseous
product, indicate that the rate-determining stage in
the thermolysis of AZT is decomposition of the azido
group. Presumably, next follows rupture of the glyco-
side bond to form thymine and products of decom-
position of the carbohydrate moiety (Scheme 1). All
these reactions constitute the first macroscopic stage
of the thermolysis process, and the kinetics of just
that stage have been studied in the present work. On
the other hand, as follows from Fig. 1, more profound
thermal transformations are possible at higher tem-
peratures.
This study was financially supported by the Rus-
sian Foundation for Basic Research (project no. 01-
03-32531) and by the Ministry of Education of the
Russian Federation (project no. T02-09.3-3222).
REFERENCES
1. Malin, A.A. and Ostrovskii, V.A., Russ. J. Org.
Chem., 2001, vol. 37, p. 759; Pathak, T., Chem. Rev.,
2002, vol. 102, p. 1623.
RUSSIAN JOURNAL OF ORGANIC CHEMISTRY Vol. 39 No. 5 2003

730
NEDEL’KO et al.
2. Malin, A.A., Shcherbinin, M.B., Poplavskii, V.S.,
Kononov, A.V., and Ostrovskii, V.A., Russ. J. Org.
Chem., 1997, vol. 33, p. 549.
7. Lopez, R., Gutierrez, M.D., Quijano, M., Nogu-
eras, M., and Sanchez, A., Thermochim. Acta, 1989,
vol. 153, p. 113.
3. Korchevskaya, E.V., Malin, A.A., Shcherbinin, M.B.,
and Ostrovskii, V.A., Russ. J. Org. Chem., 2000,
vol. 36, p. 1369.
8. Olafsson, P.G., Bryan, A.M., Davis, G.C., and Ander-
son, S., Can. J. Biochem., 1974, vol. 52, p. 997.
9. Gal’perin, L.N. and Kolesov, Yu.R., Izm. Tekhn.,
4. Geiseler, G. and Konig, W., Z. Phys. Chem. (Leipzig),
1981, no. 4, p. 23.
1964, vol. 227, p. 81.
10. Ostrovskii, V.A., Poplavskii, V.S., Shcherbi-
nin, M.B., Malin, A.A., Koshtaleva, T.M., Khokh-
ryakova, N.R., Bezklubnaya, E.V., and Kono-
nov, A.V., Russian Patent no. 2135512, 1997;
Byull. Izobret., 1999, no. 24.
5. Stepanov, R.S., Kruglyakova, L.A., and Buka, E.S.,
Kinet. Katal., 1986, vol. 27, p. 479.
6. Tanaka, Y., Takeda, N., and Fujita, D., Thermochim.
Acta, 1987, vol. 115, p. 281.
RUSSIAN JOURNAL OF ORGANIC CHEMISTRY Vol. 39 No. 5 2003