Thermal decomposition behavior of sodium arsenite in the presence of carbonized β-cyclodextrin

Article abstract of DOI:10.1021/jp900550d

The thermal decomposition behavior of sodium arsenite (SA) was greatly affected by carbonized β-cyclodextrin (CD), possessing the releases of four forms of gaseous arsenic molecules according to gas chromatography coupled to time-of-flight mass spectromet

Full text of DOI:10.1021/jp900550d

4998
2009, 113, 4998–5000
Published on Web 03/25/2009
Thermal Decomposition Behavior of Sodium Arsenite in the Presence of Carbonized
ꢀ-Cyclodextrin
Le Xin Song* and Zheng Dang
Department of Chemistry, UniVersity of Science and Technology of China, Jin Zhai Road 96, Hefei, China 230026
ReceiVed: January 19, 2009; ReVised Manuscript ReceiVed: February 28, 2009
The thermal decomposition behavior of sodium arsenite (SA) was greatly affected by carbonized ꢀ-cyclodextrin
(CD), possessing the releases of four forms of gaseous arsenic molecules according to gas chromatography
coupled to time-of-flight mass spectrometry with programmed temperature operation. Also, the decomposition
temperature of ꢀ-CD was lowered in the presence of SA.
Owing to their unique properties, cyclodextrins (CDs)
nowadays are very fascinating for supramolecular chemists in
various fields. ꢀ-CD contains seven R-D-glucose units and has
the form of a hollow truncate cone with one ring wider than
the other.1-3 It can form inclusion complexes or adducts with
many organic drugs,4,5 but there are few reports about the
interaction between ꢀ-CD and traditional inorganic drugs, such
as sodium arsenite (NaAsO2, SA), in solution, especially in
powder states. SA, an important anticancer drug, is widely
introduced into the treatment of a human acute promyelocytic
leukemia and some other cancers.6 Due to its high toxicity,
during the course of treatment, there are strong side effects on
human bodies.6 It is well-known that introduction of ꢀ-CD into
pharmaceutical processes on the molecular level can improve
the properties of included drugs evidently, so as to be applied
as a drug carrier.7,8 In our recent works, it was found that both
phase-change temperatures of organic guests and the thermal
decomposition behavior of ꢀ-CD changed upon inclusion.3,9 Can
inorganic drugs influence the thermal decomposition behavior
of ꢀ-CD? How does the existence of carbonized ꢀ-CD10 cause
this to occur sooner for decompositions of inorganic drugs with
a high thermal stability, if it does at all? Clearly, each of these
questions has a significant value both in chemical research and
in industrial application.
The molecule-ion interaction between ꢀ-CD and SA in
aqueous solution was proven by means of UV-vis absorption
spectra since the absorption intensity of SA varied continuously
with changing concentration of ꢀ-CD.11 The binding constant
(279.8 ( 7.5 mol-1 ·dm3) of ꢀ-CD to SA was determined using
the Benesi-Hildebrand equation12 in terms of 1H nuclear
magnetic resonance titration data.11
Four solid samples (1:1, initial molar ratio), including a
physical mixture 1, a ground mixture 2 (with a ground time of
20 min), a common adduct 3 (stirring for 10 h at 298 K and
dried into a powder in vacuo), and a hydrothermal adduct 4 (in
an autoclave, heated at 393 K for 4 h, and dried into a powder
in vacuo) were prepared. As a comparison, we also prepared
the adduct of R-CD and SA (1:1, initial molar ratio); the
preparation method of the adduct was the same as that of sample
4.11 Then, the solid samples were characterized by Fourier
transform infrared spectra (FTIR), powder X-ray diffraction
(PXRD), and thermogravimetry (TG) analyses.11
Initially, the stretching vibration band of OH groups due to
ꢀ-CD at about 3427 cm-1 in the FTIR spectra of the samples
2, 3, and 4 was found to red shift to a less extent compared to
that of OH groups of free ꢀ-CD, and the shift extent increased
in the order of 1 < 2 < 3 < 4.11 This suggest that the hydrogen
bondings between ꢀ-CD molecules were weakened upon
interaction with SA.13 Similarly, with the change in the FTIR
spectrum of ꢀ-CD before and after adduct, the stretching
vibration band of OH groups of R-CD also indicated a small
red shift (about 16 cm-1) upon adduct with SA.11
Next, PXRD analyses showed that the existence of SA has
affected the structural arrangement of ꢀ-CD molecules to a
different extent dependent on the preparation methods of the
samples. For example, several main peaks due to ꢀ-CD in 4
shifted toward a lower 2θ angle in comparison with those in 1
and 2, meaning that the introduction of SA has led to an increase
of the intermolecular distance of ꢀ-CD.11 The increase can be
explained as the effect of the introduction of SA into the
openings of the ꢀ-CD cavity. This observation is in good
accordance with the result of FTIR. In addition, the molecule-ion
interaction between ꢀ-CD and SA in the powder state was
demonstrated by the fact that there were two new peaks at the
2θ values of 28.0 and 34.1° in the spectra of 3 and 4 which did
not appear in the PXRD spectra of free ꢀ-CD, SA, 1 and 2.11
Also, according to the comparison of PXRD patterns between
R-CD and its adduct, the stacking form of R-CD has been
changed by the presence of SA in the adduct, from a cage form
to a channel form. Nevertheless, the significant difference in
stacking forms was not found between ꢀ-CD and its adduct of
SA, suggesting that the molecule-ion interaction between SA
and R-CD might be stronger than that between SA and ꢀ-CD.11
Subsequently, TG profiles (Figure 1) of samples 2, 3, and 4
indicated that the rapid release temperature (349 K) of water in
samples 2 and 3 was highly similar to that of water in free
ꢀ-CD,3 but the adduct 4 lost its water molecules at a higher
temperature of about 367 K (∆T1 ) 17 K). The phenomena
implied that there existed different distribution forms of water
in the two adducts, which was also found in some inclusion
* To whom correspondence should be addressed. E-mail: solexin@
custc.edu.cn. Fax: +86 551 3601592. Tel: +86 551 3601804.
10.1021/jp900550d CCC: $40.75
2009 American Chemical Society

Letters
J. Phys. Chem. B, Vol. 113, No. 15, 2009 4999
Figure 3. Mass spectra of 4 at 31.35 (A) and 35.62 min (B).
Figure 1. TG profiles of samples 2, 3, and 4 at the heating rate of 5.0
K·min-1
.
Figure 4. MID curve of the fragment at m/z 299.685.
Figure 2. TIC curves of free ꢀ-CD (A) and adduct 4 (B).
compared to those of ꢀ-CD inclusion complexes formed by
organic amines reported before.9 Besides, there were two new
peaks at 31.52 and 35.74 min in Figure 2B, corresponding to
the temperature at 773 K, which was much higher than the
decomposition temperature of ꢀ-CD but still much lower than
the decomposition temperature (mp, 888 K) of SA. This finding
revealed that a redox reaction occurred between carbonized
ꢀ-CD and SA.
complexes of CDs.14 Further, in the temperature range from 406
to 509 K, both free ꢀ-CD and the ground mixture 2 did not
lose any mass, but the two adducts 3 and 4 lost their respective
masses of about 2%. Moreover, the decomposition behavior of
ꢀ-CD in the three samples is rather different from each other.
After 613 K, the residue mass (RM) of the remains increased
in the order of 2 < 3 < 4 at any temperature, and the three curves
in the range from 613 to 773 K exhibited a larger slope than
free ꢀ-CD. This may be caused by the different decomposition
modes of ꢀ-CD.
The analyses of Figure 1 gave a strong impression that the
presence of SA influenced the thermal decomposition behavior
of ꢀ-CD and that there was a large difference in decomposition
behavior between ꢀ-CD and 4. Therefore, it is necessary to
further evaluate the direct effect of the molecule-ion interaction
on the thermal decomposition of the two reactants.
Gas chromatography coupled to time-of-flight mass spec-
trometry (GC-TOF-MS) is a soft ionization technique, which
therefore allows for the transfer of noncovalent adducts from a
liquid phase into a gaseous phase.9,15 GC-TOF-MS measure-
ments of free ꢀ-CD and 4 with the same programmed temper-
ature were carried out, with the objective of trying to find details
regarding the decomposition process of SA.
Initially, the plots of the variation of total ion current (TIC)
in intensity with the heating time of free ꢀ-CD and the adduct
4 are shown in panels A and B, repectively, of Figure 2. The
decomposition processes of the samples were controlled by the
temperature program as described in Supporting Information.3
According to Figure 2A, in the temperature range from 373 to
573 K, free ꢀ-CD did not lose any mass, but from Figure 2B,
4 began to release some of its mass at 17.52 min, corresponding
to a small peak at 553 K. This finding indicated that in the
presence of SA, some of the ꢀ-CD molecules seemed to have
a lower decomposition temperature. It can be explained that
some of the covalent bonds of the released ꢀ-CD molecules
were weakened by the molecule-ion interaction between them
and SA.
Next, as can be seen in Figure 3, on the one hand, the mass
spectra of 4 at 31.35 (773 K) and 35.62 min (kept about 6 min
at 773 K) were rather similar to each other because both of
them were involved in the four fragments at m/z 74.923 (As+),
149.844 (As2+), 224.765 (As3+), and 299.685 (As4+).16 These
values reflected that the liberation of arsenic from SA to the
gaseous phase was presented in four kinds of forms, and the
relative abundance (RA) ratios of the four simple substances
are 1.0:2.2:1.2:20 at 31.35 min and 1.0:3.1:1.2:19 at 35.62 min.
+
+
The stabilities of them decreased in the order As4 > As2
>
As3+ > As+. On the other hand, differently from Figure 3B, the
fragment at m/z 18.011 in Figure 3A was ascribed to H2O+. It
is interesting that it appeared at 31.35 min but disappeared at
35.62 min, showing that the carbonization process of ꢀ-CD is
finished before 35.62 min. Additionally, the fragment at m/z
43.990 (CO2+) had the highest RA of 100% in the two spectra.
Clearly, its occurrence at 35.62 min can be attributed to the
result of the redox reaction between SA and carbonized ꢀ-CD.
Theoretically, tetratomic arsenic vapor has the highest sym-
metry and stability among the four fragments. Actually, the
liberation of arsenic from SA was mainly in the form of As4,
with As2 next. As and As3 were unstable because of the existence
of an unpaired electron. A proposed chemical reaction describing
the decomposition behavior of SA in the presence of carbonized
ꢀ-CD was given in eq 1 as follows
4NaAsO2 + 3C f 2Na2CO3 + As4 + CO2
(1)
To investigate the decomposition of SA in detail, the
formation of As4 was traced based on the multiple ion detection
(MID, Figure 4) curve of the fragment at m/z 299.685, which
is referred to as time-varying values. Interestingly, the curve in
Figure 4 could be divided into three stages clearly. The first
was at 25.57 min, corresponding to a relatively low temperature
Further, free ꢀ-CD produced a double peak at a range of about
25-30 min, but the main peak of 4 was a single one at 25.13
min. It should be noted that the signal of 4 rose a little earlier

5000 J. Phys. Chem. B, Vol. 113, No. 15, 2009
Letters
ꢀ-CD and SA, and the comparison experiments of R-CD also
supported the viewpoint. The observation provided important
insight into the distinction between an adduct and an inclusion
complex. It can be expected that the advanced decomposition
behavior of SA in the presence of ꢀ-CD as well as the
unpredictable decomposition of adducted ꢀ-CD in the presence
of SA has important value in the application of CD in industrial
production.
Figure 5. A schematic sketch illustrating the proposed molecule-ion
interactions between SA and ꢀ-CD.
Acknowledgment. We would like to acknowledge the
assistance of H. Yin in the GC-TOF-MS measurements of this
investigation and useful discussions.
of 653 K, and just at the end of the sharp decomposition of
ꢀ-CD. The fastest liberation process of As4 to the gaseous phase
was at 31.49 min, with the temperature just reaching 773 K. It
should be noted that all of these release temperatures were below
the melting point of SA as well as the sublimation temperature
(886 K) of arsenic, implying that without the decomposition
and carbonization of ꢀ-CD, arsenic would not be released from
the sample 4.
Supporting Information Available: UV-vis absorption
spectra and 1H nuclear magnetic resonance titration data for the
system of ꢀ-CD and SA in solution, as well as the preparation,
FTIR spectra, and PXRD data of solid adducts 3 and 4. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Furthermore, according to eq 1, it was reasonable that the
thermal decomposition of sodium arsenite and the release of
arsenic would be observed in the presence of not only CDs such
as R- and ꢀ-CD but also many other oligosaccharides. The
release process of arsenic should be associated with the
molecule-ion interactions occurring between SA and ꢀ-CD,
which was in agreement with the result of TG measurements.
The interactions between inorganic guests and ꢀ-CD were
considerably weak in solution17,18 and in the solid state.19 As
shown in Figure 5 and in the formation process of the adduct
4, the effects of the interactions on electrostatic forces between
References and Notes
(1) Rekharsky, M. V.; Inoue, Y. Chem. ReV. 1998, 98, 1875–1917.
(2) Xu, P.; Song, L. X. Acta Phys. Chim. Sin. 2008, 24, 729–736.
(3) Song, L. X.; Wang, H. M.; Guo, X. Q.; Bai, L. J. Org. Chem.
2008, 73, 8305–8316.
(4) Lysik, M. A.; Wu, P. S. J. Pharm. Sci. 2003, 92, 1559–1573.
(5) Marquesa, J.; Anjoa, L.; Marquesb, M. P. M.; Santosa, T. M.;
Almeida Paza, F. A.; Braga, S. S. J. Organomet. Chem. 2008, 693, 3021–
3028.
(6) Niu, C.; Yan, H.; Yu, T.; Sun, H. P.; Liu, J. X.; Li, X. S.; Wu, W.;
Zhang, F. Q.; Chen, Y.; Zhou, L.; Li, J. M.; Zeng, X. Y.; Ou Yang, R. R.;
Yuan, M. M.; Ren, M. Y.; Gu, F. Y.; Cao, Q.; Gu, B. W.; Su, X. Y.; Chen,
G. Q.; Xiong, S. M.; Zhang, T. d.; Waxman, S.; Wang, Z. Y.; Chen, Z.;
Hu, J.; Shen, Z. X.; Chen, S. J. Blood 1999, 94, 3315–3324.
(7) Davis, M. E.; Brewster, M. E. Nat. ReV. Drug DiscoVery 2004, 3,
1023–1035.
(8) Machut-Binkowski, C.; Hapiot, F.; Cecchelli, R.; Martin, P.;
Monflier, E. J. Inclusion Phenom. Macrocyclic Chem. 2007, 57, 567–572.
(9) Song, L. X.; Xu, P. J. Phys. Chem. A 2008, 112, 11341–11348.
(10) Denicourt-Nowicki, A.; Roucoux, A.; Wyrwalski, F.; Kania, N.;
Monflier, E.; Ponchel, A. Chem.sEur. J. 2008, 14, 8090–8093.
(11) See Supporting Information.
(12) Yoshikiyo, K.; Matsui, Y.; Yamamoto, T.; Okabe, Y. Bull. Chem.
Soc. Jpn. 2007, 80, 1124–1128.
(13) Magyar, A.; Szendi, Z.; Kiss, J. T.; linko, I. P. J. Mol. Struct.:
THEOCHEM 2003, 666, 163–168.
(14) Petrovski, Z.; de Matos, M. R. P. N.; Braga, S. S.; Pereira, C. C. L.;
Matos, M. L.; Goncalves, I. S.; Pillinger, M.; Alves, P. M.; Romao, C. C.
J. Organomet. Chem. 2008, 693, 675–684.
(15) Xu, P.; Song, L. X.; Wang, H. M. Thermochim. Acta 2008, 469,
36–42.
Na+ and AsO2 were different from one another. We thought
-
that it was the different interactions in the adduct of ꢀ-CD that
resulted in different environments around SA20 as well as ꢀ-CD
molecules, which is considered the fundamental cause of the
difference between a supramolecular inclusion complex and a
molecule-ion adduct.
GC-TOF-MS measurements of the adduct of R-CD and SA
were performed as comparison experiments, and the release of
arsenic in four forms was also observed. However, the release
temperature of the arsenic forms was greatly changed when
compared with those of sample 4. The largest release of As4 to
a gaseous phase was delayed to 36.45 min. The result is ascribed
to different interaction intensities between the adducts of SA
with R-CD and ꢀ-CD.
Conclusions
The molecule-ion interaction between ꢀ-CD and SA in
aqueous solution was proven by UV-vis absorption spectra and
1H nuclear magnetic resonance titration. The adduct interaction
was also characterized by the aid of FTIR and PXRD in the
solid state. TG, especially TOF-GC-MS, confirmed that the ther-
mal behaviors of ꢀ-CD and SA were affected by the formation
of the adduct between them. This effect could be ascribed to
different molecule-ion interactions existing in the adduct of
(16) Lynch, D. C. Metall. Trans. B 1980, 11B, 623–629.
(17) Matsui, Y.; Ono, M.; Tokunaga, S. Bull. Chem. Soc. Jpn. 1997,
70, 535–541.
(18) Yamashoji, Y.; Fujiwara, M.; Tanaka, M. Chem. Lett. 1993, 1029–
1032.
(19) Chierotti, M. R.; Gobetto, R. Chem. Commun. 2008, 1621–1634.
(20) Yan, X. L.; Chen, T. B.; Liao, X. Y.; Huang, Z. C.; Pan, J. R.; Hu,
T. D.; Nie, C. J.; Xie, H. EnViron. Sci. Technol. 2008, 42, 1479–1484.
JP900550D