Thermogravimetric investigations on the mechanism of decomposition of Pb compounds on a tungsten surface

Article abstract of DOI:10.1016/S0040-6031(00)00578-5

Tungsten coils are used as electrothermal atomizers in atomic absorption spectrometry (ETA-W), but despite successful applications the atomization mechanism is still obscure. The use of the thermogravimetry (TG) can bring some information on decomposition pathways for atom formation, simulating the drying and pyrolysis steps employed in typical heating program cycles. Thermogravimetric curves of Pb salts with and without chemical modifiers, such as ascorbic, oxalic, tartaric acids, and glucose, or interferents (e.g. NaNO₃) were obtained using a lab-made tungsten sample holder in an 90% Ar plus 10% H₂ atmosphere usually employed in ETA-W. Intermediaries and residues of the pyrolysis step were characterized by X-ray diffraction. These data allows a better establishment of parameters for ETA-W and can help to find proper conditions to decrease interferences. It was showed that the general decomposition route is the formation of PbO by heating, oxide reduction by H₂ in the purge gas followed by volatilization of Pb. (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S0040-6031(00)00578-5

Thermochimica Acta 362 (2000) 161±168
Thermogravimetric investigations on the mechanism of
decomposition of Pb compounds on a tungsten surface
Â
Daniele M. Santos, Pedro O. Luccas, Joaquim A. Nobrega,
*
Â
Eder T.G. Cavalheiro
Â
Departamento de Quõmica, Universidade Federal de SaÄo Carlos, Caixa Postal 676,
SaÄo Carlos, SP, 13560-970, Brazil
Received 3 March 2000; accepted 29 June 2000
Abstract
Tungsten coils are used as electrothermal atomizers in atomic absorption spectrometry (ETA-W), but despite successful
applications the atomization mechanism is still obscure. The use of the thermogravimetry (TG) can bring some information on
decomposition pathways for atom formation, simulating the drying and pyrolysis steps employed in typical heating program
cycles. Thermogravimetric curves of Pb salts with and without chemical modi®ers, such as ascorbic, oxalic, tartaric acids, and
glucose, or interferents (e.g. NaNO3) were obtained using a lab-made tungsten sample holder in an 90% Ar plus 10% H2
atmosphere usually employed in ETA-W. Intermediaries and residues of the pyrolysis step were characterized by X-ray
diffraction. These data allows a better establishment of parameters for ETA-W and can help to ®nd proper conditions to
decrease interferences. It was showed that the general decomposition route is the formation of PbO by heating, oxide reduction
by H2 in the purge gas followed by volatilization of Pb. # 2000 Elsevier Science B.V. All rights reserved.
Keywords: Thermogravimetry; Lead; Atomization pathways; Tungsten surface
1. Introduction
owing to its design that allows reaching isothermal
furnace conditions. Open atomizers, such as metallic
coils and strips, also attained excellent sensitivities for
most metals [2,3] however due to its non-isothermal
ambient the occurrence of interferences is a critical
issue that limits their applicability [4].
The tungsten coil electrothermal atomizer (ETA-W)
was proposed by Lund and Larsen [5,6] and after a
long period without large use it was reinvestigated by
Berndt and Schaldach [7] using 150 W tungsten coil of
projector lamps. Subsequently, the ETA-W was
mainly used in studies dealing with the determination
of trace metals, e.g. Cd in biological materials [8], Cr
in hair and urine [9], Pb in hair and paints [10], and Pb
in blood [11,12]. In all these studies the quanti®cation
Atomic absorption spectrometry with electrother-
mal atomizers is one of the best techniques for deter-
mination of trace levels of metals in complex samples,
such as body ¯uids and tissues. The most conventional
atomizer is the graphite tube, that after a critical stage
in the 1970s attained maturity in the 1980s with the
proposal of the stabilized temperature platform fur-
nace concept [1]. In addition to its superb sensitivity,
this closed atomizer is virtually interference free
* Corresponding author. Tel.: ‡55-16-260-8209;
fax: ‡55‡16-260-8350.
Â
E-mail address: cavalheiro@dq.ufscar.br (E.T.G. Cavalheiro).
0040-6031/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 6 0 3 1 ( 0 0 ) 0 0 5 7 8 - 5

162
D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
was performed by matrix matching or standard addi-
tions method. The use of reference solutions prepared
in acid medium was unsuitable owing to the severe
occurrence of matrix interference processes. A separa-
tion procedure was also proposed for removing inter-
ferences before determination of Pb in residual waters
[13]. More recently an on-line electrodeposition pro-
cedure was implemented to remove interferences
before Pb determination in waters [14].
up to 9008C in a Ar/H2 dynamic atmosphere and
the characterization of the solid residues by powder
X-ray diffraction generated useful data to a better
comprehension of the processes occurring during
dry and pyrolysis steps in the ETA-W.
2. Experimental
2.1. Reagents
Despite all the practical analytical dif®culties for
obtaining accurate results, the ETA-W is evolving and
it may have a market as a dedicated analytical instru-
ment. The research developed by Jones and coworkers
[15,16] is an important route towards the development
of a practical instrument arrangement.
Analytical grade Pb(NO3)2 was used after recrys-
tallization. The salts PbCl2 and PbSO4 were prepared
by reacting Pb(NO3)2 with HCl and H2SO4, respec-
tively. All the salts were dried for 4 h at 808C in a
vacuum oven.
Mixtures of Pb(NO3)2 and organic compounds
(tartaric, oxalic and ascorbic acids, and glucose) were
also prepared by dissolving them in 1:2 mol ratio
(inorganic:organic) in water. To obtain a solid mixture
containing both compounds the solvent was elimi-
nated by heating. The solid formed was dried in an
oven at 1208C, except the glucose mixture which was
dried at 608C under reduced pressure. A mixture
containing NaNO3 plus Pb(NO3)2 was similarly
prepared.
To complement the development of the instrumen-
tation, it is need to investigate the mechanism of
atomization to try to understand the interference pro-
cesses. The studies about mechanisms occurring on
metallic atomizers are scarce, and the few researches
described dealt with tungsten tube atomizers [17±21].
Nevertheless, despite the use of the same metal, this
closed atomizer is completely different of the ETA-W.
The tungsten tube atomizer is an isothermal system as
demonstrated by Chakrabarti et al. [22]. However,
both tungsten atomizers are critically dependent on
the presence of hydrogen in the purge gas to promote
reduction of oxides involving the analyte and to
protect the metallic surface from fast oxidation and
destruction.
The mechanism of action of chemical modi®ers in
the ETA-W is also practically unknown, despite some
advantages coming from its use, such as extended coil
lifetime and better repeatibility [23,24]. In the work
related to the atomization of Pb, Bruhn et al. [23]
demonstrated that interferences caused by Na are
condensed phase processes. This is in opposition to
the expected since the non-isothermal ambient should
aggravate gas phase recombination and condensation
processes. Thus, considering that condensed phase
processes are operative for Pb, the use of thermal
analysis could bring a better understanding of them.
Alvarado used thermogravimetry to investigate
mechanisms of atomization of Fe, Co, and Ni in a
graphite tube atomizer [25].
2.2. Instrumentation
The thermogravimetric (TG) and derivative thermo-
gravimetric (DTG) curves were recorded on a TG-
2050 thermogravimetric module coupled to a 2100
thermal analyser, both from TA Instruments under a
gas ¯ow of 90 ml min 1 (90% Ar‡10% H2 mixture v/
1
v) in a lab-made tungsten crucible at 108C min
heating rate and using samples of about 5 mg of each
compound. The sample holder was prepared by fold-
ing a piece of 10Â15 mm of a tungsten foil (tungsten
foil 99.9% of purity, 0.05 mm thick, Aldrich). The
folded foil was then perforated at each extremity and a
Ni±Cr wire was used as a hanging loop. The ®nal
masses of such sample holders was around 300 mg.
The ®nal temperature was 9008C for all the TG
experiments.
X-ray diffractograms (XRD) were obtained in the
D5000 Siemens diffractometer. The XRD powder
diffractograms were recorded using the CuKa (radia-
tion at 2y angle range from 25 to 75).
The present work evaluated the thermal decompo-
sition of Pb2‡ salts alone or in media containing
chemical modi®ers. The thermogravimetric curves

D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
163
3. Results and discussions
ETA-W experiments. On the one hand, the typical
heating rates of TG is signi®cantly lower than those
employed in ETA-W even during dry and pyrolysis
steps. On the other hand, the sample mass is appre-
ciably lower in this latter technique, for example a
The electrothermal heating conditions were simu-
lated in the thermogravimetric experiment employing
the lab-made tungsten sample holder under an Ar‡H2
atmosphere. This gaseous mixture is usually em-
ployed as purge gas in ETA-W for extending the coil
lifetime and for increasing ef®ciency of the atom
formation.
1
sample volume of 10 ml for a lower than 10 mg l
solution, i.e. an analyte mass of 0.1 ng. Therefore, the
results must be carefully interpreted to avoid erro-
neous inferences.
The choice of the Pb salts was made considering the
acids commonly used for sample preparation, i.e.
H2SO4, HNO3 and HCl. The presence of organic
compounds as concomitants or chemical modi®ers
is also usual [26,27].
3.1. Decomposition of Pb(NO3)2, PbSO4, and
PbCl2
The thermal behavior of these inorganic compounds
in the tungsten surface and under N2 atmosphere was
studied for comparison purposes.
Table 1 summarizes the mass losses, temperature
ranges, and thermal events observed in each step of
TG curves. The TG curves for Pb compounds are
presented in the Fig. 1 under Ar/H2. The products
formed in each step are proposed on basis of their X-
ray diffraction patterns characterization.
The thermal decomposition of Pb(NO3)2 was simi-
lar under both N2 or Ar/H2 atmospheres. A crepitation
process was observed in both cases between 180 and
3808C, previously to the decomposition. Data
obtained by XRD for the solid residues formed at
5328C, con®rmed the formation of PbO and W5O14 in
the sample holder surface. The ®nal residue obtained
at 6338C was characterized by XRD as Pb0.
The application of heating rates varying from 5 to
1
408C min led to similar decomposition pro®les for
PbC2O4, with minor variations of the initial and ®nal
decomposition temperatures. The heating rate of
1
108C min
trying to reach the best resolution for overlapping
was selected for all TG experiments
thermal events.
The effect of sample mass was investigated in the
1±20 mg range. A distortion was observed in the TG
curves with 1 mg sample due to the compensation
system of the thermobalance. Sample masses of
3±5 mg showed suitable behavior and the last one
was used in all TG runs, in order to obtain enough
sensitivity in low mass changes.
Both heating rate and sample mass parameters
require careful analysis when comparing TG and
Table 1
Mass losses and temperature range corresponding to the decomposition of the Pb compounds
Process
TG data temperature interval (8C)
Mass loss or residue (%)
TG
Calculated
90% Ar‡10% H2 mixture
PbCl2(s)!PbCl2(g)‡H2(g)!Pb(g)‡2HCl(g)
Pb(g)!Pb(l)d
e
442±558
52.2a
28.3a
±
±
Starting at 710
Pb(NO3)2!PbOd
191±512
66.4b
63.3b
67.4
62.6
±
PbO‡H2(g)!Pb0 (l)‡H2O
512±671
Pb0 (l)!Pb0 (g)c
Starting at 700
PbSO4!PbOd
594±753
27.7a
23.4a
±
±
PbO‡H2(g)!Pb0 (l)!Pb0 (g)c
Starting at 753
a Relative to mass loss.
b Relative to residue.
c Vapor pressure.
d Characterized by XRD.
e Not calculated, mixture.

164
D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
[29]. Finally, the Pb0 volatilizes above 7508C. The
process is summarized by the following equations
PbCl2ꢀs† ! PbCl2ꢀg†
(1)
PbCl2ꢀg† ‡ H2ꢀg† ! Pb0ꢀg† ‡ 2HCl
(2)
(3)
(4)
550ꢁC
Pb0ꢀg† ! Pb0ꢀl†
>750ꢁC
Pb0ꢀl† ! Pb0ꢀg†
The decomposition of PbSO4 under N2 showed that
the salt decomposes only above 8008C. On the other
hand, under Ar/H2, two steps were observed. The ®rst
one was attributed to the formation of PbO (594±
7538C) and the second one is related to the oxide
reduction and formation of Pb0.
The anticipation of PbSO4 decomposition and the
reduction of the PbCl2 evidenced the in¯uence of the
reductive character of the atmosphere and its effects
on ETA-Watomization processes. These processes are
important to understand the atom formation in ETA-
W. As previously mentioned, this is a non-isothermal
atomizer and consequently the formation of atoms is
dependent on thermochemical processes involving
compounds generated during dry and pyrolysis steps.
Contrary to the graphite tube atomizer, processes
promoted by thermal dissociation in the gas phase
are less operative due to the low temperature. There-
fore, the atom formation processes seem to be caused
by H2 related reactions. In this sense the purge gas
could be named as the atomization gas in addition to
H2 action to increase tungsten coil lifetime.
Fig. 1. TG (solid)/DTG (dot) curves obtained for some Pb salts on
the W crucible, under a dynamic Ar/H2 atmosphere (gas ¯ow of
90 ml min 1, sample mass around 5 mg, 108C min 1 heating rate).
3.2. Decomposition of Pb(NO3)2 in the presence of
the organic modi®ers
Recently, it was studied the determination of
Pb in urine. Alkaline and alkaline-earth elements
caused a negative interference on Pb atomization.
The authors also investigated the effect of ascorbic,
oxalic, and tartaric acids as organic modi®ers. All
these compounds can be a source of C and reductive
gases as CO formed during their decomposition. Such
conditions may contribute to the oxides reduction and
formation of Pb0 in the condensed or gaseous phase
[30].
Lead chloride was completely removed by volati-
lization in a N2 atmosphere. However, under Ar/H2 the
decomposition occurred in two steps. According to
XRD data, at the end of the second step (5588C) the
sample holder was totally recovered by Pb0. In addi-
tion the amount of experimental residue was smaller
than that calculated at 5588C for metallic lead. It is
known that in a graphite tube the PbCl2 can react with
H2 in the gas phase generating Pb0 and HCl [28]. We
proposed that in ETA-W the PbCl2 is partially
removed by the purge gas, while some amount is
reduced to Pb0 (g) which condenses on the tungsten
sample holder surface since Pb0 is liquid at 327.48C
The TG curves for Pb mixtures are presented in
the Fig. 2 under Ar/H2. The mass losses, temperature
ranges and ®nal decomposition products, observed in

D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
165
Lead nitrate and each organic modi®er were mixed
in solution and the dry step was simulated out of the
thermobalance, in order to prevent mass losses by
crepitation. As expected, the presence of the organic
compounds induced the reduction of PbO to Pb0 at
lower temperatures than observed for Pb(NO3)2 alone.
The temperatures in which Pb0 is formed in the
tungsten crucible surface in presence of each modi®er
are presented in Table 3. These attributions re¯ect a
combination of data obtained by TG and XRD. The
results are in agreement with ETA-W data showing the
increment of sensitivity in media containing organic
modi®ers evidencing the importance of the presence
of such compounds in the condensed phase, to pro-
mote the formation of the atomic cloud. Once again,
these processes con®rmed that thermochemical pro-
cesses must be understood to improve the ef®ciency of
atomization in the ETA-W. However, the presence of a
high content of C cannot be fully exploited analyti-
cally to increment sensitivity owing to the reduction of
the tungsten coil lifetime caused by the formation of
brittle and less refractories tungsten carbides.
The TG curve pro®le for the thermal decomposition
of PbC2O4 is similar to the TG decomposition curve of
the mixture Pb(NO3)2‡H2C2O4. The formation of
PbC2O4 could be proposed on the basis of its Ks of
12
8.3Â10
[29] and the mixture preparation proce-
dure. After decomposition of the excess of organic
acid (25±1608C), the mixture showed only one mass
loss (320±4108C), resulting in PbO/Pb at 4358C
according to the XRD data.
3.3. Decomposition of Pb(NO3)2 in the presence of
NaNO3
Sodium is a severe interference in the Pb determi-
nation using ETA-W [23]. In the Fig. 3, TG and DTG
curves are presented for NaNO3 and its mixture.
In a Pt sample holder the decomposition of NaNO3
was expected to occur via a thermal decomposition
process (8008C) such as [31]:
Fig. 2. TG curves for mixtures of Pb(NO3)2 and organic modi®ers
(solid) and for the pure modi®ers (dot), on the W crucible, under a
dynamic Ar/H2 atmosphere (gas ¯ow of 90 ml min 1, sample mass
4NaNO3 ! 2Na2O ‡ 2N2 ‡ 5O2
(5)
1
However, on a W sample holder the TG residue at
4288C is 65.4% while the calculated for the Na2O is
36.5%.
The XRD results for the sample holder surface
showed the presence of Na2O and W2O3 and W2O5
around 5 mg, 108C min heating rate).
each step of TG curves, are presented in Table 2. The
products formed in each step are proposed on basis of
their XRD patterns characterization.

166
D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
Table 2
Mass losses and temperature range corresponding to the decomposition of the Pb mixtures
Process
TG data temperature interval (8C)
Mass loss or residue (%)
TG
Calculated
90% Ar‡10% H2 mixture
H2C2O4Á2H2O!H2C2O4‡2H2O
27.0±92.4
92.4±168
29.4a
No residue
28.6a
f
H2C2O4!total decomposition
±
Pb(NO3)2‡oxalic acid mixture
H2C2O4Á2H2O!H2C2O4‡2H2O
H2C2O4 decomposition
32.4±60.7
81.5±146
3.33a
11.2a
23.8a
50.2b
±
±
±
±
PbC2O4!PbOd‡CO‡CO2
322±410
PbO*‡H2!Pb0 (l)d‡H2O!Pb(g)c
Starting at 700
Ascorbic acid!3H2O
170±221
221±800
29.7a
46.2a
30.7
±
Ascorbic acid decomposition
Pb(NO3)2‡ascorbic acid mixture
Ascorbic acid!3H2O
120±222
222±604
34.9a
29.8a
29.4b
±
±
±
Pb(NO3)2‡ascorbic acid!PbOd‡Pb(l)d
Pb0 (l)!Pb(g)c
Starting at 604
Tartaric acid!CO2‡CO‡H2Oe
151±552
1.2b
±
Pb(NO3)2‡tartaric acid!Pb0 (l)d
112±530
79.5a
12.0b
±
±
Pb0 (l)!Pb(g)c
Starting at 715
Glucose!2H2O
146±233
14.5a
10.6b
49.6a
5.91a
30.1b
16.7
Glucose decomposition
Pb(NO3)2‡glucose!PbOd‡Pb0
PbO!Pb0 (l)d
233±730
±
±
±
±
121±416
416±601
Starting at 700
Pb0 (l)!Pb (g)c
NaNO3!Na2O‡tungstates
298±397
31.0a
±
d
d
d
Pb(NO3)2‡NaNO3!Na2WO4 ‡Na2W2O7 ‡PbWO4
327±471
545±900
29.6a
52.3b
±
±
d
d
Na2WO4‡Na2W2O7‡PbWO4!Na2WO4 ‡NaWO3
a Relative to mass loss.
b Relative to residue.
c Vapor pressure.
d Characterized by XRD.
e Ref. [33].
f Not calculated, mixture.
even in the reductive atmosphere (5008C). The pre-
sence of the tungsten oxydes explains the difference.
In the presence of NaNO3 the temperature on which
decomposition of Pb(NO3)2 begins shifted to a higher
value (3278C), in comparison with the salt alone
(1908C). X-ray diffraction data showed the presence
of Na2W2O7, Na2WO4 and the PbWO4, on the
tungsten sample holder surface. These compounds
exerted a signi®cant in¯uence on the Pb(NO3)2 de-
composition since these W oxides are reduced only
above 9008C.
Table 3
Effects of organic compounds on the temperature of Pb0 formation
According to Mellor [32]:
Modifier
T (8C)
Tartaric acid
Ascorbic acid
Glucose
527
600
650
700
900ꢁC
Na2WO4 ‡ 4H2 ! 2Na ‡ W ‡ 4H2O
(6)
(7)
1100ꢁC
Oxalic acid
Na2WO4 ‡ 3H2 ! W ‡ 2NaOH ‡ 2H2O

D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
167
to overcome interferences still need further experi-
ments. It should be mentioned that contrary on the
paradigms of the electrothermal atomization based on
STPF conditions, the interference processes studied
are mainly related to condensed phase reactions that
generate compounds that cannot be easily thermally
decomposed by the low gas phase temperature.
Probably the open W coil atomizer should be modi®ed
to create a hotter gas medium with enough thermal
energy to promote thermal decomposition of complex
compounds originated during dry and pyrolysis
steps.
References
[1] W. Slavin, D.C. Manning, G.R. Carnrick, Atom. Spectrom. 2
(1981) 137.
[2] M. Suzuki, K. Ohta, Prog. Analyt. Atom. Chem. 6 (1983)
49.
Â
[3] J.A. Nobrega, M.M. Silva, P.V. Oliveira, F.J. Krug, N. Baccan,
Â
Quõm. Nova 18 (1995) 555.
[4] D.J. Butcher, J. Sneddon, A Practical Guide to Graphite
Furnace Atomic Absorption Spectrometry, Wiley, New York,
1998.
Fig. 3. TG (solid)/DTG (dot) curves obtained for mixture of
Pb(NO3)2‡NaNO3 and for NaNO3, on the W crucible, under a
dynamic Ar/H2 atmosphere (gas ¯ow of 90 ml min 1, sample mass
[5] W. Lund, B.V. Larsen, Anal. Chim. Acta 70 (1974) 311.
[6] W. Lund, B.V. Larsen, Anal. Chim. Acta 72 (1974) 57.
[7] H. Berndt, G. Schaldach, J. Anal. At. Spectrom. 3 (1988)
709.
1
around 5 mg, 108C min heating rate).
 Â
[8] M.F. Gine, F.J. Krug, V.A. Sass, B.F. Reis, J.A. Nobrega, H.
Berndt, J. Anal. At. Spectrom. 8 (1993) 246.
These oxides are refractaries and additionally the W
oxides compete with Pb for H2 in the atmosphere and
reductor compounds in the condensed phase, retarding
the Pb atomization.
Â
[9] C.G. Bruhn, J.Y. Neira, M.I. Guzman, M.M. Darder, J.A.
Â
Nobrega, Fresenius J. Anal. Chem. 364 (1999) 273.
[10] C.L. Sanford, S.E. Thomas, B.T. Jones, Appl. Spectrosc. 50
(1996) 174.
Â
[11] F.J. Krug, M.M. Silva, P.V. Oliveira, J.A. Nobrega, Spectro-
chim. Acta 50B (1995) 1469.
[12] P.J. Parsons, H. Qiao, K.M. Aldous, E. Mills, W. Slavin,
Spectrochim. Acta 50B (1995) 1475.
4. Conclusion
The experimental data obtained combining TG and
XRD were the ®rst evidences of the processes
involved in the atoms formation of Pb in the ETA-
W. The effects of the purge gas as an atomization gas
was clearly proved and the bene®cial effects caused by
reductor compounds generated in the condensed phase
on the sensitivity were explained. The severe inter-
ference caused by Na on Pb atomization was also
understood and the formation of refractories tung-
states that trapped the Pb compounds seems to be
the operative process. The data can be used to drive
both the heating program and the chemical conditions
towards better ef®ciency of atomization, but the route
Â
[13] M.M. Silva, F.J. Krug, P.V. Oliveira, J.A. Nobrega, B.F. Reis,
D. Penteado, Spectrochim. Acta Part B 51 (1996) 1925.
[14] F. Barbosa Jr., F.J. Krug, E.C. Lima, Spectrochim. Acta 54B
(1999) 1155.
[15] K.A. Wagner, K.E. Levine, B.T. Jones, Spectrochim. Acta
53B (1998) 1507.
[16] A. Salido, B.T. Jones, Talanta 50 (1999) 649.
[17] V. Sychra, D. Kolihova, O. Vyskocilova, R. Hlavac, P.
Puschel, Anal. Chim. Acta 105 (1979) 263.
[18] V. Sychra, O. Vyskocilova, D. Kolihova, P. Puschel, Anal.
Chim. Acta 105 (1979) 271.
[19] P. Puschel, Z. Formanek, R. Hlavac, D. Kolihova, V. Sychra,
Anal. Chim. Acta 127 (1981) 109.
Â
[20] E. Krakovska, Spectrochim. Acta 52B (1997) 1327.
Â
[21] E. Krakovska, J. Anal. Atom. Spectrom. 5 (1990) 205.

168
D.M. Santos et al. / Thermochimica Acta 362 (2000) 161±168
[22] C.L. Chakrabarti, A.H. Delgado, S.B. Chang, H. Falk, V.
Sychra, J. Dolezal, Spectrochim. Acta 44B (1989) 209.
Â
[23] C.G. Bruhn, J.Y. Neira, G.D. Valenzuela, J.A. Nobrega, J.
[29] D.R. Lide (Ed.), Handbook of Chemistry and Physics, 73rd
edn., CRC Press, Boca Raton, 1993.
[30] P.O. Luccas, Thesis, Department of Chemistry, University
Federal of SaÄo Carlos, 1999.
 Ã
[31] J.D. Lee, Quõmica Inorganica naÄo taÄo concisa, 1st ed., Edgard
Anal. At. Spectrom. 13 (1998) 29.
Â
[24] C.G. Bruhn, N.A. San Francisco, J.Y. Neira, J.A. Nobrega,
Talanta 50 (1999) 967.
 Â
[25] J. Alvarado, Quõm. Analõtica 15 (1996) 173.
BluÈcher, SaÄo Paulo, 1997, 127 pp.
[32] J.W Mellor (Ed.), A Comprehensive Treatise on Inorganic
and Theoretical Chemistry, Longmans, London, V-XI-W,
1952, 793 pp.
[26] J.P. Byrne, C.L. Chakrabarti, G.F.R. Gilchrist, M.M. Lamour-
eux, P. Bertels, Anal. Chem. 65 (1993) 1267.
[27] S. Imai, Y. Hayashi, Anal. Chem. 63 (1991) 772.
[28] W. Frech, A. Cedergren, Anal Chim. Acta 82 (1976) 93.
[33] R.C. Mackenzie, Differential Thermal Analysis, Academic
Press 2, 1972, 505 pp.