Performance evaluation of a TG/MS system

Article abstract of DOI:10.1016/S0040-6031(98)00496-1

A procedure for the evaluation of the performance of a TG/MS system is proposed. This procedure includes tests for monitoring the mass-flow stability, the gas transfer delay, the evolved-gas condensation effect and the interferences of the coupling system on TGA and MSD specifications. The procedure was applied on an in-house constructed TG/MS system. The results of such evaluation procedure can be used for monitoring the performance of a specific TG/MS system and, thus, can be a basis for design improvements. In addition, they may be used for comparison of the results of different TG/MS systems, or for interlaboratory comparisons.

Full text of DOI:10.1016/S0040-6031(98)00496-1

Thermochimica Acta 322 (1998) 167±173
Performance evaluation of a TG/MS system
*
M. Statheropoulos , S. Kyriakou, N. Tzamtzis
National Technical University of Athens (NTUA), Department of Chemical Engineering, Sector I, 9 Iroon Polytechniou Str.,
Athens 157 73, Greece
Received 20 April 1998; received in revised form 22 June 1998; accepted 22 July 1998
Abstract
A procedure for the evaluation of the performance of a TG/MS system is proposed. This procedure includes tests for
monitoring the mass-¯ow stability, the gas transfer delay, the evolved-gas condensation effect and the interferences of the
coupling system on TGA and MSD speci®cations. The procedure was applied on an in-house constructed TG/MS system. The
results of such evaluation procedure can be used for monitoring the performance of a speci®c TG/MS system and, thus, can be
a basis for design improvements. In addition, they may be used for comparison of the results of different TG/MS systems, or
for interlaboratory comparisons. # 1998 Elsevier Science B.V.
Keywords: Coupling; Evaluation; Performance; TG/MS
1
. Introduction
purge gas. The evolved gases are introduced in the MS
detector (MSD) through a coupling system, which acts
both as an MSD inlet and a pressure-reduction system
[6,10]. The coupling system should meet several
speci®cations such as: rapid evolved-gas transfer from
the sample pan in the TG analyser (TGA) to MSD; no
memory effects; no degradation or condensation of
evolved gases, no interferences on the speci®cations of
the TGA and the MSD; no air insertion during TGA
furnace opening; simplicity in construction; easy sys-
tem decoupling and cleaning; low cost; and contin-
uous monitoring and versatility with respect to TGA
and/or MSD [6,10]. In addition, if the system is used
for EGA, then high sensitivity and repeatable ¯ow
conditions are necessary.
The study of compound degradation by thermogra-
vimety (TG) can be greatly enhanced when evolved-
gas analysis is used [1±4]. This can be achieved by
combination of TG with other techniques capable of
providing qualitative (evolved-gas detection ± EGD),
or both qualitative and quantitative information
(
evolved-gas analysis ± EGA) [1,5±9]. It appears that
the most popular technique used in combination with
TG, is mass spectrometry (MS) on account of its
sensitivity, versatility and fast analysis time.
In a TG/MS system, the thermobalance provides,
among other things, with precise and reproducible
heating conditions. The gaseous products generated
by volatilisation, sublimation or chemical reaction, are
A great variety of TG/MS systems have been pre-
sented in the literature during the past decade [6].
Therefore, a need appears for the development of a
procedure for quantitatively estimating the speci®ca-
tions of a TG/MS system. The purpose of this work
¯
ushed out of the furnace chamber with the help of a
*
Corresponding author. Tel.: +30-1-7723109; fax: +30-1-
7
0
723188; e-mail: stathero@orfeas.chemeng.ntua.gr
040-6031/98/$ ± see front matter # 1998 Elsevier Science B.V. All rights reserved
P I I : S 0 0 4 0 - 6 0 3 1 ( 9 8 ) 0 0 4 9 6 - 1

1
68
M. Statheropoulos et al. / Thermochimica Acta 322 (1998) 167±173
was to develop and apply various performance eva-
luation criteria on an in-house constructed TG/MS
system. The results of such evaluation procedure can
be used for monitoring the performance of a speci®c
TG/MS system and, thus, can be a basis for design
improvements. Furthermore, it may be used for com-
paring the results of different TG/MS systems, or for
interlaboratory comparisons.
coupling system used was an in-house constructed
one-stage type capillary system [11]. The TG/MS
system con®guration is shown in Fig. 1(a). The sys-
tem consists of (1) a computer-controlled TGA, (2) a
heated coupling system and (3) a computer-controlled
MSD. A rough diagram of the TG/MS coupling
system is shown in Fig. 1(b). It consists of a TG/
MS interface and a transfer line. The transfer line
consists of a deactivated 2.17 mÂ0.15 mm i.d.
Chrompack capillary and a heating thermomantle.
The TG/MS interface and the transfer line can be
heated up to 2508C. The TG/MS interface operates as
a split system. Port A is connected to vacuum. Port B
has two modes of operation: (i) the sampling mode,
when port B is connected to vacuum in order to direct
evolved gases to the head of the capillary during the
TG/MS analysis and (ii) the idle mode, when port B is
2
. Experimental
2
.1. Instruments
The TGA used was a TA Instruments TGA 2050,
using the 1.5B Universal data analysis system. The
MSD used was a Hewlett±Packard 5972. The TG/MS
Fig. 1. (a) TG/MS system configuration: (1) thermogravimetric analyser; (2) coupling system; and (3) mass spectrometric detector. (b) TG/
MS coupling system consisting of a TG/MS interface and a deactivated capillary with a thermomantle (transfer line).

M. Statheropoulos et al. / Thermochimica Acta 322 (1998) 167±173
169
Fig. 1. (Continued )
connected to a gas purge in order to prevent air
insertion during TGA furnace opening. The vacuum
in ports A and B establish laminar ¯ow conditions
through the interface. In this way, evolved-gas mixing
is avoided, thus resulting in a narrow, initial sample
mass distribution [10].
2.3.2. Evaluation of the interferences of the coupling
system on TGA and MSD performance
(
a) Calcium oxalate samples, with 10 mg weight,
were analysed with, and without, the coupling
system for the evaluation of possible TGA
deterioration of performance. TGA was weight
calibrated using TA standard weights, and the
temperature calibrated using TA nickel and Alumel
Curie-point standards. The calibration was per-
formed under the actual experimental conditions.
Heating rate was 508C/min up to 9008C.
A mixture of benzoic acid and NaOH in a 1/3 mole
ratio was analysed for monitoring MSD perfor-
mance. The mixture was heated with a heating rate
of 508C/min from ambient up to 6008C.
2
.2. Materials
The following compounds were used: KHCO3,
CaCO , and ¯akes of NaOH (pro analysis Merck),
3
calcium oxalate standard (TA), LDPE standard (CDS
pyrolysis unit), and benzoic acid 98% (Fluka).
2
2
.3. Methods
.3.1. Evaluation of mass-flow repeatability and
evolved-gas transfer delay
Three different quantities, ranging from 1 to 15 mg
2.3.3. Evaluation of evolved-gas condensation
Low-density polyethylene (LDPE) was heated at a
heating rate of 508C/min, from ambient to 7008C.
In all the foregoing cases, the coupling system was
heated at 1908C. The volumetric ¯ows through ports A
from each of KHCO and CaCO , were heated in TGA
3
3
under the following conditions: heating rate 208C/min
up to 3008 and 9008C, respectively. Mass loss gen-
erates a differential thermogravimetry (DTG) curve of
positive sign.
�
1
and B were adjusted to 60 and 39 ml min , respec-
tively, and TGA furnace was purged with Helium at

1
70
M. Statheropoulos et al. / Thermochimica Acta 322 (1998) 167±173
�
1
1
00 ml min under 0.5 bar. The mass-spectrometric
detector was operated in a total ion-current mode
TIC) at 6.1 scans/s, while GC/MS interface was
maintained at 2008C.
Volumetric-¯ow repeatability depends on the tem-
perature and pressure control of the coupling system.
The estimated standard deviations of the time values
corresponding to the maximum of the DTG curve and
(
MSD signal, when CO was monitored from KHCO3
2
3
. Results and discussion
and CaCO decomposition, are presented in Table 2.
3
Both compounds decompose in one stage, with
KHCO3 simultaneously liberating H O and CO .
3
.1. TG/MS system
2
2
The results of an F-test on these values show that
there is no signi®cant variance introduced in the
presence of the coupling system, at a 95% con®dence
level. This is indicative of repeatable temperature and
pressure conditions in the coupling system. Evolved-
gas density depends on parameters such as: weight
loss rate; nature of gas eliminated from the sample;
and the nature and ¯ow rate of purge gas. The repeat-
ability of these parameters can be assessed by the
evaluation of the interferences caused of the coupling
system on TGA performance, as presented below.
The ability of the MSD in the TG/MS system to
monitor on-line the TGA events can be measured by
the time hysteresis of the MSD signal in relation to the
DTG signal. An estimation of the time delay can be
made by calculating the difference between the time at
the maximum of the DTG curve and the maximum of
the MSD signal. A mean time difference of 0.1 min
was determined for both, CaCO and KHCO TG/MS
analysis. The time difference is the sum of the times
required for the gases evolved to be transferred: (a)
from TGA sample pan to the entrance of the capillary;
and (b) to the MSD ion source through the capillary. In
the ®rst case, the time needed depends on port B
vacuum, sampling-tube inner diameter, and the dis-
tance between the sample pan and the capillary. In the
second case, the time needed depends on the capillary
The TGA furnace has a vertical con®guration with a
horizontal purge gas ¯ow. The presence of a quartz
liner minimises memory effect problems. As evolved
gases are purged out of the TGA furnace, their tem-
perature decreases rapidly. This might result in a
severe condensation problem. Gas condensation
may be partially avoided by properly heating the
TG/MS coupling system, using a heating thermo-
mantle.
The described TG/MS system presents several
advantages, such as: low cost; versatility with respect
to MSD; and simplicity in construction. However,
fractionation of some compounds in the evolved-gas
stream may occur in the capillary. This sets some
limitations to the applications of the system, espe-
cially when a gas mixture with high molecular-weight
compounds is analysed [5,10]. Table 1 summarises
the built-in characteristics of the TG/MS system for
meeting certain speci®cations.
3
3
3
.2. Evaluation of mass-flow repeatability and
evolved-gas transfer delay
Mass-¯ow repeatability is an important factor in
quantitative determinations. It is in¯uenced by volu-
metric ¯ow and evolved-gas density stability [9,10].
Table 1
TG/MS system characteristics for meeting certain specifications
Specifications
TG/MS system characteristics
Quick transfer from TGA to MSD
No memory effects
coupling system with small dead volumes
TGA quartz liner
No evolved-gas condensation or degradation
No air insertion during furnace opening
Simplicity of construction, easy decoupling, low cost
Continuous monitoring
heated coupling system. Use of inert materials
inert-gas insertion from Port B
one-stage coupling system
MSD triggered by TGA event switch
Versatility with respect to MSD
High sensitivity
easy capillary connection to any GC/MS interface
small evolved-gas dilution. No evolved-gas mixing
stable temperature and pressure conditions in the TG/MS coupling system
Repeatable flow conditions

M. Statheropoulos et al. / Thermochimica Acta 322 (1998) 167±173
171
Table 2
Standard deviations of time values corresponding to the maximum of DTG and MSD curves
Compound
Standard deviation of time
at DTG curve maximum (min)
Standard deviation of time
at MSD curve maximum (min)
CO
CO
2
2
(KHCO
(CaCO
3
)
1.580
1.009
1.550
0.975
3
)
length and inner diameter, the pressure at the entrance
of the capillary and the dynamic viscosity of evolved
gases [12].
number of carbon atoms. The sample used was a
mixture of benzoic acid and sodium hydroxide in a
1 : 3 mole ratio. The general appearance of the ben-
zene mass spectrum, resulting from the decarboxyla-
tion reaction at 3908C, is in agreement with those
presented in various mass-spectra compilations [13].
The calculated value for the number of carbon atoms,
based on M‡1 and M isotopic abundances, is in good
agreement with the theoretical value. This indicates
that the coupling system does not affect the ion-source
performance.
3
.3. Evaluation of the interferences of the coupling
system on TGA and MSD performance
The deterioration of the TGA performance can be
evaluated by monitoring the precision and accuracy of
weight loss and decomposition temperatures of a
standard compound, in the presence of the coupling
system. Calcium oxalate can be used as a standard, as
it decomposes in three well-resolved stages ranging
from 1808 up to 7208C. Table 3 shows the mean values
and estimated standard deviations of weight loss and
decomposition temperatures, in the presence, and
absence, of the coupling system. The values of weight
loss and decomposition temperatures without the cou-
pling system were used as standard values. The results
of a t-test and F-test, at 95% con®dence level, indicate
that there is no signi®cant systematic error or variance
introduced in TGA weight and temperature measure-
ments in the presence of the TG/MS coupling system.
Mass-spectrometer electron impact (EI) ionisation-
mode performance can be evaluated by monitoring
carbon-isotope abundances in the spectrum from var-
ious compounds. Benzene was selected for various
reasons, such as strong molecular ion and suf®cient
3.4. Evaluation of evolved-gas condensation
Evolved gas condensation is possible when high
boiling point compounds are present in the evolved-
gas stream. In order to evaluate the extent of con-
densation in a TG/MS system, a compound which
produces a long range of known thermal degradation
compounds is needed. Low-density polyethylene
decomposition results in a series of normal paraf®n
and ole®n products with known boiling points [14]. By
monitoring the mass spectra of those compounds, it is
possible to identify the product with the highest m/z
ratio that has been successfully recorded. Fig. 2(a)
shows the weight loss, DTG and temperature curves
recorded from LDPE thermal degradation. Fig. 2(b
and c) show the corresponding MS signal curve and
Table 3
Mean and estimated standard deviations of % weight loss and temperatures of calcium oxalate decomposition in the presence, and absence, of
the TG/MS coupling system
Calcium oxalate
% Weight loss
% Weight loss
Decomposition
Decomposition temperature
decomposition stage
(without coupling system) (with coupling system)
temperature (8C)
(8C) (with coupling system)
(
without coupling system)
mean
SD
mean
SD
mean
SD
mean
SD
1
2
3
12.59
19.56
29.85
0.088
0.080
0.083
12.66
19.80
29.92
0.021
0.110
0.056
180.3
500.0
714.9
1.830
1.990
2.250
182.8
500.7
722.1
2.304
1.147
5.790

1
72
M. Statheropoulos et al. / Thermochimica Acta 322 (1998) 167±173
Fig. 2. (a) Weight loss, DTG and temperature curves from LDPE thermal degradation. (b) MS signal from thermal degradation of LDPE. (c)
Mass spectrum at 10.50 min, corresponding to dodecene.

M. Statheropoulos et al. / Thermochimica Acta 322 (1998) 167±173
173
the mass spectrum at 10.50 min, respectively, using
References
the TG/MS coupling system. The mass spectrum
corresponds to dodecene, indicating that the coupling
system allows the detection of compounds with boil-
ing points nearly up to 2148C [15]. However, it should
be emphasised that the results of this test can be
greatly affected by parameters such as electron multi-
plier voltage, electron impact voltage, ion-source
pressure, as well as by MSD maintenance.
[1] G. Szekely, M. Nebuloni, L.F. Zerilli, Thermochim. Acta 196
1992) 511.
(
[
2] T. Leskel aÈ , M. Lippmaa, L. Niinist oÈ , P. Soininen, Thermo-
chim. Acta 214 (1993) 9.
[
[
3] M.R. Holdiness, Thermochim. Acta 75 (1984) 361.
4] D. Dollimore, G.A. Gamlen, T.J. Taylor, Thermochim. Acta
75 (1984) 59.
[
[
[
5] K. Jaenicke-R oÈ ûler, G. Leitner, Thermochim. Acta 295
(
1997) 133.
6] K.G.H. Raemaekers, J.C. Bart, Thermochim. Acta 295 (1997)
1.
4
. Conclusions
7] W.M. Groenewood, W. de Jong, Thermochim. Acta 286
1996) 341.
(
[
[
8] M. Mittleman, Thermochim. Acta 166 (1990) 301.
It is possible to evaluate quantitatively the perfor-
9] M. Maciejewski, A. Baiker, Thermochim. Acta 295 (1997)
mance of a TG/MS system. The evaluation procedure
includes measurements on various parameters such as
the mass-¯ow stability, the evolved-gas transfer delay
and the evolved-gas condensation effect. The proce-
dure can also quantitatively determine the extent of
interferences of the TG/MS coupling system on TGA
and MSD speci®cations. It appears that the evaluation
procedure can be used for monitoring the performance
of a speci®c TG/MS system and, thus, can be a basis
for design improvements. In addition, it may be used
for comparing the results of different TG/MS systems
or for interlaboratory comparisons.
95.
[
[
10] E. Kaisersberger, E. Post, Thermochim. Acta 295 (1997) 73.
11] W.H. McClennen, R.M. Buchanan, H.L.C. Meuzelaar, Anal.
Chem. 65 (1993) 2819.
[
[
[
12] C.A. Cramers, G.J. Scherpenzeel, P.A. Leqlerq, J. Chroma-
togr. 203 (1981) 207.
13] Eight Peak Index of Mass Spectra, 1st edn., 1970, Vol. 1,
Table 1.
14] M.P. Stevens, Techniques and Methods of Polymer Evalua-
tion, Characterization and Analysis of Polymers by Gas
Chromatography, Chap. 4, Marcel Dekker, New York, 1969.
15] CRC Handbook of Chemistry and Physics, 52th edn., Section
C, 1971.
[