61

CHAPTER 2" T H E R M O G R A V I M E T R Y

2.1 Introduction

T h e r m o g r a v i m e t r y (TG), the technique in which the mass of a

sample is m o n i t o r e d against time or temperature, is p e r f o r m e d
with a T h e r m o G r a v i m e t r i c A n a l y s e r (TGA) or thermobalance.
Recently, a survey of this technique and the available
commercial equipment was given by W u n d e r l i c h [i].
Important differences b e t w e e n the balances are-
- the type of balance, vertical or horizontal furnace systems
(the horizontal furnace TGA needs a correction for the
influence of the thermal e x p a n s i o n on the length of the
balance-arm),
- the sensitivity in c o m b i n a t i o n with the m a x i m u m sample
weight (a typical example is the Perkin Elmer TGA-7 with a
sensitivity of 0.0001 mg and a m a x i m u m sample weight of 200
milligramme.
- the temperature range and the temperature accuracy.
The sample mass d e t e r m i n a t i o n and the sample temperature
measurement of the TGA has to be calibrated using calibrated
weights and the ferromagnetic transition (Curie) temperatures
of calibration metals.

A Perkin Elmer TGA-7 (vertical furnace) system is used for the

TGA experiments d e s c r i b e d in this chapter. The balance is
purged with 40 ml/minute of nitrogen. A second n i t r o g e n purge
gas stream of 20 m l / m i n u t e is entering the system via the
sample purge entrance, see Figure 6.5. Both gas streams purge
the furnace part of the TGA. The sample purge can be switched
from nitrogen to air w i t h the aid of a software c o n t r o l l e d
gas-selector.

The automated two-point t e m p e r a t u r e c a l i b r a t i o n p r o c e d u r e is

performed with the c a l i b r a t i o n standards alumel and perkalloy.
Subsequently, the Curie temperatures of four c a l i b r a t i o n
standards are measured to check the whole TGA temperature
range of interest. A typical c a l i b r a t i o n result m e n t i o n e d
below shows that a t r a n s i t i o n temperature accuracy of about •
I~ can be obtained u s i n g the standard temperature c a l i b r a t i o n
procedure"

alumel - 163oc, Curie temperature = 163~

nickel : 356~ " " = 354~
perkailoy: 597oc, " " - 597~
iron : 787~ " " - 787~

The TG technique is (just like the DSC) very p o p u l a r in

polymer reseach, in p a r t i c u l a r to study the thermal stability
of polymeric systems u n d e r a p p l i c a t i o n conditions. An example
of the straightforward use of this technique for p o l y p r o p y l e n e
(PP) is given in chapter 2.2. A series of TG experiments on PP
catalyst systems under special conditions (an oxygen- and
moisture-free sample l o a d i n g procedure) is d e s c r i b e d in
 62

chapter 2.3. A number of TGA applications in combination with

FTIR and MS analysis of the released purge gas is, finally,
mentioned in chapter 6.

2.20ligomers content and thermal stability of polypropylene

2.2,1 The non-isothermal thermal stability d e t e r m i n a t i o n

A quick impression of the thermal stability of a polymer is
obtained by m e a s u r i n g its mass with a TGA as a function of the
temperature at a constant heating rate. Figure 2.1 shows the
results of four TGA scans on ~'P samples using a heating rate
of l~ These results illustrate that"
- the thermal stability of PP considerably decreases if oxygen
is present and,
- that the added stabilisation system only works under inert
conditions.
These non-isothermal TG experiments offer a simple and
relative quick analytical p o s s i b i l i t y to compare the thermal
stability of different polymers or different polymer batches.

A polymer is thermally stable untill the d e c o m p o s i t i o n process

starts. Two (main) types of thermal d e c o m p o s i t i o n processes
are usually recognised for polymers, chain d e p o l y m e r i s a t i o n
and random decomposition. Chain d e p o l y m e r i s a t i o n is the
release of m o n o m e r units from a chain end or at a weak link
and is essentially the reverse process of polymerisation. It
is often called d e p r o p a g a t i o n or unzipping. Random degradation
occurs by chain rupture at random points along the chain,
giving a disperse mixture of fragments. These two different
processes may occur separately or in combination; the latter
case is rather normal [2]. Both processes cause sample mass
losses which can be measured with a TGA.

The thermal stability of a p o l y m e r is often expressed by its

Td(0.5)-value. This is the temperature at which the loss of
mass during p y r o l y s i s (at a constant heating rate) reaches 50
% of its final value. This Td(0.5)-value is measured in an
inert atmosphere and is, according to Van Krevelen [2],
determined by the polymer's chemical structure. It is
important to realise, however, that the physical properties of
a polymer are c h a n g e d / d e c r e a s e d significantly at the moment
that the Td(0.5) temperature is reached. We prefer, therefore,
to characterise the thermal stability of a polymer by the
Td(o)-value i.e. the temperature at which a loss of mass
during heating just starts. The d e t e r m i n a t i o n of the Td(o)-
value can be hampered, however, by mass loss effects due to a
low molecular weight fraction (oligomers) and/or residual
monomer/solvent in the polymer sample. Commercial PP grades,
for example, always contain a small (0.I %wt. - 0.3 %wt.)
oligomer fraction (C6 - C39).

A series of isothermal TG experiments was used to determine

the oligomer content of a commercial PP grade. In addition,
the Td(o) -value of pure i.e. n o n - s t a b i l i s e d PP was determined.
 63
e,i~-
I--
./" U
,/
/
!
/ i
I
~1 ~
W
/ n-
m
Q.W

N
-i
"r r ~ _
m
w
Zr
W
0

0
7
i;i w
Z
R
o
w Lu o
"r N ILl ,--
F-
o
0
I-- Z
_
o~ " " .....
I

R
_-J
Q~
c56
B
| i
- - ,., = ...... - . . . . . . . . . .
I IHrdl3M % i
 64
2.2,2 The isQ~hermal thermal stability determination.
Isothermal mass/time curves of n o n - s t a b i l i s e d PP powder
samples were measured during I000 minutes at temperatures
b e t w e e n 160~ and 280~ The PP TGA samples (about I0 mg.)
were flushed with nitrogen during one hour at 30~ before the
experiment was started.

Oligomers and polymer are separated, during these TGA

experiments, on basis of the boiling temperature (and thus the
m o l e c u l a r weight) of the different oligomer fractions. The
total amount of oligomers determined during these isothermal
TGA experiments will increase, therefore, with increasing
m e a s u r i n g temperatures.

The mass/time curve measured at T(isothermal) = 160~ shows a

non-linear mass loss effect during the first 250 minutes, see
Figure 2.2. No further mass losses were detected during the
remaining 750 minutes of this experiment. This mass loss
effect of about 0.3 %wt. is assumed to be caused by the
evaporation of an oligomeric fraction.

The TGA experiments at T(isothermal) ~ 190~ show a

continuous, nearly linear with the time, d e c r e a s i n g sample
mass after the first (non-linear) mass losses due to
evaporation of the oligomers fraction. The slopes of the
linear part of these curves increase with isothermal measuring
temperatures. This effect is thought to be caused by the
thermal degradation of the polymer matrix. The mass/time
curves were extrapolated, subsequently, as indicated in Figure
2.2 to determine the oligomers fractions [wl, w2, w(n)]. The
curves in Figure 2.2 clearly illustrate that the extent of the
oligomers fraction increases with increasing m e a s u r i n g
temperatures. The oligomer content values and slopes of the
curves which are thought to represent the rate of the' thermal
degradation process of the PP matrix are listed in Table 2.1.

Nib samples (i.e. the extruded and stabilised material) were

investigated, subsequently, in the same way. The results of
this series of experiments is listed in Table 2.2. The
oligomer contents m e a s u r e d for both the powder and the nib
samples are plotted as a function of the isothermal measuring
temperature in Figure 2.3. The oligomer content of the non-
stabilised powder sample increases nearly linearly with the
temperature. The oligomer contents measured on the
(stabilised) nib samples are lower and scatter considerably
more than the powder sample values. These values have to be
lower due to e v a p o r a t i o n losses during the nib extrusion
process. Differences in the cooling rates of single nibs after
the extrusion process might be the reason for the considerable
amount of scatter. The broken line indicates thus a kind of
m a x i m u m oligomer content for nib samples.

The loss rates due to thermal degradation of the PP matrix are

p l o t t e d in Figure 2.4 as a function of the reciprocal absolute
temperature. The first detectable loss rate value was measured
for the non-stabilised PP sample at T(isothermal) = 191~
 -100.0
I00.099.
9 I ................
- 99.9

g9. 8 99. 8
T(lsot, h.) - t68~
. . . . . l ~z
99. 7
. . . . . . . . . . . . . . . . . . . . . . . . -- . . . . . . - - _ _ _ _ _ L

o~" 99, r T ( t soth. ) - 198 ~

99. 6 o~
"" 99,5 99. 5 v

r- r-
---~ U
~: 99,4 t3 99. 4
T(t=oth.) - 228~
9g. 3 -
, %
99. 3

99. ~ 4 - 99.2

99. ] - - 99.1

99.0 - " 99.0

98, 9 - - 98.9
I . . . . . I I ...... 1 I I i '1 ' 1 1
0.0 100. 0 200. 0 300. 0 400. 0 500. 0 600. 0 700. 0 800. 0 900. 0 1000. 0
Time (minutes)

Figure 2.2
Isothermal TGA mass/temperature curves of a PP powder sample (non-stabilised)
 66

Table 2.1 R e s u l t s of i s o t h e r m a l T G A e x p e r i m e n t s on
n o n - s t a b i l i s e d PP p o w d e r s a m p l e s
~,..~[_ ,, . . . . .

T (isoth. ) oligomer thermal degradation

fraction loss rate,
oC w (n) , %wt.
,
% wt./s
161 . .
0.288
. . .
0.0
171 0.353 0.0 = , ,

181 0.278 0.0

191 0.413 6.9E-7
201 0.467 1.4E-6
211 0.463 1.4E-6
221 0.585 2 .IE-6
232 0.675 4.9E-6
251 0.783 1.0E-5

Table 2.2 R e s u l t s of isothermal TGA experiments on

stabilised PP nib s a m p l e s
. . . . 11 I J I I --7 i~ I

T (isoth. ) oligomers thermal degradation

fraction lOSS rate,
oC w(n), % wt. % wt./s
171 ...... , ,
0.282 , ,
0.0
182 0.286 0.0
192 ,,
0.350 0.0
202 ,
0.261, , ,
7.4E-7 ,, ....

222 0.423 .
9.0E-7
242
, =
0.534 7.8E-7
250 0.425 1.5E-6 ,, ,,

262 0.568 4.7E-6

,
270 0.496 1.6E-5
277 m 0.412 3.3E-5
9 7
 67

Figure 2.3
Oligomer content/temperature relation
(nonstabilised powder/stabilised nibs)
+ non- ~ stab.
stab. mibs

0.80
-I-

0.70
+

0.60
J

A J
0.50 J A
c
0
-I-
J
o

E
JA A
0
0.40 J A
0
J
-t-

0.30 J
_r
A

0.20 I _ _ I I I I

160 180 2OO 220 240 26O 28O

T(isothermal), ~
 68

Figure 2.4
The thermal degradation loss rate of PP as a function of 1000/T

+ non- ~ stab.
stab. n/bs,

0.00001 +

+ +

U')
o

le-06

5e-Q7 I I I I I I I

1.80 1.85 1.90 1.95 2.00 2.05 2.10 2.15 2.20

1000/T, K -1
 69

Hence, the Td(o)-value of pure (non-stabilised) PP is about

190~ It is also clear that the small amounts of oligomers
presen t in these samples do not hamper a proper Td(o)-value
determination. This Td(o)-value of 190~ increases to about
240~ due to the addition of the stabiliser system (ionol and
irganox). The activity of this stabiliser system is not only
temperature but also in time limited. Figure 2.5 shows that
this stabiliser system stays active (at 250~ for about I000
minutes i.e. long enough to withstand the thermal treatments
during all different PP processing procedures.

Usually, non-isothermal TGA experiments are used to determine

the 'standard' Td(0.5)- and Td(o)-values. The non-stabilised
PP sample was measured, therefore, at different heating rates
in order to match a non-isothermally determined Td(o)-value
with the isothermally determined Td(o)-value of 190~ This
match was obtained, see Figure 2.6, for a heating rate of
0.1~ The slope of this mass/temperature curve
increases due to the evaporation of oligomers between 75~ and
184~ and clearly decreases then in the temperature region
between 184oC and 204~ The presence of the latter
temperature region indicates that there is sufficient
separation between the oligomer evaporation process and the
start of the polymer matrix degradation. The Td(o)-value of PP
determined in this way is estimated at 194~ and is called the
semi-static Td(o)-value.

2.3 The TG analysis of a PP catalyst system

2,3.1 A 'plastic w r a p p e d ' TGA

TIC14, the active component of a Ziegler-Natta PP catalyst
system, has to be coordinated on the MgCI2 carrier material
together with an internal electron donor like di-isobutyl
phthalate (DIBP). The close presence of TiCl4 and DIBP might
also result, however, in the formation of a TiCI4.DIBP
complex, which would influence directly the catalyst activity.

Different analytical techniques like NMR and IR have been used

to study these catalyst systems, but they did not resolve the
doubts about the state of the TIC14, the DIBP and possibly the
TiCI4.DIBP complex in these catalysts.

Terano et.al, used TG experiments to investigate a TiCl4/ethyl

benzoate system (EB) [3]. Their results suggested that the
TiCl4 and EB exist seperately on the MgCI2 surface. Based on
these results, a series of non-isothermal TGA measurements was
performed to study the TiCI4/DIBP system in the same way. A
5~ heating rate was used (Terano used 17oC/minute) to
improve the resolving power of this technique. Besides, the
TGA sample loading procedure had to be optimised.

These types of catalysts have to be handled in an moisture and

oxygen free atmosphere and are stored in a dry-box. Special
precautionary measures are, therefore, necessary for the TGA
100.0 - lO0. 0

gg. 5 - 99. 5
PP N I B S A M P L E
stabilised
gg. 0 - 99. 0

98.5 - 98. 5

98.0 - e.. g8. 0

. = . .

97.5 - gT. 5
PP P O W D E R S A M P L E
non.stabilised

97.0 - 97.0

T i m e (minutes)
96.5 - g6. 5
0.0 500. 0 1000. 0 1500. 0 2000. 0 2500. 0 3000. 0 3500. 0 4000. 0

Figure 2.5
The mass/time curves of a non-stabilised and a stabilised PP sample during
isothermal TGA experiments at 250~ (nitrogen atm.)
 PERKIN-ELMER 7 Series Thermal Analysis System
100.0 - '. . . . .
i,i i ,, , ,,,,,,,/,, ,,, ,,, , ,,,,,, , ,, , , , ,', , , , , | i

99.8 -

9g. 6 - TCA F i l e Nome: t2472 "

Somple We19ht: 10.714 m 9
Sun H o t 31 14.49:23 lggl ~ ' ~
9g. 4 -

Unstob~lIsed PP powder" ex P e r n l s CPO

o~ 9 9 . 2 -

9g.o-
,,,3

.,m,
QI

N 9a.a-
nitrogenatmosphere
TEMP 1: 30.0~ TIME 1: 0 . 0 m i n RATE 1: 0.1~
98.6 - TEMP 2: 450.0~

98.4 -

98. 2 -
184[~ 2@41~ /
98.0 -
, , . . . . . t . . . . ~ I .... ~ ' ' -I ............ I 'I ' "
50.0 75. o ~oo. o ~25. o ~5o. o ~75. o 2oo. o 225.o 250. o 27~.o 3oo. o
Figure 2.6 Temperature(~
PP TGA experiment. Heating rate 0.1~
 72

sample loading operation.

The TGA sample pan was filled with about I0 mg. of catalyst
sample and placed in a closed, n i t r o g e n flushed, small bottle.
This happened in the dry-box used to store the catalyst. The
TGA sample pan was transported in this bottle to the TGA
sample loading table. The whole TGA balance system was packed,
subsequently, in a big p o l y e t h y l e n e bag (Sigma A3533 Atmosbag
36" x 51" x 58", provided with two gloves) This bag was
closed as good as possible around the electrical cables and
blown up with an excess of nitrogen. The relative humidity in
the bag decreased in about one hour from 50 - 70 % to less
than 5 % due to this nitrogen purge. The catalyst sample was
then, using the gloves of the Atmosbag, taken out of the
bottle and placed in the TGA hang-up wire. Subsequently, the
TGA furnace was closed and the m e a s u r e m e n t was started.

2.3.2 TG analysis of a MaCl2-suppor~ed. TiCI4/DIBP catalyst

_ v

Figure 2.7 shows the TG m a s s / t e m p e r a t u r e curves of the

reference systems MgCI2, DIBP and a mixture of MgCI2/DIBP. The
d e c o m p o s i t i o n of MgCI2 starts at temperatures > 400~ The
total mass loss between 420~ and 730~ was 17.7 %wt. The DIBP
as such evaporated completely b e t w e e n II0~ and 250~ The
MgCI2/DIBP mixture m a s s / t e m p e r a t u r e curve clearly shows a two-
step mass loss effect due to e v a p o r a t i o n of the DIBP as such
and due to evaporation or possibly d e c o m p o s i t i o n of DIBP
absorped on the MgCI2 surface. A mass loss effect of 25 %wt.
is measured between II0~ and 300~ (evaporation) while 11.3
%wt. weight loss is measured between 510~ and 584~
(evaporation or decomposition). The total weight loss due to
DIBP release (after correction for the MgCI2 weight losses) is
thus 36.3 %wt. This value fairly agrees with the DIBP content
of 34.7 %wt. reported for this system.

Two other reference systems, M g C l 2 - s u p p o r t e d TIC14 and

TiCI4.DIBP complex were measured in the same way. The TIC14
decomposes in two steps with DTGA minima at I19~ and 221~
The total weight loss at 420~ (the starting temperature of
the MgCI2 mass loss process) was 18.5 %wt. The calculated
weight loss due to a conversion of the TIC14 to Ti is 20.0
%wt.

The main mass loss effects of the T i C I 4 . D I B P complex occur

between 128~ and 288~ with at least t w o DTGA minima at about
184~ and about 220~ The main evaporation process of DIBP
both from T i C I 4 . D I B P complex and from the MgCI2/DIBP mixture
thus occurs in the same temperature region which was
disappointing. On the other hand, the TiCI4.DIBP complex shows
a small but clearly present effect at about 400~ which is not
present in the M g C I 2 / D I B P mixture results. Besides, the MgCl2/
DIBP mixture shows a strong effect between 510~ and 584~
(Figure 2.7) which is hardly present in the TiCI4.DIBP complex
results.
 100. 0 ~ : " 1e
L
~'C',-"c"---~ ~ ' : ':L=-i -' ~"-----~ ""-'-~-~

go. 0

85.0

~ -I HgCl 2 / D I B P mixture Commercial M g C 1 2 ~

..; BO. 0
1?.7 ~ = t , , loss
~
= 75. o
_1[
3 ~8
"
,
between 4 2 B - 7 3 0 C
c
75.8 ~.
70. 0 -

65.0 -

60. 0 -

55. 0 - DIBP, I B 8 ~'.=t. 1o==

se4 c
betueen t 1(]-25(] C. 55.9 ~
50. 0

,1~ .................. l j[ l ..... 1. . . . . . . . . 1 ''l I . . . . I I--

1oo. o 200. o 300. o 400. o 5oo. o son. o ~oo. o

Figure 2.7 Temperature (~
Non-isothermal TGA mass/temperature curves of DIBP, MgC12 and a mixture of
MgC12/DIBP measured in a nitrogen atmosphere
 74

Figure 2.8 shows the TG m a s s / t e m p e r a t u r e curves and the

corresponding DTGA curves of the MgCl2-supported, TiCI4/DIBP
catalyst in duplo. This catalyst contained 15.8 %wt. of TIC14
and 20.8 %wt. of DIBP. These two m a s s / t e m p e r a t u r e curves have
roughly the same shape as the M g C I 2 / D I B P mixture mass/
temperature curve shown in Figure 2.7. This is an indication
that an important part of the DIBP in this catalyst sample is
present as DIBP and not as TiCI4.DIBP complex. The "blown-up"
DTGA curves in Figure 2.9 illustrate that the TiCI4.DIBP
complex DTGA m i n i m u m at 390~ is not present in the catalyst
samples. These results are, therefore, suggesting strongly
that the DIBP is present as DIBP in this catalyst system and
not as T i C I 4 . D I B P complex.

References

i. B. Wunderlich. Thermal Analysis, A c a d e m i c Press Inc., New

York, 1990.
2. D.W. van Krevelen- Properties of Polymers, Third edition,
Elsevier, Amsterdam, 1990.
3. M. Terano and T. Kataoka, Makromol. Chemie, i__~, (1987), p.
1477 - 1487.
 100. 0

95, 0 9-0, I

go. 0 "-0. 2

85. 0
I
/
#
"-0. 3

8o.o //
I !
"-0. 4

--
J::
75.0-
/I
"-0. 5
.,,..

(9
.,,1
7o.o- -0. 6 U1

65, 0 - -0~ 7
k

60. 0 -0. 8

55, 0 -
1
528~ -0. 9

234~
50, 0 -1.0
........ ~. . . . . . . . . . . . I ......... I ..... I i . . . . . . i ....... t
I00. 0 200.0 300. 0 400. 0 500. 0 600, 0 700, 0

Figure 2.8 Temperature (~

Results of two non-isothermal TGA measurements on PP catalyst BR900718 in a
nitrogen atmosphere
 -0. 1 t
-0. 1

I
-0. 2 -0. 2
\ \
i/ |

' / .J
-0, 3 | -0. 3

.c_ -0. 4
E
o~
t

/ \., ;

[
I
"-~ -0. 5
/
t -0. 5
ID

I
~ -0. 6
.~_
L_

E3 - 0 . 7
I

ri 390~
15.BY. T I C I 4
-0, 6

-0. 7 ,..3
! 28. B~. DIBP

-0.8
f I
-0. B

4.BX T t C I 4
-0. 9

-1.0
i 9 2 2 . 5 ~ DIBP
-0. g

.0

-1.1
i
!
5aeoc -1.1

~o0. o 2oo. o a~o. o ,od. o sod. o coo. o 7od. o

9 Temperature (~
Figure 2.9
DTGA minima of two PP catalyst samples and a TiC14.DIBP complex sample
measured at 5~ in a nitrogen atmosphere