Available online at www.sciencedirect.

com

Polymer Degradation and Stability 93 (2008) 800e810

www.elsevier.com/locate/polydegstab

The thermal degradation of poly(vinyl acetate) and

poly(ethylene-co-vinyl acetate), Part I: Experimental study
of the degradation mechanism
B. Rimez a, H. Rahier a,*, G. Van Assche a, T. Artoos a, M. Biesemans b, B. Van Mele a
a
Research Unit Physical Chemistry and Polymer Science (FYSC), Vrije Universiteit Brussel (VUB), Pleinlaan 2, B-1050 Brussels, Belgium
b
Research Unit High Resolution NMR Spectroscopy (H-NMR), Vrije Universiteit Brussel (VUB), Pleinlaan 2, B-1050 Brussels, Belgium
Received 9 September 2007; received in revised form 5 December 2007; accepted 14 January 2008
Available online 20 January 2008

Abstract

The thermal degradation mechanism of poly(vinyl acetate) (PVAc) and poly(ethylene-co-vinyl acetate) (EVA) copolymers was investigated
with solid-state NMR, thermogravimetry coupled with mass spectrometry and differential thermal analysis. Between 300 and 400  C acetic acid
is eliminated (deacetylation), leaving a highly unsaturated residue or polyene. The deacetylation of PVAc is autocatalytic. Upon incorporation of
ethylene entities into the polymer backbone, autocatalysis disappears. Between 400 and 500  C, the polyene will degrade further by chain scis-
sion reactions in inert conditions or aromatise in an oxidative environment into a char, and oxidised eventually into CO2 beyond 500  C.
In inert conditions, the deacetylation step as well as the chain scission reaction shows endothermic effects. In an oxidative environment, large
exothermal effects are found for each degradation step. This indicates the occurrence of additional oxidation reactions during deacetylation, an
important reorganisation of the polyene during char formation and oxidation of the latter into CO2.
Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: PVAc; EVA; Thermogravimetric analysis; Mass spectrometry; Differential thermal analysis; Solid-state NMR

1. Introduction acetate side chain of PVAc is eliminated from the polymer

The degradation path upon heating of semi-crystalline
poly(ethylene-co-vinyl acetate) (EVA) copolymers with a vinyl
acetate content below 50% has already been well described
and used in flame retarded systems [1e4]. Upon degradation,
the acetate side groups are first eliminated from the polymer
backbone in the so-called deacetylation step, leaving an unsat-
urated polymer backbone or polyene [1,2]. This polyene
undergoes allylic chain scission reactions to reach full degra-
dation [5,6]. In an oxidative environment, the degradation of
the formed polyene into an aromatised structure at elevated
temperatures is detected upon addition of flame retardants [7].
The thermal degradation of poly(vinyl acetate) (PVAc) in
inert conditions has a similar degradation pathway [5]. The
* Corresponding author. Fax: þ32 2629 32 78.
E-mail address: hrahier@vub.ac.be (H. Rahier).
backbone in an autocatalytic reaction [8,9]. On the oxidative
degradation no information was found.
The systematic study of the degradation mechanism of any
material requires information on both volatiles and solids
formed as a function of temperature during degradation. By
coupling thermogravimetric analysis (TGA) with mass spec-
trometry (MS) information concerning produced volatiles
can be obtained [10e17]. The identification of solid material
formed during degradation can be achieved by solid-state
NMR analysis of partially degraded residues [5,18]. Isother-
mal TGA measurements are used to quantify the kinetics of
the chemical degradation reaction [9]. By coupling TGA
with differential thermal analysis (TGAeDTA), qualitative in-
formation on the thermochemistry of certain degradation steps
can be obtained [16,18e20].
In this experimental study, a mechanistic investigation is
performed on the thermal degradation of PVAc and both

0141-3910/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymdegradstab.2008.01.010
 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810
semi-crystalline and amorphous EVA with mass percentage vinyl
acetate in the polymer backbone ranging from 9 to 73 wt.% in inert
and oxidative conditions. A coherent mechanistic description for
the full degradation path of all (co)polymers is aimed at.
This descriptive thermal degradation mechanism will en-
able a quantitative modelling of the kinetics of the different
degradation steps, i.e. deacetylation, polyene formation and
degradation, and char formation and degradation, as described
in a second paper [21]. For this reason, the proposed mecha-
nistic model describes the degradation of groups of molecules
rather than defining the degradation reaction of each individual
molecule. Moreover, the degradation model in combination
with the same experimental methodology will facilitate the in-
terpretation of the activity and efficiency of flame retardants
for the chosen polymer systems.
2. Experimental
Poly(vinyl acetate) (PVAc) with a molar mass of
100 000 g mol1 from SigmaeAldrich was used as-received.
EVA emulsions Airflex EN 1035 and Airflex EAF 60 (55
and 60% solids in water, respectively) from Air Products con-
tain 73 and 60 wt.% of vinyl acetate and are denoted as EVA
73 and EVA 60, respectively. They were solution-casted and
dried at 100  C. EVA copolymers Escorene UL00309,
UL00319 and UL00333 from Exxon contain 9, 19 and
33 wt.% of vinyl acetate, respectively. Films of all polymers
with a thickness of 100 mm were pressed.
Thermogravimetric analysis coupled with mass spectrome-
try (TGAeMS) was performed on a TA Instruments Q5000IR
under nitrogen or dried air flow with a base purge flow of
25 ml min1, using sample masses of 1.0  0.2 mg in 50 ml
platinum crucibles, starting from 200  C (to eliminate water
and monomer) to 600  C (inert conditions) or 650  C (oxida-
tive conditions). Mass spectrometry (MS) measurements
were performed on a Balzer Thermostar Mass Spectrometer
using Quadstar 422 Software to analyse detected m/e values.
The produced gases were measured in the MS equipment by
continuously scanning bargraph cycles at a sampling rate of
0.1 s per m/e and a spectral width ranging from 10 to 200 m/e.
Isothermal TGA experiments were performed on a TA
Q5000IR TGA with nitrogen or dried air with a purge flow
of 25 ml min1. A sample of 1  0.2 mg in a 50 ml platinum
crucible was held isothermal at 230  C for 120 min to elimi-
nate emulsifier and residual monomer. Afterwards, the sample
was heated at 20  C min1 to the desired isothermal tempera-
ture of the measurement and maintained until the weight re-
mained constant. Inertia of the TGA was checked with MS
by the evolution of the m/e 32/28 (N2/O2) ratio. After
10 min spooling prior to the measurement, the m/e 32 signal
reached a baseline level below 0.1% of the m/e 28 signal.
Note that TGAeMS and TGA measurements were per-
formed on powdered samples or thin films (100 mm) with
a mass of 1  0.2 mg to avoid the occurrence of other interac-
tions than the expected chemical degradation reactions (slow
diffusion of gas trough the sample, gas trapping, etc.).
801
TGA coupled with differential thermal analysis (TGAe
DTA) measurements were performed on a TA Instruments
SDT Q600. The applied heating rate was 20  C min1. Sample
masses of 3  1 mg were used with nitrogen or dried air with
a base purge flow of 100 ml min1.
Samples for NMR analysis were degraded in a TA Instru-
ments 2950 TGA, with a heating at 20  C min1 to the desired
temperature, at which the sample was held isothermally until
the sample weight remained constant (time interval ranging
from 10 to 900 min). The amount of residue needed for the
NMR measurements was about 10 mg. Cross-polarized magic
angle spinning 13C NMR (CP-MAS 13C NMR) spectra were re-
corded on a Bruker Avance 250 spectrometer, operating at
62.90 MHz for the 13C nucleus. ZrO2 rotors (4 mm) were
used at a spinning rate of 4000 Hz. The chemical shift reference
was set with adamantane. Spectra were acquired with typically
4 K data points over a spectral width of 25 000 Hz, a relaxation
delay of 4 s, a contact time of 1 ms and 4000 scans.
3. Results and discussion
3.1. The degradation steps of PVAc
In Fig. 1, TGAeMS measurements performed on PVAc un-
der inert and oxidative environments are shown. From an MS
measurement, the most significant m/e signals for the degrada-
tion are chosen. For example, acetic acid evaporates as m/e 60
but also decomposes in the gas phase into smaller fragments
like m/e 15, 17, 18, 41, 42, 43, 44. Acetic acid is called the
‘‘mother molecule’’; the smaller fractions, possessing the
same gas-evolving pattern, help to attribute m/e 60 to this
molecule.
The inert degradation of PVAc as measured by TGA con-
sists of two steps as can be seen in Fig. 1A I. The first and
most intense degradation step, between 300 and 400  C, was
identified as the deacetylation step, since both m/e 60 and
smaller fragments corresponding to the acetic acid mother
molecule are detected in this temperature range (Fig. 1A II).
Note that during this step also (CH)n fragments evaporate
from the polymeric material (Fig. 1A III). Molecules such as
CHeCHeCH (m/e 39), benzene (m/e 78), toluene derivates
(m/e 91) and naphthalene (m/e 128) arise during the deacety-
lation process, with the largest fragments evaporating at the
end. This indicates the chain scission reactions occurring at
the end of the polymer main chain during deacetylation. The
formation of the aromatic volatiles slows down at the end of
the step. A stable product is formed after deacetylation.
In order to identify the solid residue after deacetylation,
PVAc was partially degraded by holding it isothermally at
280  C for 900 min or at 400  C for 15 min to reach full com-
pletion of the elimination of acetic acid in both samples. The
chemical structure of the resulting materials was analysed by
performing solid-state 13C NMR. Non-degraded PVAc shows
four resonances in its 13C NMR spectrum (Fig. 2). The poly-
mer backbone has resonances at 40 and 70 ppm. The carboxyl
function has a resonance at 170 ppm and the ester function at
20 ppm. After isothermal degradation at 280  C, most of the
802 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810

100 2 100 2
90 A I 90 B I

Der. Weight / %.°C-1

Der. Weight / %.°C-1
80 1.5 80 1.5
70 70
Weight / %

Weight / %
60 1 60 1
50 50
40 0.5 40 0.5
30 30
20 0 20 0
10 10
0 -0.5 0 -0.5
200 250 300 350 400 450 500 550 600 650 200 250 300 350 400 450 500 550 600 650
Temperature / °C Temperature / °C

1.E-09 1.E-09
II II
1.E-10 1.E-10
Ion Current / A

Ion Current / A
1.E-11 1.E-11

m/e 60 m/e 44

1.E-12 1.E-12 m/e 60

m/e 44
m/e 15
1.E-09 1.E-09
m/e 78 III III
m/e 91
1.E-10 1.E-10 m/e 78
Ion Current / A

Ion Current / A

1.E-11 1.E-11 m/e 91

m/e 128

1.E-12 1.E-12

m/e 178
m/e 128
1.E-13 1.E-13
200 250 300 350 400 450 500 550 600 650 200 250 300 350 400 450 500 550 600 650
Temperature / °C Temperature / °C

Fig. 1. TGAeMS of PVAc in (A) inert and (B) oxidative conditions at 20  C min1. I e weight and derivative of weight versus temperature; II e evolving mass
fractions m/e 15, 44 and 60, representative for deacetylation; III e evaporation of aromatic residues. Mass fraction curves are shifted vertically to an ion current
base level of 5E  12 A.

resonances of the carboxyl function and ester group disappear degradation of PVAc, but also a third step is introduced. The
and a new very intense peak appears at 130 ppm, which is typ- first degradation step in oxidative conditions is identical to
ical for carbon double bonds. The resonance at 40 ppm re- the inert degradation: the deacetylation is similar in both envi-
mains. Spinning side bands of the intense C]C peak also ronments, as can be seen by the evaporation of acetic acid (m/e
arise at 200 and 60 ppm (overlapping with some residual 60) and its fragments (Fig. 1B II). To evaluate the deacetyla-
CeO resonance at 70 ppm). A similar spectrum is obtained tion step under oxidative conditions, solid-state 13C NMR mea-
for the sample scanned at 400  C. Both experiments indicate surements were also performed on polymer partially degraded
the decomposition of PVAc into a highly regular unsaturated at 280 and 400  C (Fig. 2B). Since no CeO bond resonances
material or polyene during deacetylation. are observed when the deacetylation is fully completed, a mix-
In an inert environment the formed polyene fully degrades ture of mostly aliphatic and unsaturated Cebonds is found and,
in a subsequent degradation step between 400 and 500  C like in the inert degradation, a polyene structure is retained.
through chain scission reactions. In this temperature range The next degradation step (between 400 and 500  C) can be
the only m/e intensities found with TGAeMS are the unsatu- characterised, besides evaporating aromatic residues (see
rated and aromatic residues (Fig. 1A III) with also large frag- Fig. 1B III) and some CO2 production (Fig. 1B II), by the evo-
ments like anthracene (m/e 178). lution of m/e 17 and 18. A comparison is made from their
Comparison of the inert thermal degradation and thermo- evaporation pattern under inert and oxidative conditions as
oxidative degradation of PVAc (Fig. 1A I and B I) reveals shown in Fig. 3. After deacetylation, of which OH (m/e 17)
two major differences: not only the temperature endset of and H2O (m/e 18) also are side products, the evaporation of
the thermo-oxidation is higher than that of the thermal inert m/e 17 and 18 reaches a new maximum under airflow,
 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810 803

A B
H H 4
1 2
C C

Room O
3 H n Room
Temperature 2
O
Temperature
3C 1
4 CH
3 PVAc

Isothermal
at 280 °C Isothermal
at 280 °C

polyene

Scan to 400 °C Scan to 400 °C

polyene

240 200 160 120 80 40 0 ppm

Scan to 500 °C

char

240 200 160 120 80 40 0 ppm

Fig. 2. Solid-state 13C NMR of non-degraded (room temperature) and partially degraded PVAc under (A) inert and (B) oxidative conditions. Samples were heated
at 20  C min1 to reach the desired temperature.

indicating a dehydrogenation reaction. As seen in the 13C
NMR spectra, no oxygen is left in the polyene structure,
thus the formation of H2O and OH can only be explained by
additional subtraction of hydrogen atoms from the polymer
chain by oxygen present in the gas phase. This second step
is assigned as the charring step in an oxidative environment
of PVAc.
Evaporation of aromatic residues (Fig. 1B III) occurs be-
tween 400 and 500  C, but not in the ultimate degradation
step above 500  C. This is an indication for the formation of
a thermally stable product during the second decomposition
step. Comparison of the product formed isothermally at
500  C with the polyene structure retained after deacetylation
by 13C NMR shows that all aliphatic C-bonds have disappeared
(Fig. 2B). Only the unsaturated resonance at 200 ppm and its
spinning side bands are detected, indicating a fully aromatised
structure or char. Note that in this case a ‘‘char’’ is defined as
a fully aromatised stable material formed during a thermo-
oxidative degradation.
The last degradation step, above 500  C, is fully character-
ised by the evaporation of mass fraction 44 (CO2, see Fig. 1B II),
indicating the oxidation of the aromatic structures formed
during the charring step. No water and OH formation is de-
tected in this last step, see Fig. 3. The TGA results show
that, due to the aromatisation process, PVAc forms a more sta-
ble material during degradation under oxidative environments
than in inert conditions. Therefore, PVAc can be regarded as
a useful material for flame retardant composites.
With the combination of thermogravimetric and differential
thermal analysis (TGAeDTA) qualitative information can be
804 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810

m/e 18
3.2. Isothermal deacetylation of PVAc

2E-9 A To develop the degradation model, the deacetylation step of

PVAc was measured isothermally at 280  C in an oxidative
and inert environment. Note that this temperature is well be-
Ion Current / A

low the onset temperature of the deacetylation when measured

1E-9 A
200 300 400
Temperature / °C
Fig. 3. Selected MS signals m/e 17 and 18 of the degradation of PVAc at a heat-
ing rate of 20  C min1 under inert (d) and oxidative conditions ( ) as mea-
sured with TGAeMS at 20  C min1.
obtained on the heat released or absorbed by the polymer dur-
ing the degradation step. Under inert conditions the black
curve in Fig. 4, PVAc, shows a small endothermic effect dur-
ing deacetylation. In an oxidative environment, the grey curve
in Fig. 4 shows a large exothermal effect during each degrada-
tion step of PVAc. The degradation of the char by environmen-
tal oxygen into CO2 is related to the most prominent
exothermal signal in TGAeDTA. It means that in extreme
conditions like fire, PVAc clearly adds extra heat to its sur-
roundings during degradation.
100
75
A 2.5
non-isothermally at a heating rate of 20  C min1. The weight
m/e 17 signal was rescaled from 0 to 100 wt.%, calculated by Eq. (1),
with mb the initial mass, me the mass at the end of the deace-
tylation step, and mt the measured weight at time t. The rescal-
ing of the derivative of weight, d(weight)/dt, is calculated by
Eq. (2). Note that the data for the degradation step are repre-
sented in a normalized graph according to the relationship:
500 600
conversion ¼ 100  wt%.
ðmt  me Þ
Weight ð%Þ ¼  100% ð1Þ
ðmb  me Þ
dðweight ð%ÞÞ 1 dðmt Þ
 ¼   100% ð2Þ
dt ðmb  me Þ dt
Weight (%) and rescaled derivative of weight of the deace-
tylation process are plotted in Fig. 5. The deacetylation
process of PVAc has an autocatalytic behaviour, since the deg-
radation rate increases over time towards a maximum. The
presence of environmental oxygen gives rise to a faster deace-
tylation process since the maximum rate occurs earlier in the
isothermal segment.
In literature, two different reactions are used to describe
the autocatalytic deacetylation of PVAc in inert conditions
[9]. First, virgin PVAc undergoes a non-catalytic deacetyla-
tion reaction. The elimination of any following vinyl acetate
(VAc) entity in the polymer main chain is catalysed by the
3 conjugation of the p-electrons of the first formed double
bond. This is illustrated in Scheme 1. The catalysing double
Der. Weight / %.°C-1

bond is denoted as an active unsaturation in the polymer

50 2 main chain, Uact. After the elimination of acetic acid in a cat-
Weight / %

alytic deacetylation reaction, a new active double bond Uact

25 1.5
is formed and the former catalysing double bond is
0 1

-25 0.5 100

2
Difference Temperature / °C.mg-1

-50 0 80
-d(weight)/dt / %.min-1

1.35
B 60
1.5
1.1
Weight / %

40
0.85
20 1
0.6

0.35 0
exo 0.5
0.1

-0.15 0
200 300 400 500 600 700 0 50 100 150 200 250 300 350 400
Temperature / °C Time / min

Fig. 4. Combined TGAeDTA of PVAc in an inert ( ) and oxidative environ- Fig. 5. Rescaled weight and derivative of weight to time of the isothermal de-
ment ( ) at 20  C min1. (A) Weight loss and derivative weight loss signal; acetylation at 280  C of PVAc in inert (black curve) and oxidative conditions
(B) difference temperature curve. (grey curve). Both samples were heated at 20  C min1 to reach 280  C.
 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810 805

CH3 CH3 CH3

O O O
2
H O H O H 1O Uact
3

. ..

. ..

. ..
. ..
CH CH CH CH CH CH HC CH

CH CH HC CH HC CH HC CH
acetic acid acetic acid

. ..

. ..

. ..
2
. ..

H 3 O Uact Uact UB
1
O

CH 3
catalytic process
non-catalytic

Scheme 1. Proposal for the non-catalytic deacetylation and its catalytic propagation in the polymer chain of PVAc, Uact stands for the double bond catalysing the
deacetylation of the neighbouring VAc entity, UB for an inactive double bond.

simultaneously deactivated and therefore denoted as UB (see
Scheme 1).
The deactivation process from Uact into UB also occurs by
reaching the chain ends or an already deacetylated entity [8],
and is added in this work as a third step to describe the auto-
catalytic deacetylation process (see further for Section 3.4). A
representation of the non-catalytic, catalytic, and deactivation
processes occurring during deacetylation of PVAc is shown in
Scheme 2. Note that the catalytic reactions are assumed to
occur unidirectional [8], indicated by horizontal arrows in
Scheme 2. When the deacetylation is completed the remaining
polyene consists only of UB entities.
3.3. The degradation steps of poly
(ethylene-co-vinyl acetate)
As mentioned in Section 1, EVA copolymers with vinyl ac-
etate content below 50 wt.% (semi-crystalline EVA) and PVAc
degrade according to a similar pattern [1e7]. Also for higher
vinyl acetate contents this is the case, as demonstrated by
TGAeMS and 13C NMR in Figs. 6e8. TGAeMS measure-
ments of EVA 9, EVA 19, EVA 33, EVA 60, EVA 73, and
PVAc are shown for inert and oxidative conditions in Figs. 6
and 7, respectively. The most significant mass fractions are de-
picted. Deacetylation occurs for all (co)polymers in the tem-
perature region between 300 and 400  C, as is seen by the
m/e 60 trace. After an inert or oxidative deacetylation of
EVA 60 and EVA 73, respectively, a polyene is formed, as
shown by 13C NMR in Fig. 8.
The inert degradation of the polyene of all EVA copolymers
is characterised by the balance between the formation of aro-
matic and aliphatic volatiles, CO2 and H2O. The aromatic vol-
atiles (represented by m/e 78) originate from deacetylated VAc
entities; aliphatic volatiles are formed by chain scission reac-
tions of ethylene entities and are represented by m/e 70
(C5H10). When the amount of VAc increases in the (co)poly-
mers, the aromatic traces become more important than the
aliphatic ones.

VAc VAc VAc VAc VAc VAc VAc VAc VAc VAc VAc VAc

Non-catalytic
deacetylation

VAc VAc Uact VAc VAc VAc VAc VAc VAc VAc VAc VAc
Catalytic Non-catalytic
deacetylation deacetylation

VAc Uact UB VAc VAc VAc VAc Uact VAc VAc VAc VAc

Catalytic Non-catalytic Catalytic

deacetylation deacetylation deacetylation

Uact UB UB VAc Uact VAc VAc UB Uact VAc VAc VAc

Deactivation at Catalytic Catalyitc

chain ends deacetylation deacetylation

UB UB UB Uact UB VAc VAc UB UB Uact VAc VAc

Deactivation at Catalytic
deacetylated sequence deacetylation

UB UB UB UB UB VAc VAc UB UB UB Uact VAc

UB UB UB UB UB UB UB UB UB UB UB UB polyene

Scheme 2. Schematic representation of the deacetylation of PVAc. VAc stands for vinyl acetate entities, Uact for the double bonds catalysing a following deace-
tylation reaction, and UB for formed inactive double bonds. The horizontal arrows indicate the direction of the unidirectional catalytic process.
806 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810

100 6 100 4

A I B I
3.5
75 5 75

Der. Weight / %.°C-1

Der. Weight / %.°C-1
3
50 4 50

Weight / %
Weight / %

2.5

25 3 25 2

EVA 9 1.5
0 2 0 EVA 60
EVA 19 EVA 73 1
EVA 33 1 -25 PVAc
0.5

0 -50 0
200 250 300 350 400 450 500 550 600 650 200 250 300 350 400 450 500 550 600 650
Temperature / °C Temperature / °C

1.E-09 1.0E-08
II: m/e 60 II: m/e 60
1.0E-09
1.E-10
Ion Current / A

Ion Current / A
1.0E-10
1.E-11
1.0E-11

1.E-12
1.0E-12

1.E-13 1.0E-13
III: m/e 70 III: m/e 70
1.E-09
1.E-10
Ion Current / A

Ion Current / A

1.E-10
1.E-11
1.E-11

1.E-12
1.E-12

1.E-13 1.E-13
IV: m/e 78 IV: m/e 78
1.E-09
1.E-10
Ion Current / A

Ion Current / A

1.E-10
1.E-11
1.E-11

1.E-12
1.E-12

1.E-13 1.E-13
200 250 300 350 400 450 500 550 600 650 200 250 300 350 400 450 500 550 600 650
Temperature / °C Temperature / °C

Fig. 6. TGAeMS at 20  C min1 under inert conditions of (A) (semi-crystalline): EVA 9, EVA 19, and EVA 33; (B) (amorphous): EVA 60, EVA 73 and PVAc.
Weight loss and derivative weight loss signal in I, selected m/e traces: 60 (II), 70 (III), and 78 (IV).

In oxidative conditions after deacetylation semi-crystalline the same temperature regions (Fig. 7B). Moreover, the chem-
EVA degrades in a single degradation step between 400 and ical structure of the residue after charring, as determined by
13
550  C (Fig. 7A). The peak value of the degradation rate of C NMR (Fig. 8B for EVA 73), is assigned to a fully aroma-
this step shifts to higher temperatures with increasing amount tised residue or char, as for PVAc.
of VAc (Fig. 7A I). These copolymers do not form a stabilised In Fig. 9, isothermal TGA measurements at 300  C of EVA
aromatic structure or char as was observed for PVAc. The ox- 33, EVA 60 and EVA 73 are compared to PVAc. By the intro-
idative degradation of the polyene formed after deacetylation duction of ethylene entities in the polymer backbone, the auto-
of EVA 60 and EVA 73 on the other hand is similar to the deg- catalytic effect becomes less pronounced and even disappears:
radation of PVAc, with typical mass fractions evaporating in the derivative of weight of EVA 73 still shows some increase
 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810 807

A 100 8 B 100 3
I I
7
75 75 2.5

Der. Weight / %.°C-1

Der. Weight / %.°C-1

EVA 60
6
EVA 73
50 50 2
Weight / %

Weight / %
5 PVAc

25 4 25 1.5
EVA 9 3
0 EVA 19 0 1
EVA 33 2
-25 -25 0.5
1

-50 0 -50 0
200 250 300 350 400 450 500 550 600 650 200 250 300 350 400 450 500 550 600 650
Temperature / °C Temperature / °C
2.E-11 1.0E-08
Current / A

Current / A
II: m/e 60 1.0E-09 II: m/e 60
1.E-11
Ion

Ion
1.0E-10
5.E-12 1.0E-11
1.0E-12
1.0E-13
Current / A

III: m/e 18 III: m/e 18

Current / A
7.E-09 Ion 1.8E-08
Ion

1.6E-08
5.E-09
1.4E-08
1.2E-08
IV: m/e 44 1.0E-08 IV: m/e 44
Current / A

Current / A

2.E-08
Ion

Ion

1.0E-09
1.E-08

5.E-09 1.0E-10
V: m/e 70 V: m/e 70
Current / A

Current / A

1.E-10
2.00E-11
Ion
Ion

1.E-11
1.00E-11
1.E-12
0.00E+00
VI: m/e 78 VI: m/e 78
Current / A

Current / A

8.10E-12 1.E-11
Ion

Ion

4.10E-12 1.E-12

1.00E-13 1.E-13
200 250 300 350 400 450 500 550 600 650 200 250 300 350 400 450 500 550 600 650
Temperature / °C Temperature / °C

Fig. 7. TGAeMS at 20  C min1 under oxidative conditions of (A) (semi-crystalline): EVA 9, EVA 19, and EVA 33; (B) (amorphous): EVA 60, EVA 73, and
PVAc. I e weight loss and derivative weight loss signal; representative m/e traces 60 (II), 18 (III), 44 (IV), 70 (V), and 78 (VI).

over time, while EVA 60 and EVA 33 do not show any auto- ethylene entity can be regarded as such a deficiency. A sche-
catalytic behaviour. matic representation of the deacetylation reaction in an EVA
The autocatalytic deacetylation reaction can only occur copolymer is shown in Scheme 3.
when sequences of two or more VAc monomer units undergo
deacetylation. EVA 33 has a reported number-average chain 3.4. Mechanistic description of the degradation steps
length of VAc entities between 1.14 and 1.21 [22]. This indi-
cates that almost all VAc entities occur in an EeVAceE se- A mechanistic description valid for the full degradation
quence, with E standing for an ethylene entity. Therefore no path of PVAc and EVA copolymers is summarized in Scheme 4.
catalytic reactions are expected in this case. With an increas- It incorporates the three major degradation steps: deacety-
ing amount of VAc in the copolymers, the number-average lation with formation of a polyene (A), degradation of the
VAc sequence length increases and as such the autocatalytic polyene under inert conditions (B), and formation and degra-
nature of the deacetylation. The influence of the amount of dation of a char under oxidative conditions (C).
ethylene entities on the rate of the deacetylation process can The autocatalytic process of the deacetylation is described
be expressed by a similar deactivation process as for PVAc, by a non-catalytic reaction step producing acetic acid (Ac) and
i.e. the encounter of a deficiency in the VAc sequence. Any an active double bond (Uact), in combination with a catalytic
808 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810

A Inert (EVA 60) B Oxidative (EVA 73)

Isothermal at 280 °C Isothermal at 280 °C

polyene

Scan to 400 °C Scan to 400 °C

polyene

240 200 160 120 80 40 0 ppm 240 200 160 120 80 40 0 ppm

Scan to 500 °C

char

280 240 200 160 120 80 40 0 ppm

Fig. 8. Solid-state 13C NMR of partially degraded residues of EVA 60 (inert deacetylation) and EVA 73 (oxidative deacetylation followed by charring). The
samples were heated at 20  C min1 to reach the desired temperature.

one producing Ac and an inactive double bond (UB), Scheme

100 4A. To describe any interruption in the autocatalytic process,
90 i.e. the encounter of a deficiency in the VAc sequences by
80 EVA 73 reaching a chain end or an already deacetylated entity, a (third)
70 EVA 60 deactivation process is added. This step converts Uact into UB
Weight / %

60 EVA 33 without consumption of VAc units. The rate of deactivation

50 against catalytic reaction increases rapidly if more ethylene
40 units (E) are present in the EVA copolymer.
30 The polyene formed after deacetylation of PVAc and EVA
20
PVAc
copolymers consists of a combination of deacetylated VAc or
10 unsaturated entities UB and ethylene entities (E), depending
0 on the initial (co)polymer composition. The inert degradation
3.5 PVAc of this polyene by chain scission processes is described mech-
anistically by two parallel reaction steps (Scheme 4B). The
-d(weight)/dt / %.min-1

3
UB entities degrade into aromatic volatiles and ethylene enti-
2.5 EVA 73 ties into aliphatic volatiles. The balance between aromatic and
2 EVA 60 aliphatic volatiles evolved in this step thus depends on the ini-
1.5
tial amount of VAc in the (co)polymers; a polyene resulting
EVA 33 from the deacetylation of PVAc consists only of UB entities
1
and therefore only aromatic residues are formed.
0.5 The oxidative degradation of the polyene formed out of
0
PVAc, EVA 73 and EVA 60 occurs in two consecutive steps.
0 25 50 75 100 125 150 During the charring step an aromatised solid residue or char
Time / min is formed from the UB entities (designated as AB), with the
Fig. 9. Rescaled weight and derivative of weight to time of the isothermal de- production of OH, H2O, and aromatic volatiles (Scheme 4C).
acetylation at 300  C of PVAc, EVA 73, EVA 60, and EVA 33; time zero was During charring the ethylene entities undergo chain scission
chosen at the start of the isothermal segment. reactions into aliphatic volatiles. The solid residue AB
 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810 809

VAc E VAc VAc VAc E VAc E VAc VAc E VAc

Non-catalytic
deacetylation

VAc E VAc Uact VAc E VAc

VA E VAc VAc E VAc
Catalytic Non-catalytic
deacetylation deacetylation

VAc E Uact UB VAc E Uact E VAc VAc E VAc

Deactivation at Deactivation at
ethylene entity ethylene entity

VAc E UB UB VAc E UB E VAc VAc E VAc

UB E UB UB UB E UB E UB UB E UB polyene

Scheme 3. Schematic representation of the deacetylation of EVA copolymer. Uact stands for active double bonds catalysing the deacetylation of the neighbouring
vinyl acetate entity (VAc). E stands for ethylene entities and UB for formed inactive double bonds. The horizontal arrows indicate the unidirectional catalytic
process.

remaining after charring degrades into CO2 in a consecutive
step. Again, the balance between the amount residue AB
and the production of aliphatic volatiles depends on the initial
amount of VAc in the (co)polymer. In the case of PVAc no
chain scission reactions of ethylene sequences occur and the
maximum amount of AB residue is left after charring. EVA
copolymers with low weight% VAc have no charring step,
the reaction E / aliphatic volatiles and the oxidation reaction
occur simultaneously, as observed in Fig. 7.
Scheme 4 will be used for the quantitative modelling of the
degradation kinetics of PVAc and EVA copolymers in a follow-
ing paper [21].
4. Conclusions
By combining information on both volatiles and solids
formed as a function of temperature by means of TGAeMS,
TGAeDTA and solid-state analysis of partially degraded sam-
ples (13C NMR), a complete degradation model of poly(vinyl
acetate) and poly(ethylene-co-vinyl acetate) copolymers with
a vinyl acetate weight% ranging from 9 to 73% is obtained.
Regardless of the environment, these (co)polymers undergo
a deacetylation reaction, which is autocatalytic for sufficient
amounts of vinyl acetate in the polymer backbone. The auto-
catalytic reaction is described in the model by three processes.
A deactivation process is introduced enabling the quantifica-
tion of deacetylation. Under inert conditions, the polyene
structure formed by the deacetylation undergoes chain scission
reactions to reach full degradation. The polyene formed out of
PVAc, EVA 73 and EVA 60 forms a char in an oxidative en-
vironment, which is oxidised into CO2 above 500  C. The
model takes the experimental considerations into account, en-
abling quantification of the degradation of all (co)polymers.
Moreover, an analogous model will be extended to describe

A Deacetylation and formation of polyene

VAc (s) Uact (s) + Ac (g) Non-catalytic deacetylation
VAc (s) + Uact (s) → Uact (s) + UB (s) + Ac (g) Catalytic deacetylation
Uact (s) UB (s) Deactivation

B Inert degradation of polyene

Chain scission reactions
UB (s) aromatic volatiles
E (s) aliphatic volatiles

C Oxidative degradation of polyene

Charring
UB (s) AB (s) + OH (g), H2O (g), aromatic volatiles
E (s) aliphatic volatiles
Oxidation of the char
AB (s) CO2 (g)

Scheme 4. Proposed mechanistic model for deacetylation (A), inert polyene degradation (B) and oxidative polyene degradation (C) of PVAc and EVA. VAc: vinyl
acetate entities; Uact: double bonds catalysing a following deacetylation reaction; UB: deacetylated vinyl acetate entities or inactive double bonds; Ac: acetic acid; E:
ethylene entities; AB: aromatised solid residue or char; (s): solid state; (g): gas phase. Note: (C) is only valid for EVA copolymers with high weight% VAc and PVAc.
810 B. Rimez et al. / Polymer Degradation and Stability 93 (2008) 800e810
and quantify the interactions of the polymer model systems
with solid-state flame retardants.
Acknowledgements
TA Instruments (Benelux) and Agfa-Gevaert (Belgium) are
kindly acknowledged for the use of TGAeMS and TGAe
DTA.
References
[1] Zanetti M, Camino G, Thomann R, Mülhaupt R. Synthesis and thermal
behaviour of layered silicateeEVA nanocomposites. Polymer 2001;42:
4501.
[2] Riva A, Camino G, Fomperie L, Amigouët P. Fire retardant mechanism
in intumescent ethylene vinyl acetate compositions. Polym Degrad Stab
2003;82:341.
[3] Allen NS, Edge M, Rodriguez M, Liauw CM, Fontan E. Aspects of the
thermal oxidation, yellowing and stabilisation of ethylene vinyl acetate
copolymer. Polym Degrad Stab 2001;71:1.
[4] Allen NS, Edge M, Rodriguez M, Liauw CM, Fontan E. Aspects of the
thermal oxidation of ethylene vinyl acetate copolymer. Polym Degrad
Stab 2000;68:363.
[5] Häussler I, Pompe G, Albrecht V, Voigt D. Bestimmung des vinylacetat-
gehaltes in EVA copolymeren. J Therm Anal 1998;52:131.
[6] Costache MC, Jiang DD, Wilkie CA. Thermal degradation of ethylene-
vinyl acetate copolymer nanocomposites. Polymer 2005;46:6947.
[7] Carpentier F, Bourbigot S, Le Bras M, Delobel R, Foulon M. Charring of
fire retarded ethylene vinyl acetate copolymer e magnesium hydroxide/
zinc borate formulations. Polym Degrad Stab 2000;69:83.
[8] Simon P. Polymer degradation by elimination of small molecules. Angew
Makromol Chem 1994;3876:187.
[9] Mazon-Arechederra JM, Sanchez-Chavez M, Arranz F, Barrales-
Rienda JM. Cinetica de la degradacion termica del poliacetato de vinilo.
Anales Quim 1982;78:189.
[10] Lüftl S, Archodoulaki VM, Seifler S. Thermal-oxidative induced degra-
dation behaviour of polyoxymethylene (POM) copolymer detected by
TGA/MS. Polym Degrad Stab 2006;91:464.
[11] Snyder A, Tripathi A, Dworzanski JP, Maswadeh WM, Wick CH. Char-
acterization of microorganisms by thermogravimetric analysisemass
spectrometry. Anal Chim Acta 2005;536:283.
[12] Zhai RS, Das A, Hsu CK, Han CC, Canteenwala T, Chiang LY, et al.
Polymeric fullerene oxide films produced by decomposition of hexani-
tro[60]fullerene. Carbon 2004;42:395.
[13] Shibasaki Y, Yang M. Thermal degradation mechanism of methacryloni-
trileestyrene copolymers on flash pyrolysis. J Therm Anal 1997;49:71.
[14] Russel JD, Kardos JL. Modeling the imidization kinetics of AFR700B
polyimide. Polym Compos 1997;18:64.
[15] Groenewoud WM, de Jong W. The thermogravimetric analysereFourier
transform infrared/mass spectrometry technique. Thermochim Acta
1996;286:341.
[16] Keuleers R, Janssens J, Desseyn HO. Thermal analysis of oxamates, thi-
ooxamates and their complexes. Part 2. The Cu(II) complexes. Thermo-
chim Acta 1998;311:155.
[17] Bugajny M, Bourbigot S, Le Bras M, Delobel R. The origin and nature of
flame retardance in ethylene-vinyl acetate copolymers containing hosta-
flam AP 750. Polym Int 1999;48:264.
[18] Xie R, Qu B. Synergistic effects of expandable graphite with some hal-
ogen-free flame retardants in polyolefin blends. Polym Degrad Stab
2001;71:375.
[19] Eychenne V, Mouloungui Z, Gaset A. Thermal behaviour of neopentyl-
polyol esters. Comparison between determination by TGAeDTA and
flash point. Thermochim Acta 1998;320:201.
[20] Salam LA, Matthews RD, Robertson H. Pyrolysis of polyvinyl butyral
(PVB) binder in thermoelectric green tapes. J Eur Ceram Soc 2000;20:
1375.
[21] Rimez B, Artoos T, Van Assche G, Van Mele B. The thermal degrada-
tion of poly(vinyl acetate) and poly(ethylene-co-vinyl acetate): Part II.
Modelling the degradation kinetics. Polym Degrad Stab, submitted for
publication.
[22] Bugada DC, Rudin A. Molecular structure and melting behaviour of
ethylene-vinyl acetate copolymers. Eur Polym J 1992;28:219.