Content uploaded by Christian Laurent
Author content
All content in this area was uploaded by Christian Laurent on Apr 17, 2015
Content may be subject to copyright.
On the nature of the luminescence emitted by a polypropylene film after interaction
with a cold plasma at low temperature
F. Massines, P. Tiemblo, G. Teyssedre, and C. Laurent
Citation: Journal of Applied Physics 81, 937 (1997); doi: 10.1063/1.364186
View online: http://dx.doi.org/10.1063/1.364186
View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/81/2?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Luminescence quenching in low doped polymer light-emitting diodes
J. Appl. Phys. 104, 093513 (2008); 10.1063/1.3006442
Contribution of bottom-emitted radicals to the deposition of a film on the SiO 2 sidewall during CHF 3 plasma
etching
J. Vac. Sci. Technol. A 23, 713 (2005); 10.1116/1.1946712
Effect of inductively coupled plasma damage on performance of GaN–InGaN multiquantum-well light-emitting
diodes
Appl. Phys. Lett. 86, 102104 (2005); 10.1063/1.1882749
Significant improvement of electrical and thermal properties of low dielectric constant plasma polymerized
paraxylene thin films by postdeposition H 2 + He plasma treatment
J. Appl. Phys. 89, 1402 (2001); 10.1063/1.1335826
Polymer solution light-emitting devices
Appl. Phys. Lett. 74, 2081 (1999); 10.1063/1.123764
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
On the nature of the luminescence emitted by a polypropylene film
after interaction with a cold plasma at low temperature
F. Massines,a) P. Tiemblo, G. Teyssedre, and C. Laurent
Laboratoire de Ge
´nie Electrique, Universite
´Paul Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex,
France
~Received 18 September 1996; accepted for publication 14 October 1996!
The light emitted by an insulating material once its surface has been submitted to a cold plasma,
plasma-induced luminescence is investigated on polypropylene films at low temperature. An
analysis of the integral and wavelength resolved light is carried out. The kinetical evolution of the
spectrum after plasma interaction are reported and analyzed. Investigation of the photo- and
chemiluminescence spectra of the material shows that plasma-induced luminescence has three
spectral components, each having different excitation mechanisms and thus different kinetics. The
fastest is due to the ultraviolet irradiation during plasma contact, the second is dominated by radical
chemistry producing carbonyl groups, and the third, with the slowest kinetics, is due to carrier
recombination on the most conjugated chromophores which are preferentially charged by the
plasma. To confirm the interpretation of plasma induced luminescence spectrum, the first results
concerning short and long plasma interaction are considered. © 1997 American Institute of
Physics. @S0021-8979~97!05902-1#
I. INTRODUCTION
Surface treatments by cold plasma are being increasingly
used as a nonpolluting method for surface transformations.
The interaction of a cold discharge with a polymer induces
several kinds of transformations such as the formation of
new chemical groups, crosslinking, formation of short chains
and electron, as well as ion implantation. These modifica-
tions can be considered as degradation or, when they are
controlled, as surface properties improvement to increase ad-
hesion, for example. In both cases, it is important to identify
these transformations directly in the treatment chamber right
from the initial stages of the process to its end.
It has been shown that after the end of the interaction
with the plasma, luminescence of the polymer is
measured.1–7 Light emission has been observed with plasmas
obtained in various conditions:1–5 at high and low pressures,
with radio- and low-frequency excitations, in O2,N
2
, Ar,
H2, and He. Polymer films with or without chromophores in
their monomer unit and thin-film deposits have been
analyzed.1–3,5,6 Since plasma-induced luminescence ~PIL!is
easy to detect in situ in the plasma cell, its measurement
could be the basis for a powerful diagnostic tool for surface
transformations induced by cold plasma. The development of
such a tool implies the establishment of correlations between
luminescence features and surface chemical and physical
properties. This is the aim of the present article.
Most works regarding plasma-induced luminescence
concern thermoluminescence. Plasma treatment is done at
liquid-nitrogen temperature, then the sample is heated and
the luminescence intensity recorded as the temperature in-
creases. Comparing PIL and radio-thermoluminescence,2,3,5
most authors attribute luminescence to electron–hole radia-
tive recombination, with thermoluminescence peaks mostly
related to polymer relaxations. The mechanism involved is
the following:4
~i!Treatment leads to charge creation and trapping;
~ii!electrons are released from the traps by simple ther-
mal detrapping or by trap disappearance since traps in
polymers are related to specific chain configurations
which are temperature dependent;
~iii!electrons recombine with luminescent centers in the
polymer which were positively charged during treat-
ment.
Recently, surface luminescence has been measured in
situ in the plasma cell.4–7 This procedure allows the analysis
of the isothermal PIL transient. The influence of treatment
duration and plasma atmosphere on luminescence decay of
polyethylene has been reported.5At 77 K, luminescence is
still measurable 30 min after treatment4and spectra of dif-
ferent polymers have been recorded just at the end of the
plasma.7The main emission component has been associated
with the polymer phosphorescence or fluorescence depend-
ing on the chemical composition of the monomer. Polyeth-
ylene has been more frequently studied6and a second com-
ponent, which peaked at a higher wavelength than
photoinduced phosphorescence, has been observed but not
explained. Up to now, isothermal PIL has not been clearly
attributed to any particular mechanism. Only the following
hypotheses have been assumed:4During and after the inter-
action with the discharge the polymer and more specially its
surface is out of equilibrium, i.e., unstable chemical radicals
are created and electrical charges are implanted. Carrier re-
combinations and chemical transformations induce a release
of energy which is partly transformed into light by the poly-
mer chromophores.
The aim of this work is to identify the origin of the
chromophores involved in the emission and their excitation
processes: Are they molecular entities already present in the
fresh material or new chemical species created by plasma
a!Electronic mail: massines@lget.ups-tlse.fr
937J. Appl. Phys. 81 (2), 15 January 1997 0021-8979/97/81(2)/937/7/$10.00 © 1997 American Institute of Physics
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
interaction? Is luminescence induced by UV irradiation dur-
ing the discharge? Is it due to radiative electron–hole recom-
bination or to chemiluminescence involving formation of
emitting entities during chemical reactions? To shed some
light on these questions, we studied PIL decay and spectral
features in polypropylene as a function of the time elapsed
from the plasma–polymer interaction. To identify the differ-
ent spectrum components, we compared PIL with photo- and
chemiluminescence of the same polymer. Then the first re-
sults on the effect of treatment duration are presented and
analyzed.
II. METHODOLOGY AND MATERIAL
The PIL measurements were done with the sample in the
plasma treatment chamber. This avoided exposure of the
sample to air which is often made during its transfer from the
treatment cell to the surface analysis system. The discharge
was a silent discharge obtained between two plane parallel
electrodes. The sample was deposited on the lower one and
the upper one was coated with a layer of quartz in such a
way that the discharge was initiated between two dielectrics.
The plasma gap of 5 mm was powered at a frequency of 5.5
kHz under a voltage of about 1.5 kV rms. The discharge was
produced in helium at atmospheric pressure. Before intro-
ducing the helium, the chamber was pumped down to 1025
Pa. The sample temperature was variable from 2180 to
1170 °C. At the end of the plasma generation the upper
electrode was removed so that the luminescence of the poly-
mer surface could be analyzed in situ. The delay between the
end of treatment and the beginning of measurement was
about 2 s.
Integral and spectral light analysis were performed si-
multaneously. Wavelength-resolved detection was carried
out using a liquid-nitrogen-cooled charge-coupled-device
~CCD!camera and a low-dispersive monochromator ~4.5 nm
in resolution!. Coupling between the polypropylene surface
and optical detectors was achieved with quartz lenses. The
minimum time lag between two spectra was 0.5 s. Integral
detection was done with a photomultiplier operating in pho-
ton counting mode. Optical analysis covered the range be-
tween 300 and 800 nm. Photoluminescence was studied in
the same cell. The sample was excited by a xenon source
coupled to an irradiation monochromator. The excitation
wavelength was varied from 250 to 310 nm. The bandwidth
of the irradiation window was about 35 nm. The detection
devices were the same as those used for plasma-induced lu-
minescence. Finally chemiluminescence, heating the sample
in an oxidant atmosphere, was also done in this reactor.
All the results presented here were obtained with a com-
mercially available biaxially oriented isotactic 25-
m
m-thick
polypropylene film. Additive free polypropylene was also
characterized by photoluminescence to determine the anti-
oxidant contribution to PIL measurements. Samples were
consecutively treated five times for 5 s each and finally for 3
min. Luminescence measurements were made after each
treatment, as long as signal was observable. Treatments as
well as measurements were performed at a temperature of
2180 °C.
III. PLASMA-INDUCED LUMINESCENCE ANALYSIS
A. Spectrum features
Whatever the treatment conditions at 2180 °C, the emit-
ted light level allows measurements with the CCD camera
over 20 min after the end of the discharge provided the ex-
posure time is changed from 1 s during the first8sto300s
15 min later. Typical evolution of the spectra with time
elapsed from the end of the plasma is shown in Figs. 1 and 2.
Of course, the magnitude decreased with time ~Fig. 1!but its
shape also changed. To follow this evolution, the amplitudes
were normalized at 400 nm ~I4005100!in Fig. 2. During the
first 5 s after the treatment, the luminescence peak with a
maximum at 400 nm becomes narrower with time as if a
component centered at about 450 nm decreased faster than
the main one. This point is confirmed later on. Thereafter,
the amplitude of the 400 nm component decreases ~Fig. 1!
and another component appears around 500 nm. The lifetime
of the 400 nm component calculated from the peak decay is
about 5 s. Note that the peak corresponding to the so-called
‘‘400 nm component’’ is not symmetrical. It rises quickly
from 370 to 400 nm and then decreases slowly with still
large components around 450 nm. The overall spectrum evo-
FIG. 1. PIL spectra during the first 10 min followinga5streatment.
FIG. 2. PIL spectra during the first 8 s followinga5streatment ~normalized
to 400 nm emission!.
938 J. Appl. Phys., Vol. 81, No. 2, 15 January 1997 Massines
et al.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
lution with time lets one suppose that there are at least three
different emission mechanisms with three different kinetics.
In the following we confirm this assumption, first by looking
at the total light emission and then by comparing PIL spec-
trum with photo- and chemiluminescence spectra.
B. Kinetics
Figure 3 shows a typical luminescence decay measured
with the photomultiplier after the end of a plasma treatment.
For times shorter than 20 s, the best fit of the experimental
data by simple mathematical expressions is obtained with
two exponential functions having life times of about 2 and
5.5 s. For longer times, the best fit of the luminescence decay
is obtained with an expression of the type
I5I0~11
a
t!2m,~1!
where Iis the luminescence amplitude, I0is the amplitude at
the end of the plasma treatment, tis the time elapsed from
the end of the treatment,
a
and mare adjustable coefficients.
The order of magnitude of
a
is 0.1 s21and that of m'1.
The reciprocal time dependence of the light amplitude
can be attributed to electron tunneling to positively charged
luminescence centers.8The distribution of hole–electron pair
separation has been derived by Hama et al.9from the lumi-
nescence decay of the form given by the above equation in
the case of
g
-irradiated polyethylene terephthalate. In case of
luminescence induced by
g
irradiation, Eq. ~1!describes lu-
minescence measurements over a wide range of time scales10
~up to 10 decades!permitting an accurate determination of
the parameters. For PIL, both spectral and kinetic analysis
show that several processes act simultaneously. The tunnel-
ing mechanism becomes dominant for times longer than 30 s
which constitutes a limitation for an accurate determination
of multiparameter fits. Hence, the parameter mwas set at 1.
This implies that the excitation mechanism is electron–hole
recombination with no preferential distance between
carriers.11 It must be kept in mind that electrons and holes
are implanted during plasma treatment.
Results of integral and spectral light are in good agree-
ment. The 5.5 s lifetime agrees pretty well with the 5 s life-
time found from the amplitude decrease at 400 nm, so the
second exponential can be associated with the 400 nm com-
ponent. The I0~11
a
t!21behavior which is the main mecha-
nism for a time longer than 1 min is unequivocally associ-
ated with the spectrum measured after 40 s which peaked at
500 nm. The 2 s lifetime can be attributed to the fastest
spectrum modification, i.e., the decrease of the component
peaked at about 450 nm. The luminescence kinetics are sum-
marized in Table I.
These kinetics are in accordance with the spectra pre-
sented in earlier articles.4,6,7 In the beginning we worked4
with a photomultiplier and 50 nm width filters. The detector
was very sensitive but spectral components were recorded
for 30 s one after the other from 300 to 800 nm. Since the
400 nm filter was the third one, no signal was recorded at
this wavelength and only the 500 nm component was ob-
served. Then we used an optical multichannel analyzer with
photodiodes as sensitive elements, a detector system less
sensitive than the CCD. It was only possible to record a
single spectrum after the end of the plasma and the optimum
integration time was 30 s. When the time lag between the
end of the excitation and the beginning of the recording was
lower than 4 s the spectrum was dominated by the 400 nm
component.7For longer times the 500 nm peak was also
observed.6
C. Identification of the components of plasma-
induced luminescence
1. Photoluminescence contribution
Light emission in the visible range is due to fluorescence
or phosphorescence of the chromophores in the material.
Thus, to try to identify the PIL spectrum we carried out
photoluminescence measurements. Normally, photolumines-
cence is measured during excitation with a light source.
However, at 2180 °C, both fluorescence and phosphores-
cence are detected in polypropylene but it is not possible to
clearly separate the two. We therefore made measurements
after extinction of the lamp to eliminate any fluorescence
contribution and thus clearly distinguish phosphorescence.
UV-induced photoluminescence was therefore recorded after
as short a time as possible ~less than 0.1 s!from the end of
the light irradiation. The spectrum recorded is shown in Fig.
4. Spectra obtained with commercially available and
additive-free polypropylenes are presented. Phosphorescence
maxima were at 427, 452, and 478 nm with a shoulder at
around 510 nm. The long-wavelength part of the spectrum
was enhanced in the case of the additive-free polypropylene.
These results are in good agreement with the literature.12
Two exponentials with life times of 2.2 and 4.4 s and relative
amplitudes of 2 and 1 are necessary to describe phosphores-
FIG. 3. Time dependence of PIL followinga5streatment. Luminescence
decay is the sum of three components as indicated. Circles: experimental
data.
TABLE I. PIL spectra are the sum of three components.
Component 1 2 3
Maximum
wavelength ~nm!
450 400 500
Functional time
dependence I01 exp~2t/2!I02 exp~2t/5.5!I03/~110.1t!
939J. Appl. Phys., Vol. 81, No. 2, 15 January 1997 Massines
et al.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
cence decay. Chromophores emitting in the visible range
have at least one double bond. It is not possible to associate
a wavelength directly with a chemical group since the emis-
sion of a chromophore depends on its surroundings, but it is
well known that the emission wavelength of a polyenone
increases if the polyene length increases as if complexes ~ex-
cimer, exciplex, ...!are formed.12
Theoretically, polyolefines have no chromophores emit-
ting in the visible UV range. Nevertheless, unsaturations al-
ways occur, i.e., diene or even triene. Oxidation is also pos-
sible and when the material has additives, for example,
antioxidant, these additives can be luminescent centers.12 In
our case the contribution of antioxidant was negligible since
the results obtained with commercial and additive-free
polypropylenes were quite comparable ~Fig. 4!. The broad-
ening of the spectrum at long wavelengths in the antioxidant-
free polypropylene is in agreement with a higher degree of
oxidation. Figure 5 allows a comparison to be made between
phosphorescence measured 1 s after the excitation and PIL 5
s after short and long duration treatment. The wavelength of
the phosphorescence peak is in good agreement with compo-
nent 1 of the PIL spectrum which was assumed to peak
around 450 nm. Moreover, the lifetime of this PIL compo-
nent corresponds to that of the main photoluminescence
component. The first PIL component can therefore be attrib-
uted to bulk polypropylene phosphorescence induced by UV
irradiation during the discharge.
2. Carrier recombination contribution
The phosphorescence shoulder at 510 nm ~Fig. 4!recalls
the third component of the PIL spectrum, centered around
500 nm after a short duration treatment. It is therefore tempt-
ing to associate these two peaks with luminescence of the
same chromophores. Kinetic variation of PIL indicates that
the emission is due to charge recombination on chro-
mophores. One therefore has to explain why charge recom-
bination occurs preferentially on conjugated chromophores
which can emit light at around 500 nm. Two interpretations
can be envisaged:
~i!carriers are selectively trapped in highly delocalized
traps;
~ii!chemical transformations of the sample are such that
at the end of surface modification a large fraction of
the chromophores emitting at lower wavelengths has
been destroyed or transformed into longer poly-
enones.
Since ions arriving at the surface have a relatively low ki-
netic energy of a few eV,13 they cannot penetrate deeply into
the polymer: Charge trapping and subsequent recombination
occur in the first atomic layers. Thus, the third component
characterizes the upper surface, namely, the most trans-
formed part of the sample, i.e., the one emitting at higher
wavelengths. One the other hand, it is well known that con-
jugated species are easier to ionize and constitute efficient
hole traps. Therefore, holes would be preferentially trapped
on these structures which are also chromophores emitting at
long wavelengths when excited by recombination.
3. Chemiluminescence contribution
Figure 5 clearly shows that the 400 nm PIL component
cannot be explained by polypropylene photoluminescence.
This emission has to be fluorescence or phosphorescence of
entities which were not initially present in the polypropylene
or which were not excited by photons. To determine if there
was a signature of chemical transformations, we made lumi-
nescence measurements during the thermal oxidation of the
same polypropylene. The chemiluminescence spectrum pre-
sented in Fig. 5 was obtained in air at 165 °C. During oxi-
dation a 400 nm component became superimposed on the
polypropylene phosphorescence spectrum. This feature has
been observed14 previously. The exact nature of the light
emitting reaction is still being discussed,15,16 but all the au-
thors attribute chemiluminescence to the phosphorescence of
carbonyl group created at the final stage of chemical reac-
tions. Comparison of chemiluminescence and PIL spectra
~Fig. 5!clearly shows that the 400 nm peak in PIL is related
to the chemistry occurring during luminescence measure-
ment. In the case of plasma treatment, there are still radicals
FIG. 4. Photoluminescence spectra of commercial and additive-free poly-
propylene recorded 1 s after UV excitation ~normalized to emission maxi-
mum, excitation wavelength: 270 nm!.
FIG. 5. Comparison between photoluminescence, chemiluminescence, and
plasma-induced luminescence spectra ~normalized to emission maximum!.
940 J. Appl. Phys., Vol. 81, No. 2, 15 January 1997 Massines
et al.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
at the surface when the discharge is turned off and these
radicals are transformed into more stable entities after the
end of the treatment. Here again chemical reactions have not
been clearly defined17 but the final oxide groups are ketones,
aldehides, alcohols, and peroxides. At least the first two may
emit light when they are formed.
This discussion can be summarized in the following
way: PIL measurements have revealed three main emission
mechanisms characterized by different spectra and different
kinetics. The fastest is photoluminescence induced by UV
irradiation during the plasma. The second one is due to
chemistry occurring after the end of the plasma. Its spectrum
is rather broad but has a specific feature which is an emission
at 400 nm. This emission is certainly associated with the
carbonyl groups which are not in interaction with their envi-
ronment. The third component due to carrier recombination
is dominant for a time longer than 1 min. Its spectrum cor-
responds to the long-wavelength part of the photolumines-
cence spectrum, i.e., to the more delocalized chromophores.
To confirm our interpretation of PIL, we present the first
results concerning the effect of treatment duration.
D. Effect of treatment duration
1. Results
The amplitude of the first spectrum after treatment was
halved between the first and the second 5 s treatment and
then divided by a factor of 3 when the treatment duration
was increased to 3 min. Such behavior is consistent with
results reported in the literature.5Treatment duration changes
not only the amplitude of the spectra but also their shape. To
illustrate this effect, we compare results obtained after the
first 5 s treatment ~Figs. 1 and 2!and the 180 s one ~Figs. 6
and 7!. Long plasma interaction enlarged the spectrum and
shifted the emission maximum of the first spectrum to higher
wavelengths. For the 180 s treatment, a structure with shoul-
ders at about 431, 463, and 495 nm appeared. The fit of the
total PIL signal shows that the lifetime of the two exponen-
tial processes as well as the initial amplitude of the first
component remained about the same when treatment dura-
tion changed but the initial amplitude of components 2 and 3
were significantly modified by treatment time and the third
component kinetic decreased when treatment duration in-
creases. Since the first component peaked at 450 nm and the
second one at 400 nm, the increase of the I01/I02 ratio with
treatment duration ~see Table II!explains the shift of the PIL
spectrum maximum to higher wavelengths. The lifetime of
these components is 2 and 5.5 s, respectively, so their ratio
changes during the first 3 s after the treatment in such a way
that the peak maximum is shifted from 430 to 400 nm ~Fig.
6!. 5 s after surface activation the maxima were at about the
same wavelength ~Fig. 5!for the two durations of treatment
but the peak was broader after long interaction with the
plasma.
Spectral analysis as well as total signal measurements
showed that the area of the third component ~Table II!de-
creased after each 5 s treatment but increased after the 180 s
treatment. In addition, the peak maximum was found at
around 500 nm for all 5 s treatments and it was shifted to 600
nm after 180 s treatment because luminescence was en-
hanced around 600 nm. It is to be noticed that the last two
spectra recorded after the 180 s treatment had the same
shape. This is in accordance with kinetic measurements
showing that after a few minutes luminescence is only con-
trolled by radiative recombination of carriers.
2. Discussion
The fact that the amplitude and the kinetic of the first
component are not influenced by the treatment duration is in
accordance with our interpretation of this component. We
have associated this emission with the photoluminescence
FIG. 6. PIL spectra during the first 10 min following a 180 s treatment.
FIG. 7. PIL spectra during the fist 7 s following a 180 s treatment ~normal-
ized to emission maximum!.
TABLE II. Ratio of the three PIL components for 5 and 180 s treatment
durations.
Treatment duration
~s!I01/I02 I02/I03
5 1.612 34.1
180 2.205 48.93
941J. Appl. Phys., Vol. 81, No. 2, 15 January 1997 Massines
et al.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
induced by plasma photons. UV photons penetrate deeply
into the volume of the polypropylene, so this emission is not
really a surface emission, but a signature of the bulk of the
material. The polymer bulk is essentially unaltered by the
plasma interaction. This is confirmed by photoluminescence
spectra which are exactly the same before and after treatment
whatever the duration.
Surface chemical transformations induced by the plasma
interaction have been studied by x-ray photoelectron spec-
trometry performed on samples removed from the cell. It
was shown thata5streatment changes the surface chemistry
very little, whereas after 180 s 20% of the hydrogen and
carbon atoms are replaced by oxygen or nitrogen. A fraction
of these reactions can occur when the samples are exposed to
air, but we know that plasma treatment, even if done in he-
lium, induces oxidation of the polymer because the plasma
creates radicals and there is always some oxygen available to
react with these radicals. It could be oxygen absorbed in the
polymer, chemically linked because of oxidation during stor-
age, or due to the dissociation of adsorbed water. Therefore,
oxidation of the polypropylene during or after the plasma
treatment is always possible even in helium and at low tem-
peratures. An increase in treatment duration induced larger
chemical transformations and a shift of the third component
maximum from 500 to 600 nm which means that longer
conjugated chromophores were formed. This is in agreement
with possible chemical reactions at the surface. The longer
the treatment, the higher is the density of radicals. When the
density of radicals is low, the most probable reaction is with
an impurity like O2but when the radical density increases
interaction between radicals becomes possible and allylic
radicals may be formed. Propagation reactions of this type
induce the formation of polyene following the reactions de-
scribed below; the emission wavelength of polyene increases
with the number of double bonds;
RuCH2—C
˙—R
z
CH3
→R1uCHvC—
z
CH3
R21H•
R1uCHvC—
z
CH3
R21R•→R1—C
˙—
z
CH3
CHvC—
z
CH3
R21RH
Such a mechanism explains the influence of treatment
duration on the third component. A 3 min treatment induces
an enhancement of the third component by increasing the
emission around 600 nm as reaction between radicals is fa-
vored. On the other hand, consecutive 5 s treatments induce
a decrease of the magnitude of the 500 nm component since
these conditions are in favor of oxygen mediated reactions; it
was shown that polyolefines are less luminescent when they
are oxidized.18
Another consequence of the increase of surface transfor-
mations with treatment duration is the variation of spectrum
width. The PIL peak is narrow after a short plasma interac-
tion ~see Fig. 1!and broad during thermal oxidation or after
a long plasma interaction ~see Fig. 6!, even when the spec-
trum peaks at 400 nm, i.e., when the contribution of the main
component of UV-induced phosphorescence to the shape of
the spectra is negligible. This is related to the percentage of
surface transformations. For short treatment durations ~less
than 30 s!surface transformation was slight. There was no
interaction between radicals and since the temperature was
too low to allow movements, even of side groups, there was
no interaction between the carbonyls and the rest of the poly-
mer. This is why the emission wavelength was quite sharp.
During thermal oxidation the degree of transformation was
higher so the longer-wavelength contributions were greater.
In the same way, long interaction with plasma induced a
large number of radicals, a more complex chemistry, i.e.,
peak enlargement, and a decrease of the emission at 400 nm.
Since photoluminescence has a component with a life
time of 4.4 s, part of the second PIL component ~5 s treat-
ment!could be UV excited and not due to chemilumines-
cence. Nevertheless, the amplitude of the spectrum around
400 nm cannot be modified by photoluminescence, since
UV-induced phosphorescence of polypropylene was negli-
gible at this wavelength.
These first results on the effects of treatment duration on
PIL confirm the analysis in terms of three components. The
rapid analysis which has been done shows that PIL reflects
the degree of complexity of the surface transformations as
well as the oxidation degree and the proportion of double
bonds created. A more precise and qualitative study is nec-
essary.
V. CONCLUSION
Three different mechanisms have been distinguished in
PIL. The first is due to the phosphorescence of the film in-
duced by UV irradiation during plasma contact. This small
component decreases quickly ~lifetime: 2 s!. The second is a
chemiluminescence component peaked at 400 nm. Its char-
acteristic time is about 5 s and it is the consequence of car-
bonyl group formation. The amplitude of the peak decreases
and its width increases as treatment time increases owing to
the complexity of the chemistry involved and to the interac-
tion of carbonyls with the matrix. The high-wavelength part
of this component is influenced by photoluminescence. The
third component is due to carrier tunneling to luminescent
centers. The wavelength of the associated peak is a signature
of the final transformation of the upper surface and is very
sensitive to the creation of double bonds because they are
electrically active.
It is to be noticed that although photoluminescence spec-
tra of polypropylene films before and after interaction with a
plasma are about the same, PIL spectra are quite different
since the majority of the signal is emitted by the surface. An
increase of the treatment duration induces an enlargement of
the chemiluminescence peak with a decrease of its amplitude
and a shift to higher wavelengths of the peak excited by
electron–hole recombination.
ACKNOWLEDGMENT
Our thanks are due to A. Boulanger for his help in the
technical conception of the setup.
942 J. Appl. Phys., Vol. 81, No. 2, 15 January 1997 Massines
et al.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40
1V. A. Vonsyatskii, E. P. Mamunya, and Y. S. Lipatov, Vysokomol. So-
edin. 13A, 2164–2170 ~1971!.
2V. A. Vonsyatskii and B. V. Tkachuk, Theor. Exp. Chem. 24, 712 ~1988!.
3I. V. Novoselov, V. N. Korobeinikova, Y. A. Sangalov, and V. P. Kaza-
kov, Izv. Akad. Nauk SSSR Ser. Khim. 8, 1742 ~1990!.
4F. Massines, D. Mary, C. Laurent, and C. Mayoux, J. Phys. D 26, 493
~1993!.
5A. A. Kalachev, S. Yu. Lobanov, T. L. Lebedeva, and N. A. Plate, Appl.
Surf. Sci. 70–71, 295 ~1993!.
6J. Jonsson, B. Ra
˚nby, F. Massines, D. Mary, and C. Laurent, IEEE Trans.
Diel. Elect. Insul. ~to be published!.
7D. Mary, F. Massines, C. Laurent, and C. Mayoux, in Proceedings of the
Conference on Electrical Insulation and Dielectric Phenomena, 1993, pp.
438–443.
8R. J. Fleming and J. Hagekyriakou, Radiat Prot. Dos. 8,99~1984!.
9Y. Hama, Y. Kimuo, M. Tsumara, and N. Omi, Chem. Phys. 53, 115
~1980!.
10P. Cordier, J. F. Delouis, F. Kieffer, C. Lapersonne, and J. Rigaut, C. R.
Acad. Sci. A 279, 589 ~1974!.
11A. Charlesby, Radiat. Phys. Chem. 17, 399 ~1981!.
12L. Zlatkevitch, Luminescence Techniques in Solid State Polymer Research
~Springer, New York, 1987!.
13F. Massines, R. Ben Gadri, Ph. Decomps, A. Rabehi, P. Se
´gur, and C.
Mayoux, in Phenomena in Ionized Gases, AIP Conf. Proc. Vol. 363 ~AIP,
New York, 1996!, pp. 306–315.
14L. Matisova-Rychla, I. Rychly, and M. Vavrekova, Eur. Polym. J. 14,
1033 ~1978!.
15L. Audouin-Jirackova and J. Verdu, J. Polym. Sci.: Chem. Ed. 25, 1205
~1987!.
16M. Celina, G. Georges, D. L. Lacey, and N. C. Billingham, Polym. Deg.
Stabil. 47, 311 ~1995!.
17F. Poncin Epaillard, B. Chevet, and J. C. Brosse, J. Adhes. Sci. Technol.
8, 455 ~1994!.
18D. M. Ryder, O. Olliff, C. Laurent, F. Massines, and C. Mayoux, in
Proceedings of the Conference on Electrical Insulation and Dielectric Phe-
nomena, 1995, pp. 93–96.
943J. Appl. Phys., Vol. 81, No. 2, 15 January 1997 Massines
et al.
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.120.98.205 On: Fri, 17 Apr 2015 14:19:40