On the effect of treating poly(acrylic acid) with argon and tetrafluoromethane plasmas

15 Pages·2002·1.04 MB·English

On the Effect of Treating Poly( acrylic acid) with Argon and
Tetrafluoromethane Plasmas: Kinetics and
Degradation Mechanism
JOHANNES C. A. TERLINCEN,' CIJSBERT A. J. TAKENS,' FREDERIK 1. VAN DER CAAG,'
ALLAN S. HOFFMAN,' and JAN FElJEN*
'Department of Chemical Technology, University of Twente, P.O. Box 21 7, 7500 AE Enschede, The Netherlands;
'Ctr. for Bioengineering FI-20, University of Washington, Seattle, Washington 9819 5
SYNOPSIS
Poly (acrylic acid) ( PAAc) films were treated with either an argon or a tetrafluoromethane
(CF,) plasma and subsequently analyzed with X-ray photoelectron spectroscopy (XPS).
PAAc films were decarboxylated during both types of plasma treatments. In addition, during
the CF4 plasma treatment, the PAAc films became fluorinated. The plasma phase during
the argon plasma treatment of PAAc films was investigated with optical emission spec-
troscopy. It was shown that during this plasma treatment carbon dioxide, water, and possibly
hydrogen were liberated from the PAAc surface. By covering the surface of PAAc films
with different materials (lithium fluoride, UV fused silica, and glass) during the plasma
treatment, it was possible to differentiate between photochemically induced and particle-
induced changes of the surface. This method was used to show that decarboxylation during
the argon plasma treatment was caused by vacuum UV radiation (wavelength < 150 nm)
and the decarboxylation/ fluorination during the CF4 plasma treatment was induced by
reactive fluorine-containing species from the plasma phase. Furthermore, during both pro-
cesses, etching of the PAAc surface occurred. Based on these mechanisms, kinetic models
were derived that could be used to describe the measured kinetic data adequately. 0 1994
John Wiley & Sons, Inc.
INTRODUCTION To create carboxylic acid groups on a polyethyl-
ene (PE) surface that can be used for the covalent
Plasma techniques have been widely used for the coupling of bioactive molecules, we tried to immo-
introduction of functional groups on polymer sur- bilize preadsorbed layers of poly ( acrylic acid)
faces.' Although several groups like hydroxyl groups (PAAc) on PE by applying an argon or tetrafluo-
are readily formed by treating a polymer surface with romethane ( CF4) plasma treatment.13 Although the
a water plasma,' the selective introduction of car- concept of immobilizing a preadsorbed layer on
boxylic acid groups by, e.g., plasma polymerization polymer surfaces by using a plasma treatment has
of acrylic acid is diffi~ult.~W-~he n a polymer is been successfully used for the introduction of
treated with an oxidative plasma, a wide spectrum sulfate and amine groups at hydrophobic poly-
l4 l5
of oxygen-containing functional groups is introduced mers, the immobilization of PAAc layers on PE was
at the surface.cg To specifically obtain carboxylic not successful. The preadsorbed PAAc was etched
acid groups on polymer surfaces, other strategies like more rapidly than it became immobilized on the
plasma-induced grafting"." or grafting of pread- surface. From the gravimetrically determined etch-
sorbed carboxylic acid group containing monomers ing rates, it was concluded that PAAc is much more
like acrylic acid have been successfully applied." sensitive to argon and CF4 plasmas as compared
to PE.I3
Clark and Dilks showed that argon plasmas emit
* To whom correspondence should be addressed.
substantial amounts of hard UV radiation, with
Journal of Applied Polymer Science, Vol. 52.39-53 (1994)
Q 1994 John Wiley & Sons,I nc. CCC 00Z1-8995/94/010039-15 wavelengths starting from 90 nm.16 Several authors
39
40 TERLINGEN ET AL.
have shown that this UV radiation induces chemical 2 99.995% ) : Hoekloos, Amsterdam, The Nether-
changes of polymer surfaces during plasma treat- lands.
ment~.''-'~ Because carboxylic acid-containing
compounds are known to be sensitive to UV radia-
Methods
tion with wavelengths of around 200 nm, it seems
2o
possible that the sensitivity of PAAc to argon and All films before and after modification were stored
CFI plasmas is caused by the UV irradiation from in glass vials in the dark at ambient conditions.
the plasma phase. By covering the PAAc films with
UV transparent materials during the plasma treat- Preparation of PAAc Films
ment, the direct-ion/electron and metastable
PAAc (10 g) was dissolved in 500 mL of methanol.
bombardment-and radiative ( UV ) processes can
The solution was then cast onto two clean Petri
be discriminated. The materials form a physical
dishes (diameter 19 cm). After evaporation of the
barrier for the bombardment of reactive species from
methanol, the PAAc films were dried in vacua at
the plasma phase, and by selection of materials with
room temperature (RT).
different cutoff values, it is possible to test which
spectral region of the UV irradiation induces surface
damage. Cleaning of PE and PTFE Films
In this study, the effect of an argon or a CF4 PE and PTFE films were cleaned ultrasonically in
plasma treatment on the chemical composition of dichloromethane for 10 min. The dichloromethane
PAAc films was investigated. Surface analysis of the was refreshed and the procedure was repeated. The
modified films was performed by using X-ray pho- total cleaning procedure consisted of four times
.
toelectron spectroscopy (XPS) The kinetics and treatment with dichloromethane, four times with
the mechanism of plasma-induced surface reactions acetone, and four times with water. Subsequently,
were studied using covered and noncovered sub- the films were dried in uucw at RT and stored.
strates during the plasma treatment. To obtain some
additional qualitative information on the degrada-
Plasma Treatment
tion of PAAc during plasma treatments, the plasma
phase during the treatment of PAAc films was stud- An extensive description of the plasma treatment
ied with optical emission spectroscopy (OES) system has been given previou~ly.'I~n brief, the
.21*22
plasma system consists of a tubular reactor (internal
diameter 6.5 cm) with three externally placed ca-
pacitively coupled electrodes spaced at 10 cm dis-
MATERIALS AND METHODS
tances. One side of the tubular reactor was connected
to a turbomolecular and to a two-stage rotary vane
Materials
pump. The other side was attached to a gas inlet
Poly (acrylic acid) ( PAAc) : Aldrich Chemie, BNS- controlled by mass flow controllers. The electrodes
sel, Belgium, molecular weight 250,000 g/mol (ac- were powered through a matching network by a 13.56
cording to the supplier). PE film: low-density MHz radio-frequency generator, which was con-
poly (ethylene) thickness 0.2 mm, type 2300, DSM, trolled by a timer (function generator). This system
Geleen, The Netherlands. This PE does not contain enables an optimal control of the plasma treatment
additives. Poly ( tetrafluoroethylene ) ( PTFE ) : time.
thickness 0.5 mm, Good Fellow, Cambridge, A PAAc film (3X 3 cm) was placed in the center
England. Methanol, dichloromethane, and acetone: region of the reactor on a glass plate and the reactor
Merck, Darmstadt, Germany (purity > 99.5% ) . was evacuated (pressure < 1 X lop5 mbar). Sub-
Water: doubly deionized water. UV fused silica sequently, a gas flow ( Ar or CF4)o f 10 cm3/min was
( SOz):t hickness 3.1 mm, transparent for UV with established through the reactor. After 15 min, the
wavelengths longer than 150 nm (according to the films were plasma-treated (44 W, 0.07 mbar) for
supplier), and vacuum UV lithium fluoride ( LiF ) : different time periods ranging from 0.05 to 1000 S.
thickness 3 mm, transparent to UV with wave- Two minutes after the plasma treatment, the reactor
lengths longer than 104 nm, ABC Diagnostics, was brought to atmospheric pressure with air and
Utrecht, The Netherlands. Glass slides: thickness 1 the films were taken out.
mm, Tamson, Zoetermeer, The Netherlands. Argon To test whether UV radiation from the plasma
(purity 2 99.999% ) , tetrafluoromethane ( CF4) phase induces changes of the chemical composition
(purity 2 99.95%), carbon dioxide (CO2) (purity of the surface, PAAc films covered with a SiOz crys-
POLY(ACRYL1C ACID) TREATED WITH ARGON OR CF, 41
tal, a LiF crystal, or glass were plasma-treated. Be- was calculated by dividing the surface area under
cause the PAAc films were not completely flat, usu- the carboxylic acid peak by the total area of the Cls
ally a small gap ( 52m m) between the PAAc surface peak.
and the covering material existed. The plasma
treatment procedure used was similar to the pro- Optical Emission Spectroscopy (OES)
cedure described above. It should be noted that LiF
OES has been used to analyze the plasma phase
crystals are severely damaged by an argon plasma
during the plasma treatment of PAAc and PE. Sim-
treatment. The transparency of the crystal for UV
ilar conditions (except for a much larger film area
decreases significantly even after short plasma
of 90 cm2) as described for the plasma treatment of
treatment times. To minimize this effect, a new
PAAc were used. To increase the concentration of
crystal was used for every experiment.
degradation products, a different set of discharge
To determine whether the covering of the sub-
conditions were used 44 W, 0.24 mbar, and 3.0 sccm/
strate with the different materials was sufficient to
min argon. For comparison, also spectra of argon/
prevent the interaction of species from the plasma
water and of Ar/C02 mixtures were recorded. The
phase with covered surfaces, the spacing between
water was repeatedly (three times) frozen, degassed
two glass plates was varied (from direct contact to
to 0.01 mbar, and thawed before use. The flow rates
.
4 mm) The PAAc film to be treated was placed in
of the very small amounts of water or COP (0.3 sccm/
between these glass plates. The films were plasma-
min) were determined from the increase rate of the
treated for 1000 s according to the procedure de-
pressure in the reactor and were controlled by needle
scribed previously. At separations smaller than 5
valves. A quartz window at the vacuum side of the
mm, no plasma developed in between the glass
reactor and a spectroradiometer SR 9010-PC (Ma-
plates.
cam Photometrics, Livingston, Scotland) were used
Characterization
n
X-ray Photoelectron Spectroscopy (XPS)
Measurements
-
XPS measurements were performed with a Kratos -
XSAM-800 (Manchester, UK) apparatus using a --s8
MgKa source (15 kV, 15 mA). The analyzer was
.+-'
-
placed perpendicular to the sample surface for rou- -VI
(Y
tine analysis. During angle-dependent XPS mea- E
surements (AD-XPS), the angle between the sur-
face and the analyzer was varied between '0 (per-
pendicular) and 75'. A spot size with a diameter of
3 mm was analyzed. The spectra were recorded in
the low-resolution mode (pass energy 40 eV, fwhm
Ag3d5I2: 1.2 eV) and standard sensitivity factors
(delivered by Kratos) were used to convert the mea-
sured peak areas into atom percentages. To deter-
mine a correct empirical sensitivity factor for the
Fls peak, a clean PTFE sample was analyzed when-
ever fluorine-containing films were analyzed. For
PAAc and modified PAAc films, short analysis times
should be used, because significant radiation damage
occurs during the XPS measurements at prolonged 0
295 290 285 280
times.
Binding energy (eV)
Peak fitting of the Cls spectra was performed by
Figure 1 XPS spectra of the Cls region of argon
fitting the spectra of argon plasma-treated and un-
plasma-treated (treatment time 500 s) PAAc films. The
treated PAAc films with three Gaussian peaks. A
spectra of untreated and of argon plasma-treated films,
hydrocarbon peak at 284.6 eV, a C -0 peak at 286.1 which were either covered (with UV fused silica (SiOn)
eV, and a carboxylic acid peak at 288.6 eV were used. or lithium fluoride (LiF)) during the plasma treatment
The fraction of carboxylic acid groups on the surface or not covered, are shown.
42 TERLINGEN ET AL.
to record emission spectra (from 240 to 800 nm with Only LiF-covered PAAc films show a significant de-
a step width of 2 nm) . crease in the O/C ratio with increasing plasma
treatment time. At a plasma treatment time of 500
s, the O/C ratio decreases to a level lower than that
RESULTS of uncovered PAAc films that were plasma-treated
for a similar time period.
Because the pendant carboxylic acid groups in
Argon Plasma Treatment of PAAc
PAAc give a distinct peak (at 288.6 eV) in the Cls
In Figure 1, the XPS spectra of the Cls region of spectra, peak fitting of the Cls spectra of argon
argon plasma-treated PAAc films are shown. The plasma-treated PAAc films yields information on
peak at 288.6 eV, which is due to the pendant car- the chemical changes induced on the surface by the
boxylic acid groups of the PAAc, is lacking in the plasma treatment. In Figure 3, the fraction of car-
Cls spectrum of the argon plasma-treated (500 s) boxylic acid groups remaining on the surface for un-
PAAc films. Also, in the spectrum of LiF-covered covered and covered (with different materials) films
argon plasma-treated PAAc films, the peak at 288.6 is given as a function of the argon plasma treatment
eV is lost. For PAAc films covered with a SiOzc rystal time. The shape of the curve is similar to that of
or glass during the plasma treatment, no change in Figure 2.
the Cls spectrum compared to untreated PAAc films To investigate whether the covering of the films
is observed. is sufficient to totally block the interaction of re-
In Figure 2, the oxygen-to-carbon (O/C)ra tios active particles from the plasma phase with the
of argon plasma-treated PAAc films are given as a PAAc surface, the distance between two glass plates
function of the plasma treatment time. After a was varied. In Figure 4, the results of the XPS anal-
plasma treatment time of 1 s, the plasma induces ysis of argon plasma-treated PAAc films (1000 s),
an oxygen loss at the surface, which increases with which were placed in between two glass plates during
increasing plasma treatment time. At 50 s, the O/ the plasma treatment, are given. It is shown that
C ratio levels off to a plateau level of around 0.18. regardless of the spacing between the glass plates
Treating the PAAc films for longer time periods does no changes in the surface chemistry are observed
not induce further changes in the surface chemistry with XPS.
according to the XPS measurements. The results The results of the AD-XPS analysis of argon
for PAAc films that are covered with materials dur- plasma-treated uncovered PAAc films are given in
ing the plasma treatment are also given in Figure 2. Figure 5. The O/C ratio is given as a function of
0.60
0.50
0.40
uaurvucd
. L i P
0.30 A SioZ
a
Glass
0.20 Model
0.10
0.00
0.01 0.1 1 10 100 lo00
plssmp treatment tlw (s)
Figure 2 XPS analysis of argon plasma-treated PAAc films. The ratio of oxygen to
carbon (O/C)i s given as a function of the applied plasma treatment time. The films were
either covered (with glass, SiOp,o r LiF) during the plasma treatment or not covered. Also,
the best fit for eq. (5 ) is given for plasma-treated (not covered) PAAc films.
POLY(ACRYL1C ACID) TREATED WITH ARGON OR CF, 43
0.30
0.25
[
0.20
8X
0.15
%
E
f
0.10
c*
the cosine of the take-off angle. At a take-off angle CF4 Plasma Treatment of PAAc
of 0", a depth of approximately 100 A is probed. At
an angle of 75", the XPS analysis is more surface- Figure 6 shows the Cls spectra of CF4 plasma-
sensitive and the probed depth is in the order of 25 treated (1000 s) PAAc films. After plasma treatment
A. Within the accuracy of the XPS method (ap- of uncovered PAAc films, not only hydrocarbon
proximately lo%),t he O/C ratio is not dependent ( C -H ) moieties are present but also CF, CF2,a nd
on the probed depth (see Fig. 5). CF3 groups are formed at the surface. CF4 plasma
.- 0.7 0.30
2 .............\. . ...... + Be
0.6 0.2s =
0
P
0.20 4
*,
80
0.15
-
.I-
5
0.10
+ {1
Uncovered Lr
0.2
+0.05
0.1
0.0 0.00 - . - . - . -
0 I 2 3 4 S . - ,
U.U
Dlstance between glass plates (mm) 0.0 0.2 0.4 0.6 0.8 1.0
Figure 4 XPS analysis of argon plasma-treated PAAc Cos (take off angle)
films (1000 s), which were placed in between two glass Figure 5 Angle-dependent XPS measurements on ar-
plates during the plasma treatment. The ratio of oxygen gon plasma-treated PAAc films. The ratio of oxygen to
to carbon and the fraction of carboxylic acid groups left carbon is given as a function of the cosine of the take-off
on the surface are given as a function of the spacing be- angle for films plasma treated for different time periods.
tween the glass plates. At an angle of O", the largest depth is probed.
44 TERLINGEN ET AL.
are covered with different materials during the
plasma treatment, the O/C ratio decreases beyond
the levels found for uncovered PAAc films.
From Figure 8 it can be seen that after a plasma
treatment time of 0.1 s the F/C ratio increases with
increasing plasma treatment time. Similarly to the
O/C ratio, the F/C ratio levels off after a treatment
time of approximately 50 s. For all covered films, an
increase in the F/C ratio with increasing plasma
Plasma treated PMc,
glass covered J treatment time is also found. The F/C ratio for the
covered films approaches a value of 2 ( CFz),w hich
is significantly higher than the final F/C ratio for
Plauna treated PMc.
LiF covered uncovered films (maximal value 0.9).
Analogous to the argon plasma treatment, PAAc
films were treated with a CF4 plasma for 1000 s in
between differently spaced glass plates. The effect
of varying the distance in between the glass plates
on chemical composition of the surface was studied
Plasma treated PAAc with XPS and the results are given in Figure 9. The
not mvered
hI O/C ratio reaches a plateau at distances of more
than 1.5 mm. The F/C ratio also reaches a plateau
Untreated PMc at distances of more than 2 mm. This plateau is
significantly higher than the level found for uncov-
295 290 285 280
ered plasma-treated films. At very small distances
between the plates (10.5m m) , the chemical com-
Binding enerpg (av)
position of the modified films approaches that of
Figure 6 XPS spectra of the Cls region of CF4p lasma- unmodified films.
treated (treatment time 1000 s) PAAc films. The spectra The results of the AD-XPS measurements on
of untreated and of CF4p lasma-treatedf ilms, which were
PAAc films treated with a CF4 plasma for different
either covered (with LiF or glass) during the plasma
time periods are given in Figure 10. The F/C ratio
treatment or not covered, are given. For comparison, the
increases with decreasing depth, indicating that the
Cls spectrum of PTFE and the shifts for different fluorine
containing groups are given.23T he C- H component of top surface is enriched in fluorine during the plasma
the Cls was set on 284.6 eV, and for fluorinated films, the treatment. Because of the scatter of the data points,
Fls peak was set on 689.2 eVZ4t o correct for charging no trends for the O/C ratio can be depicted.
effects during the XPS measurements.
OES During Argon and CF., Plasma Treatments
treatment of PAAc films covered with LiF, SiOz OES was used to qualitatively study the degradation
(spectrum not given), or glass yields a totally dif- products of PAAc and PE films during argon or CF4
ferent distribution of fluorine-containing groups. In plasma treatments. Spectra of the argon plasma
this case, primarily, CF2 groups and only small treatment of PAAc and PE were recorded and these
amounts of CF and CF3 groups are found on the spectra were compared to the emission spectrum of
surface. It is quite remarkable that there are no hy- an argon plasma without any film present (spectra
drocarbon groups present at the surface of these not shown). The spectrum of the argon plasma was
films. comparable to the spectrum obtained during the ar-
The quantitative results of the XPS analysis of gon plasma treatment of PE, except for a small dif-
the CF4p lasma-treated PAAc films are given in Fig- ference in peak height at 384 nm. When the spec-
ures 7 and 8. In Figure 7, the O/C ratio, and in trum obtained during the argon plasma treatment
Figure 8, the ratio of fluorine to carbon (F/C), are of PAAc was compared with that of an argon plasma,
given as a function of the plasma treatment time. some distinct differences were observed First, the
After a plasma treatment time of approximately 1 absolute intensity of peaks due to excited argon spe-
s, the O/C ratio decreases with increasing plasma cies decreased. Furthermore, some small new peaks
treatment time. After 50 s, the O/C ratio for un- appeared in the emission spectrum obtained during
covered PAAc films levels off. For PAAc films that the argon plasma treatment of PAAc.
POLY(ACRYL1C ACID) TREATED WITH ARGON OR CF, 45
0.6
o
0.5
00
0.4 c uncovend
m L i F
4
oo rr nn A SiOZ
11 ''11 0 Glass
00
AA Model
0.01 0.1 1 10 100 loo0
plasmp treatment time (s)
Figure '7 XPS analysis of CF, plasma-treated PAAc films. The ratio of oxygen to carbon
is given as a function of the plasma treatment time. The films were either covered (with
glass, Si02,o r LiF) during the plasma treatment or not covered. Also, the best fit for eq.
( 10) is given for plasma-treated (not covered) PAAc films.
To amplify the emission lines of (possible) deg- argonlcarbon dioxide ( Ar/C02) and argonlwater
radation products with respect to the argon back- ( Ar / HzO ) mixtures have been recorded. The emis-
ground, a different set of plasma conditions was used. sion spectra obtained for argon, Ar/COZ,a nd Ar/
The power input was kept constant, but the resi- H20d ischarges and spectra obtained during the
dence time of the argon gas was increased by a factor plasma treatment of PAAc using the second set of
11 using a lower gas flow and a higher pressure. To plasma conditions are given in Figure 11. The emis-
investigate which species cause the differences be- sion spectrum obtained during the argon plasma
tween the emission spectra of an argon plasma and treatment of PE was again very similar to that of a
of an argon plasma treatment of PAAc, spectra of pure argon discharge (spectrum not given). Adding
1.5 -- 0
0 3 uacovcnd
.LiF
0
I-- A SiOZ
u 00 . 0 G h
Model
0.01 0.1 1 10 100 loo0
mmlm tmablalt (tw (a)
Figure 8 XPS analysis of CF, plasma-treated PAAc films. The ratio of fluorine to carbon
is given as a function of the plasma treatment time. The films were either covered (with
glass, SO2,o r LiF) during the plasma treatment or not covered. The best fit for eq. (14)
is given for plasma-treated (not covered) PAAc films.
46 TERLINGEN ET AL.
increasing plasma treatment time, an increasing
amount of carboxylic acid groups is lost from the
surface, which correlates directly with the decrease
in the O/C ratio. This is in agreement with the re-
sults of Hook et al., who reported that films
of poly(methy1 methacrylate)" and films of poly
( methyl methacrylate- co-methacrylic acid) 26 were
reduced/decarboxylated during argon/water plasma
treatments.
By placing a crystal or glass on the polymer film,
an attempt was made to prevent a bombardment of
.-...,
O.'I i; reactive species-except the UV radiation-from the
0.0 plasma phase to the polymer surface. Because a
0 1 2 3 4 5
small gap between the cover material and the poly-
Distance between glass plates (mm)
mer film will exist, which would still be accessible
Figure 9 XPS analysis of CF, plasma-treated PAAc
to reactive species from the plasma phase, a small
films (1000 s), which were placed in between two glass
plates during the plasma treatment. The oxygen-to-carbon
ratios and fluorine-to-carbonr atios are given as a function
of the spacing between the glass plates.
2 0.6 -
0
small amounts of water or carbon dioxide to argon 0.5 -
changes the spectrum drastically. From Figure 11 it
-
can be seen that the absolute emission intensity for 0.4
these mixtures has decreased significantly. Fur-
thermore, in the Ar / HzO spectrum, two distinct new 0.3 -
lines at 486 and 656 nm are found. The Ar/COZ -
0.2
spectrum is completely different from the Ar emis-
sion spectrum. Also, a synthesis spectrum, which is 4 I
0.1
a linear combination of the emission spectra of Ar /
COza nd Ar / HzOd ischarges, is given. A remarkable 0.0
similarity between the synthesis spectrum and the 0.0 0.2 0.4 0.6 0.8 1 .o
spectrum obtained during the argon plasma treat- Cos (take off angle)
ment of PAAc is observed.
OES has also been used to study the degradation B t=loo s
e
products during the CF, plasma treatment of PE
2
and PAAc. The emission spectra of CF4/H& and
1.5
CF4/COZh ave also been recorded (data not shown).
Even when large amounts of COz or water were \I
added to the CF, gas flow, the emission spectrum
did not change. Therefore, it was concluded that in
the case of a CF, plasma treatment of PAAc films
OES with a resolution of 2 nm was not sensitive
enough to study the degradation products in the 0.5
plasma phase.
I
0.04 ' " " " '
0.0 0.2 0.4 0.6 0.8 1.0
DISCUSSION
Cos (take off angle)
Argon Plasma Treatment of PAAc Figure 10 Angle-dependent XPS measurements on CF,
plasma-treated PAAc films. The oxygen-to-carbon ratio
Treating PAAc films with an argon plasma results (a) and the fluorine-to-carbonr atio (b) are given as a
in decarboxylation of the PAAc surface. This is con- function of the take-off angle for films plasma-treated for
cluded from the results given in Figures 1-3. With varying times.
POLY(ACRYL1C ACID) TREATED WITH ARGON OR CF, 47
0.2
150
a 0.16
3
0.12
ii
3 310
3
0.08 Ar - - 486
h
0.04 Alma h- - c
IAr~CClZ -
PAAr A
0
200 300 400 500 600 700 800
Wavelength (nm)
200 250 300 350 400 450 500 550 600 650 700
Wavelength (nm)
Figure 11 Optical emission spectra (OES) of an argon (Ar) plasma, an argon/carbon
dioxide ( Ar/COz) plasma, an argon/water ( Ar/H20)p lasma, and an argon plasma during
the treatment of a PAAc film ( PAAc). Overview spectra are given in (a) and enlarged
(enlargement factor 10) spectra are shown in (b). In (b), a synthesis spectrum consisting
of a linear combination of the Ar/H20 and the Ar/C02 spectrum is also presented. Discharge
+
conditions: 44 W, 0.24 mbar; gas flows: Ar (3 sccm/min); Ar/H20 (3 sccm/min Ar 0.3
+
sccm/min HzO); Ar/COZ (3 sccm/min Ar 0.3 sccm/min COz); PAAc, plasma treatment
of 90 cm2 PAAc (3s ccm/min Ar) . The synthesis spectrum is a combination of the following
+
spectra: 0.075 *Ar/H20 0.55*Ar/Hz0. For clarity, offsets are used for the different spectra.
flux of reactive species from the plasma phase might trum of the argon plasma consists mainly of lines
still reach the surface. To investigate this effect, the originating from excited argon atomsz7a nd not from
gap size between the surface of the substrate and ionized argon atoms" (see also Fig. 11), it seems
the cover material was deliberately increased. From likely that the decarboxylation of PAAc is caused
Figure 4 it can be seen that a flux of reactive species by UV radiation of 104.8 nm, which is the predom-
from the argon plasma phase does not influence the inant line in the vacuum UV spectrum of excited
surface chemistry of the PAAc film. argon.
Although no changes in the surface chemistry are If the surface modification would be purely pho-
observed for argon plasma-treated PAAc films, tochemical, the PAAc surface would be totally de-
which were covered by glass or SiOz during the pleted from carboxylic acid groups at long treatment
plasma treatment, decarboxylation of the LiF-cov- times. For LiF-covered films, this seems possible,
ered PAAc surface is found. This decarboxylation but for uncovered films, this is not the case (see
is caused by UV transmitted by the LiF crystal Figs. 2 and 3). Therefore, a second process should
(wavelength > 104 nm) . Because the emission spec- occur simultaneously. Although the flux of reactive
48 TERLINGEN ET AL.
species from the plasma phase does not change the cluded that during the plasma treatment of PAAc,
surface chemistry directly (see Fig. 4), this second HzO and COz are liberated from the surface. Fur-
process should still be caused by this flux. Most thermore, it seems possible that also a small amount
likely, a bombardment of ions and/or electrons to of hydrogen is formed during the argon plasma
the surface causes sputtering of the surface. The ob- treatment. Most likely, the COZ originates from car-
served carboxylic acid group levels ( and 0 / C ratios ) boxylic acid groups, which are decomposed during
at long treatment times are caused by a balance be- the plasma treatment. The liberated HzO may be
tween the sputtering-a surface renewal-and a de- due to tightly bound water in the polymer or may
carboxylation process. Explaining the surface mod- be formed during the discharge.
ification by an argon plasma in terms of a combined
photochemical and particle-induced process is in
CF, Plasma Treatment of PAAc
agreement with the model proposed by Clark and
Dilks for the argon plasma modification of fluori- The results of the XPS analysis of CF, plasma-
nated polymer surfaces.29 treated films, given in Figures 6-10, show that PAAc
From the AD-XPS measurements for argon films are decarboxylated and fluorinated during a
plasma-treated films, given in Figure 5, it can be CF4 plasma treatment. Furthermore, from Figures
seen that the decarboxylation takes place homoge- 6-8, it can be seen that regardless of the type of
neously over the probed depth. This is concluded material used to cover the PAAc film during the
from the fact that the O/C ratio of plasma-treated plasma treatment the PAAc surfaces are fluorinated
films is nearly independent of the take-off angle. and decarboxylated. The differences in surface
The photochemical decarboxylation occurs, thus, chemistry among LiF, SiOz,a nd glass-covered films
over a depth much larger than the XPS analysis after a CF4 plasma treatment are small. This indi-
depth ( 100 A). For not-plasma-treated PAAc films, cates that, in contrast to the argon plasma treat-
there seems to be a tendency to have an enrichment ment, the contribution of the photochemically in-
of oxygen in deeper surface layers. This is probably duced decarboxylation of PAAc to the surface
caused by a thin hydrocarbon overlayer of an un- chemistry is insignificant. The surface modification
known origin on the surface of the PAAc films. The is thus dominated by a reaction of active species
presence of this layer would explain the deviation from the plasma phase.
of the O/C ratio and the fraction of carboxylic acid During the CF4 plasma treatment of different
groups found for pure PAAc films from the theo- polymers, fluorine-containing groups, such as CF,
retically expected values of respectively 0.67 CFz, and CF3, are introduced at the ~urface.~'-~~
and 0.33. From the spectra of PAAc films treated for different
OES has been used to qualitatively study the time periods, it is clear that the surface is progres-
plasma phase during the argon plasma treatment of sively fluorinated. With increasing plasma treatment
PAAc. Because of the complexity of the spectra and time, higher fluorinated groups are formed. It seems
of the numerous cooperative and counteractive pro- likely that abundantly present atomic fluorine35
cesses between different (excited) species in the causes the observed fluorinationg and decarboxyl-
plasma, we have used OES only as a qualitative ation. In contrast to the UV-induced decarboxyl-
method to detect degradation products in the plasma ation, the decarboxylation/ fluorination by reactive
phase. From the emission spectrum of an argon species from the plasma phase seems to be bound
plasma, it can be seen that the spectra consist mainly to the outermost surface (see Fig. 10).
of peaks originating from excited argon atoms.27 Because the O/C ratio levels off at long plasma
During the argon plasma treatment of PAAc films, treatment times, a second process counteractive to
some new peaks are found, which seem to be caused the fluorination /decarboxylation by species from
by hydrogen (486 and 656 nm, atomic hydrogen the CF4 plasma phase occurs. This second process
line^^^,^^), possibly water3' ( 306-309 nm) ,a nd car- is most likely a combined sputter and chemical-
bon dioxide fragment^.^'^^^ The synthesis spectrum etching process. The surface chemistry observed for
[given in Fig. 11 (b)1 , which is a linear combination CF4p lasma-treated PAAc films is thus the result of
of the Ar/COZ and Ar/HZO spectrum, is similar to a fluorination /decarboxylation and sputter/etching
the spectrum obtained during the argon plasma process. When a PAAc film is covered during the
treatment of PAAc. The few differences between plasma treatment, the balance between these two
these spectra-peaks at 336 and 358 nm-are prob- processes shifts. By decreasing the distance between
ably due to a very small air leak during the plasma the covering plates, the etching rates decrease.
treatment of PAAc. Based on these results, it is con- Therefore, at small distances, the fluorinationlde-