Surface Science 251/252 (1991) 787-793 787

North-Holland

XPS study of Al/polyethylene terephtalate interface

J.F. Silvain, A. Arzur, M. Alnot, J.J. Ehrhardt and P. Lutgen *
CNRS. taboratoire Maurice &tort, F-54600 Viters /es Nancy, France

Received 1 October 1990; accepted for publication 20 December 1990

X-ray photoelectron spectroscopy has been used to study the formation of the alu~nium/polyethylene terephtafate (Al/PET)
interface. On control PET, C Is and 0 1s exhibits the usual behavior of the PET. An extra C v 1s appears on Arc-bombarded PET
whilst the C,v 1s component is decreased by about 10% compared to the control PET. On fluorine-treated PET, the C 1s spectra have
been decomposed into six different components. After annealing at 500 K, the C Is shows the usual behavior of the PET indicative of
a decrease of the fluorine concentration in the polymer surface. Al films up to 64 A were evaporated with a Knudsen cell. The
Al/PET interface exhibits the behavior of the layer by layer growth. A preferential aluminum bonding with PET is observed for
C,vls. The Al evaporated onto PET presents a complex A12p peak decomposed into Al,,,,, and Aloxtde component. The ratio
Al ,,,eta,/Aloxide changes with the PET and the temperature of the PET during Al deposition. Transmission electron microscopy
(TEM) is performed in order to correlate the morphology and the adhesion of the Al film on PET.

1. Introduction sidered first the changes of the composition of the

polymer (as produced or pretreated) as a function
Much attention has been paid recently on the of the temperature. Then the chemical composi-
tion of the interface Al/PET and the growth
metallization of polymers from a fundamental
point of view and for practical reasons, this class process were investigatedOfor coverages up to 20
of material being widely used for audio and video monolayers (1 ML = 3.2 A). Finally, some details
recording tapes, in packaging, for electric and on the morphology of the films have been ob-
electronic devices and so on. Among the various tained by TEM.
properties which have to be considered, the ad-
hesion of a metallic film is of crucial importance.
The binding modes, the composition, and the
2. Experimental set-up
structure of the interface have been extensively
studied for the deposition of metals on polyimide
[l-5] and to a lesser extent on polyesters [C-10]. The experiments have been performed in an
We report here on the first steps of the Al UHV apparatus extensively described previously
deposition on polyethylene terephtalate PET [ll]. Briefly, a sheet of polymer (10 x 20 mm, 12.5
(Mylar from Du Pont de Nemours) studied by pm thick) was fixed on a sample rod with MO
XPS and TEM. As it is well known that the strips. It can be heated up to 500 K, or cooled
surface of the polymer can be heated up to 400 K down to 77 K, and placed in front of either an
in industrial conditions of metallization, we con- hemisperical electron analyzer allowing XPS or an
Al source. This source is a Knudsen cell con-
stituted by an Al,O, crucible filled with Al wires
(purity 99.95%). The fluxes are monitored by a
* Du Pont De Nemours S.A. Luxembourg. vibrating quartz. During deposition the pressure

0039~6028/91/$03.50 0 1991 - Elsetier Science Publishers B.V. (North-Holland)

7X8 J.F. Sdvarn et al. / XPS ofAl/po!~ethylene terephtalate wrterfuce

always remains in the low lo-’ mbar range. The paratus. The number of free parameters is essen-
different samples of polymer studied here are tially limited to the peak heights.
surfaces of as produced. fluorinated (in industrial
conditions) and in situ ion-bombarded PET.
XPS spectra were recorded in the AE mode 3. Experimental results
with an instrumental resolution of about 1.1 eV.
The deconvolutions of C Is, 0 1 s and Al 2p levels 3.1. XPS studs of the surfuce of the polvmer us u
were performed using different components mod- function of the temperature
eled by symmetric Gaussian curves. The chemical
shifts are coming either from XPS data base [12,13] The assignment of the different components in
or model compounds analyzed in the same ap- the C 1s and 0 1s XPS spectra is critical for the

J
II I
I
/
I
I
PET 3OOK
Cls

4
.., ‘I \
i .‘.: , \,

sl
‘. \’ / /’
\/ I ’
._ /\ / ’
:. / \ \

-‘i-,
>
a
\

:.I/
CIV’ CIII’ CII’ ,CI’

:\_ ,_ .Y’.-
^ Fluorine PET
300K Cls
r
\ I
\ /’ I
\f
‘/ ’ I\
,I’ d\ Jt ,‘\
b

1CI3_ Fluorine PE’I

SOOK Cl s

296 293 290 287 284 E;CeV)

Fig. 1. C 1s spectra of the PET. (a) Control PET at 300 K. (b) Fluorinated PET at 300 K. (c) Fluorinated PET at 500 K.
 J.F. Silvain et al. / XPS of Al/polyethylene terephtalate interface 189

Table 1 ing energy for the three carbons appears to be

Comparative table of the energy position of the Cls (Ct. C,,, slightly too high if compared to the value reported
C,,) of the control PET at 300 K
for PET (table 1). This can be attributed to some
c, C-C-) c,, (-C-O) c,, (-0-C = 0) Ref. charge effect as no correction has been applied in
(eV) (ev) (ev) our measurements. The 01s spectra present 2
285.7 281.1 289.8 Present structures 0, at 532.4 eV (O=C-) and Or, at 534.0
285.0 286.8 289.5 t141 eV (-C-O-C-). The shape of the Cls and 01s
285.32 286.8 289.33 [I81
spectra does not change with annealing up to 490
K. A slight shift of 0.6 eV towards higher binding
energies of the full spectrum is observed. It can be
understanding of the interaction between metal attributed to some changes in the surface electric
atoms and the surface of the polymer. In this first conductivity induced either by water desorption or
paragraph, we focused on the surface chemical surface modification and thus affecting the charge
composition as a function of temperature for dif- effect [17].
ferent PET samples, taking into account the com- Fig. 1, shows the XPS spectra of the control
position of the repeat unit of the ‘PET (10 carbon PET at 300 K and of the fluorinated PET at 300
atoms and 4 oxygen atoms). and 500 K. The three typical carbons (C,, C t,,
The Cls spectra have been first recorded on a and C,,) of the PET, are present in these three
virgin PET surface. They exhibit the fine structure spectra with a slight shift due to some charge
usually found on this polymer [14,18]. They have effect. However, four more carbons (C,>, Cur,
been decomposed into three peaks called C, at C ,rr’, and C,,,) are present in the fluorinated PET
285.7 eV (-C-), C,, at 287.7 eV (-C-O-) and at 300 K (fig. lb). According to Chtaib’s [18]
C,, at 289.8 eV (-0-C=O) (fig. la). The chemical decomposition, we tentatively assigned these
shifts between the three carbons are in reasonable carbone to C-CF, CF-C, CF-CF,, and CF,-C
agreement with literature [16]. However, the bind- respectively. After annealing at 500 K (fig. lc), C,t

0.1
0.0 -
-0.1 -

-0,2 -

-0,3 -

-0.4 -

-0.5 -

-0,6 -

-0,7 -

0 1 2 3 4 5 6 7 8 9 10

Thickness <A>
Fig. 2. Attenuatton curves (reduced intensity via Al coverage) for the three usual carbons of the PET, namely C, at 285.7 eV (-C-),
C,, at 287.7 eV (-C-O-) and C,, at 289.8 eV (-0-C = 0).
740 J. F. Silvoin et al. / XPS of Al/polyethvlene terephtalate interface

and C,,, disappeared. This phenomenom has been of the Cvls peak levels up after 90 s of ion
interpreted as a decrease of the fluorine con- sputtering. Surface modifications induced by ion
centration in the outer skin of the polymer by bombardement are very complex: cross linking
either diffusion or thermal desorption when heat- [19,20]. hydrogen and oxygen bridge formation [2]
ing such a pretreated PET. and bond breaking [9] have been reported, so, the
An extra Cv Is carbon component appears at Cv component could be related to O-C-OH
291.0 eV after ion bombardement (Ar+, 2 keV, functional groups. We will only use it as a
emission of 10 mA, 15 to 150 s). Meanwhile the fingerprint of the chemical changes induced by ion
C,, peak shows a decrease of about 10%. The area bombardment.

.‘. metal

I o2

I o2

1 ML AI/PET 20 ML
I’~~~““~‘~‘~‘.“’ / 7~77--T
N(E) 1
oxide metal
(G’S)
!I
I
fi
i
i
!

l”fyyyq,
i

b ,;,~~~

--III

86 82 78 74 70 E;CeVl 86 82 78 74 70 E;CeV?

Fig. 3. Al2p XPS spectra of Al deposited on: (a) MO at 300 K (1 and 10 ML) (1 ML = 3.2 A); (b) control PET at 300 K (1 and 20
ML): (c) fluorinated PET at 300 K (I and 20 ML).
 .J.F. Sihxzin et cd. / XFS ofAi~p~~ethyl~~e terephtaiate interface

PET ‘1.. ” * ,r. “, 1 ‘. 3

N<E>
i CC/S)
metal

J
86 82 78 74 70 E&A’> 86 82 78 74 70 E&V>
Fig. 3. Continued.

3.2. Al deposition on PET a clean MO surface (fig. 3a). The spectra present
two chemical environnements which have been
3.2.1. C 1s analysis (fig. 2) attributed by using the published chemical shifts
The behaviour of the intensities of the three [12]. Deconvolution of these spectra allows an
functional groups of carbon has been investigated
during Al deposition at room temperature (fig. 2) Table 2
on biaxially stretched PET. It is worth noting Percentage of the Al metal and the Al oxide for the different
that: Al/PET systems (the chemical shift between the Al metal and
the Al oxide is 2.6 eV)
(i) the attenuation of C, and C,, components
obeys a quasi-exponenti~ law (almost parallel 1 MLa’ 10 ML =) 20 ML a)
straight lines in the logarithm representation used MO
here), % metal 60 60 60
(ii) the intensity of the C,, peak decreases much 5%Al oxyde 40 40 40
PET, controle, 20 o C
faster than the intensity of CI and C1, does. This % metal 0 35 60
is evidence of the drastic chemical change of the X Al oxyde 100 65 40
carboxylic group induced by Al deposition. PET, controle, - 200 * C
As the C,, component represents only 20% of % metal 65 60 60
the total Cls area on clean PET, it is clear that 48Al oxyde 35 40 40
PET, ion bombardment
within the experimental accuracy, the attenuation % metal 5 55 60
law can be fitted by an exponential when consid- % Al oxyde 95 45 40
ering the full C 1s peak. This point will be dis- PET, fluor, 20 o C
cussed later in terms of growth mechanism. % metal 10 35 50
% Al oxyde 60 35 30
WAIOF 20 20 15
3.2.2. Al2p analysis (fig. 3)
% Al F3 10 10 5
The A12p levels have been recorded after Al PET, fluor, - 200 o C
deposition (1, 10 and 20 ML) on biaxially streched 4: metal 35 45 50
PET, on ion-bombards PET and on fluorinated % Al oxyde 40 20 20
PET. During the Al deposition, the PET tempera- WAlOF 20 25 20
% Al F3 5 10 10
ture is kept at 290 or 77 K. They have been
compared to those obtained after Al deposition on a) 1 ML = 3.2 A,.
192 J. F. Sikurn er al. / XPS ofAl/po!veth,rlcneterephralare mrerfwr

estimate of the relative proportions of the various K treated (fluorine treatment) and non-treated
compounds. The results are summarized in table PET. For these two system. a decrease of the AGS
7A. is associated with a decrease of the PET tempera-
The Al deposition on MO shows that the Al ture: the AGS is two times smaller for the 77 K
source used in this study does not provide pure Al. PET in comparison with the 300 K PET. This
About 40% of the deposited Al appears to be decrease of the AGS can be interpreted as a
oxidized whatever the thickness of the deposit. lowering of the mobility of the Al atoms on the
The ratio between Al0 and AI”’ remains constant PET surface. We can also notice for these AI/PET
and independent of the cleaning procedure of the laminates, that the adhesion. measured by a 180 o
Al charge in the source. As the vacuum was in the peel-test technique, does not change with tempera-
low lo-’ mbar range during evaporation, it is ture.
believed that some reaction between molten Al
and the Al,0_ 3 crucible could occur giving an
unexpected flux of “oxidized” Al. As a conse- 5. Discussion
quence, our observations are not related to pure
Al but rather to a mixture of Al and oxidized Al. Three points deserve to be discussed in this
On non-treated PET (fig. 3b) and on ion- study of the formation of an Al-PET interface:
bombarded PET, for deposition at room tempera- - The analysis of the attenuation of the C 1s
ture. it has to be mentioned that the first Al signal as a function of the Al coverage can be
monolayer is fully oxidized. This is a strong evi- interpreted in terms of a pseudo layer-by-layer
dence of the formation of Al-O bounds with the growth mechanism. This is unexpected as the is-
oxygen atoms of the surface of the polymer. As. land formation has been postulated for the growth
after a deposition of about 10 ML the composi- of a metal on a polymer [21]. A possible explana-
tion of the film is close to the value observed on tion could be the low level of purity of the de-
MO, it is clear that the oxidation process should posited AI thus changing the surface properties (in
concern only a few layers. So, the interface should particular the surface free energy). For this point
be quite sharp. At low temperature, this reaction of view the study of the growth mechanism of a
is strongly inhibited and the composition of the pure Al film on PET would be very interesting.
film can be considered as independent of its thick- _ The chemical reactions at the interface have
ness. been discussed in terms of either the bound break-
The A12p spectra appear to be much more ing of the carbon oxygen bound of the carboxylic
complex for deposition on fluorinated PET. On group giving rise to C-O-Al groups [22] or the
the one hand, some new components appear with formation of a charge transfer complex as it has
a chemical shift of 2.6 and 4.6 eV tentatively been claimed for metallization of polyimide [23].
attributed to AlOF and AlF3 respectively (fig. 3~). For Al on PET (the present study) some evidence
On the other hand, a film of 20 ML still has the have been obtained on the intervention of the
signature of the presence of fluorine even for carboxylic group in the mechanism (fast attenua-
deposition at low temperature. In these condi- tion of the C,, carbon) and on the formation of
tions, a diffusion process probably takes place Al-O bounds at the interface. This last observa-
leading to a diffuse interface. tion is consistent with previously published SIMS
studies [6] where an oxidized Al layer was found
at the interface. It appears that the bond breaking
4. microscopic obse~atious model is certainly the more suitable for this
peculiar metal polymer system.
Transmission electron microscopy plane section - The fluorine treated polymer presents a
are prepared by dissolving the PET in trifluoro- peculiar behaviour:
acetic acid in order to calculate the average grain (i) Departure of fluorine from the skin of the
size (AGS) of the Al film deposited on 77 and 300 polymer is observed upon heating at 500 K;
 J.F. Silvain et al. / XPS of Al/polyethylene terephralate interface 193

(ii) Fluorine atoms are involved in the binding [lo] Y. de Puydt, P. Bertrand, Y. Novis, M. Chtaib and P.
Lutgen, Colloque GFP, Le Mans, November 1988.
mechanism and a diffuse interface is produced
[ll] M. Alnot, B. Weber, J.J. Ehrhardt and A. Cassuto, Appl.
giving a metallic layer with unusual small grains. Surf. Sci. 2 (1979) 578.
It is also observed that the fluorine treatment (121 C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder and
improves the adhesion of the film on the polymer. G.E. Huilenberg, Handbook of X-ray Photoelectron Spec-
troscopy (Perkin-Elmer, Eden Prairie, 1978).
[13] C.D. Wagner and D.M. Bickham, NIST X-ray Photoelec-
tron Database (US Department of Commerce NIST,
References Washington, 1989).
1141 D. Briggs and M.P. Seah, Practical Surface Analysis (Wi-
[l] N.J. Chou and C.H. Tang, J. Vat. Sci. Technol. A 2 (1984) ley, New York, 1983) ch. 9.
751. 1151 Y. Novis, M. Chtaib, R. Caudano, J. Vohs, J.J. Pireaux. P.
[2] F.S. Ohuchi and SC. Freilich, J. Vat. Sci. Technol. A 4 Lutgen and G. Feyder, Symposium on Metallized Plastics,
(1986) 1039. Chicago. Illinois, 9 October 1988.
[3] J.W. Bartha, P.O. Hahn, F.K. Le Goues and P.S. Ho, J. ]161 M. de Koven and L. Hagans, Appl. Surf. Sci. 27 (1985)
Vat. Sci. Technol. A 3 (1985) 1390. 199.
[4] P.S. Ho, P.O. Hahn, J.W. Bartha. G.W. Ruhloff. F.K. Le ]171 P. Swift, D. Shuttleworth and M.P. Seah. Practical Surface
Goues and B.D. Silverman, J. Vat. Sci. Technol. A 3 Analysis (Wiley, New York, 1983) p. 437.
(1985) 739. ]181 M. Chtaib, private communication.
[51 N.J. Dinardo, J.E. Demuth and T.C. Clark, Chem. Phys. ]191 J.M. Burskstrand, J. Vat. Sci. Technol. 16 (1979) 1072.
Lett. 121 (1985) 239. ]201 C.A. Chang, Y.K. Kim and A.G. Scrott, ACS Symp. Ser.
]61 Y. de Puydt. P. Bertrand and P. Lutgen. Surf. Interface 440 (1990) 416.
Anal. 12 (1988) 486. 1211 Y. Momose, K. Ikawa and T. Satd. J. Appl. Polym. Sci. 33
]71 J.F. Silvain. A. Picco, J.J. Ehrhardt and P. Lutgen, ACS (1987) 2715.
Symp. Ser. 440 (1990) 453. ]221 Lj. Atanasoska, S.G. Anderson, H.M. Meyer III. Zhangda
[81 M. Chtaib, Y. Novis, R. Caudano, P. Lutgen and G. Lin and J.H. Weaver. J. Vat. Sci. Technol. A 5 (1987)
Feyder. ACS Symp. Ser. 440 (1990) p. 288. 3325.
191 Y. Novis, M. Chtaib, R. Caudano, P. Lutgen and G. 1231 R. Haight, R.C. White, B.D. Silverman and P.S. Ho, J.
Feyder, Br. Polym. J. 21 (1989) 171. Vat. Sci. Technol. A 6 (1988) 2688.