Fichier PDF

Partage, hébergement, conversion et archivage facile de documents au format PDF

Partager un fichier Mes fichiers Convertir un fichier Boite à outils PDF Recherche PDF Aide Contact



analysis of brass .pdf



Nom original: analysis of brass.pdf
Titre: PII: S0254-0584(98)00268-5

Ce document au format PDF 1.2 a été généré par 3B2 Total Publishing 5.55b/X / Acrobat Distiller 3.02, et a été envoyé sur fichier-pdf.fr le 13/03/2010 à 12:42, depuis l'adresse IP 193.253.x.x. La présente page de téléchargement du fichier a été vue 2319 fois.
Taille du document: 273 Ko (9 pages).
Confidentialité: fichier public




Télécharger le fichier (PDF)









Aperçu du document


Materials Chemistry and Physics 58 (1999) 212±220

Heat transfer to metals in low pressure oxygen plasma:
application to oxidation of the 90Cu±10Zn alloy
a

K. Draoua, N. Bellakhala, B.G. CheÂron b, J.L. Brisseta,*

Laboratoire d'Electrochimie L.E.I.C.A, UFR des Sciences et Techniques de Rouen F-76821, Mont Saint Aignan Cedex, France
b
URA 230-CORIA, UFR des Sciences et Techniques de Rouen F-76821, Mont Saint Aignan Cedex, France
Received 9 October 1998; received in revised form 5 November 1998; accepted 24 November 1998

Abstract
A low pressure inductively coupled oxygen plasma was used to oxidize the surface of a-brass (90Cu±10Zn) foils. Electrochemical and
optical (i.e, UV±Vis±Nir diffuse re¯ectance spectroscopy and FTIR) investigation techniques show that the oxide layer initially formed on
chemically cleaned samples mainly involve the copper oxides; the initial oxide layer was formed in majority by copper oxides Cu2O and
CuO. When the CuO layer is extensively grown, a ZnO layer begins growing together with CuO. The thickness and composition of the
oxide ®lms formed at the surface are governed by matter and energy transfers between the plasma and brass surface. The in¯uence of the
reactor working parameters is also considered. # 1999 Elsevier Science S.A. All rights reserved.
Keywords: Oxygen plasma; Heat transfer; Brass; Copper oxides; Zinc oxide; Spectroscopy surface analysis

1. Introduction
Low-temperature plasmas are widely used for surface
treatments, such as to prepare diamond-like layers, to
change the molecular structure of polymers or to modify
the surface of solids by conferring new properties on them or
improving the existing ones. An example of surface modi®cation of metals is given by plasma nitriding which
induces increased hardness, shorter heat treatments or thermal treatments at lower temperatures [1±3] on the treated
material. Another pertinent example is the formation of
passivating oxide layers to prevent or to limit corrosion
phenomena on the metal or to prepare insulating layers at its
surface. However, the plasma oxidation of metals is not
widely studied and little information is available. On the
other hand, the thermal oxidation of metals and alloys was
studied by many investigators [4±8] on the assumption that
an equilibrium involving the gas and the metal occurs. Also,
the oxidation of alloys constituted by two easily oxidized
components may lead to either one or two phase scales [9].
Both oxides may simultaneously nucleate and continue
growing under conditions for which the oxide of the less
noble metal is favored.
The RF discharges in oxygen were studied from various
points of view. One aim of these studies is to determine the
*Corresponding author.

composition of the bulk plasma and the kinetics related to
the different components. Another aim is to precisely study
the physical and chemical mechanisms which take place at
the surface.
Copper forms several alloys with zinc. The 90 wt.% Cu±
10 wt.% Zn alloy was thus selected as a suitable material to
illustrate the oxidizing properties of an RF oxygen plasma
and to quantify the in¯uence of the discharge parameters.
The primary key parameters to vary the plasma treatment
conditions are the distance d between the ®rst HV coil and
the sample, the gas pressure p, and the electric power P in
addition to the treatment time t. These primary parameters,
except t, are gathered in the plasma density P/n where n
refers to the number of gaseous species present in the
reactor.
The copper and zinc oxides formed by the treatment at the
sample surface are identi®ed by optical methods such as
FTIR, UV±Vis±NIR diffuse re¯ectance spectroscopy, X-ray
diffraction and by complementary electrochemical methods
(i.e., linear sweep voltammetry). The ®lm thickness is
determined by interferometry.
2. Experimental
The experimental device is the same as the one used for
the previous studies on copper and zinc [10,11]. The plasma

0254-0584/99/$ ± see front matter # 1999 Elsevier Science S.A. All rights reserved.
PII: S0254-0584(98)00268-5

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

213

Fig. 1. Experimental setup.

reactor is a quartz tube 10 cm in diameter and 30 cm in
length. The driving frequency of the 13.56 MHz generator is
transferred to the gas in the reactor by means of an impedance matching network ending in a seven turn copper coil.
Vacuum (100±4000 Pa) is sustained by a primary pump. The
samples are placed on a water cooled stainless steel quenching head normally to the gas ¯ow (Fig. 1). The sample
temperature is measured by means of a K-type chromel±
alumel thermocouple, the weld of which is set in a hole
bored in the top part of the quenching head 0.3 mm underneath the sample. Hence, the relevant uncertainty on the
temperature cannot exceed 0.18C. On account of the strong
RF noise, the thermocouple acts as an antenna during the
experiments which constrains us to record the signal after
switching off the plasma. The sample temperature Ts is then
deduced by extrapolating the relaxation signal to the initial
time (Fig. 2 is relevant to P ˆ 1100 W, p ˆ 23 102 Pa,
t ˆ 5 min).
The 0.5 mm thick brass (90Cu±10Zn) foils were purchased from Tre®metaux. The industrial grade material
contains less than 0.3% of total impurities. The samples
were chemically cleaned in 15% H2SO4 for 10 s, then
washed in absolute ethanol and nitrogen dried before being
exposed to the plasma.
The UV±Vis±NIR diffuse spectra of the treated alloy
were recorded on Perkin Elmer Lambda 9 spectrophotometer equipped with an integrating sphere (BaSO4). The
specular re¯ectance infrared spectra were recorded on

Fig. 2. Temperature relaxation of a 90Cu±10Zn foil exposed to the
oxygen plasma (P ˆ 1100 W, p ˆ 23 102 Pa, d ˆ 5 cm, t ˆ 5 min).

Nicolet FTIR 710 spectrometer (analyzed range: 5000±
225 cmÿ1, with 168 and 808 incidence angles).
The electrochemical measurements were performed at
room temperature in a three electrodes glass cell ®lled with a
1 mM NaOH solution. The potentials are referred to the KCl
saturated calomel electrode (SCE). The voltammetry curves
are recorded from 0 to ÿ1.6 V versus SCE on a PAR 273
potentiostat with a potential sweep rate of 0.5 mV sÿ1.
The X-ray diffraction patterns were recorded at a near
grazing incidence angle ( 68) with a cobalt anticathode
(Co ka1, ˆ 0.1789 nm).

214

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

3. Characterization of copper and zinc oxides
3.1. Copper oxides
Exposure of copper samples to a low pressure oxygen
plasma shows the formation of several oxides which were
spectrophotometrically and electrochemically characterized
[11±13]. The results agree with thermal oxidation studies.
The oxidation of copper starts with the formation and the
growth of a precursor oxide CuxO. This oxide has the same
crystallographic structure as Cu2O [14]. The relevant UV±
Vis diffuse re¯ectance spectra are similar to those of
metallic copper except for an absorption band in the
360±380 nm range. In addition, the oxide CuxO is electrochemically reduced at a higher potential than Cu2O [13].
The UV±Vis±NIR spectra of non stoichiometric cuprous
oxide Cu2O3 exhibit absorption bands between 400 and
550 nm. The electrochemical reduction peak takes place in
the range ÿ0.85/ÿ0.95 V versus SCE. This peak moves
towards the lower potentials (i.e., ÿ1.1 V versus SCE) when
the ®lm thickness increases. The IR re¯ectance spectra of
Cu2O/Cu system consist of two bands around 650 cmÿ1 (LO
mode) and 610 cmÿ1 (TO mode).
Czanderna [15] prepared the oxide of the composition
CuO0.67 (or Cu2O3) by low temperature oxidation of a
copper ®lm. This oxide composition is a gross defect
structure of Cu2O. A previous study [11] con®rms the
differences between the optical properties of Cu3O2 and
Cu2O which are respectively characterized by luminescence
emissions at 760 nm (intense) and 720 or 820 nm. Cu3O2
have similar vibrations to those of Cu2O (FTIR).
The optical spectrum of the cupric oxide is characterized
by a sharp rise of absorption in the range 700±800 nm. The
IR re¯ectance spectra of CuO/Cu system consist of many
bands around 605, 530, 470 cmÿ1 (TO mode) and 620, 580,
550, 510 cmÿ1 (LO mode).
3.2. Zinc oxide
Exposure of zinc foils to an inductively coupled low
pressure RF oxygen plasma leads to the formation of
ZnO thin ®lms which were spectroscopically and electrochemically characterized [10]. The UV±Vis±NIR spectrum
of ZnO presents a band at 380 nm which corresponds to the
absorption edge of the ZnO optical transition
(gap ˆ 3.2 eV). The FTIR spectrum shows a band between
570 and 590 cmÿ1 (LO mode) and two other bands at 410
and 370 cmÿ1 (TO mode). The electrochemical reduction
peak takes place in the range ÿ1.3/ÿ1.4 V versus SCE [16].
3.3. Oxidation of brass
The thermal oxidation of Cu±Zn alloys was thoroughly
studied and the results are widely reported in the literature;
the oxidation kinetics and mechanisms were studied over
the whole composition range of the alloy [17±19]. The

oxides developed on the brass foils at a high temperature
(i.e., T ˆ 873 K) are ZnO, Cu2O and CuO, disposed on a
multilayer structure, with the copper oxides at the surface
and ZnO at the oxides±metal interface.
4. Results and discussion
4.1. The plasma phase
Recent work [20±22] devoted to O2 low pressure
capacitive RF discharges shows satisfactory agreement
between experiments and numerical calculations. In
these models, electronic states of oxygen molecules X3 ,
a1 , b1 , c1 , c3 , A3 and ozone molecules O3 are
considered as the most important plasma components
besides electrons and ions O2‡, O2ÿ, Oÿ and atomic states
O(3P), O(1D), O(1S). Simulation of such RF plasma processes are very complicated tasks and require simplifying
assumptions.
In our experimental conditions ± a low pressure oxygen
plasma generated by a 13.56 MHz inductive excitation ± the
complexity of the physical situation is enhanced by the
presence of the quenching head which makes capacitive the
relaxation region of the plasma ¯ow downstream the excitation coil. To our knowledge, there is to date no theoretical
modeling of such a situation. So, considering this failure, we
will merely present the plasma-wafer heat transfer though
realistic orders of magnitude.
The major contributions to the surface thermal load are
the molecular heat transfer of translational, rotational and
vibrational energy, the catalytic reactions involved in the
atomic oxygen recombination and the radiative exchange.
The ¯ow can be regarded as a continuous medium since
the Knudsen number is lower than 0.1. Hence, the convective part of the heat transfer may be expressed as

@Tg
ˆ h…Ts ÿ Tg †
qcv ˆ ÿkg
@y s
where the indices g and s respectively refer to the gas and the
wall. kg is the thermal conductivity including the internal
energy which is due to other degrees of freedom than those
associated with translational motion. The convection heat
transfer coef®cient h can be evaluated by means of semiempirical formula [23] which expresses the Nusselt number
NuD ˆ hD/kg as a function of the Reynolds and Prandtl
numbers. In our conditions, the characteristic length D is the
quenching head diameter (D ˆ 4 cm), the temperatures are
Tg ˆ 500 K and Ts ˆ 600 K and the mean ¯ow velocity
Vq 1.2 m sÿ1. These values roughly lead to
qcv 200 W mÿ2 for the convective heat ¯ux from the
quenching head to the gas.
The heat transfer involved in catalytic reactions is currently calculated through two parameters, the `gamma' ( )
coef®cient which is the ratio of the recombined atoms to the
striking atoms at the surface and the `beta' ( ) coef®cient

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

215

which is the ratio of the energy transferred to the surface
over the total heat produced by the recombination of the
atoms.
Oxygen atomic recombination on a surface may occur by
either of the two mechanisms:
O ‡ s ! O
O ‡ O ! O 2
O 2

LangmuirÿHinshelwood …LÿH†

! O2 ‡ s

O ‡ s ! O
O ‡ O ! O 2
O 2

EleyÿRideal …EÿR†

! O2 ‡ s

Where the symbol s stands for a site, i.e., a preferential
location for the surface/gas bond. Most of what is known
about sites refers to metals where clearly de®ned potential
wells exist at speci®c locations. Sites behave like species in
homogeneous kinetics, except that they have a much higher
number of degrees of freedom. Therefore, sites may dissipate recombination energy very much like third bodies in
gas collisions.
The most probable mechanism for oxygen recombination
at low temperature (up to 973 K) is the Eley±Rideal one due
to the high coverage of the surface. The recombination
coef®cient on metals was evaluated either from the coaxial
®lament ¯ow reactor method or from the actinometric
optical emission spectroscopy, while the accommodation
coef®cient was measured by calorimetric methods
[24,25].
Comparing the details of the mechanisms with the experimental results, the ®rst-order dependence of O recombination may be explained if ®rst-order adsorption is the
controlling step. The catalytic heat release qrec to the
quenching head may be expressed as:
qrec ˆ kw ED ‰OŠ
where ED and [O] are the oxygen dissociation energy and
the oxygen number density respectively. kw is called the
`global reaction rate' of the O ‡ O ! O2 reaction. Experiments [26] have yielded kw ˆ 120 cm sÿ1 at copper surface
and kw ˆ 38 cm sÿ1 at zinc surface. In our conditions,
[O] 2.9 1015 cmÿ3 [27]. It follows that
qrec 1:9 103 W mÿ2
The third contribution to the thermal load is the net
radiative transfer qR expressed as
qR ˆ " Ts4 ÿ ER
The wafer (polished brass) radiates as a gray surface, so
its total emissivity " is equal to its total absorptivity
(" 0.1). is the Stefan±Boltzmann constant
( ˆ 5.67 10ÿ8 W mÿ2 Kÿ4). Since the plasma is optically thin, the total irradiation ER mainly originates from the
inner surface of the silica cylindrical reactor (Tp ˆ 480 K).

Fig. 3. Temperature of a 90Cu±10Zn foil treated for different values of
injected power at p ˆ 23 102 Pa, d ˆ 5 cm and t ˆ 5 min.

Accordingly,
qR " …Ts4 ÿ Tp4 † 500 W mÿ2
Therefore, the calculated value of the net thermal load
qnet to the wafer surface is
qnet ˆ qrec ÿ qcv ÿ qR 1:2 103 W mÿ2
Similar expressions applied to the stainless steel (304 L)
quenching head fairly checks the conservation of the
enthalpy between the entrance and the outlet of the cooling
water circuit. Its dependence on injected power changes is
shown in Fig. 3: Ts increases with increasing power due to
the enhancement of the ionization degree and the molecular
dissociation degree as veri®ed by recording the evolution of
the 777.196, 777.418, 777.540 and 844.65 nm oxygen line
intensities.
Let us now consider the in¯uence of pressure on the wafer
temperature. The increase of the pressure induces a shortening of the mean free path which leads to an enhancement
of the recombination in the homogeneous phase. Hence, the
number of atoms striking the wafer surface per unit time
decreases. The variation of Ts as function of the reactor
pressure is displayed in Fig. 4; it increases with increasing
pressure up to 23 102 Pa and then becomes roughly
constant. This stabilization of the temperature can be
explained by considering a balance between the decrease
of the atomic density and the charged species mean energy,
and the increase of the sticking time of the atoms at the
surface leading to an increase of the accommodation
factor mentioned above.
Table 1 summarizes the maximal temperature values of
the brass samples obtained after 8 min treatment and for
different treatment parameters.

216

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

Fig. 4. Temperature of a 90Cu±10Zn foil treated for different pressures at
P ˆ 1100 W, d ˆ 5 cm and t ˆ 5 min.

Fig. 5. Infrared reflectance spectra (808 off normal) of 90Cu±10Zn
samples exposed to the oxygen plasma (P ˆ 800 W, p ˆ 23 102 Pa,
d ˆ 5 cm) for (1) 3 min, (2) 4 min, (3) 6 min.

Table 1
Maximum temperatures Tmax (K) obtained from 90Cu±10Zn samples
exposed to 8 min plasma treatment

Table 2
Zinc±copper ratios for 90Cu±10Zn and 70Cu±30Zn samples chemically
cleaned in 15% H2SO4

d (cm)
5
10
20

P(Pa):
P(W):

2 102
1100

2 102
800

23 102
1100

23 102
800

376
358
325

338
326
312

673
618
463

563
501
402

4.2. Reaction mechanism
Focusing on the ®rst steps of the oxidation mechanism,
we paid special attention to the ®rst oxidation state of the
alloy surface at the beginning of the oxidation process. The
brass samples were thus exposed to the plasma (injected
power P ˆ 800 W, gas pressure p ˆ 23 102 Pa) at different distances d from the ®rst inductive coil (d ˆ 5 cm, 10
and 20 cm) and for different exposure times t. The IR
re¯ectance spectra of samples treated for 3 min or more
were recorded under an incidence angle of 808. The spectra
of the 3 min treated samples suggest that the initial oxidation process involves the copper oxides (Cu2O and CuO) as
the major oxide species (Fig. 5) because they show two
bands at 478 cmÿ1 and 525 cmÿ1 which characterize the
transversal optical vibration mode (TO) of CuO, and a
shoulder at 645 cmÿ1 which is relevant to the longitudinal
optical mode (LO) of Cu2O. The presence of ZnO is related
to the occurrence of a minor peak at 581 cmÿ1 which is
attributed to the LO vibration mode.
The nature of the oxides present at the surface may result
from the technique used to prepare the sample surface since
a sulfuric acid cleaning leads to a copper enriched surface

a-brass

Zn/Cu ratio
(in middle)
before cleaning

Zn/Cu ratio
(on surface)
after cleaning

90Cu±10Zn
70Cu±30Zn

0.14
0.49

0.03
0.04

due the preferential attack of zinc. Van Ooij [28] measured
by XPS the surface concentration of copper and zinc on
90Cu±10Zn and 70Cu±30Zn alloys chemically cleaned in
15% H2SO4. The results reported in Table 2 show that the
surface enriches with copper, whatever be the zinc charge of
the alloy. For exposure times longer than 3 min (i.e.,
t ˆ 4 min), the ZnO band at 593 cmÿ1 (LO mode) develops
while that of Cu2O at 645 cmÿ1 vanishes. With increasing
times of exposure (i.e., 6 min), the intensity of these bands
becomes stronger and other LO vibration modes of ZnO and
CuO appear (i.e., at 412 cmÿ1 and 395 cmÿ1 for ZnO, and at
328 cmÿ1 for CuO). In addition, the spectra shows a
shoulder at 645 cmÿ1, which is attributed to the LO mode
of Cu2O.
The UV±Vis±NIR diffuse re¯ectance study of the brass
alloy provides spectra which are more dif®cult to interpret
than the FTIR technique, due to the superimposition of the
absorption bands of zinc and copper oxides in the spectrum
range (Fig. 6). ZnO shows two main absorption zones: 1000
and 380 nm. Cu2O absorbs down to 600 nm and CuO
exhibits a board absorption band down to 850 nm. Moreover, an intense absorption peak at 280 nm due to the
specular part of the light re¯ected by metallic substrate

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

217

Fig. 6. UV±Vis±NIR diffuse reflectance spectra of 90Cu±10Zn samples
exposed to the oxygen plasma (P ˆ 800 W, p ˆ 23 102 Pa, d ˆ 5 cm)
for (1) 3 min, (2) 4 min, (3) 6 min.

appears on all the spectra, even on the diffuse re¯ectance
spectra. The spectra of samples treated for 3 min or longer
times present only the absorption band around 700 nm
speci®c of the CuO oxide, and this band increases drastically for exposures times longer than 4 min.
The results of the optical study are con®rmed by an
independent technique based on electrochemical measurements. For this purpose we the linear potential sweep
voltammetry method in pure diffusion conditions. Fig. 7
(1) illustrates the typical curves (current (i in mA) versus
applied potential (E in V)) recorded for ®xed samples
previously exposed to the plasma for 3 min. The minor
peak at ÿ0.75 V versus SCE and the shoulder at ÿ0.88 V
versus SCE are respectively assigned to the ®rst and second
reduction step of CuO: [Cu(II) ‡ e ! Cu(I)] (®rst reduction
step) and [Cu(I) ‡ e ! Cu(0)] (second reduction step). The
peak around ÿ0.95 V is related to the reduction of Cu(I)
oxide and that of ZnO is observed at ÿ1.20 V versus SCE.
Fig. 7 (2) is typical for long treatments: it shows a major
peak at ÿ1.05 V versus SCE, corresponding to the reduction
of all the copper oxides formed at the metal surface. For thin
®lms, the second reduction peak of CuO is masked by the
second reduction peak of the Cu(I) oxides. A second minor
peak is observed at ÿ1.22 V versus SCE and related to the
reduction of ZnO.
The nature of the phases present in the ®lms formed at the
90Cu±10Zn surface by the oxygen plasma treatment is
con®rmed by X-ray diffraction (Table 4). For short plasma
exposures, CuO is the major species formed (i.e., for
t ˆ 3 min), and for longer exposure times (t ˆ 6 min), all
the oxides (copper and zinc oxides) are largely developed,
particularly CuO.

Fig. 7. Cathodic reduction curves of oxides formed by an oxygen plasma
treatment of 90Cu±10Zn alloy foils (P ˆ 800 W, p ˆ 23 102 Pa,
d ˆ 5 cm) for (1) 3 min, (2) 4 min.

4.3. Modeling the reflection spectra
Infrared re¯ection spectroscopy is an appropriate method
to study and understand the complex structures such as those
developed during the oxidation of copper and its alloys
[29,30]. The optical constants and absorption bands are
calculated by using the model of independent harmonic
oscillators which is deduced from the classical theory of
bound electrons and applied to dielectric materials. Our aim
is to determine the structure of the oxide layer developed
during the oxidation of brass by oxygen plasma by comparing the spectra calculated for different structures to our
experimental spectra. However, the lack of optical data
relevant to CuO prevents the achievement of an acute
model. Hence, the theoretical spectra were determined from
a linear combination of spectra relevant to ZnO and CuO.
These spectra were provided by two particular brass samples: the ®rst one was the 70Cu±30Zn alloy, mechanically
polished and exposed for 10 min to the plasma
(P ˆ 1100 W, p ˆ 23 102 Pa, d ˆ 5 cm). ZnO is then
the main oxide present in the ®lm ((LO) 605 cmÿ1, (TO)
390 and 410 cmÿ1). The second spectrum is provided by
chemically cleaned 90Cu±10Zn brass treated for 4 min
(P ˆ 800 W, p ˆ 23 102 Pa, d ˆ 5 cm), which yields an

218

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

Fig. 8. (a) Experimental spectra of a 70Cu±30Zn foil polished mechanically and treated at P ˆ 1100 W, d ˆ 5 cm, p ˆ 23 102 Pa and t ˆ 10 min; (b)
experimental spectra of a 90Cu±10Zn foil cleaned chemically and treated at P ˆ 800 W, p ˆ 23 102 Pa, d ˆ 5 cm and t ˆ 4 min. Calculated infrared
spectra obtained at a 168 incidence angle from an oxide film (CuO ‡ ZnO) 0.7 mm thick of a paving structure and for different proportions of CuO; (c) 25%;
(d) 50%; (e) 75%.

oxide ®lm mainly composed of CuO ((LO) 525 and
475 cmÿ1). We could then calculate various spectra by
varying the ratio CuO/ZnO (25, 50 and 75%) for a total
thickness of 0.7 mm of the layer examined under quasinormal incidence angle (168 off-normal). Such spectra
correspond to a paving structure, where ZnO and CuO
are present adjacent to each other (Fig. 8) and ®t very well
with the experimental spectra. For example, the experimental spectrum of a 6 min plasma treated sample at
P ˆ 800 W, p ˆ 23 102 Pa, d ˆ 5 cm is shown in
Fig. 5 which illustrates the main bands: [(LO) 598 cmÿ1,
(TO) 395 and 412 cmÿ1 for ZnO and (LO) 520 and
475 cmÿ1 for CuO]. This spectrum agrees fairly well with
the synthetic spectrum calculated for a paving structure, if
we take into account the evolution of the absorption bands of
CuO and ZnO with the treatment parameters. Conversely,
the thermal treatments generally lead to multilayer structures [17±19], with copper oxides at the surface and zinc
oxide at the oxides±metal±oxides interface.
4.4. Influence of the working parameters
The primary working parameters are the gas pressure p,
the distance d and the feeding electric power P. An increase

in P induces an increase in the energy transferred to the
surface sample and consequently an increase in temperature. We shall focus on the effects induced by an increase in
p and d.
4.5. Influence of the gas pressure
A pressure change in the reactor modi®es the mean free
path and the vibration±excitation states of the heavy species
in the plasma phase, that is their reactivity, and additionally
the temperature of the target material. Hence, the composition of the oxide layer and the kinetics of the growth depend
on the temperature and therefore on the gas pressure. The
thickness of the oxide layer varies in the same way as the
temperature: for example, a pressure increase from 2 102
to 35 102 Pa (for given distance d ˆ 5 cm, time of exposure t ˆ 6 min and applied power P ˆ 800 W) induces an
increase in thickness from 0.1 to 0.55 mm. However there is
no direct relationship between the layer thickness e and the
pressure, since a maximum of e (i.e., e 0.7 mm) is
observed for p ˆ 23 102 Pa. This feature shows that the
®lm thickness does not follow the variations of pressure, but
those of the overheating of the material induced by the
evolution of the plasma regime. Table 3 reports the respec-

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220
Table 3
Effect of pressure on the oxidation of 90C±10Zn samples
Pressure (Pa)

TS (K)

Thickness (mm)

2

2 10

333

0.10

23 102

533

0.70

35 102

500

0.55

219

Thus, smaller the d, higher is the thermal effect induced by
the plasma, and the plasma ef®ciency follows the temperature changes as already discussed. The resulting chemical
effect, that is, the thickness of the oxide layer, increases as
the distance d decreases (i.e., e ˆ 0.7 mm, d ˆ 5 cm;
e ˆ 0.5 mm, d ˆ 10 cm; e ˆ 0.06 mm, d ˆ 20 cm) and in
addition, the composition of the layer depends on the
distance d (Table 4).

XRD analysis
Cu2O (M)
CuO (w)
Cu2O (M)
CuO (M)
ZnO (M)
Cu2O (w)
CuO (M)
ZnO (m)

5. Conclusions

At P ˆ 800 W, d ˆ 5 cm and t ˆ 6 min.
M: major, m: medium, w: weak.

The use of non destructive techniques (i.e., optical methods) and linear potential sweep voltammetry allows us to
identify the various oxides formed at the surface of a-brass
foils exposed to an oxygen plasma. The thickness of the
oxide layer yielded for short treatments shows the high
ef®ciency of the low pressure inductively coupled plasma
technique used despite the moderate temperature of the
samples compared with standard thermal oxidation treatments. An important exchange of energy between the
plasma and the brass foil takes place at the sample surface
and is assigned to the exothermic recombination of oxygen
atoms. The quantity of heat transferred depends on the
reactor parameters and governs the nature of the oxides
and the thickness of the layer.
Measurements performed on chemically cleaned brass
samples show that the oxidation of copper prevails over that
of zinc at all states of the oxidation. The ®rst steps of the
oxidation mechanism are clari®ed. Oxidation starts by
forming Cu2O which rapidly turns to the thermodynamically stable species CuO. When the CuO is extensively
grown, ZnO begins growing adjacent to CuO.

tive evolutions of pressure and the layer thickness with
temperature.
4.6. Influence of the distance
The distance d between the brass sample and the ®rst coil
controls the lifetime l of the reagents (i.e., l ˆ 0.202 s,
l ˆ 0.69 s for d ˆ 5 cm and 20 cm respectively), and consequently both the nature of the gaseous species present in
the reactor and the temperature of the brass sample. The
in¯uence of the distance is summarized in Table 4 for the
given working conditions (e.g., P ˆ 800 W; p ˆ 2300 Pa)
and for various t values. Oxidation of brass is affected by
temperature and consequently by the distance from the ®rst
coil. Reaction products and kinetic rates may be deeply
modi®ed according to the working conditions. For example,
the temperature of three samples treated at 5 cm, 10 cm and
20 cm for 6 min rises to 533, 498 and 391 K, respectively.

Table 4
Oxidation states of 90Cu±10Zn as a function of distance d between the brass samples and the first coil and of time exposure
Analytical
methods

d(cm):
t(min):

20
3

20
6

10
3

10
6

5
3

5
4

5
6

Cu2O (M)

Cu2O (M)

Cu2O (w)
CuO (M)
ZnO

Cu2O (m)
CuO (M)
ZnO (w)

Cu2O (w)
CuO (M)
ZnO (w)

Cu2O (w)
CuO (M)
ZnO (m)

Cu2O
CuO (M)
ZnO (M)

UV±Vis±NIR

±

±

CuO

CuO

CuO

CuO (M)

CuO (M)

XRD analysis

±

Cu2O (M)
CuO (M)
ZnO (w)

Cu2O (w)
CuO (M)
ZnO

Cu2O (m)
CuO (M)
ZnO (w)

Cu2O (w)
CuO (M)
ZnO (w)

Cu2O (w)
CuO (M)
ZnO (m)

Cu2O (M)
CuO (M)
ZnO (M)

Cu2O

Cu2O
CuO
ZnO

Cu2O
CuO
ZnO

Cu2O
CuO
ZnO

Cu2O
CuO
ZnO

Cu2O
CuO
ZnO

±
±
±

ND

0.06

0.35

0.50

ND

0.35

0.70

FTIR

Electrochemical analysis

Thickness (mm)
2

At P ˆ 800 W, p ˆ 23 10 Pa.
M: major, m: medium, w: weak, ND: not determined

220

K. Draou et al. / Materials Chemistry and Physics 58 (1999) 212±220

References
[1] H. Knupell, K. Brotzmann, F. Eberhard, Stall und Eisen 78 (1958)
1871.
[2] J. Takada, Y. Ohizumi, H. Miyamura, H. Kuwahara, S. Kikuchi, J.
Tamura, J. Mater. Sci. 21 (1986) 2493.
[3] H. Kuwahara, J. Takada, I. Tamura, in: K. Ajashi, A. Kinbara (Eds.),
Proc. 8th Int. Symp. Plasma Chem., Tokyo, Japan, 1987, p. 1709.
[4] E. Beucher, M. Lenglet, J. Lopitaux, Analusis 18 (1990) 547.
[5] B. Lefez, K. Kartouni, M. Lenglet, D. RoÈnnow, C.G. Ribbing, Surf.
Interface Anal. 22 (1994) 451.
[6] R. Souchet, M. Lenglet, P. Miche, S. Weber, S. Scerrer, Analusis 21
(1993) 173.
[7] E. Beucher, B. Lefez, M. Lenglet, Phys. Stat. Sol. 136 (1993) 139.
[8] S. Maroie, P.A. Thiry, R. Caudano, J. Verbist, Surf. Sci. 127 (1983)
200.
[9] C. Wagner, J. Electrochem. Soc. 99 (1952) 369.
[10] K. Draou, N. Bellakhal, B.G. CheÂron, J.L. Brisset, Mater. Chem.
Phys., 51/2 (1997) 142.
[11] N. Bellakhal, K. Draou, B.G. CheÂron, J.L. Brisset, Mater. Sci. Eng. B
41 (1996) 206.
[12] N. Bellakhal, K. Draou, B.G. CheÂron, M. Lenglet, J.L. Brisset, in:
J.V. Herberlin, D.W. Ernie, J.T. Roberts (Eds.), Proc. 12th Int. Symp.
Plasma Chem., Minneapolis, USA, 1995, p. 1583.
[13] N. Bellakhal, K. Draou, J.L. Brisset, J. Appl. Electrochem. 28 (1997)
134.
[14] J.M. Machfert, M. Lenglet, D. Blavette, A. Menand, A. D'Huysser,
Structure and Reactivity of Surfaces, Elsevier, Amsterdam, 1989, p.
625.

[15] H. Wieder, A.W. Czanderna, J. Phys. Chem. 66 (1962) 816.
[16] R.J. Brodd, J. Werth, in: A.J. Parsons, J. Jordan (Eds.), Standard
Potentials in Aqueous Solutions, Marcel Dekker, New York 1989, p.
249.
[17] G.C. Wood, B. Chattopadhyay, J. Inst. Metals 98 (1970) 117.
[18] T.L. Barr, J.J. Hackenberg, Appl. Surf. Sci. 13 (1982) 523.
[19] D.P. Whittle, G.C. Wood, Br. Corr. J. 3 (1968) 294.
[20] M. Shibata, N. Nakano, T. Makabe, J. Appl. Phys. 77(12) (1995)
6181.
[21] V.A. Feoktistov, A.V. Mukhovatova, A.M. Popov, T.V. Rakhimova, J.
Phys. D: Appl. Phys. 28 (1995) 1346.
[22] M. Tuszewski, J.T. Scheuer, J.A. Tobin, J. Vac. Sci. Technol. A13(3)
(1995) 839.
[23] F.P. Incropera, D. de Witt, Fundamentals of Heat and Mass Transfer,
Wiley, New York, 1985.
[24] B. Halpern, D.E. Rosner, J. Chem. Soc., Faraday Trans. l 74 (1978)
1883.
[25] P. Cauquot, in: P. Fauchais (Ed.), Proc. 4th Eur. Cong. on Thermal
Plasma Processes, Athens, ICHMT Begell House, N.Y. Wallingford,
UK, 1996.
[26] N. Rahmoun, S. Cavadias, J. Amouroux, in: Proc. 13th Int. Symp.
Plasma Chem., Peking, China, 1997, p. 905.
[27] K. Draou, Ph.D. Thesis, Universite de Rouen, 1997.
[28] W.J. Van Ooij, Surf. Tech. 6 (1977) 1.
[29] B. Lefez, R. Souchet, K. Kartouni, M. Lenglet, Thin Solid Films 268
(1995) 45.
[30] J.M. Machfert, M. Le Calvar, M. Lenglet, Surf. Interface Anal. 17
(1991) 13.


Documents similaires


Fichier PDF 4gjezht
Fichier PDF fichier pdf sans nom 1
Fichier PDF kevin richardson pure natural healing ebook
Fichier PDF n2o derm surg 518
Fichier PDF 13042 ftp
Fichier PDF oximed2012 741545


Sur le même sujet..