Studies of transport in oxides
on Zr-based materials
Clara Anghel
Licentiate Thesis
Division of Corrosion Science
Department of Material Science and Engineering
Royal Institute of Technology, KTH
SE-10044 Stockholm, Sweden
Stockholm 2004
ISRN KTH/MSE--04/71--SE+CORR/AVH
ISBN 91-7283-921-X
Abstract
Zr-based materials have found their main application in the nuclear field having high
corrosion resistance and low neutron absorption cross-section. The oxide layer that is formed
on the surface of these alloys is meant to be the barrier between the metal and the corrosive
environment. The deterioration of this protective layer limits the lifetime of these alloys. A
better understanding of the transport phenomena, which take place in the oxide layer during
oxidation, could be beneficial for the development of more resistant alloys.
In the present study, oxygen and hydrogen transport through the zirconia layer during
oxidation of Zr-based materials at temperatures around 400°C have been investigated using
the isotope-monitoring techniques Gas Phase Analysis and Secondary Ion Mass Spectrometry.
The processes, which take place at oxide/gas and oxide/metal interface, in the bulk oxide and
metal, have to be considered in the investigation of the mechanism of hydration and oxidation.
Inward transport of oxygen and hydrogen species can be influenced by modification of the
surface properties. We found that CO molecules adsorbed on Zr surface can block the surface
reaction centers for H2 dissociation, and as a result, hydrogen uptake in Zr is reduced. On the
other hand, coating the Zr surface with Pt, resulted in increased oxygen dissociation rate at the
oxide/gas interface. This resulted in enhanced oxygen transport towards the oxide/metal
interface and formation of thicker oxides. Our results show that at temperatures relevant for
the nuclear industry, oxygen dissociation efficiency decreases in the order: Pt > Zr2Fe > Zr2Ni
> ZrCr2 ≥ Zircaloy-2.
Porosity development in the oxide scales generates easy diffusion pathways for molecules
across the oxide layer during oxidation. A novel method for evaluation of the gas diffusion,
gas concentration and effective pore size of oxide scales is presented in this study. Effective
pore sizes in the nanometer range were found for pretransition oxides on Zircaloy-2.
A mechanism for densification of oxide scales by obtaining a better balance between inward
oxygen and outward metal transport is suggested. Outward Zr transport can be influenced by
the presence of hydrogen in the oxide/metal substrate. Inward oxygen transport can be
promoted by oxygen dissociating elements such as Fe-containing second phase particles. The
results suggest furthermore that a proper choice of the second-phase particle composition and
size distribution can lead to the formation of dense oxides, which are characterized by low
oxygen and hydrogen uptake rates during oxidation.
Keywords: Zirconium, hydrogen, oxygen diffusion, second-phase particle, oxidation,
dissociation, hydration, carbon monoxide, adsorption, porosity.
i
ii
Preface
This thesis includes the following papers:
I. Gas phase analysis of CO interactions with solid surfaces at high temperatures.
Clara Anghel, Erik Hörnlund, Gunnar Hultquist, Magnus Limbäck
Applied Surface Science 233 (2004) p. 392–401
DOI number: doi:10.1016/j.apsusc.2004.04.001
II. Influence of Pt, Fe/Ni/Cr–containing intermetallics and deuterium on the oxidation
of Zr-based materials.
Clara Anghel, Gunnar Hultquist and Magnus Limbäck
Accepted for publication in Journal of Nuclear Materials
III. A method for characterization of gas transport in porous oxides
C. Anghel, Q. Dong, J. Rundgren, G. Hultquist, I. Saeki and M. Limbäck
Manuscript
iii
Table of content
1. Introduction .................................................................................................……...... 1
2. Oxidation of Zr-based materials
2.1 Oxidation in dry environment …………………………………………….
3
2.1.1 Thermodynamic considerations …………………….………...
4
2.1.2 Zirconium-oxygen phase diagram ……………………………...
6
2.1.3 Kinetic considerations………..…………………………………
7
2.2 Oxidation in wet environment …………………………………………….
8
2.3 Influence of second-phase particles (SPP)………………………………...
10
2.4 Diffusion ………………………………………………………………….
11
3. Experimental
3.1 Materials …………………………………………………………………
12
3.2 Gas Phase Analysis Technique (GPA).………………………….……….
12
3.2.1 Outgassing ……………………………………………..………
13
3.2.2 Hydration and oxidation ………….………..…….…………….
14
3.2.3 Dissociation and exchange ………….….……….…………….
15
3.2.4 Permeation ………………………..……………..…………….
18
3.3 Secondary Ion Mass Spectroscopy (SIMS).….………………..….……..
18
3.4 X-ray Photoelectron Spectroscopy (XPS)……………………….……….
19
3.5 Scanning Electron Microscopy (SEM)…….……………………..……...
19
4. Summary of appended papers
4.1 Gas phase analysis of CO interactions with solid surfaces at high
temperatures. …………………………………………………………..
20
4.2 Influence of Pt, Fe/Ni/Cr–containing intermetallics and deuterium
on the oxidation of Zr-based materials…………………………………
22
4.3 A method for characterization of gas transport in porous oxides ……...
26
5. Conclusions ......................................................................................................…
29
6. Future work ......................................................................................................…
30
7. Acknowledgments ............................................................................................…
31
8. References .......................................................................................................….
32
Papers I-III
v
1. Introduction
Martin Heinrich Klaproth, a German chemist, discovered zirconium metal in 1789
by analyzing the composition of the mineral jargon (ZrSiO4) [1]. In 1824, Jöns Jacob
Berzelius, a Swedish chemist, isolated for the first time zirconium metal in impure form and
after 90 years, in 1914, pure Zr was prepared by D. Lely and L. Hamburger [1]. Zirconium
is usually associated with hafnium, being difficult to be obtained in pure form [2]. The high
resistance to corrosion in aggressive environments at high temperatures makes the Zr metal
to be appropriate in many applications, such as: heat exchangers, pumps, reactor vessels,
valves, surgical appliances [1,2]. Furthermore, Zr metal has a low neutron absorption crosssection (unlike Hf), therefore the nuclear power industry became the major application of
the purified Zr [1]. The necessity for better in-reactor performance generated further
materials development, and Zr-based alloys containing Sn, Fe and Cr with or without Ni,
called Zircaloy-2 and –4, have been designed for use in water-cooled nuclear reactors [2].
Other alloys like Zr-2.5%Nb have also been developed for use as pressure tube material.
A more efficient use of the nuclear fuel is limited by the corrosion of the fuel
cladding. Further development of the existing alloys by optimization of the alloying
elements chemical composition, size distribution and density for higher burn-up applications
is the subject of many recent studies [3,4,5]. Fe, Cr and Ni have a low solubility in the Zr
matrix and are therefore mainly incorporated in the form of second-phase particles (SPP)
[6]. Nodular corrosion was observed in systems were initially large size SPPs or
agglomerated small size SPPs were present [5]. In studies related to the Pt influence on the
oxidation of other metals than Zr, the same effect of enhanced oxidation in the Pt area has
been observed [7,8]. The presence of Pt is connected to the generation of more oxygen in
the form of On- (n = 0, 2) [7,8]. In the case of zirconium alloys, each alloying element has
certain effects on the oxidation and it seems plausible that one effect is related to efficiency
for oxygen dissociation. In this study, oxygen dissociation efficiency on preoxidized ZrCr2,
Zr2Fe and Zr2Ni intermetallics as well as preoxidized Zircaloy-2 and Pt is investigated upon
exposure to a mixture of oxygen isotopes at relevant temperatures for the nuclear industry.
One aim is to study the possible influence of O2 dissociation rate on the corrosion rate of
zirconium alloys. The influence of Pt on the oxidation of Zr in 20 mbar O2 at 400°C is also
investigated.
1
Another problem in the use of Zr-based materials in wet environments at high temperatures
is related to hydrogen uptake [9]. A certain fraction of the hydrogen produced during the
reaction of Zr with water, diffuses through the oxide layer and is accumulating in the bulk
metal. When the hydrogen content exceeds the solubility limit, hydrides starts to form, and
as a result, hydrogen embrittlement may occur [9-11]. Many parameters influencing
hydrogen uptake in zirconium alloys have been identified, such as: composition and size
distribution of alloying elements, corrosive environment, oxide microstructure, temperature
and irradiation [9-12]. T. Laursen et al. [10] suggested that hydrogen ingress in Zr-based
materials depends on the surface preparation. By modifying the surface properties, hydrogen
uptake rate can be influenced [10]. In this work, the modification of the surface of
preoxidized Zr by blocking the active sites for hydrogen dissociation is investigated. The
effect of CO and N2 is hereby considered.
In recent publications dealing with other metals than Zr [13-15], it has been shown that the
presence of hydrogen in the metal substrate or in the gas phase has an effect on oxidation by
increasing the metal outward transport, which eventually can generate voids at the
metal/oxide interface. One possible mechanism is based on a proton-induced high
concentration of metal ion vacancies in the oxide, which is likely to result in an increased
metal ion transport [15]. Oxidation of Zr by cation outward diffusion via cation vacancies
was already taken into consideration in early 50’s by E. A. Gulbransen and K.F. Andrew
[16]. In this study, the influence of hydrogen, present in the substrate, on the oxidation of Zr
in O2 at 400°C is investigated. Evaluation of the hydrogen uptake upon exposure of 1m long
Ar-filled Zr-based tubes to water from one end of the tubings at 370°C is also presented. G.
Hultquist et al. [13] indicated that the oxidation rate decreases if a better balance between Zr
ion and oxygen ion transport is obtained, resulting in adherent oxide growth, with low
density of pores and other defects. Obviously, healing of pores by oxide formation needs
outward Zr transport. The rate of metal consumption in corrosion is dramatically dependent
on the transport of species like oxygen and hydrogen through the oxide layer, so defects like
pores can have a significant influence on the corrosion rate. Even pores with nm-size and
their abundance, distribution and interconnectivity may affect the corrosion rate. Naturally
where one is aiming at obtaining a gas-tight barrier, the size of the interconnected pores
needs to be smaller than the size of the diffusing molecules. N. Ramasubramanian et al. [11]
suggested that the pretransition zirconium oxide layer is porous not only in the outer part,
but also within the barrier layer. Therefore, determination of the open porosity of
pretransition oxides on Zr-based materials is crucial for a better understanding of the
transport processes, which take place during oxidation.
2
Finally, a novel and relatively straightforward method to quantify the effective pore size,
gas diffusion and amount of gas present in oxide scales, with known oxide thickness, on
samples equilibrated in controlled atmosphere is presented. The method is validated in
measurements of diffusivity and solubility of He in quartz at 80°C and also applied for
characterization of an Fe-oxide and two pretransition oxides on Zircaloy-2.
The aim of this thesis is to improve the knowledge related to the transport of species like
oxygen and hydrogen through the growing Zr-oxide layer using Gas Phase Analysis and
Secondary Ion Mass Spectrometry.
2. Oxidation of Zr-based materials
2.1 Oxidation in dry environment
The formation of the oxide layer on Zr-based materials has been extensively studied during
the last decade with a wide variety of techniques [16-19]. Generally, starting with a virtually
clean metal surface, the formation of the oxide layer can be divided in stages as illustrated in
Figure 2.1 [20].
Figure 2.1- General mechanism for metal-oxygen reaction, after Kofstad [20]
In the case of Zr, in the first stage, dissociative adsorption of oxygen molecules on the Zr
metal surface takes place. The sticking probability of oxygen molecules on the clean Zr
3
surface is close to 1 for coverages up to 1 monolayer (ML) [21]. Oxidation proceeds with
the oxide nucleation and growth followed by a continuous thin film formation [20]. C. -S.
Zhang et al. [21] have studied the initial stage of oxidation of Zr by Auger Electron
Spectroscopy (AES) and nuclear reaction analysis and at 90, 293 and 473 K. They
concluded that Zr oxide is growing layer by layer at 90 K, layer by layer for approximately
2 layers followed by island growth into the metal at 293 K, and by an island growth
mechanism at 473 K [21].
Yoshitaka Nishino et al. [22] have studied the formation of suboxides on Zr and Zircaloy-2
in the early stages of oxidation in low pressure O2 and water vapor exposures at room
temperature using AES and X-ray Photoelectron Spectroscopy (XPS). They found that three
suboxides of Zr2O, ZrO and Zr2O3 as well as ZrO2 are formed on the respective surfaces.
The last three stages in Figure 2.1 define the oxide growth, in which oxidation requires that
species like oxygen and/or metal are transported through the already existing oxide film. It
is well known that Zr oxide is growing mainly by inward oxygen transport [16,17]. The
oxide is highly stressed as a result of lattice parameters mismatch and thermal expansion
difference between Zr oxide and Zr metal [23]. There is a stress gradient across the oxide
thickness: maximum stress at oxide/metal (O/M) interface and minimum at gas/oxide (G/O)
interface [23]. As a result of the high compressive stress, the tetragonal ZrO2 phase is
stabilized at O/M interface [5,6]. Away from the O/M interface, the tetragonal to monoclinic
phase transformation takes place as a result of stress relief with a volume expansion of 7%
and generates defects like cracks and pores (easy diffusion pathways) [5]. This
transformation can occur inside the barrier layer, possibly inducing porosity within the
barrier layer [11]. The porosity development in the zirconium oxide scale is considered to be
the main reason that leads to the transition from parabolic to linear oxide growth kinetics
[24].
Oxidation of metals and alloys can be considered from thermodynamic and kinetic point of
view. Both are important to consider therefore some of the most important aspects for the
case of Zr-based materials oxidation are addressed below.
2.1.1 Thermodynamic considerations
Thermodynamically, to determine the stability of Zr oxide at different temperatures and O2
pressures, the values of the free energy, G, enthalpy, H, and entropy, S, are required [20].
The chemical reaction for Zr oxidation is presented in equation (1).
Zr(s) + O2 (g) = ZrO2 (s)
(1)
4
The standard free energy change of this reaction, ∆G0T(ZrO2), for temperatures T > 298K
can be expressed as:
∆G0T = ∆H0298 – T∆S0298
(2)
where ∆H0298 and ∆S0298 are the standard enthalpy and entropy change of the reaction at
298K.
This is a linear equation, and the slope of this line is constant as long as no phase
transformation takes place [2]. A chemical reaction can occur only if ∆G0T is negative.
When a metal is exposed to an oxygen-containing atmosphere, the metal oxide will start to
form only if the partial pressure of oxygen is higher than the dissociation pressure of the
oxide at the temperature of exposure [20]. In reaction (1), Zr can only be oxidized if:
 ∆G 0 ( ZrO2 ) 
PO2 ≥ exp− T

RT


(3)
The diagram of the standard free energy of formation of oxides per mole of oxygen versus
temperature with the corresponding dissociation pressures of the oxides, called
Ellingham/Richardson diagram, is a useful tool to predict oxide stability for different
temperatures and oxygen pressures [20]. In Figure 2.2, the Ellingham/Richardson diagram
for some oxides is presented [20].
* The line for ZrO2 was
added
to
the
original
diagram using calculated
values
for
∆G0T(ZrO2)
using eq. (2) and standard
enthalpy and entropy of
formation from ref. [25].
*
Figure 2.2 - Ellingham/Richardson diagram for some oxides, after Kofstad [20].
5
It can be seen in Figure 2.2 that Zr oxide is stable, even at low pressures obtained in ultra
high vacuum, and as a result, a thin oxide layer (2-4 nm) will form on the surface [24]. In
gas mixtures of H2/H2O or CO/CO2 the required low partial pressure of oxygen may be
achieved and under these conditions the Zr metal became stable. In Figure 2.2, the stability
of Fe, Cr and Ni oxides are also presented. In the figure, lines of constant PO2 are shown. For
example, at 500°C, ZrO2, Cr2O3, NiO and Fe2O3 respectively are stable if the partial
pressure of oxygen (atm O2) is higher than approx. 10-65, 10-42, 10-23 and 10-19 respectively
(only for zirconia is presented in Figure 2.2). These results are valuable in the case of
oxidation of alloys in low oxygen pressure to predict which alloy component will be
oxidized [20].
2.1.2 Zirconium-oxygen phase diagram
Phase diagrams are important in interpretation of the mechanism of oxidation. These
diagrams show the phases that can form as a function of temperature and composition of the
system. Correlated phase equilibrium and crystallographic information are obtained [26]. In
Figure 2.3, the Zr-O phase diagram is presented [24]. It can be seen that oxygen can be in
solid solution with αZr up to 28.6 at-% oxygen at temperatures around 500°C.
Figure 2.3 – Zr-O phase diagram, after Oskarsson [24]
Kinetic factors and other parameters like, stress, impurities also influence the phases that
will appear at the respective temperatures, and therefore have to be taken in consideration.
6
2.1.3 Kinetic considerations
During the oxidation of Zr-based materials, in different stages of oxidation, different
processes can be rate limiting. Raspopov et al. [18] investigated the initial stage of Zr
oxidation (formation of Zr-O solid-solutions) in atomic and molecular oxygen at low
pressures in the temperature range 873-1123K. They [18] concluded that oxygen
chemisorption is likely to be a rate limiting step for the oxidation process at the early stages
of oxidation. Linear oxidation rates have been reported [18]. It is suggested that the
formation of Zr-oxide takes place after the surface of Zr metal is saturated in dissolved
oxygen [18]. Generally, after the formation of a continuous thin oxide layer on the Zr
surface, the oxidation rate becomes parabolic, so called pretransition period, which is
followed by a pseudo-linear regime (three short parabolic regimes which can be
approximated with a linear regime) [27]. The transition between the parabolic and linear
oxidation kinetics is called “break away” point being associated with the change from a
protective to a non-protective oxide scale [27]. The rate of the oxidation reaction is
influenced by many parameters such as: temperature, oxygen pressure, surface preparation
and metallurgical characteristics of the metal [20, 27].
Parabolic oxidation rate equation is given by [20]:
'
dx k p
=
dt
x
where x is the oxide thickness and
(4)
k p' is the parabolic rate constant.
Parabolic oxidation rate can be interpreted as diffusion limited oxidation.
Linear oxidation rate means that the time evolution of the oxide thickness, x, is given by the
equation (5) [20]:
dx
= k1 t
dt
(5)
where k1 is the linear rate constant.
Linear oxidation rate can be interpreted as regular oxide failure in combination with new
oxide formation [27]. The rate-determining step could be: a surface or grain boundary
process or a chemical reaction.
If the rate-limiting step is the adsorption of oxygen molecules on the surface, a great
dependence of the reaction with the oxygen pressure is observed [20].
7
2.2 Oxidation in wet environment
When oxidation takes place in aqueous environment, hydrogen formed during the reaction
of Zr with water plays an important role in the oxidation kinetics [22, 28]. The oxidation
proceeds also in stages, the difference between wet and dry oxidation being related to the
effect of hydrogen uptake. Hydrogen accumulates in the metal substrate and forms solid
solutions with Zr metal until the hydrogen content reaches the solubility limit. Further
hydrogen uptake generates formation of hydrides, and as a result, hydrogen embrittlement
may occur [9-11]. Hydrogen solubility in α-zirconium at different temperatures has been
extensively studied [29]. A summary of the available data is presented in Figure 2.4.
Figure 2.4 – Hydrogen solubility in Zr, after Dupin et al. [29]
The presence of alloying elements in the Zr matrix influences the solubility of hydrogen
[30]. This might cause problems for example in the case of fuel claddings in which the liner
is made by a different Zr alloy. Takagi et al. have studied the redistribution of hydrogen in
Zr-lined Zircaloy-2 claddings [31]. They suggested that a significant amount of hydrogen is
moving from Zircaloy-2 to Zr during slow cooling, although the difference in their
hydrogen solubility limit is quite small. It is reported that among the Zr alloys, Zircaloy-4
has the highest solubility for hydrogen in the α region [30].
Hydrogen solubility also depends on interstitially dissolved oxygen [32]. For evaluation of
the hydrogen solubility in zirconium-oxygen solid solutions, the Zr-O-H ternary system,
exemplified for 700°C in Figure 2.5, can be used [32]. For small oxygen content, the
thermal solubility of hydrogen in α phase increases and then decreases at higher oxygen
content [30, 32].
8
I.
α+β
II.
α+β+δ
III.
β+δ
IV.
α+δ
V.
α+δ+ε
VI.
α+ε
VII.
δ+ε
VIII.
α + ε + ZrO2
IX.
ε+ α ZrO2
X.
α + α ZrO2
Figure 2.5 - Isothermal Zr-O-H ternary system at 700°C, after Miyake et al. [32]
The driving force for hydrogen diffusion through the oxide layer is the concentration
gradient. Using a thermal desorption technique, Miyake et al. have found that the solubility
of hydrogen in zirconia is in the range of 10-5 – 10-4 mol H/mol oxide and decreases with
increasing temperature [32]. The variation of hydrogen solubility in monoclinic zirconia as a
function of temperature is shown in Figure 2.6. The diffusion of hydrogen through the oxide
scale depends on the fraction of the tetragonal ZrO2 present in the oxide [33].
[32]
Figure 2.6 – Hydrogen solubility in some oxides versus 1/T, after Miyake et al. [32]
9
2.3 Influence of second-phase particles (SPP)
Fe, Cr and Ni have a low solubility in the Zr matrix and are therefore mainly incorporated in
the form of SPP [3,5,6,34] and also partly incorporated in the oxide during oxidation. The
size distribution of SPPs in the Zr-based materials has an important effect on the in-reactor
performance of these materials [34]. The formation rate of protective oxides and also the
thickness of the barrier layer are found to be dependent on the initial SPP size distribution
[34]. Nodular corrosion was observed in systems were initially large size SPPs or
agglomerated small size SPPs were present [5]. Dissolution of small-size SPPs in Zr-based
materials during in-reactor operation under high neutron flux in combination with hydrogen
pick-up can enhance the deterioration of the barrier layer [34]. In Zircaloy-2, for example,
two types of SPPs are reported: Zr2(Fe,Ni) and Zr(Fe,Cr)2 particles [3,5,34]. The Crcontaining particles were found to be less resistant to dissolution in the matrix during
irradiation than the Ni-containing particles [5,34]. Abolhassani et al. [17] performed in situ
studies of the oxidation of Zircaloy-4 at 700°C using environmental scanning electron
microscopy (ESEM). The samples were also characterized with atomic force (AFM) and
transmission electron microscopies (TEM). They found that after the uppermost precipitates
start to oxidise, the surface oxide, above the respective precipitates, became depleted in Zr,
showing a lens-type feature. This is due to outward diffusion of Cr and Fe probably via
grain boundaries, which will then be oxidised at the O/G interface. AFM revealed granular
morphology for the pretransition oxide and the presence of randomly distributed pores with
non-uniform size and geometry [17]. In the vicinity of SPPs, cracks lying parallel to the
interface were identified. The un-oxidized SPPs, which are found deeper in the oxide layer,
may influence the electron and hydrogen transport across the oxide layer [5]. An overall
view of possible roles of the SPPs on the oxidation of Zr alloys is illustrated in Figure 2.7
[33].
Figure 2.7 – Possible effects of SPPs on the oxidation of Zr alloys, after Lim [33]
10
2.4 Diffusion
Crank [35] defined diffusion as: “the process by which matter is transported from one part
of a system to another as a result of random molecular motion”. Fick expressed the
mathematical equation for diffusion in isotropic materials as shown in equation (6), also
called Fick’s first law of diffusion:
F =− D
∂C
∂x
(6)
where F is the flux of diffusing species, C their concentration, D diffusivity and x the space
coordinate measured normal to the section [35].
The diffusivity, or the diffusion coefficient, is related to the rate of movement of species.
The solubility represents the dissolved gas concentration per unit of applied pressure of gas
[36].
Experimentally, to determine the diffusion parameters, the differential equation of diffusion
is also used. For a plate-like geometry, if the gradient is only along the x-axis, Fick’s second
law of diffusion is expressed as:
∂C
∂ 2C
=D 2
∂t
∂x
(7)
This equation can be applied to other geometries using the appropriate coordinates.
To obtain a solution of the diffusion equation, the initial and boundary conditions have to be
considered.
Isotope tracing techniques like SIMS [37], nuclear microprobe technique [38], ion beam
induced nuclear reactions [39], are widely used for diffusion studies. Analysing the
concentration profiles of the diffusing isotopes the mechanism of diffusion can be identified.
The saturation/outgassing method is also a useful technique for determination of gas
solubility and diffusivity, using solutions of the diffusion equation corresponding to the
shape and properties of the respective samples [36]. The gas phase is continuously analysed
by a mass spectrometer and the obtained results give average values for the diffusivity of
gases in solids.
11
3. Experimental
3.1 Materials
The materials used in this study together with the methods used for the characterization of
the samples are summarized in Table 3.1.
Table 3.1 – Materials used in this study
Dissociation
Oxidation
SIMS
XPS
SEM
Zr plate
x
x
-
x
x
x
1, 2
Zircaloy-2 plate
x
x
x
-
-
-
1, 2, 3
1m long Zr-based tubes
-
x
-
x
-
-
2
Fe plate
x
x
x
-
-
x
3
Pt plate
x
-
-
-
-
-
1, 2
Ni plate
x
x
-
-
-
-
1
Al plate
x
x
-
-
-
-
1
Cr plate
x
x
-
x
-
-
1
Cu plate
x
x
-
-
-
-
1
Cu-8Al
x
x
-
-
-
-
1
SS 304
x
x
-
-
-
-
1
SS 304-0.1Pt
x
x
-
-
-
-
1
Zr2Fe
x
x
-
-
-
-
2
Zr2Ni
x
x
-
-
-
-
2
ZrCr2
x
x
-
-
-
-
2
Quartz plate
x
-
x
-
-
-
3
Material
Gas diffusivity,
pore size
estimation
Paper
3.2 Gas Phase Analysis Technique (GPA)
The schematic diagram of the GPA equipment presented in Figure 3.1 consists of:
•
A 70 cm3 reaction chamber made of a silica tube and a stainless steel cross. The
reaction chamber is pumped with an ion pump via a leak valve.
In oxidation of 1m long Zircaloy-2 tubes, the silica tube in Figure 3.1 is
replaced by the Zircaloy-2 tube (paper 2).
12
In the experiments applied for evaluation of effective pore size of oxide
scales (paper 3), the silica tube has been removed giving too high
background levels.
For permeation studies, the silica tube is replaced by a double wall tube.
•
A gas handling system where pressures up to 1 atm can be used.
•
A pressure gauge to measure the total pressure inside the reaction chamber.
•
A mass spectrometer (MS), with a quadruple analyser, placed in an UHV chamber.
Hydrogen and different isotopes can be detected.
•
A tube furnace, which can be used up to 1200°C.
The GPA equipment can be used for different types of measurements: outgassing,
hydration, oxidation, dissociation, exchange and permeation. These different measurements
are shortly described bellow with some examples.
Printer
Leak
valve
Pressure
gauge
Tube furnace
UHV
QMS
Quartz tube
400°C
Ion pump
Sample
Rough
pump
P in MS, mbar
Pressure
gauge
Time, s
Computer
Gas
handling
system
Figure 3.1 – Gas Phase Analysis setup
3.2.1 Outgassing
Heating of a sample in vacuum generates release of gases that can be quantified. The sample
is placed in the reaction chamber and pumped continuously by the ion pump via a fully open
leak valve. The MS analyses the composition of the gases released by outgassing and the
computer graphically reproduces the time evolution of the partial pressure of the gas
constituents. By careful calibration of the mass spectrometer signal and ion pump rate, the
amount of gas released from the sample can be quantified. This enables, for example, the
hydrogen content in a material to be determined. In Figure 3.2, the outgassing of hydrogen
from a zirconium plate at temperatures up to 700°C is shown.
13
Hydrogen pressure in MS, mbar
1.E-06
1.E-07
1.E-08
700°C
600°C
200°C 400°C
1.E-09
0
100
200
300
Time, h
Figure 3.2 – Outgassing of hydrogen from a Zr plate
The data obtained by outgassing from air-equilibrated oxide samples at low temperatures
(20-80°C) can be used for evaluation of gas diffusivity, solubility and effective pore size in
respective oxides.
3.2.2 Hydration and oxidation
The most common way to study oxidation of metals is to measure the weight gain of the
sample after oxidation. With GPA, the changes in the gas phase are measured. The pressure
decrease of the gas in the reaction chamber upon time is continuously monitored and is
conveniently recorded on an x-t plotter. A 1.8 cm2 Zr plate covered with a naturally formed
oxide was exposed to 7.5 mbar deuterium at temperatures 200-550°C. Calculated from the
decrease of D2-pressure in the reaction chamber, the deuterium uptake versus time in the Zr
plate is shown in Figure 3.3. After charging, the D content in the Zr plates was 600wt ppm.
As seen in the figure, the uptake rate of D in Zr is increasing considerably at temperatures
above 400°C.
For study of the oxidation mechanism, the oxidation can be performed in two stages, first in
“normal” O2 followed by exposure in
18
O-enriched oxygen gas. In a subsequent analysis
with SIMS the position of the oxide growth can be found if the exchange rate between O in
the oxide and O in O2 is slow compared with O-uptake rate in the oxidation. The previously
D-charged Zr plate, labelled Zr(D), was partly coated with Pt and exposed at 400°C for 12
14
hours to 20 mbar O2 in two stages. The oxidation kinetics presented as oxygen uptake from
the gas phase is shown in Figure 3.4.
25
550°C
15
400
10
500°C
5
200
D uptake, wt ppm
D uptake, µmol D cm
-2
600
20
400°C
200°C
0
0
0
20
40
60
Time, h
Figure 3.3 – The kinetics of deuterium uptake in a Zr plate
5
1.5
4
1
3
2
0.5
18
Adition of O
1
0
0
200
400
600
-2
6
Oxygen uptake, µmolO cm
Oxygen uptake, mbar
2
0
800
Time, min.
Figure 3.4 - Oxidation of a partly Pt coated Zr sample containing 600 wt ppm D, in 20 mbar
O2 at 400°C
3.2.3 Dissociation and exchange
Dissociation measurements
The interaction between gas molecules and a surface can lead to dissociation (molecules
split into smaller constituents). The surface activity of a material for dissociation can be
quantified [40]. To measure the dissociation rate of a molecule, a mixture of its isotopes are
15
used, for ex.
1,1
H2+2,2H2 or
16,16
O2+18,18O2. The dissociation process and a subsequent
association of the dissociated species result in the formation of the mixed molecules 1,2H2 or
16,18
and
O2. The partial pressures of the gas components,
18,18
1,1
H2,
1,2
H2 and
2,2
H2 or
16,16
O2,
16,18
O2
O2, are obtained via a negligible inlet to the MS. By considering the measured
formation rate of the mixed molecules and the distance from the statistical equilibrium (no
more changes in the gas composition), the dissociation rate can be calculated. The
background dissociation rate of the silica tube must of course also be taken into account.
The composition at statistical equilibrium depends on the isotopic abundance in the gas
phase. In the case of hydrogen isotopes, for an initial 50% 1,1H2 + 50% 2,2H2 gas mixture, the
statistical equilibrium composition is: 25%
1,1
H2 + 50%
2,2
1,1
H2, 50%
1,2
H2, 25%
2,2
H2. Starting with 50%
H2 and assuming a constant dissociation rate, vdiss, the kinetics of
1,2
H2
formation is shown in Figure 3.5.
Figure 3.5 - Time evolution of 1,2H2 molecules, with an initial composition of the hydrogen
gas 50%
1,1
H2 + 50%
2,2
H2. fE is the fraction of the mixed molecules at statistical
equilibrium (t→ ∞ ).
In this figure the preferable region for measurement with good accuracy is indicated.
In the following, dissociation refers to molecular dissociation. vdiss can be calculated from
[40]:
v diss = 0.04 ⋅ n ⋅
dP
−1
⋅Vch ⋅ ( f E ⋅ B ⋅ As )
dt
(6)
where

f 

B = 1  fE 
(7)
16
In equation (6), the factor 0.04 (mol atm-1 dm-3) comes from the molar volume of an ideal
gas at standard pressure and temperature, P is the partial pressure of the “mixed” molecules
(1,2H2 or
16,18
O2) (atm), As is the sample area (cm2), Vch is the volume of the reaction
chamber (dm3). The factor n is 1 for molecules of the form XY (one X and one Y atom per
molecule) and 2 for molecules of the form X2 (two X atoms per molecule). The dissociation
rate, νdiss, given by equation (6) is expressed in mol X cm-2h-1. Equation (7) defines a
compensation factor, B, where f is the fraction of the mixed molecules in the gas phase at
time t and fE is the fraction of the mixed molecules in the gas phase at statistical equilibrium.
As an example, in Figure 3.6, the time evolution for isotopic mixtures of hydrogen in
exposure to a preoxidized Zircaloy-2, Zr2ox, (approximately 2µm oxide thickness) at 300°C
is presented.
Figure 3.6 - Pressure in reaction chamber during exposure of an 18 cm2 Zircaloy-2 sample
at 300°C in a gas mixture with 1H and 2H isotopes. The indicated statistical equilibrium
among the hydrogen molecules is based on the abundances of 1H and 2H at 2 hours.
From Figure 3.6, hydrogen dissociation rate was calculated using eq. (1). The resulted
dissociation rate is in the range 2.7 – 7.7 µmol H cm-2 h-1. The dissociation rate decreased in
time, which shows that the activity of the surface can change in time.
Exchange measurements
Exchange may take place between oxygen in the gas phase and oxygen in the oxide. This
exchange can be measured in exposure of a 16O-containing oxide to
rate of
16,18
18,18
O2. The formation
O2 in the gas phase will then be a result of exchange between 18O from the gas
phase and 16O from the oxide lattice.
17
3.2.4 Permeation
Permeation of gases from the high-pressure side of a membrane to the low-pressure side,
which is evacuated by an ion pump, can be investigated using isotopic gas mixtures. Special
setup with double wall tubes is used for this purpose.
3.3 Secondary Ion Mass Spectroscopy (SIMS)
SIMS analyses were performed with a Cameca IMS-6f apparatus, 10 keV, 50-200 nA
primary beam of Cs+ ions, which was rastered over 200x200 µm2, where ions where
detected from a centered area with a diameter of 70 µm.
The oxidation mechanism can be investigated by using two stage oxidation, fist in
16,16
O2
followed by exposure to 18O-enriched oxygen gas, in combination with SIMS depth profiles
OnMen+
O2
a. Mainly inward oxygen
transport
Men+
Men+
Men+
18O
18O
18O
Sputter time, s
O2
On-
count rate
count rate
On-
Men+
count rate
On-
metal
O2
On-
oxide
analysis. The oxide growth mode can be retrieved as can be seen in Figure 3.7.
Sputter time, s
b. Balanced transport
Sputter time, s
c. Mainly outward metal
transport
Figure 3.7 – Mechanism of oxide growth obtained by investigations of 18O SIMS depth
profiles
The exchange with the lattice has to be considered when profiles like b. and c. (Figure 3.7)
are obtained. When the exchange rate is negligible related to the oxidation rate, it can be
concluded that the mechanism of oxide growth for case b. is by balanced transport and for
case c. is mainly by outward metal transport.
18
The oxide thickness can be obtained based on sputter time when the 90Zr ion count rate has
decreased to half of its maximum value.
SIMS depth profiles of
90
Zr18O, and
90
Zr16O ions were used in addition to
18
O and
16
O
profiles to investigate the oxide layer formation on Zr surface.
An advantage of SIMS technique is the capability of detecting hydrogen-containing species.
Deuterium depth profiles give valuable information about the mechanism of hydrogen
uptake.
18
OD SIMS profiles were considered to elucidate the possible connection between
hydrogen and oxygen diffusion.
3.4 X-ray photoelectron spectroscopy (XPS)
XPS analyses have been performed with a Kratos AXIS HS spectrometer using a
monochromatic Al Kα X-ray source (1486.6 eV). The area of analysis was approximately
0.4 mm2.
The method is based on the photoelectric effect. Upon exposure of a sample to photons with
relatively high incident energy, electrons from the internal shell of the atoms are released.
The energy of these photoelectrons can be expressed as:
E kin = hν − E B −Φ
(8)
where hν is the X-ray incoming energy, EB is the binding energy of the electron and Φ is the
work function of the material. The kinetic energy of the electron depends on the incident
photon energy, but the binding energy is a physicochemical constant. Therefore, XPS can be
used to identify which elements are present at the outermost surface of a sample and also to
characterize the nature of the chemical bonds, which connects these atoms.
3.5 Scanning Electron Microscopy (SEM)
SEM uses electrons to form the image of the surface of a sample. When the incoming beam
of electrons hits the surface, photons, secondary and backscattered electrons are ejected
from the sample. The ejected electrons are used in SEM to obtain high-resolution images of
the topography or morphology of the surface. FEG-SEM analysis were performed with a
Leo 1530 Field Emission Scanning Electron Microscope equipped with a GEMINI field
emission column.
19
4. Summary of appended papers
4.1 Gas phase analysis of CO interactions with solid surfaces at high temperatures
In Paper 1, the deactivation of a surface related to adsorption and dissociation of molecules
has been examined and related to problems, which are common in industrial applications.
Using the GPA technique, the dissociation of hydrogen on preoxidized-Zr and of carbon
monoxide on oxidized Cr have been investigated. Carbon monoxide is well known for its
poisoning effect on different catalysts.
The dissociation rates of carbon monoxide on various materials exposed to 20 mbar CO at
different temperatures have been studied. The results are presented in Figure 4.1.
Figure 4.1 – CO dissociation rate on different materials exposed to 20 mbar CO at different
temperatures
High dissociation rates have been measured on materials such as pure Cr, Ni, Fe and
stainless steel SS 304. It is interesting to note that these materials often suffer from high
carbon uptake in CO containing atmospheres, a negative consequence of the high CO
dissociation rates. On the other hand, relatively low dissociation rates were found on Zr, Cu,
Al, Cu-8Al alloy, and Pt. CO adsorption on the surface of these materials can have a
20
positive effect. Carbon monoxide is identified as an effective site blocker for hydrogen
adsorption and dissociation on preoxidized Zircaloy-2 (Zr2ox) at temperatures around
400°C. The uptake of hydrogen, measured as a pressure decrease in the reaction chamber
upon exposure to Zr2ox, is shown in Figure 4.2.
Figure 4.2 – Influence of carbon monoxide addition to the gas phase on the hydrogen
pressure upon exposure to Zr2ox and time evolution of hydrogen pressure in the reaction
chamber without sample at 400°C.
It can be seen that the uptake rate of hydrogen is substantially lowered (approximately 40
times) when 2 mbar carbon monoxide is added to the gas phase. No influence of nitrogen
gas on the hydrogen uptake by Zircaloy-2 at 400°C was seen.
When Cr2O3 was exposed to carbon monoxide at 600°C, a decrease in dissociation rate was
observed upon addition of water. This can be interpreted as a blocking effect of carbon
monoxide by water. Upon a subsequent removal of water from the gas phase, the carbon
monoxide dissociation rate increased again. The results suggest that the surface activity for
carbon monoxide dissociation should be a key factor for carbon uptake in certain
applications and can be reduced by adsorbed water.
Based on the results of the present work the following rating for the tendency of adsorption
on Zr2ox and Cr2O3 at temperatures in the range 400-600°C has been made: N2 < H2 < CO <
H2O.
21
4.2 Influence of Pt, Fe/Ni/Cr–containing intermetallics and deuterium on the oxidation
of Zr-based materials
In Paper 2, the transport of oxygen and hydrogen through the growing zirconia layer in the
oxidation of Zr-based materials in O2 at 400°C and in water vapour at 370°C was
investigated using GPA and SIMS.
The paper is divided in three sections.
I.
Oxygen dissociation on Zr-based materials and on Pt
II. Influence of Pt and deuterium on the oxidation of Zr in O2 at 400°C
III. Oxidation of Zr-based tubes in water at 370°C
I. Oxygen dissociation on Zr-based materials and Pt
Oxygen dissociation rate on Pt as well as on preoxidized Zr2Fe, Zr2Ni, ZrCr2 and Zircaloy-2
in exposure to 20 mbar O2 has been measured and the results are summarized in Figure 4.3.
T,°C
600
500
400
350
Dissociation rate, mol O cm s
-2 -1
10-6
700
Pt
10-8
Zr Fe
2
Zr Ni
10-10
2
10-12
Zry-2 oxide
ZrCr
2
10-14
0.0009
0.00105
0.0012
0.00135
0.0015
0.00165
1/T,1/K
Figure 4.3 – Oxygen dissociation rate on Pt, preoxidized Zr2Fe, Zr2Ni, ZrCr2 and Zircaloy-2
in 20 mbar O2
At 400°C it can be seen that the oxygen dissociation efficiency decreases in the order: Pt >
Zr2Fe > Zr2Ni > ZrCr2 ≥ Zircaloy-2. It is important to note that the dissociation rate of O2 on
preoxidized-Zr2Fe is 103 times higher than on preoxidized-Zircaloy-2 at 400°C.
22
II. Influence of Pt and deuterium on the oxidation of Zr in O2 at 400°C
Two 1.8 cm2 Zr plates (99.9% metal basis excl. Hf) were used in this study. One sample was
charged with deuterium to 600 wtppm D. Both samples were then partly coated with 200Å
porous Pt. The samples were oxidized at 400°C in two stages, first in
16,16
O2 then in
18
O-
enriched oxygen gas for totally 12 hours.
After oxidation, the samples were characterized with SIMS, XPS, SEM and optical
microscopy.
A summary of the 18O SIMS results, expressed as integrated oxygen-counts from the second
1 1012
8 1011
Integrated O-count rates from 2
nd
stage of oxidation
stage of oxidation, is presented in Figure 4.4.
6 1011
Effect of Pt on
the oxidation of Zr
Zr
4 1011
Effect of Pt on
the oxidation of
Zr(D)
2 1011
O-spill over
~ 10 nm
O-spill over
~ mm
0
Area with
Pt particles
Zr(D)
0
2000
4000
6000
8000
Distance from Pt-area, µm
Figure 4.4 – Effect of Pt and D on the oxide growth on Zr in 20 mbar O2 at 400°C
It can be seen that an enhanced oxidation in the area with Pt particles takes place on both
samples. In the vicinity of the Pt-enriched surface area a local minimum in oxide thickness
is obtained for both samples and the thinnest oxide is obtained on the Zr sample with D in
the substrate. At the interface between the Pt-enriched area and the area without Pt-particles,
the oxygen dissociation rate decreases about 105 times as can be seen in Figure 4.3. There is
also a spillover effect by surface diffusion of the dissociated oxygen from the area with Pt
particles, which is identified as having mm-range effect. These observations suggest that in
the vicinity of the Pt-enriched area, which has a lower activity of On- compared with the area
with Pt particles, an outward Zr diffusion induced by D from the substrate balances the On23
inward flux and can explain the formation of a more dense oxide. Diffusion profiles for
oxygen and hydrogen were investigated and related to the latest findings in the field of Zr
oxidation.
We suggest that in the oxidation of Zr-based materials in water containing atmospheres,
hydrogen accumulation in the oxide near the oxide/metal interface could explain the
frequently observed deterioration of the “barrier” layer. This also indicates that Zr could
move easier in the deteriorated barrier when hydrogen is present.
Our XPS results show that at the oxide/gas interface, the Zr oxidation state varies, being
high in the Pt area (probably ZrO2 or ZrO2+x), and some suboxides are probably present
away from the area with Pt particles. For the sample with D in the substrate, the presence of
D in the oxide at the oxide/gas interface generates higher binding energies for O1s and Zr
3d5/2, especially outside the Pt-enriched area.
A general illustration of the influence of Pt and D on the oxidation of zirconium in O2 at
400°C is presented in Figure 4.5. For simplicity, the molecular transport is not shown in the
figure.
Zr
Zr(D)
Porous Pt coating
O2
Zr oxide
Zr metal
Zr metal with D in the substrate
Zrn+
Dn+
On-
Figure 4.5 - Illustration of the influence of Pt and D on the oxide growth on Zr
24
III. Oxidation of Zr-based tubes in water at 370°C
The exposure of two Ar-filled 1 m long Zr-based tubes to 22 mbar water vapour at 370°C
for 305 and respectively 630 minutes was investigated. The oxide thickness and the
hydrogen content were measured at different positions from the water inlet using SIMS
depth profiles.
Integrated SIMS count rates of oxygen and deuterium along the length of the tubes are
shown in Figure 4.6.
Related to the water inlet, the thickest oxide was obtained at 330-400 mm. Local minima in
the oxide thickness were found at 430 mm and 478 mm respectively. At these positions
there is some hydrogen in the substrate, as a result of the reaction with water. In this case,
both oxygen and hydrogen diffuse inward towards the oxide/metal interface. At 996 mm no
significant oxidation takes place but the deuterium content is slightly higher than at the
water inlet position (20°C). Deuterium, resulting from the reaction of D2O with Zr at 370°C,
diffuses faster than D2O through the Ar filled tubes.
D-counts
630 min.
250
200
O - counts
630 min.
150
100
50
O - counts
305 min.
O - counts
20°C
100
Oxygen counts, a.u.
Counts in SIMS, a.u.
Oxygen counts, a.u.
300
305 min.
Ar+H2O
50
0
415
420
425
d, mm
150
430
435
630 min.
Ar+D2O
100
50
0
400
450
500
d, mm
550
600
0
0 D - counts
20°C
200
400
600
800
1000
Distance from water inlet (d), mm
Figure 4.6 – SIMS integrated oxygen and deuterium count rates along the length of the Zrbased tubes after oxidation in 20 mbar water vapour for 305 and 630 minutes respectively
The experiment shows that the hydrogen content in the substrate influences the formation of
the oxide layer.
25
A balance between inward oxygen and outward metal diffusion has been previously
proposed as an important condition for obtaining oxides, such as zirconium oxide, with high
protective ability [13]. The inward oxygen flux can be promoted by depositing Pt or other
oxygen dissociating elements on the Zr surface. We propose that Fe-containing secondphase particles act as oxygen dissociating elements in analogue way as Pt. The outward
zirconium flux, on the other hand, can be promoted by introduced hydrogen in the oxide
lattice. We suggest that this phenomenon is beneficial for healing defects in the zirconium
oxide scale, such as pores and microcracks, which are mainly formed during oxidation
caused by the inward oxygen diffusion. Hence, a proper choice of second-phase particles
composition and size distribution can be beneficial for reducing the negative effect of
hydrogen uptake.
4.3 A method for characterization of gas transport in porous oxides
In Paper 3, a theoretical method based on gas release is presented, which allows calculation
of gas diffusivity and concentration in samples with specific geometries such as plate,
cylinder and sphere. The model is also used to estimate the effective pore size of
microporous and/or mesoporous oxides, previously equilibrated in controlled humid air.
Low outgassing temperatures are used for this purpose. This method is non-destructive and
can also be applied to characterize metal oxide scales, with known oxide thickness. The
characterization of the open micro- and/or meso- pores formed in the oxide during early
oxidation stages of metals is crucial. The gas release during outgassing is quantitatively
evaluated using a mass spectrometer having a quadrupole analyser, placed in ultra high
vacuum. The method is validated in measurements of diffusivity and solubility of He in
quartz at 80°C. Two pretransition oxides on Zircaloy-2 and one Fe-oxide have been
characterized.
Diffusion of nitrogen and water in Zircaloy-2 and Fe oxide scales
Using the solution of the differential equation of diffusion for plate-like geometry and
assuming that the initial concentration is uniform and equal with C1, and there is no flux
across L=0 (metal substrate), the mathematical equation for flux during outgassing was
obtained.
26
oxide
metal substrate
The outgassing from an oxide scale of thickness L is illustrated in the Figure 4.7:
F=
UHV
2 DC1
L
∞

∑ exp − (2n − 1) π
2
2
n =1
Dt 
4 L2 
L
Figure 4.7 - Outgassing from an oxide scale of thickness L
The experimental data for water and nitrogen flux evolution in time and the integrated flux
over time are used for determination of the diffusivity (D) and concentration (C1) of these
gases in Zircaloy-2 and Fe-oxide scales. The results are presented in Table 4.1 and 4.2
respectively.
Table 4.1 – Transport parameters for H2O and N2 in oxide scales on Zircaloy-2 at 80°C
H2 O
1.3 µm
Oxide thickness, L
2
D, cm /s
N2
3.4 - 5.2 10
3
C1, µmol/cm
3 µm
-14
8.6 10
190
-14
- 1.1 10
119
1.3 µm
-13
2.4 - 2.8 10
3 µm
-13
2.1 10
3.8
-13
5.0
Table 4.2 - Diffusivity and concentration of H2O and N2 in Fe oxide scale at 20°C
Oxide thickness, L
2
H2 O
N2
300 µm
300 µm
3-6 10
D, cm /s
3
C1, µmol/cm
-9
90
5 10-9
10
Effective pore size estimation
We assume that the pores are long cylinders with circular cross section of diameter Φ. The
samples used for this study were equilibrated with controlled humid air at room
temperature. Under these conditions, the water molecules, which have higher tendency for
adsorption than N2 molecules, will stick to all surfaces, including internal pore surfaces. N2
molecules will predominantly be found in the pore’s gas phase. The H2O/N2 molar ratio,
27
which can be associated with the ratio between the surface area and the volume of the pores,
can be used to calculate the effective pore size.
The C1 values for water and nitrogen from Table 4.1 and 4.2 were used to calculate the
H2O/N2 molar ratio for the Zr and Fe oxide scales. The estimated effective pore size for the
above-mentioned oxide scales is presented in Figure 4.8.
1000000
1.0 ML
100000
0.04 ML
1000
100
10
0.1
Fe oxide
1
Zry-2
3 µ m oxide
Zry-2
1.3 µ m oxide
H2 O/N2 molar ratio
10000
0.01
0.1
1
10
100
1000
Effective pore diameter, Φ, nm
Figure 4.8 - Effective pore size estimation for oxide scales on Zircaloy-2 and Fe
Effective pore sizes in the range of 1-10nm and 100-1000nm for the Zircaloy-2 and Fe
oxides have been found.
28
5. Conclusions
The transport of oxygen and hydrogen in zirconia oxide scales at temperatures around
400°C has been investigated using the isotope monitoring technique Gas Phase Analysis and
Secondary Ion Mass Spectroscopy, and the following conclusions have been drawn:
Modification of the oxide surface properties (oxide/gas interface) can influence the
inward transport of species such as hydrogen and oxygen. We can distinguish two
main effects:
o Inhibitor effect: CO is identified as an effective site blocker for hydrogen
adsorption and dissociation on preoxidized Zircaloy-2 at temperatures around
400°C. No influence of N2 on the hydrogen uptake in Zircaloy-2 is found. Based on the results of the present work the following rating for the tendency
for adsorption on oxidized Zircaloy-2 and Cr2O3 at temperatures in the range
400-600°C has been made: N2 < H2 < CO < H2O.
o Catalytic effect: porous Pt coated on the Zr surface, catalyses the oxygen
dissociation at the gas/oxide interface, promoting the inward oxygen
transport. At 400°C, oxygen dissociation efficiency decreases in the order: Pt
> Zr2Fe > Zr2Ni > ZrCr2 ≥ Zircaloy-2. The dissociation rate of O2 on
preoxidized-Zr2Fe is 103 times higher than on preoxidized-Zircaloy-2 at
400°C.
Hydrogen present in the oxide generates enhanced outward diffusion of Zr.
Combining the effect of Pt (enhanced inward oxygen transport) and D (enhanced
outward Zr diffusion), a better balance in the oxide growth was obtained, and as
a result, a more protective oxide layer was formed.
We propose that Fe-containing second-phase particles act as oxygen dissociating
elements in the analogue way as Pt does. Hence, a proper choice of second-phase
particles composition and size distribution can be beneficial for reducing the
negative effect of hydrogen uptake.
Open porosity with effective pore diameter in nanometer range was found in
thermally-grown pretransition oxides on Zircaloy-2.
29
6. Future work
The results obtained opened new perspectives for understanding of the phenomena involved
in the transport of hydrogen and oxygen species in the Zr-oxide scales. Further investigation
of the hydrogen transport in zirconia, including new ways to reduce the hydrogen uptake
needs to be considered. Dual transport mode (atomic and molecular) for hydrogen transport
in silica was identified in studies performed by E. Hörnlund and G. Hultquist [41]. One
question arises: Can atomic and/or molecular hydrogen be transported through the zirconia
layer, and under which conditions? We found in pretransition oxides in Zircaloy-2 porosity
in the nanometer range, which can easily accommodate molecular species like hydrogen
and/or water.
Formation of hydrides is also an important issue that needs to be further clarified. To
become hydride, hydrogen needs to receive electrons, which is not difficult inside metals,
but for oxides, special reducing conditions are necessary. Close to the oxide/metal interface,
these reducing conditions are possible to appear, and thus, the oxide becomes
thermodynamically instable, oxygen can dissolve in the metal underneath the oxide, and
hydrides start to from. The formation of “substitutional hydroxides” could be the
prerequisite for hydride formation inside the oxide layer.
The correlation between the second-phase particles composition, size distribution and the
hydrogen uptake could be interesting to investigate using the GPA technique. Better
knowledge of hydrogen uptake rate by Zr intermetallics exposed in hydrogen gas or aqueous
environment could also be useful to consider for optimisation of the existing alloys.
Oxygen transport in zirconia, investigations of dual transport mode is certainly still another
subject of interest not only for nuclear energy production but also for fuel cell technology.
30
7. Acknowledgments
I would like to thank:
My supervisor, Doc. Gunnar Hultquist for guiding me in using the GPA technique, for
sharing with me valuable knowledge and for the long discussions and scientific support.
Prof. Christofer Leygraf, as an excellent leader for the “corrosion family”, for the friendly
and fruitful working atmosphere, for reviewing my manuscript and for his guiding and
moral support in difficult moments.
Magnus Limbäck, my supervisor from Westinghouse Electric Sweden AB, for organising
this project, for his kindness and friendship and for all the stimulating discussions. Thanks
also to Mats Dahlbäck and Per Tägtström for their suggestions and fruitful discussions.
Erik Hörnlund for all his help related to computers, animations and experimental questions
and for the friendly working atmosphere.
Dr. John Rundgren for his suggestions about animations and for sharing his knowledge
about diffusion.
All Zr Experts Group Friends for interesting discussions and scientific advice.
Qian Dong, Wenle He, Andreas Pettersson and Jan Sandberg for being really nice
roommates and friends.
All my colleagues from the corrosion division for their friendship. It is a real pleasure to
work with you! Anna, Johan, Qian and Gunilla thanks for the relaxing dinners after hardworking days. Klara, Inger and Sofia for the moral support and friendship. Babak for help
during my party. Jinshan and Bo for suggestions and fruitful discussions.
Peter Szakalos for helping me with the SEM and LOM and being a good friend.
Johan Angenete for guiding me in Boston and for fruitful discussions.
Financial support from Westinghouse Electric Sweden AB is gratefully acknowledged.
Finally I would like to thank my family, especially my husband, for making everything
easier, even being kilometres away.
31
8. References:
[1]
[2]
P. Patnaik, Handbook of Inorganic Chemicals, The McGraw-Hill, p. 995 (2003).
L.L. Shreir, R.A. Jarman, G.T. Burstein, Corrosion, vol. 1, Third Edition, Oxford:
Butterworth-Heinemann (1994).
[3]
P. Barberis, E. Ahlberg, N. Simkic, D. Charquet, C. Lemaignan, G. Wikmark, M.
Dahlbäck, M. Limbäck, P. Tägtström, B. Lehtinen, Zirconium in the Nuclear
Industry: Thirteenth International Symposium, ASTM International, STP 1423, p.
33-58 (2001).
[4]
D. Sen, S. Mazumder, R. Tewari, P.K. De, H. Amenitsch, S. Bernstorff, J. of Alloys
and Compounds 308, p. 250–258 (2000).
[5]
P. Rudling, G. Wikmark, J. Nucl. Mater. 265, p. 44-59 (1999).
[6]
C. Lemaignan, Zirconium in the Nuclear Industry: Thirteenth International
Symposium, ASTM STP 1423, Annecy, France, p. 20-29 (2001).
[7]
G. Hultquist, E. Hörnlund, Q. Dong, Corrosion Science, vol. 45, p. 2697-2703
(2003).
[8]
G. Hultquist, G.I. Sproule, S. Moisa, M.J. Graham and U. Södervall, J. Electroch..
Soc. 150(10), p. G617-G623 (2003).
[9]
M. B. Elmoselhi, J. of Alloys and Compounds, 231, p. 716-721 (1995).
[10]
T. Laursen, G.R. Palmer, J.E. Hayson, J. Nolan and R.L. Tapping, J. Nucl. Mater.
209(1), p.52-61 (1994).
[11]
N. Ramasubramanian, P. Billot, S. Yagnik, Zirconium in the Nuclear Industry:
Thirteenth International Symposium, ASTM International, STP 1423, p. 222-243
(2001).
32
[12]
B. Cox, Y. -M. Wong, J. Nucl. Mater. 270, p. 134-146 (1999).
[13]
G. Hultquist, B. Tveten, E. Hörnlund, M. Limbäck and R. Haugsrud, Oxidation of
Metals 56(3/4), p.313-346 (2001).
[14]
G. Hultquist, B. Tveten, E. Hörnlund, Oxidation of Metals 54(1/2), p.1-10 (2000).
[15]
B. Tveten, G. Hultquist, T. Norby, Oxidation of Metals 52(3/4), p. 221-233 (1999).
[16]
E.A. Gulbransen, K.F. Andrew, Journal of Metals, pg. 394-400, April (1957).
[17]
S. Abolhasssani, M. Dadras, M. Leboeuf and D. Gavillet, J. Nucl. Mater. 321: p. 7077 (2003).
[18]
S. A. Raspopov, A.G. Gusakov, A.G. Voropayev, A.A. Vecher, M.L. Zheludkevich,
O.A. Lukyanchenko, A.S. Gritsovets, V. K. Grishin, J. Chem. Soc. Faraday Trans.,
93(11): p. 2113-2116 (1997).
[19]
X. Iltis, M. Viennot, D. David, D. Hertz, H. Michel, J. Nucl. Mater. 209(2): p. 180190 (1994).
[20]
P. Kofstad, High Temperature Corrosion, Elsevier Applied Science London and
New York, p. 2-10 (1988).
[21]
C.-S. Zhang, B.J. Flinn, P.R. Norton, Surf. Sci. 264, p.1-9 (1992).
[22]
Yoshitaka Nishino, Alan R. Krauss, Yuping Lin, Dieter M. Gruen, J. Nucl. Mater.
228, p. 346-353 (1996).
[23]
A.P. Zhilyaev, J.A. Szpunar, J Nucl. Mater. 264, p. 327-332 (1999).
[24]
Magnus Oskarsson, Study on the mechanisms for corrosion and hydriding of
Zircaloy, doctoral thesis, KTH Stockholm, ISBN 91-7170-637-2, p. 10-19 (2000).
33
[25]
CRC Handbook of Chemistry and Physics 80th edition, David R. Lide (Ed.) p.5-23
(1999).
[26]
W. Wong-Ng, R.S. Roth, T.A. Vanderah, H.F. McMurdie, J. Research of NIST,
Nov. (2001).
[27]
E.A. Garcia, G. Bérarger, ARN PI-1 p.1-10 (1997).
[28]
A. Yilmazbayhan, A.T. Motta, R.J. Comstock, G.P. Sabol, B. Lai, Z. Cai, J. Nucl.
Mater. 324, p. 6-22 (2004).
[29]
N. Dupin, I. Ansara, C. Servant, C. Toffolon, C. Lemaignan, J.C. Brachet, J. Nucl.
Mater. 275, p. 287-295 (1999).
[30]
S. Yamanaka, M. Miyake, M. Katsura, J. Nucl. Mater. 247, p. 315-321 (1997).
[31]
I. Takagi, S. Shimada, D. Kawasaki, K. Higashi, J. Nucl. Sci. Tech. 39(1), p. 71-75
(2002).
[32]
M. Miyake, M. Uno, S. Yamanaka, J. Nucl. Mater. 270, p.233-241 (1999).
[33]
B.H. Lim, H.S. Hong, K.S. Lee, J. Nucl. Mater. 312, p.134-140 (2003).
[34]
P. Tängstrom, M. Limbäck, M. Dahlbäck, T. Andersson, H. Pettersson, Zirconium
in the Nuclear Industry: Thirteenth International Symposium, ASTM STP 1423,
Annecy, France, p. 96-118 (2001).
[35]
J. Crank, The Mathematics of Diffusion, Clarendon, Oxford (1976).
[36]
J. E. Shelby, Permeation, Diffusion and Solubility Measurements, in Handbook of
Gas diffusion in Solids and Melts, ASM International (1996).
[37]
C. –S Zhang, B. Li, P.R. Norton, J. Alloys and Comp. 231, p. 354-363 (1995).
34
[38]
P. Berger, R. El Tahhann, G. Moulin, M. Viennot, Nucl. Instrum. Methods Physics
Research Section B 210, p. 519-525 (2003).
[39]
I.S. Woolsey, J.R. Morris, Corrosion 37, p. 575-585 (1981).
[40]
E. Hörnlund, Appl. Surf. Sci. 199, p.195-210 (2002).
[41]
E. Hörnlund, G. Hultquist, J. Appl. Phys., 94, p. 4819-4823 (2003).
35
36
Paper I
Applied Surface Science 233 (2004) 392–401
Gas phase analysis of CO interactions with solid
surfaces at high temperatures
Clara Anghela,*, Erik Hörnlunda, Gunnar Hultquista, Magnus Limbäckb
a
Division of Corrosion Science, Department of Materials Science and Engineering, Royal Institute of Technology,
Drottning Kristinas väg 51, S-100 44 Stockholm, Sweden
b
Westinghouse Electric Sweden AB, S-721 63 Västerås, Sweden
Received in revised form 5 April 2004; accepted 5 April 2004
Available online 28 May 2004
Abstract
An in situ method including mass spectrometry and labeled gases is presented and used to gain information on adsorption of
molecules at high temperatures (>300 8C). Isotopic exchange rate in H2 upon exposure to an oxidized zicaloy-2 sample and
exchange rate in CO upon exposure to various materials have been measured. From these measurements, molecular dissociation
rates in respective system have been calculated. The influence of CO and N2 on the uptake rate of H2 in zirconium and oxidized
zicaloy-2 is discussed in terms of tendency for adsorption at high temperatures. In the case of oxidized Cr exposed to CO gas
with 12 C, 13 C, 16 O and 18 O, the influence of H2O is investigated with respect to dissociation of CO molecules. The presented data
supports a view of different tendencies for molecular adsorption of H2O, CO, N2, and H2 molecules on surfaces at high
temperatures.
# 2004 Elsevier B.V. All rights reserved.
PACS: 82.65.P; 82.30.L
Keywords: CO; Chromium; Zirconium; Zircaloy-2; Adsorption; Dissociation; High temperature; Gas phase analysis
1. Introduction
The importance of the fundamental interactions of
gas molecules with surfaces is reflected by the vast
quantity and diversity of papers published on adsorption [1–8]. Traditional surface science techniques such
as AES, XPS, UPS, LEED, HREELS, IRAS and TPD
continue to be used to reveal new insights of molecular–solid surface interactions. However, in most of
*
Corresponding author. Tel.: þ46-8-7906670;
fax: þ46-8-208284.
E-mail address: [email protected] (C. Anghel).
the published papers the temperatures considered are
scarcely above room temperature. This reflects the
experimental difficulties to measure surface–molecular interaction at elevated temperatures. For example,
there is a large amount of data on surface coverage and
sticking probabilities of molecules on surfaces below
room temperature while there is a striking lack of this
information for higher temperatures. While waiting
for experimental methods to directly measure adsorption at high temperatures, theoretical calculations on
the subject have been performed [9–12]. Our experimental approach to the high temperature problem is to
focus not directly on the sample itself but instead on
0169-4332/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.apsusc.2004.04.001
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
the gas phase surrounding the sample. Isotopic labeled
gases are used in the technique called gas phase
analysis (GPA), where processes on a sample surface
are monitored by analyzing changes in the gas phase
with mass spectrometry. By mixing isotopic labeled
gases and measure changes in the gas phase over time,
information about surface reactions such as dissociation and exchange can be retrieved. Information about
these processes is crucial in many scientific fields
involving elevated temperatures such as high temperature oxidation [13], hydration, metal dusting (catastrophic carburization) and catalysis [14] to name a
few.
In this paper, the following two issues are considered:
1. The influence of CO on the hydrogen uptake in Zrbased materials is examined and relates to the
important technological problem of hydrogen
embitterment of the Zr-based materials used in
the nuclear power industry.
2. A summary of dissociation rates of CO on some
materials is presented. Specifically the influence
of H2O on CO dissociation on oxidized Cr is
investigated and relates to applications, where
metal dusting is at risk on stainless steel.
393
2.2. GPA equipment
Fig. 1 shows the experimental apparatus used in
the GPA studies in this work. The equipment consists
of essentially three parts. A mass spectrometer (MS)
placed in ultrahigh vacuum (UHV), an approximately
70 cm3 virtually closed reaction chamber (consisting
of a quartz tube and a cross made of stainless steel)
and a gas handling system. By means of a pressure
gauge the total pressure is measured in the reaction
chamber. The equipment allows inlet of gas mixtures
with various isotopes from the gas handling system
into the reaction chamber. The partial pressures of the
gas constituents in the reaction chamber can be
analyzed over time with the MS (placed in UHV)
via a leak valve. Gas consumption in reaction of a
sample with a gas is measured as pressure decrease in
the closed reaction chamber. Heating of a sample in
the reaction chamber during evacuation with the ionpump via a fully opened leak valve provides a method
for out-gassing samples from mostly hydrogen. A
calibration of mass spectrometer signal and ionpump rate gives the opportunity also to quantify
the amount of hydrogen and other gases released
in out-gassing [15].
2.3. Secondary ion mass spectroscopy analysis
SIMS analyses were performed with a 10 keV, 50–
200 nA primary beam of Csþ ions, which was rastered
over 200 mm 300 mm, where ions where detected
from a centered area with a diameter of 70 mm.
2. Experimental
2.1. Materials
All samples were polished down to 2400 mesh SiC
paper and rinsed in ethanol before the experiments. All
pure metals investigated had purities higher than
99.9% and the composition of the alloys are presented
in Table 1.
Leak valve
Enclosed volume
P
Table 1
Composition of alloys
MS
Alloys
Composition (wt.%)
Zircaloy-2
Cu–8Al
SS 304
1.5Sn, 0.14Fe, 0.1Cr, 0.06Ni, balance Zr
8Al, balance Cu
18.5Cr, 10.1Ni, 1.2Mn, 0.43Si, 0.41Mo,
0.25Cu, balance Fe
18.5Cr, 10.1Ni, 1.2Mn, 0.43Si, 0.41Mo,
0.25Cu, 0.1Pt, balance Fe
SS 304-Pt
Quartz
tube
10 -12 atm
Ion
pump
Tube
Sample
furnace
Rough
pump
Gas handling
system
Fig. 1. Equipment used in gas phase analysis (GPA) experiments.
394
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
2.4. Molecular dissociation
2.4.1. Calculation method
The GPA technique described above has been used
to quantify molecular dissociation rates in the following way. A mixture of isotopic labeled gases (1;1 H2 þ
2;2
H2 or 12 C16 Oþ13 C18 O or 16;16 O2 þ 18;18 O2 ) is introduced into the reaction chamber. The exchange of
isotopes between the molecules is then measured
over time. After a sufficiently long exposure time,
the composition of the gas phase will reach to statistical equilibrium (fully scrambled system where no
further changes in the gas composition takes place).
The composition at statistical equilibrium depends on
the isotopic abundance in the gas. The composition at statistical equilibrium of CO molecules,
12 16
C O, 13 C16 O, 12 C18 O, 13 C18 O, is calculated by
combinatorial analysis using the abundance of 12 C,
13
C, 16 O, and 18 O. For an initial 50% 12 C16 O þ 50%
13 18
C O gas mixture, the statistical equilibrium composition is: 25% 12 C16 O, 25% 13 C16 O, 25% 12 C18 O,
25% 13 C18 O.
In the case of hydrogen isotopes, 1 H and 2 H are
combined in the form of 1;1 H2 , 1;2 H2 and 2;2 H2 . For
an initial 50% 1;1 H2 þ 50% 2;2 H2 gas mixture, the
statistical equilibrium composition is: 25% 1;1 H2 ,
50% 1;2 H2 , 25% 2;2 H2 . Starting with 50%
1;1
H2 þ 50% 2;2 H2 and assuming a constant dissociation rate, ndiss , the kinetics of 1;2 H2 formation
is shown in Fig. 2. In this figure the preferable
Fig. 2. Time evolution of 1;2 H2 molecules, with an initial
composition of the hydrogen gas 50% 1;1 H2 þ 50% 2;2 H2 . fE is
the fraction of the mixed molecules at statistical equilibrium
(t ! 1).
region for measurement with good accuracy is
indicated.
In the following, dissociation refers to molecular
dissociation. ndiss can be calculated from [16]:
ndiss ¼ 0:04n
dP
Vch ðfE BAs Þ1
dt
(1)
where
B¼1
f
fE
In Eq. (1), the factor 0.04 (mol atm1 dm3) comes
from the molar volume of an ideal gas at standard
pressure and temperature, P is the partial pressure of
the ‘‘mixed’’ molecules (1;2 H2 , 12 C18 O, 13 C16 O and
16;18
O2 respectively) (atm), As the sample area (cm2)
and Vch is the volume of the reaction chamber (dm3).
The factor n is 1 for molecules of the form XY (one X
and one Y atom per molecule) and 2 for molecules of
the form X2 (two X atoms per molecule). The dissociation rate, ndiss , given by Eq. (1) is expressed in
mol cm2 h1. B defines a compensation factor, where
f is the fraction of the mixed molecules in the gas phase
at time t and fE is the fraction of the mixed molecules
in the gas phase at statistical equilibrium. A thorough
description of the method used can be found in [16].
2.4.2. Experimental examples
The time evolution for isotopic mixtures of hydrogen in exposure to a preoxidized zicaloy-2, Zr2ox,
(approximately 2 mm oxide thickness) at 300 8C and
isotopic mixtures of CO in exposure to Cr with native
oxide at 600 8C is shown in Figs. 3 and 4, respectively.
After the gas consumption in the MS measurements
was taken into account, the uptake rate of carbon was
just about to be measurable (approximately
0.1 mmol C cm2 h1) as seen in Fig. 4. Slopes of
the ‘‘mixed’’ molecules are measured at different
times in respective graph. The compensation factor,
B, is calculated and by the use of Eq. (1) the dissociation rates of H2 and CO are evaluated. The values for
the dissociation rate of H2 on Zr2ox and the hydrogen
uptake rate in Zr2ox at 3008 at different exposure times
are presented in Table 2. The data from Table 2 shows
that the hydrogen dissociation rate exceeds the hydrogen uptake rate, which is plausible since the hydrogen
should be absorbed in atomic form. For CO on Cr it is
shown in Fig. 4 that the dissociation rate is about
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
Table 2
Dissociation rates of H2 on Zr2ox calculated with Eq. (1) and
hydrogen uptake rate in Zr2ox at 300 8C
12
1,1
H+
1,2
2,2
H+ H
2
2
10
8
1,2
H
2
4
2,2
H
2
2
1,1
H
Hydrogen
dissociation rate
(mmol H cm2 h1)
Hydrogen
uptake rate
(mmol H cm2 h1)
0.017
0.1
0.2
0.5
6.8
7.7
6.8
2.7
1.5
0.7
0.5
0.3
Data from Fig. 3.
2
0
Exposure time (h)
Statistical
equilibrium
6
2
H pressure, mbar
2
0
1
395
t
2
∞
Time, h
3. Results and discussion
Fig. 3. Pressure in reaction chamber during exposure of an
18 cm2 zicaloy-2 sample at 300 8C in a gas mixture with 1 H
and 2 H isotopes. The indicated statistical equilibrium among
the hydrogen molecules is based on the abundances of 1 H and 2 H
at 2 h.
3.1. Influence of CO on hydrogen uptake in Zr-based
materials
The hydrogen uptake rate at 400 8C in an 18 cm2
Zr2ox sample was investigated with and without CO
present in the gas phase. The oxide on the Zr2ox was
approximately 2 mm thick. The uptake of hydrogen
measured as a pressure decrease in the reaction chamber is shown in Fig. 5. It is seen that the uptake rate of
hydrogen is substantially lowered (approximately 40
times) when 2 mbar CO is added to the gas phase. In
10 mmol C cm2 h1 at 2 h and decreases slightly. The
dissociation rates from the background (the quartz
tube) obtained from exposure without any sample in
the reaction chamber were subtracted to obtain the
relevant data from the sample.
12
(axis)
12
Dissociation rate
10
8
12
C O
16
13
C O
12
C O
13
C O
10
16
18
8
18
6
6
Uptake rate
2
0
0
5
10
15
-1
Statistical
equilibrium
(axis)
-2
4
4
Rate, mol C cm h
Pressure in reaction chamber, mbar
14
2
∞
0
Time, h
2
Fig. 4. Pressures in reaction chamber during exposure of 2.3 cm Cr at 600 8C in a gas mixture of CO containing 12 C, 13 C, 16 O, and 18 O. Also
calculated CO dissociation and carbon uptake rates are shown. The indicated statistical equilibrium among the CO molecules is based on the
abundances of 12 C, 13 C, 16 O, 18 O at 15 h.
396
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
40
blank=quartz tube, 0 mbar CO
Pressure of H , mbar
35
2
30
25
Zr2ox, 0 mbar CO
20
Zr2ox, 2 mbar CO
15
10
0
200
400
600
800
1000
1200
Time, min.
Fig. 5. Influence of CO addition to the gas phase on the hydrogen pressure in exposure of preoxidized zicaloy-2 and time evolution of
hydrogen pressure in reaction chamber without sample at 400 8C.
Table 2 is seen that hydrogen adsorbs dissociatively on
Zr2ox at a rate of about 5 mmol H cm2 h1 at 300 8C,
whereas the dissociation rate of CO on Zr at 400 8C
has been measured to 0.1 mmol H cm2 h1 (Fig. 7).
Hence, we believe that CO acts as a blocking entity for
hydrogen adsorption and dissociation.
In Fig. 5, it can also be seen that the uptake rate of
hydrogen at 400 8C in a quartz tube (reaction chamber
material) exposed to approximately 35 mbar H2 is
negligible.
A comparison between the effects of additions of N2
and CO on the hydrogen uptake in a Zr sample at
400 8C is presented in Fig. 6. The hydrogen uptake
rate is high during the first 25 min when no CO or N2 is
present in the gas phase, corresponding to the slope of
the first five data points of hydrogen pressure
(enlarged scale from 0 to 90 min in Fig. 6). The
hydrogen uptake rate was then decreased in three
consecutive additions of CO to the gas phase as seen
by the reduced slopes of the hydrogen pressure. After
50
2
7
CO
H
2
40
H
6
2
5
30
4
2
H
2
20
3
2
CO
N
2
1
10
N
N
2
2
0
0
0
Pressure of N , mbar
Pressure of H and CO, mbar
8
30
60
90
1000
2000
3000
0
4000 4080
4140
0
4200
Time, min.
Fig. 6. Influence of CO and N2 additions to the gas phase on the hydrogen uptake in Zr at 400 8C. The X-axis has been expanded between 0
and 90 min and after 4000 min.
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
397
on the stainless steel SS304 as well as on the pure
metals, Fe, Cr, and Ni, which are main components
of the steel. On the other hand, relatively low dissociation rates were found on Zr, Cu, Al, Cu–8Al
alloy, and Pt.
Two basic exchange mechanisms contribute to the
measured isotopic exchange:
4000 min in Fig. 6, the gas was changed to pure
hydrogen, which resulted in a slowly increasing
hydrogen uptake rate. Such an increase can be interpreted as the result of a ‘‘memory effect’’ from the
previous exposure to 7.4 mbar CO. Two consecutive
additions of N2 to the gas after 4140 min caused no
significant change in the hydrogen uptake rate, which
is believed to indicate a negligible tendency for
adsorption of N2 (an instant drop of hydrogen pressure
upon the first addition of N2 can be seen in Fig. 6, but
this is trivial and only due to a slight escape of
hydrogen into the gas handling system upon the N2
addition). From the experimental results in Fig. 6
concerning CO, N2 and H2, we suggest the following
tendency for molecular adsorption on Zr-based materials surfaces: N2 < H2 < CO.
The results presented regarding uptake of hydrogen
in preoxidized zicaloy-2 are the basis for a recently
obtained patent by Westinghouse Electric Sweden AB
[17].
1. exchange of oxygen and carbon in between
molecules taking place on the surface:
12
C16 O þ 13 C18 O ! 13 C16 O þ 12 C18 O
2. exchange of oxygen in CO with oxygen in a
metal-oxide:
C18 O þ Mex 16 Oy ! C16 O þ Mex 18 Oy
If mechanism 1 dominates, the abundances of the
different isotopes in the gas phase should be constant
over time. If mechanism 2 dominates, the fraction of
18
O in the gas phase decreases if the oxide on the
sample contains mainly 16 O and 18 O is used in the gas
phase. However, the ‘‘native’’ oxide formed during
sample preparation of most metals and alloys is in the
order of a few nanometer [18] and the sensitivity of the
equipment does not allow measurements of the small
amounts of 16 O in such thin oxides on the samples
(area 2–7 cm2). Therefore, it is difficult to distinguish
3.2. Dissociation of CO on different materials
The dissociation rates of CO on various materials
exposed to 20 mbar CO at different temperatures have
been studied with the technique described above and
summarized in Fig. 7. The dissociation rates are high
o
Temperature, C
900
800
700
600
500
400
-2
Dissociation rate, mol C cm h
-1
100
Cr
SS 304-0.1Pt
10
Ni
Fe
1
SS 304
Al
0.1
Pt
Zr
Cu-8Al
0.01
Cu
0.001
0.8
1
1.2
1.4
1.6
-1
1000/T, K
Fig. 7. Summary of CO dissociation rates in 20 mbar CO on some materials at different temperatures.
398
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
Table 3
Sample condition, gas composition, exposure time and total weight gain after exposure
Sample
Gas exposure at 600 8C
Exposure
time (h)
Total weight
gain (mg cm2)
Cr (native oxide on the surface)
outgassed (<5 wt. ppm H)
10 mbar 12 C16 O þ 10 mbar 13 C18 O
10 mbar 12 C16 O þ 10 mbar 13 C18 O þ 10 mbar H2 16 O
10 mbar 12 C16 O þ 10 mbar 13 C18 O
10 mbar 12 C16 O þ 10 mbar 13 C18 O þ 10 mbar H2 16 O
1
21.5
2.5
195
400
between the two mechanisms. A straightforward way
to gain information about mechanism 2 is to preoxidize a sample in 16;16 O2 and subsequently expose the
oxidized sample to 13 C18 O. The formation rate of
13 16
C O is then a measure of mechanism two. Analog
experiments where exchange between O2 and ZrO2,
Ag(O), or oxidized Pd were performed, have been
presented elsewhere [19].
From Fig. 7, it is concluded that the dissociation rate
of CO varies considerably on different materials. It is
interesting to note that the materials, which often
suffer from high carbon uptake in CO containing
atmospheres, exhibit the highest dissociation rate
among the materials investigated here. An addition
of Pt to many alloys can be beneficial for the protection against carbon uptake where both O2 and CO are
present in the gas phase. This is expected since Pt
dissociates O2 at a very high rate [13] but dissociates
CO at a very low rate (Fig. 7) and thereby the formation of an oxide rather than carbon uptake is promoted.
The alloy Cu–3Al is known to have a good oxidation
resistance [20]. In Fig. 7, it is seen that a similar alloy,
Cu–8Al, has a low CO dissociation rate and the uptake
of carbon have also been found to be low. The
combination of good oxidation resistance and low
carbon uptake, makes the Cu–8Al alloy interesting
in certain applications.
3.3. H2O influence on CO dissociation on Cr at
600 8C
3.3.1. CO dissociation on Cr studied in situ with
GPA
A Cr sample (with native oxide) was exposed at
600 8C to dry and wet CO, respectively. The exposure
was performed in four steps. The experimental conditions are summarized in Table 3 together with the
weight gain of the sample in the exposure.
The effect of H2 16 O addition on the formation rate
of CO ‘‘mixed molecules’’ 13 C16 O þ 12 C18 O upon
exposure to the Cr sample at 600 8C can be found
in Fig. 8 (only the first 35 h of the last exposure shown
in the figure). The water inlet also somewhat affected
the relative abundances of the CO molecules as can be
seen in the figure but is not of importance for the
objective of the experiment. An abrupt decrease of the
13 16
C O þ 12 C18 O formation rate after water additions
at 1 and 25 h is reflected by the change in slope of
(13 C16 O þ 12 C18 O). After these additions of water the
sample is gradually oxidized by H2 16 O and the H2 16 O
content in the gas phase decreases with a concomitant
increase in formation rate of 13 C16 O þ 12 C18 O (seen at
40–45 h). The dissociation rate of CO was calculated
from the formation rate of 13 C16 O þ 12 C18 O, corresponding to the slopes indicated in the figure, by using
Eq. (1) and are found in Table 4.
It is believed that the decreased dissociation rate of
CO due to water adsorption is caused by a blocking of
surface reaction centers. It is earlier reported [21] that
water can block dissociation sites for hydrogen on an
oxidized Zr surface. Therefore, the water molecules
should have higher tendency for molecular adsorption
than both CO and H2.
Table 4
CO dissociation rates at different times on Cr upon exposure to dry
and wet CO gas mixtures, calculated with Eq. (1) from slopes in
Fig. 8
Time (h)
Gas mixture in
the exposure
Dissociation rate
(mmol C cm2 h1)
1
1
25
25
45
CO
CO þ H2O
CO
CO þ H2O
CO
13.2
5.3
2.6
1.3
2.5
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
399
Pressure in reaction chamber, mbar
20
12
16
13
18
C O+ C O
15
10
Addition of
H
16
2
Addition of
O
H
16
2
O
5
13
16
12
18
C O+ C O
0
0
10
20
30
40
50
60
Time, h
Fig. 8. Influence of H2O on CO dissociation on Cr (with native oxide) at 600 8C (composition of the gas in exposure in Table 3). Times for
water additions are indicated.
3.3.2. Reaction product analyzed ex situ with SIMS
The mechanism of Cr oxidation in exposure to the
CO gas mixtures at 600 8C (Fig. 8) has been further
interpreted after characterization of the reaction product with secondary ion mass spectroscopy. When
using oxygen isotopes (16 O, 18 O), this technique can
give answers about where the oxide growth takes
place. The SIMS profiles of oxygen and carbon are
presented in Fig. 9.
The original ‘‘native’’ oxide formed during sample
preparation can be used in the analysis of the SIMS
profiles since it acts as a marker for the position of the
original oxide–metal interface. The CO-gas contained: 10 mbar 12 C16 O þ 10 mbar 13 C18 O (16 O/18 O
Count rate normalized to
52
Cr
120
100
16
O
80
16
18
60
O from water
O
40
16
16
O from
sample
preparation
18
( O - O)
20
12
13
( C + C) x 10
0
0
2000
4000
6000
8000
Sputter time, s
Fig. 9. Carbon and oxygen SIMS profiles on Cr sample exposed according to Table 3. Position of (16 O–18 O) peak at 7000 s corresponds to the
original Cr interface.
400
Table 5
Total counts of
sputter profiles
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
12
C,
13
C,
16
O, and
18
O calculated from SIMS-
Sample
Cr
12
1.5 1.6 1.3 6.6 0.02
C counts
C counts
16
O counts
18
O counts
P P
C/ O
13
107
107
109
108
from CO ¼ 1), (see also Table 3). By subtracting
the 18 O signal from the 16 O signal we obtain the
(16 O–18 O) ‘‘signal’’ which shows three peaks. The
(16 O–18 O) peaks at 3900 and 2500 s, correspond to
the first and second water containing exposure, respectively (see Table 3). The peak originating from the
native oxide can then be identified as the one at the
inner part of the reaction product (the peak near sputter
time 7000 s). This implies that mainly outward Cr
transport has taken place during the formation of the
reaction product, which seems plausible when comparing previous results concerning oxidation of pure
Cr [22].
The (16 O–18 O) peaks near 2500 and 3900 s appear
at the same sputter depths as minima in the 18 O profile
in Fig. 9. This shows that when H2 16 O is added to the
system, oxidation of Cr takes place mainly by water
consumption, and when water is consumed, it continues with consumption of O from CO. An accumulation of carbon takes place at the substrate interface
as can be seen in the figure.
Carbon and oxygen counts have been calculated by
integration of respective SIMS count rates down to a
sputter depth where the count rate was approximately
1% of its maximum value. The integrated counts are
presented in Table 5, where as expected, similar counts
of 12 C and 13 C are found.
4. Summary
A method for measuring dissociation of molecules
on solid surfaces and uptake of gas components in
materials, based on isotopically labeled gases and
mass spectrometry, is presented. The hydrogen uptake
in pure Zr and a preoxidized commercial zicaloy-2 is
measured and the influence of CO and N2 in the gas
phase on the hydrogen uptake rate is studied. Dissociation of CO on some pure metals and alloys is
measured and the influence of water on the CO dissociation on Cr (native oxide) is studied. The following main results have been obtained.
4.1. Influence of CO on hydrogen uptake in
zirconium-based materials
Addition of CO to H2 reduces the hydrogen uptake
both in Zr and preoxidized zicaloy-2. Approximately
10% addition of CO to H2 gives 40 times lower uptake
rate of hydrogen in Zr2ox at 400 8C. Approximately
20% addition of CO to H2 gives 10 times lower uptake
rate of hydrogen in Zr at 400 8C and further additions
of CO show the same effect of a decreasing hydrogen
uptake rate. In contrast, the decrease in the uptake rate
of hydrogen in Zr upon addition of 43 mbar N2 to
6 mbar H2 is less than 10%.
4.2. CO dissociation, influence of water and
carbon uptake
The dissociation rate of CO varies within a factor of
104 between materials studied. Knowing such rates a
forecast of the uptake of C can be made for respective
materials. It was found that upon addition of water to
the CO gas, the dissociation rate of CO on Cr
decreased which can be interpreted as a blocking
effect by water. It was also found that upon a subsequent removal of water from the gas phase the CO
dissociation rate increased again. It is shown that the
surface activity for CO dissociation should be a key
factor for carbon uptake in certain applications and
can be reduced by adsorbed water.
4.3. Tendency for molecular adsorption
Based on the results of the present work the following rating for the tendency of adsorption at high
temperatures (400–600 8C) is made: N2 < H2 <
CO < H2 O.
Acknowledgements
Financial support from the Competence Center
for High Temperature Corrosion (HTC), KME and
C. Anghel et al. / Applied Surface Science 233 (2004) 392–401
Westinghouse Electric Sweden AB is gratefully
acknowledged.
References
[1] K. Hadjiivanov, J.C. Lavalley, FT-IR spectroscopic study of
CO adsorption on monoclinic zirconia of different hydroxylation degrees, Catal. Commun. 2 (3/4) (2001) 129–133.
[2] M. Pykavy, et al., Adsorption of CO on Cr2O3(0 0 0 1), Surf.
Sci. 479 (1–3) (2001) 11–25.
[3] M.P. Hooker, J.T. Grant, The use of Auger electron spectroscopy to characterize the adsorption of CO on transition
metals, Surf. Sci. 62 (1) (1977) 21–30.
[4] H. Ogasawara, et al., Structure and bonding of water on
Pt(1 1 1), Phys. Rev. Lett. 89 (2002) 276102.
[5] A. Föhlisch, et al., How carbon monoxide adsorbs in different
sites, Phys. Rev. Lett. 85 (2000) 3309–3312.
[6] T. Becker, et al., Adsorption probability of CO on a metal
oxide: the case of oxygen-terminated ZnO and the influence
of defects, Phys. Rev. B 61 (7) (2000) 4538–4541.
[7] C. Morterra, G. Cerrato, S.D. Ciero, IR study of the low
temperature adsorption of CO on tetragonal zirconia and
sulfated tetragonal zirconia, Appl. Surf. Sci. 126 (1/2) (1998)
107–128.
[8] K. Hadjiivanov, L. Dimitrov, IR spectroscopy study of CO
and NOx adsorption on a Cu/Zr-HMS catalyst, Micropor.
Mesopor. Mater. 27 (1) (1999) 49–56.
[9] M.A. van Daelen, M. Neurock, R.A. van Santen, Reactivity
of diatomic molecules on Cu(1 0 0), Surf. Sci. 417 (1998)
247–250.
[10] H. Cheng, et al., Role of transition metal oxides in metal
dusting: density functional study, AIChE J. 44 (1) (1998)
188–196.
401
[11] K. Honkala, K. Laasonen, Oxygen molecule dissociation on
the Al(1 1 1) surface, Phys. Rev. Lett. 84 (2000) 705–708.
[12] D.C. Sorescu, et al., First-principle calculations of the
adsorption, diffusion, and dissociation of a CO molecule on
the Fe(1 0 0) surface, Phys. Rev. B 66 (2002) 35416.
[13] G. Hultquist, B. Tveten, E. Hörnlund, M. Limbäck, R.
Haugsrug, Self-repairing metal oxides, Oxid. Met. 56 (3/4)
(2001) 313–346.
[14] G. Sedmak, S. Hocevar, J. Levec, Kinetics of selective CO
oxidation in excess of H2 over the nanostructured
Cu0.1Ce0.9O2y catalyst, J. Catal. 213 (2003) 135–150.
[15] T. Åkermark, G. Hultquist, Q. Lu, Release of water and
hydrogen during outgassing of some materials, J. Mater. Eng.
Perform. 5 (1996) 516–520.
[16] E. Hörnlund, Studies of dissociation of diatomic molecules
with isotope spectroscopy, Appl. Surf. Sci. 199 (1–4) (2002)
195–210.
[17] Swedish Patent Application No. 0203815-6 (2002).
[18] F.P. Fehlner, M.J. Graham, Thin oxide film formation on
metals, in: P. Marcus, J. Oudar (Eds.), Corrosion Mechanisms
in Theory and Practice, Marcel Dekker, 1995.
[19] T. Åkermark, G. Hultquist, L. Gråsjö, Probabilities of oxygen
exchange in O2 in contact with a solid surface, J. Trace
Microprobe Tech. 14 (2) (1996) 377–388.
[20] N.J. Tannyan, G. Plascencia, T.A. Utigard, High temperature
oxidation of copper and copper aluminium alloys, Can.
Metall. Q. 41 (2) (2002) 213–218.
[21] G. Hultquist, Hydrogen pick-up kinetics and inter-film
transport, in: P. Rudling, G. Wikmark (The Expert Group),
A Review with Perspective on Performed Work, Advanced
Nuclear Technology, Uppsala, 2000.
[22] G. Hultquist, B. Tveten, E. Hörnlund, Hydrogen in chromium: influence on the high-temperature oxidation kinetics in
H2O, oxide-growth mechanisms, and scale adherence, Oxid.
Met. 54 (1/2) (2000) 1–10.
Paper II
Influence of Pt, Fe/Ni/Cr–containing intermetallics and deuterium on the
oxidation of Zr-based materials
Clara Anghel*,1, Gunnar Hultquist1 and Magnus Limbäck2
1
Division of Corrosion Science, Department of Materials Science and Engineering, Royal Institute
of Technology, Drottning Kristinas väg 51, S-100 44 Stockholm, Sweden
2
Westinghouse Electric Sweden AB, S-721 63, Västerås, Sweden
Abstract
An in-situ Gas Phase Analysis technique and the 18O-SIMS technique are used to evaluate
the transport of oxygen and hydrogen in oxidation of Zr-based materials. At 400°C, it is
found that oxygen dissociation efficiency decreases in the order: Pt > Zr2Fe > Zr2Ni > ZrCr2
≥ Zircaloy-2 .Two Zr-plates partly coated with 200 Å porous Pt, with and respectively
without D in the substrate, were oxidized in two stages at 400°C. SIMS depth profiles in the
Pt area show that an enhanced oxidation takes place mainly by inward oxygen transport. A
minimum in the oxide thickness was found near the Pt area on both Zr plates. Two Ar-filled
Zircaloy-2 tubes with ZrSn liner were exposed at 370°C to 22 mbar water, filled in from one
side. Our experimental results suggest that a proper choice of the SPP composition and size
distribution can lead to reduced hydrogen uptake during oxidation of Zr-based materials in
water.
Keywords: Zr oxidation, SPP, Ni/Cr/Fe intermetallics, Zircaloy-2, oxidation, oxygen and
hydrogen diffusion, GPA, SIMS
* Corresponding author: Email: [email protected], Tel. +46 8 7906670, Fax. +46 8 208284
1. Introduction
The corrosion behaviour of Zr-based materials has been extensively studied during the last
decade with a wide variety of techniques [1-10]. The role of a surface oxide layer is to
protect the metal substrate against the corrosive environment. A protective metal oxide
should be gas-tight and therefore only ions should be transported in the oxide. This situation
is often assumed but rarely (if ever) confirmed. It is generally considered that in the
oxidation of the Zr-based materials, the oxide is growing mainly by inward oxygen transport
[1,4,5]. One possible factor related to the transport of molecules is the existence of pores
inside the oxide. It was found that the zirconium oxide layer is porous not only in the outer
part, but also within the barrier layer [8,11]. The distribution of ZrO2 phases in the oxide
layer is also changing in depth. Tetragonal and monoclinic ZrO2 phases have been reported
[3,6,8,12]. The fraction of tetragonal phase is higher at the oxide/metal, O/M, interface
(stabilized by high compressive stress) and is irregularly distributed in the oxide [3]. The
tetragonal to monoclinic phase transformation takes place as a result of stress relief with a
volume expansion of 7% and generates defects like cracks and pores (easy diffusion
pathways) [3]. This transformation can occur inside the barrier layer, possibly inducing
porosity within the barrier layer [8]. The porosity development in the zirconium oxide scale
is considered to be a main reason that leads to the transition from parabolic to linear oxide
growth kinetics [5]. A.P. Zhilyaev and J.A. Szpunar [12] proposed a model for oxygen
diffusion through the oxide scale considering stress development.
Enhanced oxidation of Zr-based materials in atomic oxygen enriched gas obtained by
microwave discharge has been reported in the literature [9,10]. This shows that increased
availability of dissociated oxygen at the oxide/gas, O/G, interface generates enhanced
oxidation. S. A. Raspopov and co-workers [9] suggested that oxygen chemisorption is likely
to be a rate-limiting step for the oxidation process at the early stages of oxidation. To
understand how the dissociation rate of oxygen influences Zr oxidation at the temperatures
relevant for the nuclear industry, we have to consider in which form oxygen is transported
through the oxide layer. For oxygen dissociation to take place an activation energy has to be
overcome. At 300-400°C, the thermal energy is low for oxygen dissociation on ZrO2 but Pt
and second-phase particles, SPP, may catalyze the dissociation. The presence of SPP
incorporated, partly un-oxidized, in the oxide layer may therefore influence the ratio
between molecular and dissociated oxygen transport across the oxide. One aim with this
paper is to study O2 dissociation rates on Zr-based materials and their possible influence on
2
the corrosion rate. The dissociation of O2 on Pt, pre-oxidized Zircaloy-2 and preoxidized
Fe/Ni/Cr containing intermetallics has been measured and will be discussed in this paper.
Also the influence of Pt, a very effective oxygen dissociating element, ODE, on the
oxidation of Zr is presented here.
Isotope monitoring techniques have been used to investigate the mechanism of oxygen and
hydrogen incorporation in the growing zirconia layer and substrate [11,13-15]. Many studies
related to the influence of hydrogen on Zr oxidation have been published lately
[3,6,8,11,16,17]. Recent publications dealing with other metals than Zr [16,18,19] have
concluded that the presence of hydrogen in the metal substrate or in the gas phase has an
effect on oxidation by increasing the metal outward transport, which eventually can generate
voids at the metal/oxide interface. One possible mechanism is based on a proton-induced
high concentration of metal ion vacancies in the oxide, which is likely to result in an
increased metal ion transport [19]. Oxidation of Zr by cation outward diffusion via cation
vacancies was already taken into consideration in early 50’s by E. A. Gulbransen and K.F.
Andrew [1]. The Zr-metal and oxygen mobilities during oxidation of Zircaloy-4 at 633K in
water and in dry O2 were investigated by A. Grandjean and Y. Serruys [20] using Xenon
marker atoms implanted into the pre-formed oxide and the Rutherford backscattering
spectrometry (RBS) technique. They interpreted their results that, before transition, the
oxide grows only by oxygen inward transport, both in water and in oxygen exposures.
However in their experiments performed in water, big losses of Xenon occurred, in contrast
to the exposure to dry oxygen when the Xe marker did not move during exposure [20]. G.
Hultquist et al. [16] indicated that the oxidation rate decreases if a better balance between Zr
ion and oxygen ion transport is obtained, resulting in adherent oxide growth, with low
density of pores and other defects.
In this paper, the effect of deuterium and the combined effect of Pt and deuterium (D) on the
oxidation of Zr at 400°C in 20 mbar O2 are investigated. This is done by oxidation of partly
Pt coated Zr samples with 600 wtppm D and without D, respectively. Finally, experiments
where Ar-filled Zr-based tubes with ZrSn liner are exposed to water from one end of the
tubings are presented. The results show that the deuterium concentration in the oxide and
substrate of two Zr-based tubes varies as well as the oxide thickness upon exposure in
H2O/D2O at 370°C.
3
2. Experimental
2.1 Gas Phase Analysis (GPA) technique
The schematic diagram of the GPA equipment presented in Figure 1 consists of:
•
A 70 cm3 reaction chamber made of a silica tube and a stainless steel cross. The
reaction chamber is pumped with an ion pump via a leak valve. In oxidation test of
the 1m long Zr-based tubes, the silica tube in Figure 1 is replaced by the Zr-based
tube.
•
A gas handling system where pressures up to 1 atm can be used.
•
A pressure gauge to measure the total pressure inside the reaction chamber.
•
A mass spectrometer (MS), with a quadruple analyser, placed in an UHV chamber.
Hydrogen and different isotopes can be detected.
•
A tube furnace which can be used up to 1200°C.
The GPA equipment can be used for different types of measurements: outgassing,
hydration, oxidation, dissociation and exchange. These different measurements are shortly
described bellow with some examples.
2.1.1 Outgassing
Heating of a sample in vacuum generates release of gases that can be quantified with a
relevant calibration. This enables, for example, the hydrogen content in a material to be
determined. In Figure 2, the outgassing of hydrogen from a Zr plate at temperatures up to
700°C is shown. The outgassing method has previously been described in detail [21].
2.1.2 Hydration and oxidation
The most common way to study oxidation of metals is measurement of the weight gain of
the sample. With GPA, the changes in the gas phase are measured. The pressure decrease of
the gas in the reaction chamber upon time is continuously monitored and is conveniently
recorded on an x-t plotter. A 1.8 cm2 Zr plate covered with a naturally formed oxide was
exposed to 7.5 mbar deuterium at temperatures 200-550°C. Calculated from the decrease of
D2-pressure in the reaction chamber, the deuterium uptake versus time in the Zr plate is
shown in Figure 3. After charging, the D content in the Zr plates was 600 wtppm. As seen in
4
the figure, the uptake rate of D in Zr is increasing considerably at temperatures above
400°C.
For study of the oxidation mechanism, the oxidation can be performed in two stages, first in
“normal” O2 followed by exposure in 18O-enriched O2. In a subsequent analysis with SIMS
the position of the oxide growth can be found if the exchange rate between O in the oxide
and O in O2 is slow compared with O-uptake rate in the oxidation. The previously Dcharged Zr plate, labelled Zr(D), was partly coated with Pt and exposed at 400°C for 12
hours to 20 mbar O2 in two stages. The oxidation kinetics presented as oxygen uptake from
the gas phase is shown in Figure 4. (SIMS profiles of the reaction products are presented in
the results section of the paper).
2.1.3 Dissociation and exchange measurements
The interaction between gas molecules and a surface can lead to dissociation (molecules
split into smaller constituents). The surface activity of a material for dissociation can be
quantified [22]. To measure the dissociation rate of a molecule, a mixture of its isotopes are
used, for ex.
16,16
O2+18,18O2. The dissociation process and a subsequent association of the
dissociated species result in the formation of the mixed molecules
pressures of the gas components,
16,16
O2,
16,18
O2 and
18,18
16,18
O2. The partial
O2, are obtained via an inlet to the
MS. The inlet is small enough not to influence the total oxygen pressure in the reaction
chamber. By considering the measured formation rate of the mixed molecules and the
distance from the statistical equilibrium (no more changes in the gas composition), the
dissociation rate can be calculated. The background dissociation rate of the silica tube must
of course also be taken into account. A detailed description of the measurements has been
presented elsewhere [22].
Exchange may take place between oxygen in the gas phase and oxygen in the oxide. This
exchange can be measured in exposure of a 16O-containing oxide to
rate of
16,18
18,18
O2. The formation
O2 in the gas phase will then be a result of exchange between 18O from the gas
phase and 16O from the oxide lattice.
2.2 Characterization of the exposed samples
In Table 1 samples and analysis methods used in this study are collected.
5
Secondary Ion Mass Spectrometry analyses
SIMS analyses were performed with a Cameca IMS-6f apparatus, 10 keV, 50-200 nA
primary beam of Cs+ ions, which was rastered over 200x200 µm2, where ions where
detected from a centered area with a diameter of 70 µm.
X-ray photoelectron spectroscopy, XPS
After oxidation, the outermost layers on the Zr samples were analyzed with X-ray
photoelectron spectroscopy, XPS, with Kratos AXIS HS. Detailed scans (pass energy of 80
eV) of C 1s, Zr 3d, O 1s, Pt 4f were obtained with a monochromatic AlKa(alfa) X-ray
source (1486.6 eV). The area of analysis was approximately 0.4 mm2. The binding energies
for O1s and Zr3d were corrected related to C1s peak (285 eV).
FEG-SEM analysis was performed with a Leo 1530 Field Emission Scanning Electron
Microscope equipped with a GEMINI field emission column.
3. Results and discussion
3.1 Oxygen dissociation on Zr-based materials and Pt
Westinghouse Electric Sweden AB has supplied Zircaloy-2 and Zr2Fe, Zr2Ni, ZrCr2intermetallics used in this work.
Before dissociation experiments, Zr2Fe, Zr2Ni, ZrCr2 intermetallics have been oxidized for
80 minutes in approximately 65 mbar 16,16O2 at 400°C resulting in 19, 0.8 and 0.4 µm oxide
thickness respectively. Zircaloy-2 was oxidized for 90 hours in approximately 200 mbar
16,16
O2 at 400-500°C resulting in 2.5 µm oxide thickness.
In Figure 5, the formation rate of the mixed oxygen molecules, 16,18O2, during the exposure
of a 2 cm2 preoxidized Zr2Fe intermetallic sample at 400°C and a 28 cm2 preoxidized
Zircaloy-2 sample at 500°C to approximately 50%
18
O-containing oxygen gas mixture are
presented. Though at lower temperature, and with a much smaller surface area, the
formation rate of the mixed molecules is higher on the Zr2Fe sample. This implies that the
Zr2Fe has a considerably higher oxygen dissociation efficiency than Zircaloy-2.
Oxygen dissociation rate on Pt as well as on preoxidized Zr2Fe, Zr2Ni, ZrCr2 and Zircaloy-2
in exposure to 20 mbar O2 has been measured and the results are summarized in Figure 6.
At temperatures around 400°C, oxygen dissociation efficiency is decreasing in the order Pt
6
> Zr2Fe > Zr2Ni > ZrCr2 ≥ Zircaloy-2. (The accuracy in the measurement of ZrCr2 sample
only allowed us to determine the dissociation rate of O2 on ZrCr2 to be lower than 2.10-12
mol O cm-2 s-1).
Fe, Cr and Ni have a low solubility in the Zr matrix and are therefore mainly incorporated in
the form of SPP [8,23] and also partly incorporated in the oxide during oxidation. The size
distribution of SPPs in the Zr-based materials has an important effect on the in-reactor
performance of these materials [23]. The formation rate of protective oxides and also the
thickness of the barrier layer are found to be dependent on the initial SPP size distribution
[23]. Dissolution of small-size SPPs in Zr-based materials during in-reactor operation under
high neutron flux in combination with hydrogen pick-up can enhance the deterioration of
the barrier layer [23].
The chemical composition of SPPs is also important [23]. In Zircaloy-2, for example, two
types of SPPs are reported: Zr2(Fe,Ni) and Zr(Fe,Cr)2 particles [3, 8, 23, 24]. The Crcontaining particles were found to be less resistant to dissolution in the matrix during
irradiation than the Ni-containing particles [3, 23]. Each alloying element has certain effects
on the oxidation and it seems plausible that one effect is related to efficiency for oxygen
dissociation. Thus the results from Figure 6 suggest that among Zr2Fe, Zr2Ni and ZrCr2, the
Fe-containing particles are the most effective in dissociation of O2 at 400°C. It is reported
that the addition of up to 150 ppm Fe to pure Zr is beneficial for the corrosion resistance
[24]. Our suggested explanation for this observation is linked to a synergistic effect of
oxygen dissociating elements, such as Fe, and hydrogen and is discussed in the section
below.
3.2 Influence of Pt and deuterium on the oxidation of Zr in O2 at 400°C
Two 1.8 cm2 Zr plates were outgassed at temperatures up to 700°C (Figure 2). The hydrogen
content in the samples after outgassing was less than 5 wtppm. One sample was charged
with deuterium to 600 wtppm D (see Figure 3). Both samples were then partly sputtercoated with 200Å porous Pt.
The samples were placed in the GPA equipment and the reaction chamber was evacuated at
temperatures up to 400°C for 12 hours. Oxidation was carried out in two stages at 400°C.
First in
16,16
18
O2 then in O-enriched oxygen gas (85% 18O for Zr(D) sample and 52% 18O for
Zr sample) for totally 12 hours exposure to approximately 20 mbar gas. The experimental
conditions are summarized in Table 2 and the samples are shown in Figure 7. The oxygen
7
uptake from the gas phase was used to calculate the average oxide thickness in the
oxidation. In 12 hours of oxidation a 0.6 µm oxide was formed on the D-containing sample
(see Figure 4 for oxygen uptake rate) whereas a 0.8 µm oxide was formed on the D-free
sample.
After oxidation, the samples were characterized with SIMS, XPS, SEM and optical
microscopy. In Figure 8, SEM images of the area with Pt particles before and after oxidation
of the Zr sample are shown.
3.2.1 SIMS results – Oxygen transport in oxide
SIMS depth profiles (SIMS analysis points seen in Figure 7) and XPS were used to
characterize the oxide on areas with Pt particles and at different distances from the Pt area.
In Figure 9, the SIMS integrated oxygen count rates from the second stage of oxidation are
presented for the two Zr samples. Obviously in the Pt-area a considerably higher oxidation
rate has taken place. A possible reason is that in the Pt area, oxygen dissociates on the Pt
particles with high efficiency (see Figure 6). The On- (n = 0, 2) species resulted from
dissociation can then:
•
associate and form molecular O2
•
diffuse laterally via surface diffusion
•
diffuse through the oxide layer to the O/M interface
In the Pt area, the lateral diffusion length corresponds to the distance between the Pt
particles, which is 10 to 20 nm as can be seen in Figure 8. In the Pt area, the Onconcentration is high at the O/G interface, with a high concentration gradient of On- over the
oxide thickness and therefore facilitating a high inward oxygen flux. At the interface
between the area with Pt-particles and area without Pt-particles, the oxygen dissociation rate
decreases about 105 times as can be seen in Figure 6. Especially for the D-containing
sample, a mm-ranged effect of Pt is observed in Figure 9.
In the Pt area, high oxidation rates by inward oxygen transport generate thick oxides with
defects, which also may enhance molecular transport. The thinnest oxide is found near the
Pt area on the Zr(D) sample in Figure 9. A local minimum in the oxygen uptake can also be
found near the Pt area on the Zr sample. These observations suggest that in the vicinity of Pt
area, which has a lower activity of On- compared with the Pt area, an outward Zr diffusion
induced by D from the substrate balances the On- inward flux. The balanced growth
facilitates a self-repairing mechanism to be operative and can explain the formation of a
more dense oxide. At all positions the oxide on the Zr(D) sample is thinner than at
8
corresponding positions on the Zr sample. This can be seen as a combined effect of Pt and D
on Zr oxidation. Exchange between 18O from the gas phase and 16O from the oxide formed
in the first stage is almost negligible at 400°C.
To compare the oxidation mechanism inside the Pt area and far away from the Pt area,
SIMS profiles of Zr18O and Zr16O +Zr18O are considered and presented in Figure 10a and
10b where also Pt count rates are shown.
The oxidation mechanism can be retrieved by considering the Zr18O SIMS profiles
(oxidation in
18
O containing gas mixture was carried out in the second stage). High Zr18O
count rates indicate where oxide growth mainly takes place. As clearly seen in Figure 10, in
the Pt area on both samples an enhanced oxidation takes place mainly by inward oxygen
transport. Pt is enriched at the O/G interface.
Relatively high Zr18O count rates are found for the Zr(D) sample far away from the Pt area
at the O/G interface, which could be related to a D-induced outward Zr diffusion.
3.2.2 SIMS results – Hydrogen transport
The SIMS profiles of deuterium for the Zr(D) sample in different positions from the Pt area,
after 12 hours oxidation are shown in Figure 11. In this figure, arrows indicate the position
of the O/M interface based on sputter time when the 90Zr ion count rate has decreased to half
of its maximum value.
The area close to the O/G interface is magnified in Figure 11 to show the difference
between the D maxima at different distance from Pt area. For all positions except in the Pt
area a local maximum appears near the O/G interface. This can result from a suppressed D
outward diffusion in the Pt area. A local maximum in the D count rate appears also near the
O/M interface at all analyzed positions. D outward transport is slightly increasing as the
distance from the Pt area increases. The highest concentration of D in the oxide is found at
the longest distance away from the Pt area, +6662 µm.
It is reported in the literature [25] that hydrogen atoms dissolved in the Zr substrate prefer
tetrahedral interstitial sites and they diffuse interstitially. The radius of these tetrahedral
interstitial sites is approximately 0.036 nm [25]. The radius of hydrogen atom is about 0.04
nm [25], which means that the strain induced by hydrogen dissolution in Zr is negligible
[25]. C. –S. Zhang et al. [25] have found that hydrogen tends to segregate to the Zr surface.
When dissolved oxygen is present in the Zr substrate, hydrogen segregation to the Zr
surface is enhanced [25]. The concentration of dissolved oxygen in Zr varies from the bulk
9
to the surface and is much higher underneath the oxide, accommodating also more hydrogen
in this area.
In our Zr samples, prior to oxidation the hydrogen content was < 5 wtppm H and (<1 wtppm
H + 600 wtppm D) respectively. SIMS profiles of H, D and (H+D) are presented in Figure
12 for the two Zr samples at +3543 µm, Zr, and +5871 µm, Zr(D), away from respective Pt
area. From Figures 11 and 12 it is observed that:
•
A high H concentration at the G/O interface, which is likely due to “contamination”
and/or 1H SIMS background.
•
D contamination is negligible at the G/O interface, which indicate that accurate
information requires the use of D instead of H.
•
The 1H SIMS depth profile on the sample with <5 wtppm H shows a maximum
inside the metal at the O/M interface but no corresponding maximum is found on the
sample with <1 wtppm H + 600 wtppm D.
•
The D profile shows a local maximum inside the oxide, near the O/M interface at all
analysed positions. As going away from the Pt area, a small but increasing
accumulation of D appears at the G/O interface.
A. Stern et al. [26] have reported that a maximum in hydrogen concentration is obtained in
the metal region immediately under the oxide after oxidation of pre-hydrated Zr-based
materials. They suggested that an interaction between the interstitially dissolved H
(occupying the tetrahedral sites) and O (occupying the octahedral sites) under the oxide
layer is responsible for these sub-oxide peaks in the hydrogen profile. However the peak in
D in our experiments is located in the oxide near the O/M interface and this fact is further
considered below.
a. Hydrogen concentration in the Zr substrate is low and is close to the dissolved
oxygen concentration in the metal region immediately under the oxide (CH ≈ CO). H
from the substrate is diffusing from the bulk to the surface during oxidation and
short range bonds between dissolved O and H can take place at the O/M interface.
This results in a maximum in the H profile in the metal region immediately under the
oxide.
b. Hydrogen concentration in the Zr substrate is high and greatly exceeds the dissolved
oxygen concentration in the metal region immediately under the oxide (CH >> CO).
A small fraction of the outward diffusing H from the bulk will be trapped in shortrange bonds with the dissolved O, but a main part will diffuse into the oxide where it
can be accommodated in tetrahedral sites [27]. This can explain the maximum that
10
appears in the oxide near the O/M interface on the sample with 600 wtppm D in the
substrate.
It is reported that in zirconia, protons are not centred in a normal interstitial position but are
closely bonded to oxygen ions [27]. The “substitutional hydroxide” term is frequently used
to refer to this defect [27]. The presence of hydrogen in zirconia influences the bonds of the
neighbouring atoms, Zr and O. We suggest that in the oxidation of Zr-based materials in
water containing atmospheres, hydrogen accumulation in the oxide near the O/M interface
could explain the frequently observed deterioration of the “barrier” layer. This also indicates
that Zr could move easier in the deteriorated barrier when hydrogen is present. Partially
“degraded” oxide within the barrier layer due to hydrogen was also considered by M. B.
Elmoselhi [28], in the exposure of a pre-oxidized Zircaloy-2 to a low pressure of deuterium
(D) at 380°C for 15 days.
In our case, for the Zr(D) sample in the Pt area, the enhanced oxygen inward transport can
also generate more dissolved oxygen, which can trap more deuterium. The outward
transport of deuterium (D) to the G/O interface is also lowered due to the longer diffusion
length (thicker oxide) obtained in the Pt area.
Deuterides are expected to be present in the Zr(D) sample, the D content being above the
solubility limit. It is reported [7] that the thermal solid solubility (TSS) of H in Zr at 400°C
is approximately 200 wtppm H.
Analysis of the
18
OD and D SIMS profiles away from Pt area, shown in Figure 13, are
considered to elucidate the connection between deuterium and oxygen during Zr oxidation.
The sample without D is also shown in Figure 13 for comparison of 18OD and D count rates.
The O/M interface for Zr(D) sample is indicated in the figure. Close to this interface, a
distinct maximum in 18OD can be seen as well as a local maximum in D which may confirm
formation of “substitutional hydroxides” in the oxide close to the O/M interface.
3.2.3 XPS results
XPS was used to characterize the outermost oxide layer at different positions from the Pt
area.
In Figure 14, the O1s and Zr 3d5/2 binding energies are shown for different distance from the
Pt area. The binding energies were corrected related to C1s peak (285 eV). Comparing the
area with and without Pt particles, a chemical shift in the O1s binding energy of
approximately 1.5 eV is obtained for both Zr samples. For both samples, the O1s binding
11
energy inside Pt area has the highest value of 531.6 eV. The deuterium is not influencing the
O1s binding energy in the Pt area significantly, as indicated previously since D outward
diffusion at the G/O interface is suppressed in this area. Away from the Pt area, the change
in O1s binding energy is more pronounced for the Zr(D) sample. The Zr 3d5/2 binding
energy in Figure 14b on the Zr(D) sample is rather different from the Zr sample.
The formation of Zr suboxides was investigated by Y. Nishino et al. [29] in the initial
oxidation stages of Zr and Zircaloy-2 with oxygen and water vapor at room temperature
using Auger and X-ray photoelectron spectroscopy. The binding energy values for Zr 3d5/2
used in their study [29] were for Zr in ZrO2, 182.9 eV and for Zr in Zr2O3, 181.8 eV. Our
values are ranging from 182.1 eV close to Pt area to 182.6 eV inside Pt area for the Zr
sample. It seems plausible that at the O/G interface, the Zr oxidation state varies, being high
in the Pt area (probably ZrO2 or ZrO2+x), and some suboxides are probably present away
from the Pt area. The presence of D in the Zr(D) sample generates higher binding energies
for O1s and Zr 3d5/2, especially outside Pt area, possibly due to the presence of D at the G/O
interface.
3.3 Oxidation of Zr-based tubes in water at 370°C
Two Zr-based tubes with ZrSn liner were used in experiments where the tubes were Arfilled and, subsequently, exposed to water from one end of the tubings. The experimental
set-up is schematically shown in Figure 15. The tubes were evacuated for 24 hours at 370°C
before exposure and then filled at 370°C with approximately 700 mbar Ar and exposed from
one end to an inlet of approximately 22 mbar H2O for 305 minutes (first tube) and 22 mbar
D2O for 630 minutes (second tube).
After exposure, the tubes were visually inspected. The oxide thickness seemed to have a
local minimum at a certain distance from the water inlet. This was observed in both tubes
but at slightly different distances from the water inlet. SIMS analyses at different positions
from the water inlet were used to characterize the oxide thickness and the hydrogen content
after oxidation of the tubes.
Integrated SIMS count rates of oxygen and deuterium along the length of the tubes are
shown in Figure 16. Thickest oxide was obtained at 330-400 mm from the water inlet. Local
minima in the oxide thickness are found at 430 mm and 478 mm respectively from the water
inlet. At these positions there is some hydrogen in the substrate as a result of the reaction
with water. In this case, both oxygen and hydrogen diffuse inward towards the O/M
12
interface. At 996 mm from the water inlet position no significant oxidation takes place but
the deuterium content is slightly higher than at the water inlet position (20°C). Deuterium,
resulted from the reaction of D2O with Zr at 370°C, diffuses faster than D2O through the Ar
filled tubes. The deuterium concentration underneath the oxide in the second tube is
presented in Figure 17. In the area where the oxide is thickest, around 330 mm from the
water inlet, the concentration of deuterium underneath the oxide is the highest. There is a
tendency for local minima in the D content at positions close to 478 mm from the D2O inlet.
The experiment shows that the hydrogen content in the substrate influences the formation of
the oxide layer. After starting the oxidation but before water reaches to 430 mm (tube 1) and
respectively 478 mm (tube 2) from the water inlet position, the D content in the substrate in
these positions increases. At 370°C, the oxide formed at these positions is denser, being
characterized by a slow oxygen and deuterium diffusion. We suggest that a certain amount
of hydrogen seems to be beneficial for the formation of dense Zr-oxide. This may be
understood by a more balanced O/Zr transport, where the Zr-transport is induced by
hydrogen.
3.4 Interpretation of the experimental results
In oxidation of Zr-based materials, the oxide growth takes place mainly by inward oxygen
transport, which creates a network of micropores, or rather nano-pores. These pores can
easily act as diffusion paths for molecular species. Hydrogen present in the substrate or in
the corrosive environment diffuses in the oxide layer and tends to generate deterioration of
the barrier layer. However, hydrogen also generates outward diffusing Zr, which can react
with oxygen or water molecules within the oxide and/or at the G/O interface. Obviously
healing of pores by oxide formation needs outward Zr transport.
In the case of inward hydrogen transport (Figure 17), the area where the oxide thickness
shows a local minimum, the D content underneath the oxide is relatively low. By blocking
the molecular diffusion pathways, relatively more ionic transport pathways are used for the
transport of species through the oxide layer, which is a slower transport near 400°C,
resulting in growth of more protective and thinner oxides. Related to the “self repairing”
concept, a small but significant fraction of Zr outward diffusion was identified in the 600
wtppm D containing Zr sample in oxidation in dry O2 at 400°C.
13
4. Summary
The transport of oxygen and hydrogen through the growing zirconia layer in the oxidation
of Zr-based materials in O2 at 400°C and in water at 370°C is investigated using GPA and
SIMS.
Oxygen dissociation on Zr-based materials and Pt
•
The dissociation rate of the O2 on preoxidized-Zr2Fe is 103 times higher than on
preoxidized-Zircaloy-2 at 400°C.
•
At temperatures around 400°C, oxygen dissociation efficiency decreases in the
following order: Pt > Zr2Fe > Zr2Ni > ZrCr2 ≥ Zircaloy-2.
Influence of Pt and deuterium on the oxidation of Zr in O2 at 400°C
The influence of Pt on the oxidation of Zr, with and respectively without D in the substrate,
at 400°C in dry O2 was investigated. After oxidation, the samples were characterized with
SIMS, XPS, SEM and optical microscopy.
SIMS results
•
An enhanced oxidation rate is observed in the Pt area for both samples. The
deuterium outward diffusion is suppressed in the Pt area.
•
In the vicinity of the Pt area a local minimum in oxide thickness is obtained for both
samples and thinnest oxide is obtained on the Zr(D) sample.
•
At areas away from the Pt area, the D SIMS depth-profile shows a maximum at the
G/O interface, which is more pronounced as the distance from the Pt area increases.
The formation of “substitutional hydroxides” at the O/M interface is proposed as a
mechanism for oxide deterioration. Hydrogen may also induce Zr diffusion in the oxide,
resulting in new oxide formation with fewer defects. If the hydrogen content is high the
dissolution rate may become higher than the healing capacity by oxide growth within the
oxide and lead to deterioration of the oxide.
XPS results
The XPS results reveal that Zr has different oxidation states in the oxides formed in the Pt
area and outside Pt area. A high oxidation state is obtained in the Pt area.
14
Oxidation of Zr-based tubes in water at 370°C
The exposure of two Ar-filled 1 m long Zircaloy-2 tubes with ZrSn liner to 22 mbar water at
370°C for 305 and respectively 630 minutes was investigated. The oxide thickness and the
hydrogen content were measured at different positions from the water inlet using SIMS
depth profiles.
Local minima in oxide thickness are obtained at 430 mm and respectively 478 mm from the
water inlet. Hydrogen gas produced in the oxidation reaction is diffusing “ahead of the
steam front” and is taken up by the tube material. The two positions corresponding to local
minima in oxide thickness have a certain amount of hydrogen present in the substrate before
being oxidized by water. The D content underneath the oxide shows also that D uptake is
small where the oxidation rate is low.
Main result
For a content of 600 wtppm D in the Zr substrate, a beneficial effect of Pt (high efficiency in
oxygen dissociation) on the Zr oxidation in O2 at 400°C is found. Considering the O2
dissociation efficiency at 400°C on Pt and Zr2Fe, we believe that an analogue beneficial
effect of Fe-containing SPPs is operative when the chemical composition, size distribution
and volume fraction of SPPs harmonize with the hydrogen concentration. We suggest that
H-induced Zr outward diffusion is beneficial for healing defects like pores and microcracks
formed in the oxide during oxidation. The formation rates of these defects, generated by an
exclusive inward oxygen transport, can be counter battled by healing generated by outward
Zr diffusion.
Acknowledgments
Financial support from Westinghouse Electric Sweden AB is gratefully acknowledged.
15
References:
[1]
E.A. Gulbransen, K.F. Andrew, J. of Metals, April 1957, p. 394-400.
[2]
B. Cox, Y.-M. Wong, J. Nucl. Mater. 218(1995), p. 324-334.
[3]
P. Rudling, G. Wikmark, J. Nucl. Mater. 265(1999), p. 44-59.
[4]
S. Abolhasssani, M. Dadras, M. Leboeuf, D. Gavillet, J. Nucl. Mater. 321(2003), p.
70-77.
[5]
A. Yilmazbayhan, A.T. Motta, R.J. Comstock, G.P. Sabol, B. Lai, Z. Cai, J. Nucl.
Mater. 324(2004), p. 6-22.
[6]
B. H. Lim, H.S. Hong, K.S. Lee, J. Nucl. Mater. 312(2003), p. 134-140.
[7]
M. Oskarsson, E. Ahlberg, U. Södervall, U. Andersson, K. Pettersson, J. Nucl.
Mater. 298(2001), p. 315-328.
[8]
C. Lemaignan, Zirconium in the Nuclear Industry: Thirteenth International
Symposium, ASTM STP 1423, Annecy, France, 2001, p. 20-29.
[9]
S. A. Raspopov, A.G. Gusakov, A.G. Voropayev, A.A. Vecher, M.L. Zheludkevich,
O.A. Lukyanchenko, A.S. Gritsovets, V. K. Grishin, J. Chem. Soc. Faraday Trans.
93(1997), p. 2113-2116.
[10]
X. Iltis, M. Viennot, D. David, D. Hertz, H. Michel, J. Nucl. Mater. 209(1994), p.
180-190.
[11]
N. Ramasubramanian P. Billot, S. Yagnik, Zirconium in the Nuclear Industry:
Thirteenth International Symposium, ASTM STP 1423, Annecy, France, 2001, p.
222-243.
[12]
A.P. Zhilyaev, J.A. Szpunar, J Nucl. Mater. 264(1999), p. 327-332.
16
[13]
P. Berger, R. El Tahhann, G. Moulin, M. Viennot, Nucl. Instrum. Methods Physics
Research Section B 210(2003), p. 519-525.
[14]
I.S. Woolsey, J.R. Morris, Corrosion 37(1981), p. 575-585.
[15]
I. Takagi, S. Shimada, D. Kawasaki, K. Higashi, J. Nucl. Sci. Tech. 39(2002), p.7175.
[16]
G. Hultquist, B. Tveten, E. Hörnlund, M. Limbäck, R. Haugsrud, Oxidation of
Metals 56(2001), p.313-346.
[17]
B. Cox, Y. -M. Wong, J. Nucl. Mater. 270(1999), p. 134-146.
[18]
G. Hultquist, B. Tveten, E. Hörnlund, Oxidation of Metals 54(2000), p.1-10.
[19]
B. Tveten, G. Hultquist, T. Norby, Oxidation of Metals 52(1999), p. 221-233.
[20]
A. Grandjean, Y. Serruys, J. Nucl. Mater. 273(1999), p. 111-115.
[21]
T. Åkermark, G. Hultquist, Q. Lu, J. Mater. Eng. Perf. 5(1996), p.516-520.
[22]
E. Hörnlund, Appl. Surf. Sci. 199(2002), p.195-210.
[23]
P. Tängstrom, M. Limbäck, M. Dahlbäck, T. Andersson, H. Pettersson, Zirconium in
the Nuclear Industry: Thirteenth International Symposium, ASTM STP 1423,
Annecy, France, 2001, p. 96-118.
[24]
P. Barberis, E. Ahlberg, N. Simic, D. Charquet, C. Lemaignan, G. Wikmark, M.
Dahlbäck, P. Tägtström, B. Lehtinen, Zirconium in the Nuclear Industry: Thirteenth
International Symposium, ASTM STP 1423, Annecy, France, 2001, p. 33-58.
[25]
C. –S Zhang, B. Li, P.R. Norton, J. Alloys and Comp. 231(1995), p. 354-363.
17
[26]
A. Stern, D. Khatamian, T. Laursen, G.C. Weatherly, J.M. Perz, J. Nucl. Mater.
148(1987), p. 257-265.
[27]
T. Norby, Protons in ZrO2; A search for effects of water vapour on the electrical
conductivity and M/T phase transformation of undoped ZrO2, in: S. Meriani and C.
Palmonari (Eds.), Zirconia’88 – Advances in Zirconia Science and Technology,
Elsevier Applied Science London and New York, 1988, p. 209-218.
[28]
M. B. Elmoselhi, J. Alloys and Comp. 231(1995), p. 716-721.
[29]
Y. Nishino, A.R. Krauss, Y. Lin, D.M. Gruen, J. Nucl. Mater. 228(1996), p. 346353.
18
Figure captions
Figure 1 – GPA equipment for outgassing, hydration, oxidation, dissociation and exchange
experiments.
Figure 2 – Outgassing of hydrogen from a Zr-plate at temperatures up to 700°C.
Figure 3 – Kinetics of deuterium uptake in a Zr plate.
Figure 4 – Oxidation of a partly Pt coated Zr sample containing 600 wtppm D, Zr(D), in 20
mbar O2 at 400°C.
Figure 5 – Oxygen dissociation measurements on Zr2Fe and Zircaloy-2.
Figure 6 – Oxygen dissociation rates on Pt as well as pre-oxidized Zr2Fe, Zr2Ni, ZrCr2 and
Zircaloy-2 in exposure to 20 mbar O2 versus reciprocal temperature. The value of oxygen
dissociation at 400°C on preoxidized Zircaloy-2 is extrapolated.
Figure 7 – Partly Pt coated Zr samples with visible craters from SIMS sputtering
a. Zr(D) sample and b. Zr sample
Figure 8 – SEM images of the area with Pt particles on Zr substrate.
a. before oxidation and b. after 12 hours oxidation in O2 at 400°C
Figure 9 - SIMS integrated oxygen count rates from second stage of oxidation on different
positions on the two samples after two-stage oxidation at 400°C.
Figure 10 – SIMS depth profiles inside Pt area (-) and far away from Pt area (+) for
a. Zr(D) sample and b. Zr sample
Figure 11 - SIMS depth profiles of D at different positions on the Zr(D) sample after
oxidation. The oxide/metal interface is indicated by arrows.
Figure 12 – H and D SIMS depth profiles far away from the Pt area for the two Zr samples
after 12 hours oxidation in 20 mbar O2 at 400°C. The arrows indicate the oxide/metal
interface (grey for Zr(D) sample and black for Zr sample). Profiles of H and (H+D) for Zr
sample virtually coincide.
19
Figure 13 – SIMS 18OD and D count rates for the Zr samples after oxidation in 20 mbar O2
at 400ºC far away from the Pt area (+6662 µm from the Pt area for Zr(D) sample and
respectively +6534 µm from the Pt area for Zr sample). The arrow indicates the position of
the oxide/metal interface.
Figure 14 - XPS analysis: a. O 1s binding energy and b. Zr 3d 5/2 binding energy, at
different positions on the two samples. The C1s was set to 285 eV, which was used as an
internal reference.
Figure 15 - Experimental set-up for the exposure of Ar-filled Zircaloy-2 tubes with ZrSn
liner to 22 mbar H2O (D2O) at 370°C.
Figure 16 – SIMS analysis of the Zr-based tubes: integrated oxygen and deuterium count
rates at different positions from the water inlet.
Figure 17 – SIMS analysis of the Zr-based tubes: deuterium count rate in the metal
underneath the oxide vs. position in tube. In the range 260 → 1000 mm the tube temperature
is approximately 370°C.
20
Tables
Table 1– Samples, analysis methods with respective figure numbers
Pt
coating
Temperature
(°C)
Analysis methods
Sample
Hydrogen
content
Zr2Fe
< 5 wtppm H
no
400
x
Zr2Ni
< 5 wtppm H
no
400
x
ZrCr2
Zry-2
plate
Pt plate
< 5 wtppm H
no
400
x
< 5 wtppm H
no
500-700
x
-
-
400-700
x
GPA
SEM
SIMS
XPS
Figure
Oxygen dissociation measurements:
5, 6
Pt and D influence on Zr oxidation:
Zr(D)
plate
<1wtppmH
+600 wtppm D
yes
400
x
Zr
plate
< 5 wtppm
hydrogen
yes
400
x
x
x
x
x
x
3, 4, 7, 9,
10, 11, 12,
13, 14
7, 8, 9, 10,
12, 13,14
Oxidation of Zr-based tubes in water:
Zr-based
tube 1
(inside)
Zr-based
tube 2
(inside)
< 1 wtppm H
no
370
x
x
15, 16
< 1 wtppm H
no
370
x
x
15, 16, 17
Table 2 – Experimental conditions for the oxidation of Zr(D) and Zr samples
Sample
Zr(D)
Zr
Gas phase composition in oxidation, % Average oxide
(oxidation time)
thickness*,
First stage
Second stage
µm
16
18
16
85% O + 15% O
99.9% O
0.6
<1wtppm H +600 wtppm D
(130 min.)
(590 min.)
99.9% 16O
52% 18O + 48% 16O
0.8
< 5 wtppm hydrogen
(130 min.)
(590 min.)
-3
* for conversion an average density of 6 g cm was considered for the oxide
Hydrogen content
Printer
Leak
valve
Pressure
gauge
Tube furnace
UHV
QMS
Quartz tube
400°C
Ion pump
Sample
Rough
pump
P in MS, mbar
Pressure
gauge
Gas
handling
system
Time, s
Computer
Figure 1
Hydrogen pressure in MS, mbar
10-6
10-7
10-8
400°C
600°C
700°C
200°C
10-9
0
100
200
Time, h
Figure 2
300
25
20
550°C
15
400
10
500°C
5
200
D uptake, wt ppm
D uptake, µmol D cm
-2
600
400°C
200°C
0
0
0
20
40
60
Time, h
Figure 3
1.5
4
1
3
2
0.5
18
Adition of O
1
0
0
200
400
Time, min.
Figure 4
600
0
800
Oxygen uptake, µmolO cm
Oxygen uptake, mbar
5
-2
6
2
3.6
2
Zr Fe (2 cm ) +
2
quartz background
400°C
16,18
O2, %
3.2
2.8
2.4
2
Zircaloy-2 (28 cm )+
quartz background
500°C
2
0
20
40
60
80
100
Time, min.
Figure 5
T,°C
600
500
400
350
Dissociation rate, mol O cm s
-2 -1
10-6
700
Pt
10-8
Zr Fe
2
Zr Ni
10-10
2
10-12
Zry-2 oxide
ZrCr
2
10-14
0.0009
0.00105
0.0012
0.00135
1/T,1/K
Figure 6
0.0015
0.00165
Porous
Pt
area
-
3 mm
0
Porous
Pt
area
-
+
0
+
b. Zr sample
a. Zr(D) sample
Figure 7
Figure 8 a.
Figure 8 b.
8 1011
Integrated O-count rates from 2
nd
stage of oxidation
1 1012
6 1011
Effect of Pt on
the oxidation of Zr
Zr
4 1011
Effect of Pt on
the oxidation of
Zr(D)
2 1011
O-spill over
~ 10 nm
Zr(D)
O-spill over
~ mm
0
Area with
Pt particles
0
2000
4000
6000
Distance from Pt-area, µm
Figure 9
8000
18
Zr O, -1924 µm
18
Zr O, +2298 µm
Count rates, a.u.
16
18
16
18
Zr O+Zr O, -1924 µm
195
Pt, -1924 µm
Zr O+Zr O, +2298 µm
195
Pt, +2298 µm
1
Zr(D)
Pt-induced
oxide growth
0.5
Pt-induced
oxide growth
at the substrate
interface
0
0
500
1000
Pt enrichment
1500
2000
2500
Sputter time, s
Figure 10 a.
18
Zr O+Zr O, -2043 µm
18
Zr O+Zr O, +3543 µm
Zr O, - 2043 µm
Count rates, a.u.
Zr O, +3543 µm
1
16
18
195
16
18
195
Zr
Pt-induced
oxide growth
Pt, -2043 µm
Pt, +3543 µm
Pt-induced
oxide growth
at the substrate
interface
0.5
0
0
500
Pt enrichment
1000
1500
Sputter time, s
Figure 10 b.
2000
2500
+230 µm, close to Pt area
+1244 µm, close to Pt area
-1924 µm, inside Pt
D count rate, counts s -1
D count rate in SIMS, counts s
-1
1500
1000
+5871 µm, far away from Pt
+6662 µm, far away from Pt
300
200
100
0
0
125
250
375
500
Sputter time, s
500
0
0
500
1000
Sputter time, s
Figure 11
1500
2000
H, +5871µm, Zr(D)
D, +5871µm, Zr(D)
H+D, +5871µm, Zr(D)
H, +3543µm, Zr
D, +3543µm, Zr
H+D, +3543µm, Zr
5 10 3
Count rate, counts s
-1
4 10 3
3 10 3
2 10 3
1 10 3
0 10 0
0
500
1000
1500
2000
2500
Sputter time, s
Figure 12
4
2 10
D - Zr(D)
18
OD - Zr(D)
3
5.2 10
4
18
3
3.5 10
4
1 10
3
3
5 10
1.8 10
0
0
18
0
OD - Zr
500
D - Zr
1000
1500
Sputter time, s
Figure 13
2000
2500
D counts/s
OD counts/s
1.5 10
O1s binding energy, eV
531.5
531
Zr(D)
530.5
Zr
530
-4 Area -2
with
0
Pt particles
2
4
6
8
Distance from Pt area, mm
Figure 14 a.
Zr 3d5/2 binding energy, eV
182.7
182.5
Zr(D)
182.3
Zr
182.2
182
-4
Area-2with
Pt particles
0
2
4
6
Distance from Pt area, mm
Figure 14 b.
8
Zircaloy 2 tube
Ar
No water inlet
0 mm
1000 mm
Ar
H2O
H2O
0 mm
Time = 0 min.
Time = 305 min.
Water diffusionfront after 305 minutes
D2O
D2O
1000 mm
Ar
Time = 630 min.
Water diffusionfront after 630 minutes
Figure 15
200
O - counts
630 min.
150
100
50
O - counts
305 min.
O - counts
20°C
100
Oxygen counts, a.u.
D-counts
630 min.
250
Counts in SIMS, a.u.
Oxygen counts, a.u.
300
305 min.
Ar+H2O
50
0
415
420
425
d, mm
430
435
150
630 min.
Ar+D2O
100
50
0
400
450
500
d, mm
550
600
0
0 D - counts
20°C
200
400
600
Distance from water inlet (d), mm
Figure 16
800
1000
D normalized counts/s
0.1
D normalized counts/s
0.08
0.06
0.08
370°C
630 min. Ar+D2 O
0.04
0
460
480
500
520
d, mm
0.04
20°C,
630 min.
Ar+D2 O
0.02
370°C
630 min. Ar+D2 O
0
0
100
200
300
400
500
600
700
Distance from water inlet (d), mm
Figure 17
800
900
1000
Paper III
A method for characterization of gas transport in porous oxides
C. Anghel*, Q. Dong, G. Hultquist
Division of Corrosion Science, Department of Materials Science and Engineering, Royal Institute of
Technology, Drottning Kristinas väg 51, S-100 44 Stockholm, Sweden
J. Rundgren
Theory of materials, Department of Physics, Royal Institute of Technology, Drottning SCFAB Building 11
S-106 91 Stockholm, Sweden
I. Saeki
Muroran Institute of Technology, Department of Materials Science and Engineering, 27-1 Mizumotocyo
050-8585, Muroran, Japan
M. Limbäck
Westinghouse Electric Sweden AB, S-721 63, Västerås, Sweden
Abstract
A theoretical model of gas release from samples with different geometries: plate, cylinder
and sphere is presented and forms the basis for evaluation of diffusion and concentration of
gases in oxides. The model is also used for evaluation of nm ranged oxide porosity. When
the model is applied for experimentally obtained outgassing data, a novel and relatively
straightforward non-destructive method for evaluation of diffusivity, concentration,
permeability and effective pore size is obtained. The sample to be characterised can simply
be air equilibrated. The gas release during outgassing is quantitatively evaluated using a
mass spectrometer having a quadruple analyser, placed in ultra high vacuum. The method is
validated in measurements of diffusivity and solubility of He in quartz at 80°C and applied
for characterization of two Zr- and an Fe-oxide.
Keywords: diffusion, oxide film, porosity, flux, Gas Phase Analysis
Corresponding author: Anghel Clara email: [email protected], Tel. +46 8 7906670, Fax. +46 8 208284
1. Introduction
A better understanding of the transport properties of gases in oxides is certainly very
important in many applications1. For example, in solid oxide fuel cells (SOFC) the “fully”
dense oxygen conducting zirconia-based electrolyte requires operation at high temperatures
(900-1000°C) for high ionic conductivity2. The reduction of the operation temperature to
moderate temperatures would generate higher efficiency in energy conversion, prolonged
lifetime and production of cheaper energy2. In the case of metals, a general protection
measure against corrosion implies formation of a dense surface metal oxide layer3, so called
“barrier layer”4 which should be gas-tight. This situation is often assumed but rarely
confirmed. The characterization of the micro- and/or meso- pores formed during the early
oxidation stage of metals is crucial. Attempts to visualize these ultrafine pores have been
made by using electron microscopies like Scanning and Transmission Electron
Microscopies (STM and TEM)5 but there are limitations and possible artefacts are
introduced in cross sectioning of samples and by the electron beam.
Many methods can be used for characterization of oxide porosity, like densitometry based
on volume displacement by Archimedes’s principle6, mercury and Brunauer-Emmett-Teller
(BET) porosimetry6, ellipsometric porosimetry7, small-angle neutron scattering (SANS),
positron annihilation spectroscopy (PAS)6,7 or Scanning Probe Microscopy (SPM)8. These
methods are appropriate for various ranges of pore sizes and distributions:
•
Densitometry gives the average density of the sample6. The method is not sensitive
enough for detection of ultrafine pores and requires large samples.
•
Mercury and BET porosimetry measurements give reliable values for open porosity,
but information about pore size distribution is limited. The methods are questioned
for ultrafine pores6 and require utilization of large samples7; furthermore, mercury is
toxic.
•
Ellipsometric porosimetry allows measurements of pore size distribution in
mesoporous (pore size 2-50 nm) thin films deposited at room temperature on smooth
solid surfaces like Si 7.
•
SANS is sensitive for pore sizes 1-100 nmREF6.
•
PAS detects ultrafine pores ranging from vacant lattice sites to larger voids and gives
information about bulk porosity6.
2
•
SPM, including scanning tunneling microscopy (STM) and atomic force microscopy
(AFM), was recently used for characterization of microporous (pore size <2 nm) and
mesoporous materials. Its use is limited to the outermost surface of the sample, and
the pore size detection limit depends on the tip diameter8.
This paper presents a novel and relatively straightforward method for characterizing gas
release from an oxide previously equilibrated in a controlled atmosphere. The geometry of
the oxide sample is assumed to be a plate, a rod, or a sphere. The plate can be selfsupporting or constitute a scale on a substrate. When the size of the oxide is known, a
desorption experiment determines the diffusivity and the amount of the gas contained. A
mathematical manual for the calculation of diffusivity and gas amount is given for the oxide
geometries under consideration. The method is validated by an experiment on helium in
quartz at 80°C. To demonstrate the feasibility of the method, we determine water and
nitrogen diffusivity and pore size in oxidized Zircaloy-2 and in one iron oxide by water
vapour and nitrogen desorption.
2. The mathematics of outgassing from plate, cylinder and sphere.
2a
a
UHV
UHV
UHV
UHV
UHV
UHV
2a
Figure 1. Geometry of sorption/desorption specimens: (a) plate of thickness 2a, (b) cylinder
of radius a, and (c) sphere of radius a.
Consider outgassing (gas release) from (a) a plate of thickness 2a, (b) a long cylinder of
radius a, and (c) a sphere of radius a, see Fig. 1. The plate has a normal coordinate x
extending from x=-a to x=a, and the cylinder and the sphere have a radial coordinate r, r≤a
The initial concentration is assumed to be uniform and equal to C1. From Crank9 one finds
that the concentration distribution in the material varies with spatial coordinate x (or r) and
time t in the following way:
3
(a) C ( x, t ) =
2C1 ∞ (−1) n −1
cos k n x exp(−k n2 Dt ),
∑
a n =1 k n
(b) C (r , t ) =
2C1
a
∞
∑k
n =1
J 0 (k n r )
exp(− k n2 Dt ),
n J 1 (k n a)
kn =
kn =
2C1 ∞ (−1) n −1 a
(c) C (r , t ) =
sin k n r exp(− k n2 Dt ),
∑
a n =1 k n r
(2n − 1)π
2a
j 0,n
(2)
a
kn =
(1)
nπ
a
(3)
where J0 and J1 are Bessel functions of order 0 and 1, respectively, and jo,n is the nth root of
the equation J0(kna)=0. In particular, j0,1=2.4048REF 10. The flux from (a) a single surface of
the plate, (b) the cylinder, and (c) the sphere is expressed by the formula
F=
2 DC1 ∞
∑ exp(−k n2 Dt )
a n =1
(4)
where D denotes diffusivity. In case (a) there is no flux across x=0 because of the symmetry
of the initial condition, which means that formula (4) describes the actual case of desorption
from a scale attached to an impenetrable substrate.
The terms of the series in eq. (4) decrease with n, and in a long time experiment only the
first term is significant. This means that in an e log[F ] versus t diagram (bracket [] signifies
dimensionless value) the flux approaches an asymptote in the form of a straight line of slope
1 dF
F dt
= −k12 D
(5)
t →∞
when t gets very long. The time evolution of the flux is illustrated in Figure 2.
Time is horizontal. Flux is vertical and is equipped with scales on both sides of the diagram,
for clarity. On the right-hand side there is the true logarithmic scale given in terms of
dimensionless e log[F ] (or ln[F ] ) so as to be compatible with equations (4) and (5). On the
left-hand side the logarithmic scale is translated into a direct F scale expressed in powers of
ten and in units of µmol cm-2 s-1. For instance, the tick mark − 20 on the right-hand scale
4
gives rise to a flux value e −20 = 2.06 × 10 −9 µmol cm-2 s-1 on the left-hand side. In the
10 -7
LogF, dimensionless
-16.1
-2
F, µmol cm s
-1
following the flux diagrams are drawn exclusively with such a power-of-ten scale.
10 -8
2
2 -1
m =slope of dashed line = π D(4a )
-18.4
1
10 -9
-20.7
e
10 -10
m = F = 2DC a
2
-1
1
-23
10 -11
10 -12
-25.3
Time, s
Figure 2. Flux versus time. The vertical axis has flux F with tick marks in powers of ten
corresponding to the dimensionless scale elogF.
A first measurement value of the outgassing experiment is the absolute value of slope (5),
 π 2 /(4a 2 )
(a)

2
2
−1
m1 = D × k1 = D × 5.7831 / a (s ) in case (b)
 π 2 / a2
(c)

(6)
Referring to Fig. 2, we draw the considered straight line of slope (5) backward in time until
it intersects the e log[F ] axis. Raising the exponential e to the value of intersection
e
log[2 DC1 / a ] , we get a second measurement value,
m2 = 2 DC1 / a (mol cm-2 s-1)
(7)
For a given size a, C1 and D are determined by the values m1 and m2. A third experimental
measurement value m3 is the integrated flux,
5
 a (mol cm −2 )
(a)
 2
−1
m3 = C1 ×  πa (mol cm ) in case (b)
(4π / 3)a 3 (mol)
(c)

(8)
The integration is performed over time and (a) plate thickness, (b) cylinder cross section,
and (c) sphere volume. The three measurement values m1 , m2 , and m3 are connected by the
relationship
0.8106
(a)

m2

−1
=  0.1101 / a (cm ) in case (b)
m1 m3 
2
−2
(c)
0.04838 / a (cm )
(9)
which can be used for estimating the coherence of the measuring procedure. The flux
Integrated F, µmol cm
-2
integration is illustrated in Fig.3.
C a=m
1
0
t
1
3
Time, s
Figure 3. Integrated flux over time for an oxide scale of thickness a
The measurement is carried out to time t1 , when the slope of the asymptote in Fig. 2 is well
determined. The remaining amount of outgassed substance is estimated from the first term
of the sum in eq. (4). Integration of (2 DC1 / a ) exp(− m1t ) from t1 to infinity gives rise to the
rest contribution (m2 / m1 ) exp(− m1t1 ) .
For completeness, measurement values are given for two-sided desorption from a plate of
thickness L on the assumption of a uniform initial concentration C1 . On doubling the flux in
eq. (4) one obtains for the desorption from a self-supporting plate,
m1 = D × π 2 / L2
(10)
m2 = 8 DC1 / L
(11)
6
3. Experimental
3.1 Gas Phase Analysis (GPA) and materials
The experimental setup for GPA presented in Figure 4 consist of:
a. A reaction chamber made by a stainless steel cross; the reaction chamber is pumped
with an ion pump via a leak valve.
b. A gas handling system, including a rough pump; controlled humidity gasses can be
used.
c. A pressure gauge which measures the total pressure in the reaction chamber;
d. A Mass Spectrometer, equipped with a quadrupole analyser, placed in UHV.
The reaction chamber material, stainless steel, ensures low background levels. The use of
silica as reaction chamber material is not always appropriate, especially for small samples,
giving too high background levels.
The sample is placed in the reaction chamber and pumped continuously by the ion pump via
a fully open leak valve. The MS analyses the composition of the released gases and the
computer graphically reproduces the time evolution of the partial pressures of the gas
constituents. By careful calibration of the mass spectrometer signal and ion pump rate, the
amount of gas released from the sample can be quantified. The theoretical model presented
above is used to calculate the diffusivity and the gas content in the sample.
Pressure Gauge
Mass spectrometer
Leak valve
in UHV
SS cross
Temperature
80°C/20°C
Sample
Gas handling
system
Figure 4. GPA setup for study of gas diffusion, concentration and effective pore size in
oxide scales.
7
The materials used in this study together with the methods used for the characterization of
the samples are summarized in Table 1.
Table 1. Characterization of the samples used in this study
Sample
2 nm oxide on
Zircaloy-2
1.3 µm oxide
on Zircaloy-2
Analysis methods
Optical
GPA SEM
microscopy
Shape
Plate
x
-
-
Plate
x
-
-
3 µm oxide on
Zircaloy-2
Plate
x
x
x
300 µm oxide
on Fe plate
Plate
x
x
-
Figures
Remark
8 – 12
Naturally form oxide on Zircaloy-2
at RT
The 1.3 µm and 3 µm oxides were
obtained by exposure of Zircaloy-2
plates to approximately 800 mbar
air at 450°C and respectively 500°C
using GPA
13 – 15
Optical microscopy was performed using a Leica DM IRM optical microscope.
FEG-SEM analysis was performed with a Leo 1530 Field Emission Scanning Electron
Microscope equipped with a GEMINI field emission column.
4. Validation of the method. He in Quartz.
Using the above-described method, the transport of He at 80°C in a quartz plate of thickness
2a = 0.1 cm, after equilibration in an exposure to 250 mbar He at room temperature, was
investigated and the results were compared with the existing literature data. The raw He flux
data are shown in Figure 5, where the data at long time is extrapolated.
-3
10
-4
10
-5
10
-6
10
-7
10
-8
10
-9
Background
Quartz sample plus background
µmol He s
-1
F
10
x A
10
-10
0
10000
20000
30000
40000
50000
60000
70000
Time,s
Figure 5. Outgassing raw data for He in quartz and for the stainless steel background.
8
After subtraction of the background and normalising to area unit, the time evolution of He
flux from quartz at 80°C is shown in Figure 6.
The parameters m1 and m2 are obtained from Figure 6. He diffusivity is thereby calculated
using eq. (10), (11) and (8) and the results are presented in Table 1. By integrating the flux
data from Figure 6 over time, Figure 7 is obtained. The concentration of He in quartz at
80°C is calculated using eq. (8) and the result is presented in Table 2.
It can be seen in Table 2 that the diffusivity and permeability (permeability = concentration
gradient times diffusivity) of He in quartz at 80°C evaluated using this method are in
agreement with the literature data1. Therefore, it indicates that the method is valid and can
be used to characterize the gas transport in oxides.
10
-4
F
-5
10
-6
10
-7
10
-8
10
-9
-2
µmol He cm s
-1
10
10
-10
10
-11
Quartz
0
10000
20000
30000
40000
50000
60000
70000
Time,s
Figure 6. Flux data for He outgassing at 80°C after background subtraction
Intergrated F over time
0,015
C x 0.1cm
0,012
µmol He cm
-2
Quartz
0,009
0,006
0,003
0,000
0
10000
20000
30000
40000
50000
60000
70000
Time, s
Figure 7. He integrated flux over time for a quartz plate of thickness 0.1 cm at 80°C
9
Table 2 – Transport parameters for He in quartz at 80°C
Transport parameters
Diffusivity
cm2 s-1
Concentration
µmol cm-3
Permeability
atoms s-1 cm-1 atm-1
Our results from this method
Results from references1
1.2 ⋅ 10-7
1.2 ⋅ 10-7
Dm1
Dm2
(1-2) ⋅10-7
0.12
-
3.5 ⋅1010
(4-6) ⋅1010
5. Characterization of oxide scales formed in thermal oxidation of
Zircaloy-2 and Fe
5.1 Diffusion of nitrogen and water in zirconium oxide scale at 80°C
Three Zircaloy-2 samples, with oxide thickness (L) of approximately 2 nm, 1 and 3 µm
respectively were outgassed at 80°C. Optical microscopy- and SEM images of the sample
with oxide thickness 3 µm are presented in Figure 8 and 9 respectively. The oxide thickness
was calculated previously from the oxygen uptake during oxidation in GPA to be 1.3 and
3.4 µm respectively, assuming the density of the oxide being 6 g cm-3.
Figure 8. Optical microscopy of preoxidized Zircaloy-2 with a 3 µm oxide
Figure 9. SEM of preoxidized Zircaloy-2 with a 3 µm oxide
10
After equilibration in dry gas containing nitrogen, the Zircaloy-2 samples were outgassed at
80°C. The raw flux data are presented in Figure 10a for water and 10b for nitrogen.
2 nm oxide + SS background
1.3 µm oxide + SS background
3 µm oxide + SS background
2
Outgassing rate FxA, µmol H O s
-1
10 -2
10 -3
10 -4
10 -5
10 -6
0
100000
200000
300000
400000
Time, s
Figure 10a. Raw data from outgassing of H2O from oxidized Zircaloy-2
2 nm oxide + SS background
1.3 µm oxide + SS background
3 µm oxide + SS background
2
Outgassing rate FxA, µmol N s
-1
10 -4
10 -5
10 -6
10 -7
10 -8
0
100000
200000
300000
400000
Time, s
Figure 10b. Raw data from outgassing of N2 from oxidized Zircaloy-2
The outgassing data obtained for the sample with the naturally formed oxide were used as
background for the other two samples. In this way, the gas molecules that are desorbed from
the macroscopic surface of the samples and from the reaction chamber’s surface are
subtracted. The resulted flux data, presented in Figure 11a for water and 11b for nitrogen
respectively, correspond to desorption of the molecules, which have been adsorbed on the
internal oxide surfaces.
11
10 -5
2
F, µmol H O cm s
-2 -1
10 -4
10 -6
3 µm oxide
10 -7
1.3 µm oxide
10 -8
0
100000
200000
300000
400000
Time, s
Figure 11a. Outgassing of H2O from oxidized Zircaloy 2
10 -7
2
-2
F, µmol N cm s
-1
10 -6
10 -8
3 µm oxide
10 -9
1.3 µm oxide
10 -10
0
100000
200000
300000
400000
Time, s
Figure 11b. Outgassing of N2 from oxidized Zircaloy 2
From Figure 11a and 11b, the parameters m1 and m2 have been obtained and used to
calculate the diffusivity for H2O and N2 in zirconium oxide scale at 80°C. The results are
summarised in Table 3.
Integrating the flux data from Figure 11 over time, the concentration of water and nitrogen
in the oxide scales can be calculated by using the parameter m3 obtained from Figure 12 a
(water) and 12b (nitrogen). The results are summarized in Table 3.
12
0.04
-4
C x 3 10 cm
0.03
3 µm oxide
2
µmol H O cm
-2
1
-4
0.02
C x 1.3 10 cm
1
1.3 µm oxide
0.01
0
/
0
150000
300000
t→∞
450000
Time, s
Figure 12a. H2O integrated flux over time for Zircaloy-2 oxide scales
0.0016
0.0012
1
3 µm oxide
2
µmol N cm
-2
-4
C x 3 10 cm
0.0008
-4
C x 1.3 10 cm
1
0.0004
1.3 µm oxide
0
/
0
150000
300000
450000
t→∞
Time, s
Figure 12b. N2 integrated flux over time for Zircaloy-2 oxide scales
Table 3. Transport parameters for H2O and N2 in oxide scales on Zircaloy-2 at 80°C
H2O
Oxide thickness, µm
2
Dm1, cm s-1
Diffusivity
2
Dm2, cm s-1
3
Concentration, µmol cm-
1.3
-14
5.2 . 10
-14
3.4 . 10
190
N2
3
-14
8.6 10
-13
1.1 . 10
119
.
13
1.3
-13
2.8 . 10
.
-13
2.4 10
3.8
3
-13
2.1 10
-13
2.1 . 10
5.0
.
From Table 3 it is evident that there is a good agreement between the diffusivity values Dm1
and Dm2 for both, water and nitrogen.
5.2 Diffusion of nitrogen and water in Fe-oxide scale at 20°C
A 1cm2 oxidised Fe sample, with oxide thickness of 300 µm was outgassed at 20°C. The
oxide was previously exposed in air at RT for 100 hours. For these conditions, water
coverage of 1ML has been considered. A SEM cross-section micrograph of the oxidized-Fe
sample is presented in Figure 13.
300 µm
Figure 13. SEM cross-section of oxidized-Fe sample with oxide thickness of 0.3 mm
The outgassing of H2O and N2, at RT from the 0.3 mm thick Fe-oxide is presented in Figure
14 and 15 respectively.
1.E-02
µ molH2 O cm s
-2 -1
Flux of H2 O
1.E-03
1.E-04
1.E-05
1.E-06
0
40000
80000
120000
Time, s
Figure 14a. H2O flux vs. time for a 0.3 mm thick Fe-oxide after background subtraction
14
3
Integrated flux of H2 O
µ mol H2 O cm
-2
2.5
2
1.5
1
0.5
0
0
40000
80000
120000
Time, s
Figure 14b. H2O integrated flux over time for a 0.3 mm thick Fe-oxide
1.E-03
Flux of N2
µ mol N2 cm s
-2 -1
1.E-04
1.E-05
1.E-06
1.E-07
0
40000
80000
120000
Time, s
Figure 15a. N2 flux vs. time for a 0.3 mm thick Fe-oxide after background subtraction
15
0.3
µ mol N2 cm
-2
Integrated flux of N2
0.2
0.1
0
0
40000
80000
120000
Time, s
Figure 15b. N2 integrated flux over time for a 0.3 mm thick Fe-oxide
From these figures, the parameters m1, m2 and m3 for H2O and N2 have been obtained and
by applying the equations (6) – (8), the diffusivity and concentration of H2O and N2 in the
Fe-oxide at 20°C have been calculated. The results are summarized in Table 4.
Table 4. Diffusivity and concentration of H2O and N2 in Fe oxide scale at 20°C
Oxide thickness, µm
Diffusivity
H2O
N2
300
300
2 -1
-9
6 10
5 10-9
2
3 10-9
5 10-9
90
10
Dm1, cm s
Dm2, cm s-1
3
Concentration, µmol cm-
In the case of the 0.3 mm thick Fe-oxide, the diffusivity constants for both, water and
2
nitrogen, are in the range of 10-9 cm s-1.
The agreement between Dm1 and Dm2 reflects the overall consistency in between C1, L and
D.
16
6. Effective pore size
6.1 Calculation of effective pore size based on H2O/N2 mol ratios
Data on water and nitrogen content in a sample can be used for estimation of effective size
of open meso/micropores.
For simplicity, we make the assumption that the pores are long cylinders with circular cross
section of diameter Φ. The samples are equilibrated in water and nitrogen containing gas
near room temperature. Under these conditions, the water molecules, which have higher
tendency for adsorption than N2 molecules10, will stick to all surfaces, including internal
pore surfaces. N2 molecules will predominantly be found in the pore’s gas phase. If the pore
size is about two molecular diameters of water, they will block the pores, so N2 cannot be
accommodated in the pores. For oxides with an open network of micropores and mesopores,
the water molecules are mainly adsorbed on the pore walls, the amount of water vapour
present in the gas phase being negligible. As the pore size increases, the amount of water
vapour from the gas phase increases and cannot be neglected. The H2O/N2 molar ratio,
which can be associated with the ratio between the surface area and the volume of the pores,
can be used to calculate the effective pore size. This is illustrated in Figure 16 where a
monolayer of water is adsorbed on the pore surface. It can also be seen in Figure 16 that the
nitrogen molecules are present in the pore’s gas phase and therefore the total volume of the
nitrogen gas released by outgassing can be used to calculate the total volume of the pores. In
Figure 17, the relation between H2O/N2 molar ratio and effective pore diameter is shown for
water coverages of 0.01, 0.04, 0.3 and 1 ML.
Φ pore
H2O
N
O2
Figure 16. Schematics of an air-containing pore (effective diameter Φ) with one monolayer
of adsorbed water
17
1000000
1.0 ML
0.3 ML
0.04 ML
0.01 ML
100000
H2O/N2 molar ratio
10000
1000
100
10
1
0.1
0.01
0.1
1.0
10.0
100.0
1000.0
Effective pore diameter, Φ , nm
Figure 17. Graph for estimation of effective pore diameter of a sample based on known
H2O/N2 molar ratio and water coverage
6.2 Effective pore size in Fe- and Zr oxides
For the oxide scales on Zircaloy-2, with the data from Table 3, the H2O/N2 molar ratio
obtained for 1.3 µm oxide thickness is 50 and for 3 µm oxide thickness is approximately 24.
A water coverage of approximately 0.04 ML is obtained from amount of water released
from the 2nm oxide plus background with known surface area. Using these values, the
effective pore diameters are then found to be in 1 -10 nm range as indicated in Figure 18.
For the Fe-oxide, the SEM image in Figure 18 reveals the presence of big pores and cracks,
which seems to be isolated at the used magnification. H2O/N2 molar ratio obtained from
Table 4 is 9. For water coverage of approximately 1 MLREF11 the effective pore diameter is
approximately 500 nm as indicated in Figure 18.
The effective pore diameters for these oxides Φ are estimated from Figure 18. As can be
seen in Figure 18, the effective pore diameters for the evaluated oxide scales are in
nanometer range.
18
1000000
1.0 ML
100000
0.04 ML
1000
100
10
0.1
Fe oxide
1
Zry-2
3 µ m oxide
Zry-2
1.3 µ m oxide
H2 O/N2 molar ratio
10000
0.01
0.1
1
10
100
1000
Effective pore diameter, Φ, nm
Figure 18. Effective pore size estimation for oxide scales on Zircaloy-2 and Fe
6. Summary
A novel and relatively straightforward method to quantify diffusion and amount of gas
present in oxide scales on samples equilibrated in controlled atmosphere is presented.
Theoretical outgassing kinetics from plate, cylinder or sphere geometries are given. From
outgassed amounts of water and nitrogen the H2O/N2 molar ratio is used to estimate an
effective pore size in the oxide layers.
The method is validated in measurements of
diffusivity and concentration of He in quartz at 80°C and applied for characterisation of two
Zr oxides and one Fe oxide. The diffusivity of N2 at 80°C in a 1 µm Zr-oxide was in order
of 10-13 cm2s-1 and in a 0.3mm thick Fe-oxide at 20°C it was in the order of 10-9 cm2s-1. The
effective pore size was 1-10nm and 100-1000nm for the Zr-and Fe oxide respectively.
Acknowledgements
Financial support from Westinghouse Electric Sweden AB, SFS and HTC/KME is
gratefully acknowledged.
19
References
1. J. E. Shelby, Permeation, Diffusion and Solubility Measurements, in Handbook of
Gas diffusion in Solids and Melts, ASM International.
2. U. Brossmann, G. Knöner, H.-E. Schaefer, R. Würschum, Oxygen transport in
nanocrystalline ZrO2, Rev.Adv.Mater.Sci. 6 (2004), p. 7-11.
3. P. Kofstad, High Temperature Corrosion, Elsevier Applied Science London and
New York, 1988.
4. B. MacDougall, M. Graham, Growth and Stability of Passive Films, in: Corrosion
Mechanisms in Theory and Practice, P. Marcus (Ed.), Second Edition, revised and
expanded, Marcel Dekker Ink. New York, p. 189-216 (2002).
5. Yuquan Ding and Derek O. Northwood, Characterization of the oxides formed on
Zr-2.5wt.% Nb during high temperature corrosion, J. Alloys Comp. 179 (1-2), 1992,
p. 207-217.
6. R.W. Siegel, Nanostructures of metals and ceramics, in: Nanomaterials: Synthesis,
properties and applications, AS Edelstein and RC Cammarata (Eds), Institute of
Physics Publishing Bristol and Philadelphia, pp. 201- 217.
7. M. R. Baklanov, K. P. Mogilnikov, V. G. Polovinkin, and F. N. Dultsev,
Determination of pore size distribution in thin films by ellipsometric porosimetry, J.
Vac. Sci. Technol. B 18(3), May/Jun 2000, p. 1385-1391.
8. J.I. Paredes, A. Martίnez-Alonso, J.M.D. Tascόn, Application of scanning tunneling
and atomic force microscopies to the characterization of microporous and
mesoporous materials, Microporous and Mesoporous Materials 65 (2003), p. 93–
126.
8. J. Crank, The Mathematics of Diffusion, Clarendon, Oxford, 1976.
9. M. Abramowitz, I. A. Stegun, Handbook of Mathematical Functions, NIST, 1972.
20
10. C. Anghel, Erik Hörnlund, Gunnar Hultquist, Magnus Limbäck, Gas phase analysis of
CO interactions with solid surfaces at high temperatures, Applied Surface Science 233
(2004), 392 –401.
11. C. Leygraf, Atmospheric corrosion, Electrochemical Society Series, John Wiley & Sons Inc.,
p.9-12 (2000).
21