Corrosion at high temperature

The corrosion at high temperature is the degradation of metals by the environment at high temperatures (higher than 500  °C); it is a complex phenomenon which takes place in the engines, boilers and engines. The combustion gases indeed have a complex composition because of composition of fuel and the air: N2, O2, CO2 and H2O of course, but very often also S 2, SO2, Cl2, NaCl, and various Oxide S (V 2O5…).

One then distinguishes two type of degradations:

  • the corrosion known as “dry”, which results from the oxidation of metal by the gases (O2, S2, SO2, H2O); one also speaks about oxidation at high temperature ;
  • and corrosion known as “hot”, or “fluxing”, which result from a dissolution of oxide by molten salts (Na2SO4) and oxides which settle (there can also be a fusion Eutectique, mechanism similar to salt on the ice).

In certain situations, there is molten metal and solid metal cohabitation (the temperature is thus necessarily high). It is for example the case of the foundry; but the molten metals are sometimes used as fluids, such as for example the Sodium in the Nuclear plant Super-Phenix (Creys-Malville, France). These situations involve particular phenomena of corrosion.

Corrosion dries (oxidation at high temperature)

When one puts a metal in the presence of dioxygene, this one is adsorbed (i.e. fixes itself) on surface and reacts to form an oxide coating. With room temperature, the diffusion in the solid is negligible; either the oxide coating is compact and protective (alumina on aluminum or chromine on the stainless steels) and metal does not move, or it is porous or not adherent (rust), and metal is degraded by a growth of the oxide coating to the detriment of metal. The mechanisms which come into play are the migration in the medium external (diffusion, convection, electric field) and the reactions of surface.

Beyond 400  °C, the diffusion in solid phase, which is activated thermically, between concerned, and even a compact layer will be able to degrade itself (the oxide forms a crust which is cracked).

To simplify, the following study relates to the action of dioxygene alone.

Mechanism of degradation

In certain cases, the oxide is volatile (case for example of PtO2), or is fragile, porous, does not adhere to the substrate. In this case, the mechanism of degradation is obvious, the dioxygene reacts with metal to form oxide and this oxide evaporates or is scaled.

In the case of an adherent and compact oxide, the mechanism of degradation was described by J. Bénard. Degradation is done in five stages:

  1. Adsorption and dissociation of dioxygene on the surface of metal;
  2. reaction enters the adsorbed oxygen atoms and metal to form oxide germs;
  3. side growth of the germs until the junction, formation of a continuous film;
  4. growth of oxide film in thickness by diffusion in film;
  5. rupture of oxide film by the Forced S induced by its growth and the defects.


Oxidation at high temperature: mechanism of degradation of an adherent oxide coating and compacts

Balance thermodynamic: oxide, the stable shape of metal

Let us note M the atom of metal, whatever its nature (Fe, Ni, Al, Cr, Zr…), and let us note M N O2 the corresponding oxide; the coefficients were selected to simplify the writing by considering the reaction with a molecule of whole dioxygene, that can be Fe2O3, Al2O3, Cr2O3 ( N = 4/3), Fe3O4 ( N = 3/2), FeO, NiO ( N = 2), ZrO2 ( N = 1)… Molar enthalpy partial (free energy of Gibbs) Δ G M N O2 of the reaction of oxidation

N M + O2 → M N O2
is written:
Δ G M N O2 = Δ G 0M N O2 + RT ·ln P O2
where P O2 is the pressure partial of dioxygene expressed in atmospheres, R is the constant of perfect gases and T is the absolute temperature expressed in Kelvin (K). The enthalpy is represented in the diagram of Ellingham-Richardson, where one traces Δ G 0 ( T ).

The diagram is built by supposing thermodynamic balance, of the pure solid phases (activities equal to one), the fugacity of dioxygene equal to its partial pressure, and that Δ G 0 depends linearly on the temperature; Δ G refer to the reaction for a mole of O2. Oxidation can take place only if

Δ G M N O2 < 0
that is to say
Δ G 0M N O2 < RT ·ln P O2
If one defines the free enthalpy of the dioxygene
Δ G O2  =  RT ·ln P O2
and that one traces - Δ G O2 ( T ) in this diagram, one obtains a line passing by 0; the intersection of this line and the right-hand side representing Δ G M N O2 defines the zone of temperatures where the oxide is thermodynamically stable for the pressure partial of dioxygene given. For the usual conditions, the stable shape of metals is the oxidized form.

----

Note : in any rigor, it would be necessary to note “- RT ·ln ( P O2/ P 0)” where P 0 is the pressure being used to define Δ G 0 (or “normal pressure”), in fact 1  atm.
----

Adsorption of dioxygene and formation of the oxide small islands

The molecule of dioxygene O2 binds to metal then dissociates in two separate oxygen atoms. The oxygen atoms occupy of the preferential sites of adsorption, in general the sites having the greatest number of close metal atoms. The distribution of these sites depends on the crystallographic structure of surface, therefore in particular of the orientation of the Cristallite (or grain).

Certain authors suggest that in the case of alloys, the atoms place themselves preferentially in the vicinity of the least noble atoms, for example iron in the case of an alloy Fe-Al. This has three consequences:

  • the structure of the incipient oxide germ adapts to the structure of the substrate, and in particular to the orientation of the metal crystal (epitaxy);
  • according to their orientation, certain oxide germs grow more quickly than others, the oxide film is thus likely to have a texture (preferential orientation of crystallites);
  • it can have a rearrangement of the metal atoms there by diffusion of surface (crystallite changes form), to reduce the energy of interface, which can lead to the formation of facets or scratches.
Part of adsorbed oxygen dissolves in metal and diffuses (i.e. the oxygen atoms slip between the atoms of metal and progress towards the interior of the part), which in certain cases can lead to an internal oxidation (cf further).

Side growth of the oxide small islands

Oxide the small islands, very thin, grow laterally until joining. This growth is done by diffusion of surface; the speed of diffusion thus depends on the atomic density of surface. Thus, according to the crystalline orientation of the substrate, certain oxide germs grow more quickly than others. The initial oxide film can thus have a texture (preferential crystallographic orientation).

Growth of oxide film in thickness

When the layer is adherent and compact, the oxide now isolates metal from the atmosphere. The atoms of dioxygene are thus adsorbed on oxide. When the oxide is compact and adherent, one can consider two mechanisms of growth:

  • the oxygen adsorbed on oxide dissociates, passes in solution in oxide, diffuse towards the interface metal/oxide, and combines with this interface with the metal atoms; the creation of oxide is thus done with the interface metal/oxide, one speaks about “growth towards the interior”;
  • metal with the interface metal/oxide passes in solution in oxide, diffuse towards the interface/gas oxidizes, and combines with this interface with adsorbed oxygen; the creation of oxide is thus done with the interface oxidizes/gas, one speaks about “growth towards outside”.
One can also have a combination of both, with the oxide which is formed in the middle of the oxide coating.

One usually considers that the oxide M N O2 is an ionic compound O2-/M m +, m respecting the neutrality of the loads ( m × N = 4); the bond oxide is in fact more complex, but this approximation simplifies calculations of diffusion. The diffusion of the species is thus done also in ionic form, primarily in interstitial or lacunar form; the presence of defects of antisites OM m + and MO m + N • in oxide is not considered because of energy which it would be necessary to create them (one uses the Notation of Kröger and Vink, recommended by Iupac).

Growth towards outside

In the case of a diffusion towards outside, the metal ions therefore leave behind them gaps. There is thus a contraction of the surface layer of the metal which creates constraints. When the concentration in gap is sufficient, they condense to form pores (principle similar to the precipitation). One thus notes frequently pores with the interface metal/oxide. This formation of pores causes a stress relaxation, but gives place to concentrations stresses.

The growth towards outside can be done in two manners:

  • the gas reacts with oxide (the dioxygene is reduced), it collects metal atoms of oxide, leaving gaps as well as holes of electron;
    O2 + MM (oxide) → 2 O2- + V M (oxide) + 2 H
    the gaps and the holes of electron diffuse towards the interior, and with the interface metal/oxide, of the atoms of metal oxidize (collect the holes of electron) to become an ion M m + and passes in oxide, leaving a lacune
    M (metal) + m H → M m • (metal)
    M m • (metal) + V M (oxide) → MM (oxide) + V M (metal) ;
    in some kinds, the reduction of dioxygene creates a metal deficit of ion in the layer, which “aspires” the atoms of metal;
  • an atom of metal passes in insertion in the oxide coating, diffuse towards outside and reacts with dioxygene on the surface of oxide.

Growth towards the interior

In the case of a diffusion towards the interior, the oxygen ions “are encrusted” in metal and thus create a dilation, which generates constraints.

The growth towards outside can is done in the following way:

  1. with the interface metal/oxide, the atoms of metal reacts with the O2- ions of oxide (metal oxidizes), leaving gaps as well as electron free;
    N MM (metal) + 2OO (oxide) → 2 N MM (metal) + 2OO (oxide) + V O + me'
    to some extent, the ions oxygenates OO oxide play the part of Catalyseur;
  2. the gaps and the electrons diffuse towards outside, and with the interface/gas oxidizes, of the molecules of gas are reduced (collect the electron) to become O2- ions and pass in the oxyde
    O2 + 4e' → 2O2-
    O2- + V O (oxide) → OO (oxide)
In some kinds, the oxidation of metal creates a deficit of ion oxidizes in the layer, which “aspires” the atoms of gas.

The other situation (diffusion of an atom or an oxygen ion in intersticiel) is not very probable, oxygen being “a large” atom.

Kinetics of oxidation

Kinetics of adsorption

The Adsorption of dioxygene can be described by two phenomena: initially a Physical absorption: the O2 molecule binds to metal by a Force of van der Waals, a reversible way, then a Chimisorption, thermically activated reaction

O2 + < < s> > = < < O2-s> > physical absorption
< < O2-s> > + < < s> > = 2< < O-s> > dissociation (chemisorption)
“S” indicates a site of adsorption, and the doubles brackets < < … > > indicate that the species is with the interface metal/gas. Several models describe the isothermal kinetics of adsorption:
  • full-course adsorption: Hill, Hill and Everett, multi-layer Langmuir
  • adsorption: Study Bureau (Brunauer, Emmet and Teller), theory of the blade (Frenkel, Hasley and Hill), potential Polyani.
but they are seldom used within this framework. Indeed, in our case, we can retain the following assumptions:
  • the diffusion in gas is fast and is not a limiting factor;
  • one is almost instantaneously with balance adsorption ↔ desorption, the processes of adsorption being thermically activated.
If the limiting factor of the phenomenon is a reaction of surface, there is then linear kinetics: being with balance, the gas contribution on surface and the quantity of matter being desorbed are constant, therefore the concentrations in réactants are constant. Consequently, the quantity of matter reacting is determined by the quantity of matter arriving on surface and while leaving. This flow being constant (balance), one concludes from it that the reaction follows linear kinetics:
m ox = kl . T
where m ox is the oxide mass, kl the linear coefficient of oxidation and T is time.

The kinetics of adsorption plays whenever one has a nonprotective oxide coating (porous or not adherent, or volatile oxide): if the layer is protective, the diffusion in the oxide coating is much slower than adsorption and it is thus the kinetics of diffusion which controls the phenomenon. However, the kinetics of adsorption controls the first minute of oxidation, during the germination of oxide and the side growth of the grains; certain authors raised linear kinetics in the first minutes of oxidation even in the case of a compact and adherent oxide.

Adherent and compact layer

The initial training of oxide film depends only on the gas supply of surface, and is thus overall linear. Once this formed film, it constitutes a barrier between metal and gas, provided that this film is adherent and compact. There is thus a deceleration of corrosion.

All in all, corrosion is done by diffusion through oxide. The thicker the film is, the more time of diffusion is long. A fast analysis shows that the thickness E oxide film, and thus the catch of mass of the part, varies like the square root of time:

e \ alpha \ sqrt {T} ;
m_ {OX} \ alpha \ sqrt {T} .
The first model with parabolic kinetics was proposed by Tamman in 1920.

In 1933, Carl Wagner made a finer analysis and obtains to him also parabolic kinetics. It poses like assumptions that:

  • the migration utilizes, in addition to the diffusion by random jumps, the effect of the chemical Gradient of Potentiel as well as the effect of the local Electric field created by the burden-sharing;
  • the oxide has a composition close to the Stœchiométrie;
  • at any moment, the oxide is locally with chemical balance;
  • the circuit is open, i.e. the total Electric current is null and thus that flows of species charged are coupled.
The theory of Wagner is of the interest to connect the constant speed (constant of proportionality between the thickness and the square root of time) to the fundamental parameters of material (like the coefficients of diffusion). In the facts, that gives rather bad results, the assumptions of Wagner being too far away from reality (he is unaware of in particular the role of the grain boundary in the diffusion).

But one notes a growth however well in experiments of square root of time.

See the detailed article Theory of the kinetics of oxidation of Wagner.

Rupture of the layer or porous layer

When the layer breaks, because of the generated constraints, the gas reaches directly an not-oxidized important surface. One a constant thus acceleration of the catch of mass, a rupture of the quadratic law.

When the layer is very fragile and breaks or falls apart permanently, or when the oxide is porous, to see volatile, nothing is opposed to oxidation, the law is thus linear.

Hot corrosion (fluxing)

Corrosion by molten metals

Constant physics

  • constant of the Perfect gas S R = 8,314  472  J.mol-1.K-1 ± 1,5.10-5  J.mol-1.K-1

See too

Random links:Avathar | The Community of communes of Xaintois | Claës Fredric Hornstedt | Channel Escape | Georges Gault | 420