2. Thermodynamics of high temperature oxidation
2.1 Basic thermodynamics
Most metals are generated by reducing ores. In other words, oxides, etc., are stable under natural conditions. Therefore, t o generate a metal, a charge of energy is required. Metals tend to return to a more stable state as an ore by reacting to oxygen, etc. Smelting is done artificially and corrosion occurs naturally. W ithout water, in many cases the corrosion rate at a normal temperature does not cause a particular problem. However, as the temperature rises, corrosion progresses at arate that is an engineering problem not to be ignored. Corrosion that occurs unrelated to water is called high temperature corrosion. The reaction of the high temperature corrosion in which metallic element M is transformed into oxide MαOβ is represented by formula (1-1).
To calculate the free energy (ΔG) in formula (1-1), formula (1-2) is used.
ΔG
0：Standard free energy of formation
R：Gas constant
T：Absolute temperature
α
x：Activity of x
P
O2：Oxygen partial pressure
Here, based on the ΔG value, the following can beestimated:
ΔG＜0 ; The reaction represented by formula (1-1)
is in progress. … Oxidation
ΔG＝0 ; Equilibrium
ΔG＞0 ; A reverse reaction inverse of the reaction represented by formula (1-1) is in progress. … Reduction
When ΔG = 0, PO2 is the oxygen partial pressure at which the metal and the oxide are balanced. The pressure is called dissociation pressure and represented by the formula below.
In short, if the oxygen partial pressure of the atmosphere is greater than the dissociation pressure, a metal is oxidized, and if it is lower than the dissociation pressure, the oxide is reduced to a metal. Figure 1-1 shows the dissociation pressure of NiO which is calculated by using Thermodynamic Database MALT2 (sold by Kagaku Gijutsu-Sha). The dissociation pressure increases as the temperature rises. At a temperature of 400 °C, the oxidation reaction progresses if the oxygen partial pressure is greater than 2.6 × 10-28 atm. At temperature of 800 °C, it is 1.2×10-14 atm. Judging from thermodynamic stability, the lower the temperature, the easier the oxidation reaction. However, because mass transfer generally controls the oxidation rate, the higher the temperature, the faster the oxidation rate. The kinetics will be described in the following chapter.
Fig. 1-1 also shows the dissociation pressure of Cr2O3. The graph shows that the dissociation pressure of Cr is much lower than that of Ni, and that Cr is more likely to be oxidized than Ni. In the graph, there are three areas: A, B, and C in ascending order of oxygen partial pressure. In Area A, both Ni and Cr are stable. In Area B, only Cr is oxidized. In Area C, the graph indicates that both metal oxides are stable. This thermodynamic stability is summarized in the Ellingham diagram, whose horizontal and vertical axes show temperature and ΔG
0, respectively. Figure 1-2 shows an example1). In the figure, M indicates the melting point of an element, and <span style="border: 1px solid ">M</span> indicates the melting point of an oxide. From the figure, the dissociation pressure of each oxide can be estimated. For example, at 1 000 °C, the dissociation pressure of Cr2O3 is about 10-22 atm, which is indicated by the broken line in Fig. 1-2. If the value is converted into the H2/H2O ratio, it is around 103 to 104. The lower an element is positioned in Fig. 1-2, the more stable its oxide becomes, and the harder the oxide is reduced to the element. For example, at 1 000 °C, FeO, NiO, etc are positioned higher than CO2, which means that they are less stable than CO2. In other words, if FeO coexists with C (carbon), Fe is reduced, and CO2 is generated. On the other hand, substances such as Al2O3, SiO2, and Cr2O3 are more stable than CO2. Similarly, if Al coexists with CO2, Al is oxidized, and CO2 is reduced. As shown above, the Ellingham diagram is widely used to examine possibilities of reactions.
Metals such as Al, Si, and Cr are known as elements with a good high-temperature oxidation resistance. This means that their oxides are more stable than those of Fe and Ni. An element (such as Cr) that is likely to be oxidized (stable oxide) may turn to an element that is hard to be oxidized (slow oxidation rate). Therefore, it should be understood that equilibrium and kinetics are considered separately.
Fig. 1-1 Dissociation pressures of NiO and Cr<sub>2</sub>O<sub>3</sub>
Fig. 1-2 Ellingham diagram
2.2 Thermodynamics of corrosion products
Figure 1-3 is a schematic diagram of a corrosion product (hereinafter referred to as an oxide scale). On the interface between the oxide scale and the atmospheric gas, if the gas flow is sufficient, the oxygen partial pressure is equal to that of the atmosphere. In the oxide scale, an oxygen partial pressure gradient occurs. The deeper the layer in the oxide scale, the lower the partial pressure. Then, if thermodynamic equilibrium is established on the interface between a metal and its oxide, the metal can coexist with the oxide, and they are considered to be in equilibrium with each other. Thus, the oxygen partial pressure is equal to the dissociation pressure. For Ni at 1 000 °C, the oxygen partial pressure is 4.5×10-11 atm. For Cr at the same temperature, the pressure is 9.1 × 10-23 atm. Understanding an oxygen partial pressure drop in this scale is important for understanding high temperature corrosion phenomena.
Fig. 1-3 Schematic diagram of an oxide scale
3. Kinetics of high temperature oxidation
3.1 Rate formula
The following are the typical processes of high temperature oxidation. (Actually, there are other processes such as adsorption and dissociation of oxygen. However, they are omitted here for simplicity.)
(1) Diffusion in the gas: Oxygen moves to the surface of the oxide scale, from the atmosphere.
(2) Diffusion in the scale: Ionic metal or oxygen moves in the oxide scale.
(3) Reaction: The metal reacts with oxygen, and then the oxide scale grows.
Since diffusion in a gas is generally much faster than that in a solid, diffusion in the gas does not become the rate-controlling step except in special circumstances. If an oxidation rate depends on diffusion or reaction in the gas, the oxide scale thickness (x) is proportional to the time (t), which is expressed in formula (1-4).
Formula (1-4) is called the linear law, and kl
is called the linear rate constant. The above formula is satisfied when the growth of the oxide scale is insufficient in an early stage of corrosion and diffusion in the solid is not a problem. In this case, the mechanism of controlling the corrosion reaction does not work, causing corrosion damage often, which does not fit practical use.
On the other hand, when mass transfer in a scale becomes the rate-controlling step, the thicker the scale, the longer the time the mass transfer takes. Therefore, the growth rate of the oxide scale (dx/dt) is inversely proportional to the scale thickness (x).
Integrating and arranging the above formula resultsin formula (1-6).
Formula (1-6) is the parabolic rate law which is important for engineering. kp
is called a parabolic rate constant and is used for evaluating the high temperature corrosion resistance of materials. Where x is scle thickness, the scale thickness = the amount of corrosion. Therefore, the unit of kp
is [mm·s− ½]. Generally, the mass change or reduced thickness is used to evaluate the amount of corrosion in a laboratory or in inspection using actual equipment. In this case, by selecting the unit of kp
, it is possible to use “x” (the mass or reduced thickness) according to the actual measurement system. On the contrary, when kp
is used for evaluation, be careful about unit systems.
Figure 1-4 shows the correlation of oxide scale thickness (= amount of corrosion) and time associated with formula (1-6). (The continuous scale destruction shown in the figure is described in section 4.2.) It is found that the growth rate (= corrosion rate) of the oxide scale becomes slow as time goes by, indicating that the progress of corrosion is slowed. When the scale becomes a barrier against mass transfer, it works as a protective scale and suppresses corrosion. In other words, the key to suppressing high temperature corrosion is how well to form a protective scale which can maintain the parabolic rate law.
Fig. 1-4 Schematic diagram of the time dependence of the amount of corrosion
3.2 Factors affecting the parabolic rate law
3.2.1 Diffusion coefficient
When the oxide scale grows in accordance with the parabolic rate law, diffusion in the oxide scale becomes the rate-controlling step. The diffusion equation is expressed by the product of the mobility and the driving force based on Fick’s first law.
By combining Wagner’s theory with the parabola rate law based on Fick’s first law, etc., kp can be expressed in formulas (1-8) and (1-9) 2), 3).
DM and D0 respectively represent diffusion coefficients of the metal (M) and the oxygen in the scale. μM and μO respectively represent the chemical potentials of the metal (M) and oxygen. Suffix [ ' ] and [ " ] respectively represent the values on the interfaces between the metal and the oxide scale and between the oxide scale and the atmosphere. Both the metal and oxygen move in the scale. However, since there is usually a large difference in diffusion capability between them, the movement of one of them can be ignored. Formula (1-8) represents the movement of a metal. Formula (1-9) represents the movement of oxygen. These formulas contain μM and μO that are difficult to measure. If the deviation of MαOβ from the stoichiometric composition is small, kp can be expressed with PO2 using formulas (1-10) and (1-11).
In the above formula, μ° O2 represents the standard chemical potential of O2.
By plugging the above formulas into formulas (1-8) and (1-9), formulas (1-12) and (1-13) are obtained.
kp can be expressed as the product of the driving force and the mobility. The driving force is the oxygen partial pressure gradient on the interfaces between the metal and the oxide scale and between the oxide scale and the atmosphere. The mobility is expressed by DM or Do which is the self-diffusion coefficient of the metal or oxygen in the scale. In other words, these formulas show that in order to suppress a corrosion rat oxygen partial pressure gradient or the diffusion coefficient must be small. Cr2O3, Al2O3, and SiO2 are typical oxides with a small diffusion coefficient. Heatresistant alloys are designed to form one of these oxide scales. Stainless steel is the most common heat-resistant alloy, to which at least a certain amount of Cr is added so that Cr2O3 can be formed by corrosion to provide a corrosion protection effect.
3.2.2 Oxygen partial pressure
Oxide scales contain defects and function as p- or n-type semiconductors. An n-type semiconductor has lattice defects of oxygen, and oxygen ions move in the oxide scale. In a p-type semiconductor, lattice defects of metal cause metal ions to move in the oxide scale. Figure 1-5 is a schematic diagram of defects in NiO that is a typical p-type oxide. Stoichiometrically, the ratio of Ni to O is 1:1.However, there is Ni absence part (indicated by□ in the figure); therefore the molecular formula is written as Ni1 −δO. In order to maintain electrical neutrality, some Ni2＋ molecules change to Ni3＋ ones. A part where Ni2＋ is lost is called vacancy, and Ni3＋ in which Ni2＋ is positively charged is called a positive hole. The former and latter are represented by “VM'' ” and “h･” respectively, and the number of charges are represented by [ '' ] or [ ･ ]. Figure 1-6 shows the processes of forming a vacancy and a positive hole. In Step 1, an oxygen molecule in the atmosphere reaches NiO. In Step 2, an oxygen atom is absorbed by NiO. In Step 3, the oxygen atom receives an electron from Ni2＋ and is chemically adsorbed by NiO, which forms a positive hole. In Step 4, a reaction couple of O2- and Ni2＋ is formed, and then it is absorbed by the NiO lattice. On the other hand, in the NiO lattice, the Ni2＋ site reacted becomes vacant, simultaneously forming a new positive hole.
Fig. 1-5 Schematic diagram of a p-type semiconductor
with positive holes (Ni3+) and vacancies (□ )
Fig. 1-6 Schematic diagram of forming positive holes (Ni3+)
and vacancies (□ ) in a p-type semiconductor (NiO)
If a divalent metal vacancy is formed, a series of vacancy formation reactions and a reaction equilibrium constant K are expressed in the formula below.
If VM'' overwhelmingly outnumbers other defect concentrations and its porosity is low enough to use Henry’s law, formula (1-15) can be expressed as follows by using concentration C:
Due to electrical neutrality, 2 [CVM''] = [Ch･] is established.
Therefore, formula (1-15) can be changed as follows:
As shown above, the vacancy concentration depends on the oxygen partial pressure. If metal vacancies are dominant as defects, and if the diffusion of metal ions in the oxide scale is caused by a vacancy mechanism (diffusion due to position exchanges between vacancies and adjacent atoms), the self-diffusion coefficient of the metal is proportional to the vacancy concentration.
By plugging formulas (1-17) and (1-18) into formula (1-12), formula (1-19) is obtained.
In general, since PO''2, is much greater than PO'2 formula (1-19) can be simplified as follows:
As shown above, if the substance that moves in the oxide scale is a metal, and if a divalent metal vacancy is formed, the corrosion rate is proportional to the 1/6 power of the atmospheric oxygen partial pressure.
If the oxide scale is an n-type semiconductor, the formula for the formation of a divalent oxygen vacancy is expressed in the following formula:
If the above formula is solved as well as a p-type semiconductor, the following formulas are obtained.
As shown above, when oxygen moves in a scale, it depends on the oxygen partial pressure on the interface between the metal and the oxide scale.
If a univalent vacancy is formed, coefficient 6 changes to 4 for both the p- and n-types. Typical p-types are Cr2O3, FeCr2O4, FeO, Fe3O4, and NiO, and typical n-types are Fe2O3, Al2O3 and SiO2. Formulas (1-20) and (1-23) mean that if the oxide scale is a p-type (e.g. Cr2O3 for stainless steel), reducing the oxygen partial pressure of the atmosphere can reduce the corrosion rate, depending on the 1/4 to 1/6 power of the oxygen partial pressure. On the other hand, the formulas mean that when SiO2 or Al2O3 is the protective scale, the corrosion rate does not depend on the oxygen partial pressure of the atmosphere, except in a case where the oxygen partial pressure is extremely low.
3.2.3 Temperature
Generally, a diffusion coefficient is expressed in the Arrhenius equation.
In the above formula, Q indicates activation energy. In formulas (1-12) and (1-13), kp is considered to be proportional to the diffusion coefficient and the oxygen partial pressure gradient, the following formula can be obtained to express the dependence on the temperature.
It is easy to imagine that the corrosion rate rises as the temperature rises. During high temperature oxidation, the corrosion rate increases as the diffusion rate in the oxide scale increases. Because of the compositions of the scale, reaction mechanism, etc. change due to a temperature change, formula (1-25) is often not established. However, the formula is one of the means to judge how the corrosion rate depends on the temperature. If an examination of the dependence on the temperature using the Arrhenius equation does not prove the correlation between the corrosion rate and the temperature, it is highly likely that the reaction mechanism changed under the temperature conditions. The Arrhenius equation is very helpful for studying the corrosion mechanism.
4. High temperature oxidation of practical alloys
4.1 Composition of the oxide scale
Section 3.2 described that Al2O3, SiO2, Cr2O3, etc. function as protective scales. On the other hand, the Ellingham diagram in Fig. 1-2 indicates that metals such as Al, Si, and Cr have high compatibility with oxygen and are easily oxidized. This section describes a corrosion protection mechanism of these metals that are easily oxidized.
Figure 1-7 is a schematic diagram of the oxide scale of an alloy (Fe-Cr alloy) 4). When a Fe-Cr alloy is exposed in a high-temperature oxidative atmosphere, corrosion progresses forming an oxide scale. (1) At an early stage of the corrosion, both Cr2O3 and a Fe oxide are formed. Each oxygen partial pressure on the interface (point A in the figure) between the alloy and the Fe oxide and on the interface (point B in the figure) between the alloy and the Cr2O3 is the same as its own dissociation pressure. (2) Fe ions diffuse faster than Cr ions an cover Cr2O3 particles. Meanwhile, the dissociation pressure of the Cr2O3 is lower than that of the Fe oxide.In other words, even in the inside of the Fe oxide where the Fe metal is stable, the Cr oxide has a stable phase. Near the alloy surface, dissolved oxygen oxidize Cr inside the alloy. (3) When this internal oxide layer grows to a continuous layer, the oxygen partial pressure on the interface between the alloy and the Cr2O3 decreases to the dissociation pressure of the Cr2O3, suppressing subsequent internal oxidation. It is known that when pure iron is oxidized at around 1000 °C, the scale has a three layer structure of FeO, Fe3O4, and Fe2O3 viewed from the metal side. Adding Cr to the scale reduces the scale thickness and changes the scale composition. Then, complex oxides such as FeCr2O4 are formed, the increased Cr thickens the FeCr2O4 layer, and the FeO layer gradually disappears. When the ratio of Cr increases and exceeds 18%, a layered scale containing Cr2O3 is formed. And if the ratio exceeds 23%, a monolayer of Cr2O3 is formed, suppressing the corrosion rate drastically5).Figure 1-8 shows the results of an EPMA (electron probe microanalyzer) analysis on the cross sections of the test specimens (austenitic stainless steels [Fe-Cr-Ni]) which were corroded in an atmosphere at 1000 °C for 200 hours. In the Type304 (18Cr-8Ni), an Fe oxide layer is formed in the outermost layer (gas side). Under the layer, there is another oxide layer consisting of Fe, Ni, and Cr. Under the second layer, there are cracks made during cooling. Furthermore, as the innermost layer, the oxide layer that is mainly comprised of Cr containing Fe and Ni is formed. In the entire specimen, the corrosion scale grew with a thickness of 100μm. On the other hand, in the Type310 (25Cr-20Ni), a Cr2O3 layer consists mostly of the corrosion scale and the thickness of the scale is around 20 μm, which indicates that the amount of corrosion was greatly decreased. The results show that the corrosion scale and corrosion resistance greatly vary depending on the Cr content.
Thus, when the oxide of a metal having affinity with oxygen is formed on the alloy surface, the subsequent oxidation is inhibited. Then, the slow diffusion of oxidation of the formed oxide scale inhibits corrosion. Accordingly, in order to make the oxide scale work as a protective scale, the metal oxide must have a stronger affinity with oxygen than that of the base metal. In addition, the diffusion coefficient in the oxide must be small. Typical metals having the above properties are Al, Si, and Cr. In terms of high-temperature strength, workability, etc., Fe, Ni, Co, etc. are used as a base metal for heat resistant alloys. To maintain their heat resistance, protective oxide scale forming elements such as Cr, Si, and Al are added to the base metal. Additionally, Mo and W, typical elements to enhance high-temperature strength are added, and rare earth elements are added to inhibit the detachment of corrosion scales. Thus, designs for heat-resisting alloys include high-temperature strength and corrosion resistance.
Fig. 1-7 Schematic diagram of the Fe-Cr alloy oxide scale
Fig. 1-8 SEM photos of and results of EPMA analysis on the sections of the test specimens which were corroded
in high temperature (in the atmosphere whose temperature is 1 000 °C, for 200 hours)
4.2 Stability of oxide scales
As stated in chapter 3, when a protective oxide scale is formed on an alloy, the corrosion rate decreases according to the parabolic rate law. Conversely, for corrosion prevention, it is necessary to select the materials that form protective oxide scales in the operating environment. In the case of stainless steel and the like, a steel largely containing Cr is selected according to the severity of the operating environment. However, forming a protective scale is not sufficient, and what is important is that the scale is not only not destroyed but is also firmly maintained. Fig. 1-4 schematically shows changes in the amount of corrosion caused by the scale destruction. When the scale is destroyed, the corrosion rate drastically increases at an early stage of the parabolic rate law.
If scale detachment regularly occurs, the corrosion may progress linearly as time passes, as if it followed the linear law. Figure 1-9 shows the results of the measurement of the thickness reduction of the superheater tubes in incinerators, as examples of changes over time in the thickness reduction of actual machines6). In the long-term tendency, the thickness reduction increases linearly with time. In the actual environments, the corrosion caused by this scale destruction is often a problem.
Fig. 1-9 Results of the measurement of the thickness reduction of the superheater tubes in incinerators
One of the main causes of scale destruction is thermal stress. Table 1-17), 8) shows the thermal expansion coefficients of typical metals and oxides.
Since there is a difference in the thermal expansion coefficient between oxides and metals, temperature fluctuations cause oxide scales to be destroyed. A typical cause of temperature fluctuation is start and stop of equipment. Repeating these operations increases the risk of scale destruction. The larger the difference in the thermal expansion coefficient between the base material and the oxide, the more chances of scale destruction. Therefore, in the environment in which Cr2O3 or Fe2O3 is generated, austenitic stainless steels whose thermal expansion coefficient is larger than that of ferritic stainless steels are more likely to have scale destruction due to thermal stress.Other factors such as temperature fluctuation during operation9) and soot blow10) are the thermal stress sources which promote corrosion. Erosion is also an important factor that causes scale destruction11). The continuous scale destruction caused by a physical factor such as the particles contained in combustion gases or the heat transfer tube in the fluidized bed also increases thickness reduction. Regarding the time dependence of the thickness reduction shown in Fig. 1-9, destruction and reproduction of scales were repeated by a physical factor such as start and stop or soot blow for removing attached ashes. It is considered that the linear increase of thickness reduction was shown as a result.
Table 1-1 Thermal expansion coefficients of metals and oxides
| Item |
Thermal expansion coefficient（10−6K−1） |
Temperature range（℃） |
| Fe |
14.6 |
800 |
| Ni |
16.3 |
900 |
| Cr |
9.4 |
700 |
| Al |
26.5 |
400 |
| Si |
7.6 |
0−100 |
| Type304 |
18.8 |
0−648 |
| Type410 |
11.7 |
0−648 |
| Type430 |
11.9 |
0−648 |
| FeO |
12.2 |
100−1000 |
| Fe3O4
|
16.6 |
25−1000 |
| Fe2O3
|
12.5 |
25−1000 |
| NiO |
17.1 |
25−1000 |
| Cr2O3
|
8.7 |
25−1200 |
| Al2O3
|
8.1 |
25−1200 |
| SiO2
|
3 |
300−1100 |
Because the volume of an oxide is different from that of the base metal, stress is generated in an oxide scale as the scale grows, resulting in conditions for scale detachment. The difference in volume is called the Pilling-Bedworth ratio (PBR) and expressed in the following formula:
Table 1-212), 13) shows the PRBs of typical metals. If the PBR is 1 or smaller, tensile stress is applied to the oxide scale. Therefore, the scale is easily destroyed, resulting in poor protection. However, since most metals increase their volumes by combining with oxygen, their PBRs exceed 1 and compressive stress is generated. PBRs are guidelines of the occurrence of stress in oxide scales, and are particularly important if oxygen moves in the oxide scale and causes a reaction on the interface between the metal and the oxide scale.On the other hand, if a reaction occurs on the interface between the oxide scale and the atmospheric gas, the reaction progresses on free surfaces. Therefore, stress due to the volume difference is unlikely to be generated. However, actually, in the case of Cr2O3 in which a reaction occurs on the interface between the oxide scale and the atmospheric gas, stress is applied to the scale. One of the probable causes is that some oxygen existing in grain boundaries, etc. enters into the scale, resulting in a growth of oxides14). In addition to this cause, various factors come into play. Her important factors are that (as PBRs indicate) stress is potentially generated in a metal and an oxide scale, and that an application of a physical factor to the metal and oxide scale easily causes scale detachment. Since most of the past high temperature corrosion studies are about scle destruction, it is a very critical problem. To prevent high temperature corrosion in actual environments, take measures such as avoiding temperature changes and easing the erosion conditions. In short, it is also important to pay attention to the operating conditions, for example, for easing scale destruction factors.
Table 1-2 Pilling-Bedworth ratios
| Item |
PBR |
| K-K2O |
0.45 |
| Li-LiO2
|
0.58 |
| Mg-MgO |
0.81 |
| Na-Na2O |
0.97 |
| Cd-CdO |
1.21 |
| Al-Al2O3
|
1.28 |
| Zn-ZnO |
1.55 |
| Cu-Cu2O |
1.64 |
| Ni-NiO |
1.65 |
| Fe-FeO |
1.68 |
| Ti-TiO2
|
1.73 |
| Co-CoO |
1.86 |
| Cr-Cr2O3
|
2.07 |
| αFe-Fe3O4
|
2.1 |
| αFe-Fe2O3
|
2.14 |
| Ta-Ta2O5
|
2.5 |
| Nb-Nb2O5
|
2.68 |
| V-V2O5
|
3.19 |
5. Corrosion in a real environment
5.1 Corrosion in the system which contains volatile components
It is known that Cr2O3 which is protective oxide partly changes to CrO3 and volatilizes at a high temperature of over 1 000 °C. The following formula, which is a modified formula based on formula (1-5), is proposed as a general formula for calculating the corrosion rate in the system containing such volatile components15).
ks is the constant that expresses the volatilization rate. The following is how to interpret this formula. If the growth rate of the scale is large enough at an early stage of the corrosion (k'p>>ks), the scale growth almost complies with the parabolic rate law. However, as time passes, the growth rate of the scale decreases, and the influence of volatilization cannot be ignored. Finally, the growth rate of the scale caused by corrosion is equal to the disappearance rate caused by volatilization, which causes the scale growth to stop (dx/dt = 0) and maintains the constant thickness of the scale (scale thickness limit ; x0).
Plug formula (1-28) into formula (1-5) to calculate the kinetics of the scale growth in the condition where the scale thickness is x0. The result is ks. This means that after the scale thickness reaches the limit, the corrosion progresses at constant speed ks.
5.2 Corrosion in compound gas
Actual environments may contain not a single gas atmosphere but a c ompound g as one. F or e xample, fossil fuel furnaces contain not only oxygen, but also sulfur gases such as H2S and SOx, and waste incinerators contain a minute amount of chloride gas components such as HCl. In some cases, sulfur and chlorine concentrate on the alloy surface, promoting corrosion. A corrosion behavior in such a compound gas atmosphere can be explained with the thermodynamic equilibrium diagram. Figure 1-10 i s t he equilibrium diagram o f t he F e-O-S s ystem at 5 00 ° C, in which thermodynamic database MALT2 was used for the calculation. The figure shows that the stable phase changes depending on the O2 and SO2 partial pressures.For example, Fe2O3 h as a s table p hase i n t he atmosphere in which the O2 and SO2 partial pressures are 1% and 1 ppm, respectively. When the oxygen partial pressure decreases, the stable phase is changed to Fe3O4. When the pressure decreases further, the FeS stable area appears . I f Fe is corroded in this atmosphere, Fe2O3 is formed in the outermost layer, but the partial pressure decreases in the scale. Therefore, it is presumed that Fe3O4 is formed in the scale, and the innermost layer is an area containing both FeS and Fe3O4. As for corrosion in such mixed gases, note that even if the amount of the contained gases are very small and they are in an oxide stable area in the atmosphere, a tiny amount of gas compounds may be formed in the scale, resulting in a large influence on the corrosion. In particular, compounds such as sulfide and chloride are less dense and adhesive than oxides and thus cannot be satisfactory protective scales. The formation of an oxide layer t hrough such a compound layer m ay i nhibit t he formation o f a protective oxide layer or promote scale detachment.
Fig. 1-10 Equilibrium diagram of the Fe-O-S system, and schematic diagram of an oxide scale
Especially, in the case of chloride corrosion, it was reported that chloride is formed in the innermost layer of corrosion and the following recycle reactions promote corrosion16), 17).
Chloride stable area (innermost layer of the scale) ;Me＋Cl2→MeCl2
Oxide stable area (gas side of the scale) ;MeCl2＋1/2O2→MeO＋Cl2
Although the atmosphere is an oxide stable area, chlorides are formed in the scale. Part of the formed chloride is oxidized in the oxide stable area, which generates Cl2. When the generated Cl2 affects the corrosion again, the corrosion is accelerated. On the other hand, it is known that the use of a volatilization phenomenon inhibits corrosion in mixed gases. The following case is reported: When Al is added to an alloy in the environment which contains a tiny amount of chloride gas, even if a chloride enters the alloy, AlCl3 with high vapor pressure is formed, preventing the accumulation of chlorides in the alloy, and inhibiting corrosion18). As shown above, complicated corrosion behaviors can be seen particularly in mixed gases. Accordingly, the corrosion prevention based on fully understanding of a corrosion mechanism is required.
References
1) L. S. Darken and R. W. Gurry: Physical Chemistry of Metal, McGraw Hill, p.342 (1953).
2) C. Wagner: Z. Phys. Chem., 21, 25 (1933).
3) Keizo Nishida and Toshio Narita: P.44, Introduction to high temperature oxidation of metals (published by Maruzen)(1990).
4) Japan Society of Corrosion Engineering: P.65, High Temperature Oxidation and High Temperature Corrosion of Metallic Materials (published by Maruzen) (1982).
5) Yasutoshi Saito, Toru Atake, and Toshio Maruyama: P.102, High Temperature Oxidation of Metals (published by Uchida Rokakuho) (1991).
6) Makoto Shizuma, and others: Boiler System for Highefficiency Power Generation Utilizing Heat from MSW Incinerators, Ebara Engineering Review, No. 186, 19 (2000).
7) Japan Society of Corrosion Engineering: P.81, High Temperature Oxidation and High Temperature Corrosion of Metallic Materials (published by Maruzen) (1982).
8) The Iron and Steel Institute of Japan: PP158 and 162, Tekko-Binran 3rd edition, Metal Materials-Tests and Analyses, Vol.6 (1981) (published by Maruzen).
9) Yuzo Kawahara, Masaharu Kira, and Minoru Ike: P75, Development of High Performance Corrosion Test Method under Thermal Gradient in Waste Incineration Environment:2000 Vol.39 No.1, Materia Japan.
10) M. Noguchi, H. Yakuwa, M. Miyasaka, M. Yokono, A. Matsumoto, K. Miyoshi, K. Kosaka, and Y. Fukuda:Experience of superheater tubes in municipal waste incineration plant, Materials and Corrosion, 51 [11], 774 (2000).
11) M. Noguchi, H. Yakuwa, M. Miyasaka, H. Sakamoto, S. Kosugi and T. Narita: High temperature erosion-corrosion behavior of boiler tube materials in fluidized-bed waste incinerator conditions, Proceedings of the International Symposium on High-Temperature Oxidation and Corrosion 2000, Sapporo, Japan, 573 (2000).
12) P. Hancock, R. C. Hurst: The mechanical properties and breakdown of surface oxide films at elevated temperatures, in: Advances in Corrosion Science and Technology, ed. R. W. Staehle and M. G. Fontana, Plenum Press, New York (1974).
13) Rikio Nemoto: P62, Nippon Yakin technical report No.5, Basics of High Temperature Corrosion (1996).
14) Kazuyoshi Nii: (4, 49) “High Temperature Oxidation: What are identified and not identified” published by Sumitomo Metal Mining Co., Ltd. (1997).
15) C. S. Tedmon, Jr.: The Effect of Oxide Volatilization on the Oxidation Kinetics of Cr and Fe-Cr Alloys, J. Electrochem. Soc., 133, 766 (1966).
16) Yoshiyuki Sato, Motoi Hara, Yutaka Shinada, and Toshio Narita: P841, Vol. 60 (1996), J. Japan Inst. Metals,“ Effect of a Small Amount of Hydrogen Chloride on High Temperature Oxidation of Iron-Chromium Alloys”.
17) Manabu Noguchi, Kei Matsuoka, Hiroyuki Fujimura, Yoshiyuki Sawada, and Shigeki Ueta: P75, Zairyo-to-Kankyo 51 (2002), The Corrosion Mechanism of Ni Base Cast Alloy in High Temperature Combustion Gas Environment of a Gasification and Ash Melting System.
18) Manabu Noguchi, Kei Matsuoka, Hideyuki Sakamoto, Yoshiyuki Sawada, and Shigeki Ueta: P218, Zairyo-to-Kankyo, 54 (2005), Effect of Alloying Elements for Corrosion Resistance of Ni-base Cast Alloys in Chlorine-containing High Temperature Environments.