Parabolic rate law derivation:
  Scale growth rate: dx/dt = k' / x   (growth rate inversely proportional to thickness)

  Separating variables and integrating:
    x·dx = k'·dt
    ∫₀ˣ x·dx = ∫₀ᵗ k'·dt
    x² / 2 = k'·t
    x² = 2k'·t = K_p·t    (K_p = parabolic rate constant for thickness)

  In practice, oxidation measured by MASS GAIN per unit area (Δm, mg/cm²):
    Δm² = k_p · t         [k_p in mg²/cm⁴·s or g²/cm⁴·s]

  Instantaneous oxidation rate (differentiating):
    d(Δm)/dt = k_p / (2·Δm)   → rate decreases as scale thickens

  Arrhenius temperature dependence of k_p:
    k_p = A · exp(−Q / RT)

    A = pre-exponential factor (material and gas-composition dependent)
    Q = apparent activation energy (kJ/mol) — identifies rate-limiting step:
        Q ≈ 120–160 kJ/mol → O²⁻ diffusion (inward; anion transport)
        Q ≈ 200–250 kJ/mol → cation diffusion (M²⁺ outward; cation transport)
        Q ≈ 300–350 kJ/mol → grain boundary diffusion in Al₂O₃

  Converting mass gain to oxide thickness (using oxide density ρ_ox):
    Scale thickness x (µm) = Δm (mg/cm²) × M_ox / (n × M_O × ρ_ox × 10)
    where n = number of O atoms per oxide formula unit
    For Al₂O₃: M_ox=102, M_O=16, ρ_ox=3.99 g/cm³, n=3:
    x = Δm × 102 / (3 × 16 × 3.99 × 10) = Δm × 0.534  [µm per mg/cm²]

Linear rate law:
  d(Δm)/dt = k_l   (constant)   →   Δm = k_l · t

  Conditions producing linear kinetics:
    (a) Porous scale: FeO (wüstite) above 570°C — pores allow rapid gas access
    (b) Volatile oxide: MoO₃ (bp 795°C) — oxide evaporates as fast as it forms
    (c) Liquid oxide: V₂O₅ (mp 675°C), liquid at service temperature → runs off
    (d) Mechanically unstable: scale with PBR >> 2 spalls continuously
    (e) Very thin non-adherent scale: no barrier to diffusion

  Examples of linear oxidation rate constants at 700°C:
    Fe (wüstite layer regime): k_l ≈ 0.5 mg/cm²·h
    Mo (volatile MoO₃):        k_l ≈ 2.0 mg/cm²·h  (accelerates with temp)
    W  (volatile WO₃ at >750°C): k_l ≈ rapid and catastrophic
    Nb (loose Nb₂O₅):          k_l ≈ 0.8 mg/cm²·h

  Engineering consequence: LINEAR oxidation means unlimited metal loss.
  A steel component corroding at 0.5 mg/cm²·h at 700°C consumes:
    ~0.3 mm of section thickness per 1000 hours of operation.
  This is why unprotected carbon steel cannot be used above ~570°C.

Pilling-Bedworth Ratio definition:
  PBR = (Molar volume of oxide) / (n × Molar volume of metal consumed)
      = (M_ox / ρ_ox) / (n × M_M / ρ_M)
      = (M_ox × ρ_M) / (n × M_M × ρ_ox)

  M_ox = molar mass of oxide (g/mol)
  M_M  = atomic mass of metal (g/mol)
  ρ_ox = density of oxide (g/cm³)
  ρ_M  = density of metal (g/cm³)
  n    = number of metal atoms per oxide formula unit

  Worked example — Aluminium / Al₂O₃:
    M_ox = 101.96 g/mol;  ρ_ox = 3.99 g/cm³ (corundum)
    M_M  = 26.98 g/mol;   ρ_M  = 2.70 g/cm³
    n    = 2  (two Al atoms per Al₂O₃ formula unit)
    PBR  = (101.96 × 2.70) / (2 × 26.98 × 3.99) = 275.3 / 215.4 = 1.28 ✓ protective

  Interpretation:
    PBR < 1:   Oxide volume < metal volume consumed → porous scale → non-protective
               Examples: MgO (0.81), Li₂O (0.59), Na₂O (0.57)
    PBR 1–2:   Oxide in compressive stress → continuous, adherent scale → protective
               Examples: Al₂O₃ (1.28), NiO (1.65), Cr₂O₃ (2.07, borderline)
    PBR > 2:   Excessive compressive growth stress → buckling, cracking, spallation
               Examples: WO₃ (3.35), Nb₂O₅ (2.68), VO₂ (3.19)

  PBR for iron oxide system (three layers at > 570°C):
    FeO:    PBR = (71.85 × 7.87) / (1 × 55.85 × 5.74) = 1.77  (porous despite PBR!)
    Fe₃O₄: PBR = (231.5 × 7.87) / (3 × 55.85 × 5.16) = 2.10
    Fe₂O₃: PBR = (159.7 × 7.87) / (2 × 55.85 × 5.24) = 2.14
    → Iron scale is destructive NOT because of PBR but because FeO is porous
       (stoichiometric defects: excess Fe vacancies make FeO a leaky lattice)

Oxide defect types (Kröger-Vink notation):

  p-TYPE oxides (metal-deficient; metal vacancies dominant):
    Example: NiO, Cr₂O₃, FeO, CoO
    Defect reaction: Ni(s) + ½O₂(g) → Ni_Ni + V''_Ni + O_O
                                       (cation vacancy V''_Ni forms; acceptor)
    In p-type oxides:
      · Cation vacancies are majority defects
      · Cations diffuse OUTWARD (from metal to oxide surface) via vacancy hopping
      · Scale grows at the oxide-gas interface (new oxide forms at outer surface)
      · Increasing pO₂ → more vacancies → faster oxidation
      · Doping with higher-valent cations (e.g., Cr³⁺ in NiO) reduces V''_Ni
        → REDUCES k_p → basis for chromium alloying in nickel

  n-TYPE oxides (oxygen-deficient; oxygen vacancies or interstitial cations):
    Example: ZnO, TiO₂ (low pO₂), Al₂O₃ (debated; likely mixed)
    Defect reaction: O_O → ½O₂(g) + V··_O + 2e'
                          (oxygen vacancy V··_O forms; donor)
    In n-type oxides:
      · Anion vacancies or interstitial cations are dominant
      · Oxygen diffuses INWARD (via vacancy hopping)
      · Scale grows at the metal-oxide interface
      · Reducing pO₂ → more oxygen vacancies → faster oxidation (opposite to p-type!)
      · Doping with lower-valent cations increases V··_O → increases oxidation rate

Wagner's parabolic rate constant:
  k_p = (RT / F²) × ∫[μ'_O₂ to μ''_O₂] (σ_M · σ_e) / (σ_M + σ_e) d(μ_O₂)

  σ_M = ionic (metal cation or oxide anion) conductivity
  σ_e = electronic conductivity
  Integration from metal-oxide interface (low μ_O₂) to oxide-gas interface (high μ_O₂)

  For most engineering oxides, σ_e >> σ_M (good electronic conductors):
    k_p ≈ (RT/F²) × ∫ σ_M d(μ_O₂)

  This shows: k_p ∝ concentration of mobile ionic species (defects)
  → Doping to REDUCE defect concentration → REDUCES k_p → design principle

Ellingham diagram: ΔG° vs T for oxide formation reactions (per mole O₂):
  More negative ΔG° at a given T → oxide is more stable → forms preferentially

  Selected values at 1000°C (1273 K) in kJ/mol O₂:
    SiO₂:   ΔG° ≈ −670 kJ/mol O₂    (most stable practical oxide)
    Al₂O₃:  ΔG° ≈ −840 kJ/mol O₂    (extremely stable; thermodynamically favoured)
    Cr₂O₃:  ΔG° ≈ −630 kJ/mol O₂    (stable; selective oxidation above 15% Cr)
    FeO:     ΔG° ≈ −400 kJ/mol O₂
    NiO:     ΔG° ≈ −350 kJ/mol O₂
    CoO:     ΔG° ≈ −370 kJ/mol O₂
    WO₃:     ΔG° ≈ −450 kJ/mol O₂   (moderately stable but volatile above 800°C)
    MoO₃:   ΔG° ≈ −440 kJ/mol O₂   (volatile; MoO₃ sublimes at 795°C)

  Selective oxidation condition (binary A-B alloy where oxide B_xO is more stable):
    For selective formation of B_xO that covers the alloy surface:
    The activity of B at the alloy surface must be sufficient to maintain the
    equilibrium oxygen partial pressure at the metal-oxide interface:

    pO₂,int = exp(ΔG°_B_xO / RT) / a_B^x

    If pO₂,int < pO₂,bulk → B oxidises selectively
    If B diffusion in the alloy cannot supply the surface fast enough →
       internal oxidation of B (precipitates within the alloy rather than continuous surface scale)

  Critical aluminium content for external Al₂O₃ scale formation:
    In Ni-Al alloys (diffusion coefficient D_Al in Ni ≈ 3×10⁻¹⁴ m²/s at 1000°C):
    N*_Al ≈ √(π·N_O · D_O / (3·D_Al)) ≈ 4–7 at% Al for external scale
    → MCrAlY bond coats require 10–12% Al to ensure reliable α-Al₂O₃ formation

Quantitative effect of Y on Al₂O₃ growth rate:
  k_p (FeCrAl, no RE) at 1000°C:   ~1 × 10⁻¹⁶ g²/cm⁴·s
  k_p (FeCrAl + 0.1%Y) at 1000°C:  ~2 × 10⁻¹⁷ g²/cm⁴·s  (5× reduction)
  k_p (MCrAlY, optimised) at 1000°C: ~3–8 × 10⁻¹⁸ g²/cm⁴·s (10–30× reduction)

  Practical consequence at 1000°C over 10,000 hours:
    Without RE: Δm ≈ √(1×10⁻¹⁶ × 3.6×10⁷) = 0.60 mg/cm² → TGO ~0.32 µm
    With 0.1%Y: Δm ≈ √(3×10⁻¹⁷ × 3.6×10⁷) = 0.10 mg/cm² → TGO ~0.053 µm

    The Y-containing alloy reaches the critical TGO thickness (5–8 µm)
    approximately 10–30× more slowly → component life extended by same factor

Optimal RE concentration:
  Too little (<0.02 wt% Y): insufficient grain boundary saturation — limited benefit
  Optimal (0.05–0.2 wt% Y): maximum benefit; fine, adherent scale
  Too much (>0.5 wt% Y): Y₂O₃ particles form continuous phase → embrittlement;
                           "overdoping" can INCREASE oxidation rate (fast diffusion
                           through Y-rich grain boundary phase)

Chromia volatility above 1050°C:
  2Cr₂O₃(s) + 3O₂(g) → 4CrO₃(g)    ΔG° becomes negative above ~1060°C at 0.21 bar O₂
  K_eq = pCrO₃^4 / (pO₂^3) → increases sharply with temperature

  CrO₃ vapour pressure at pO₂ = 0.21 bar:
    At  900°C: p(CrO₃) ≈ 10⁻⁶ bar → negligible loss
    At 1050°C: p(CrO₃) ≈ 10⁻⁴ bar → begins to matter in flowing gas
    At 1200°C: p(CrO₃) ≈ 10⁻² bar → catastrophic volatility; linear kinetics

  Consequence: all gas turbine hot-section components (>1050°C) must use
               alumina-forming alloys or MCrAlY coatings — NOT chromia formers.
               Chromia formers (310SS, Alloy 600, Hastelloy X) are limited to
               furnace hardware, boiler superheater tubes, and exhaust systems
               operating below 1050°C.

  Wet oxidation (steam environments):
    CrOOH(g) = H₂O + CrO₂ → effective Cr loss accelerated in high steam atmospheres
    This is why 9Cr creep-resistant steels require protection by thin Al₂O₃-forming
    overlay coatings in advanced ultra-supercritical (A-USC) steam turbines above 700°C.

Type I hot corrosion (Na₂SO₄-induced, ~900°C):

  Source of deposit:
    NaCl (sea salt, airborne) + SO₃ (from fuel combustion) → Na₂SO₄ (mp 884°C)
    At component surface ~900°C: Na₂SO₄ is LIQUID → forms continuous melt film

  Two-stage mechanism:
  Stage 1 — Initiation:
    Na₂SO₄ melt contacts protective Cr₂O₃ or Al₂O₃ scale
    Basic fluxing:  Cr₂O₃ + Na₂SO₄ → Na₂CrO₄ + SO₃ (gas)
                    Al₂O₃ + Na₂SO₄ → NaAlO₂ + SO₃
    The protective oxide dissolves into the melt as chromate/aluminate ions
    → scale is destroyed; metal surface re-exposed

  Stage 2 — Propagation (autocatalytic):
    CrO₄²⁻ in the Na₂SO₄ melt migrates outward → precipitates as Cr₂O₃ at melt surface
    S²⁻ released inward → reacts with metal → sulfide layer at metal surface
    Metal sulfides oxidise at the sulfide-melt interface → regenerate SO₃
    SO₃ dissolves back into melt → self-sustaining cycle

  Kinetics:
    Initiation period: 10–100 hours (depends on salt flux rate, temperature)
    Propagation: CATASTROPHICALLY FAST — similar to linear kinetics
    Component life reduced from thousands of hours to tens of hours

  Critical threshold:
    Salt deposition rate > ~0.1 µg/cm²·h → Type I initiates
    Temperature must be above Na₂SO₄ melting point (884°C) and below ~1000°C
    (above 1000°C: Na₂SO₄ vapour pressure too high for stable melt deposit)

  Affected alloys (in order of decreasing Type I susceptibility):
    Highly susceptible: Co-base alloys (Stellite, MAR-M509)
    Moderately susceptible: Ni-base alloys without MCrAlY coating (IN738, IN792)
    Resistant: alloys with >35% Cr (but Cr level incompatible with high strength)
    Best resistance: MCrAlY coatings (Al₂O₃ scale resists basic fluxing better than Cr₂O₃)

Type II hot corrosion (mixed sulphate eutectic, 650–750°C):
  At 650–750°C, pure Na₂SO₄ is solid (mp 884°C) → no liquid deposit → no Type I
  However: SO₃ from combustion gas reacts with Ni and Co in the alloy:
    Ni + SO₃ → NiSO₄  (mp 848°C — solid alone)
    Co + SO₃ → CoSO₄  (mp 735°C — solid alone)

  Eutectic formation:
    Na₂SO₄ – NiSO₄ eutectic: mp = 671°C → LIQUID at 700°C service temperature!
    Na₂SO₄ – CoSO₄ eutectic: mp = 565°C → liquid even at cooler components

  Type II mechanism:
    · High-SO₃ partial pressure environment required (pSO₃ > 10⁻³ bar)
    · Mixed eutectic salt melt dissolves protective Cr₂O₃ by acidic fluxing:
      Cr₂O₃ + 3SO₃ → 2CrO·SO₄ (chromyl sulphate — highly corrosive)
    · Pitting morphology characteristic of Type II: discrete pits rather than
      uniform attack (unlike Type I which is more uniform)

  Affected components:
    · Low-pressure turbine blades (cooler than HPT blades → Type II temperature range)
    · Industrial gas turbines burning heavy fuel oil (high SO₃ content)
    · Marine turbines in high-NaCl environments

  Prevention:
    · Fuel desulphurisation (reduce SO₃ partial pressure)
    · MCrAlY coatings (Al₂O₃ scale is more resistant to acidic than basic fluxing)
    · Platinum-aluminide coatings (Pt modifies Al activity; better salt resistance)
    · High-Co content alloys are MORE susceptible to Type II (CoSO₄ eutectic)

Sulfidation kinetics comparison at 700°C:
  Parabolic rate constant k_p for sulphide formation:
    FeS on iron:   k_p ≈ 10⁻⁸ g²/cm⁴·s  (10⁶× faster than Cr₂O₃ oxidation!)
    Ni₃S₂ on Ni:  k_p ≈ 10⁻⁷ g²/cm⁴·s  (catastrophic)
    Cr₂S₃ on Cr:  k_p ≈ 10⁻¹² g²/cm⁴·s  (still 100× faster than Cr₂O₃ oxidation)

  Critical eutectic melting points:
    Ni  – Ni₃S₂:  637°C  → liquid in many refinery service conditions!
    Fe  – FeS:    988°C  → liquid in furnace environments
    Co  – Co₄S₃:  880°C  → liquid in hot turbine hot corrosion + sulfidation

  Stability diagram (log pS₂ vs log pO₂) at 700°C:
    Region A (high pO₂, low pS₂): protective oxide stable
    Region B (low pO₂, high pS₂): sulphide stable → catastrophic attack
    Region C (intermediate): mixed oxide + sulphide → complex behaviour

  Alloy selection for mixed environments:
    Best resistance: alloys with high Cr (>25%) + high Si (>1.5%)
    FeCrAl alloys: Al₂O₃ more resistant to sulfidation than Cr₂O₃
    Ni-base superalloys: avoid — Ni₃S₂ eutectic is too low
    Cobalt alloys: moderate — Co₄S₃ eutectic higher than Ni₃S₂
    Ferritic 310 stainless (25Cr-20Ni): reasonable in refinery service to 650°C