Semiconductivity of steam generator tubing alloys in simulated crevice chemistries containing lead and sulphur

Materials

The test materials were Alloys 800 (from Sandvik, heat no. 516809) and 690 (UNS N066900). The elemental compositions of these alloys are listed in Table 1. The outside surface of each specimen was grounded with wet silicon carbide papers (Buehler Ltd) in the following sequence: 320, 600, 800 and 1200 grit, rinsed copiously with deionised water, acetone and ethanol, then dried in air. Specimens then were prepassivated at open circuit potential for 24 h in the SG crevice chemistry. After the prepassivation, the specimens were taken out for further analysis.

OPEN IN VIEWER

Table 1

Chemical compositions of Alloys 800 and 690/wt-%

	C	Si	Mn	P	S	Cr	Ni	Co	Ti	Cu	Al	N	Fe
Alloy 800	0.017	0.46	0.5	0.012	0.001	21.87	32.78	0.01	0.48	0.02	0.29	0.016	43.2
Alloy 690	0.02	0.5	0.5	…	…	30	Balance	…	…	…	…	…	10

The chemical compositions of the test solutions containing Pb and S with different pHs are listed in Tables 2 and 3 respectively. These solutions are designed to simulate the crevice chemistries in SG of Canada Deuterium Uranium system. The pH values of test solutions at 300°C were determined with the software SteamAnalyzer (Version 2.048, OLI Systems Inc. USA). The passivation was performed in separate autoclaves, one for the lead free chemistries and another for the lead contaminated ones.

OPEN IN VIEWER

Table 2

Simulated CANDU SG crevice chemistries*/mol L^{− 1}

ID	Na2SO4	NaCl	KCl	CaCl2	NaOH	NaHSO4	PbO	pH300°C	pH25°C	pHNeutral
Neutral	0.15	0.30	0.05	0.15				6.08	7.02	5.16
	0.15	0.30	0.05	0.15			0.0022	7.94	9.62	5.16
Alkaline	0.15	0.30	0.05	0.15	0.40			(9.73)	(12.91)	5.14
	0.15	0.30	0.05	0.15	0.40		0.0022	(9.72)	(12.90)	5.15
Acidic	0.15	0.30	0.05	0.15		0.05		(3.02)	(1.61)	5.16
	0.15	0.30	0.05	0.15		0.05	0.0022	(3.33)	(1.65)	5.16

*

The ion concentrations were calculated with OLI Systems.

OPEN IN VIEWER

Table 3

Test matrix for degradation of SG tubing materials in reduced sulphur environments

		Chemical composition
Simulated crevice environment	Test ID	Base composition	Sulphur additive(s)	pH25°C
Neutral	R1	…	0.15 mol L− 1 Na2SO4
	R2	…	0.075 mol L− 1 Na2S2O3
	NC0	0.30 mol L− 1 NaCl	…	7.00
	NC1	0.05 mol L− 1 KCl0.15 mol L− 1 CaCl2	0.15 mol L− 1 Na2SO4	7.02
	NC2	0.05 mol L− 1 SiO2*	0.075 mol L− 1 Na2S2O3	6.98
Alkaline	AKC1	0.30 mol L− 1 NaCl	0.15 mol L− 1 Na2SO4	12.91
	AKC2	0.05 mol L− 1 KCl0.15 mol L− 1 CaCl2	0.075 mol L− 1 Na2S2O3	12.90
	AKC3	0.05 mol L− 1 SiO2*0.40 mol L− 1 NaOH	0.075 mol L− 1 Na2SO4+0.0375 mol L− 1 Na2S2O3	12.90
Acidic	AC1	0.30 mol L− 1 NaCl	0.15 mol L− 1 Na2SO4	1.61
	AC2	0.05 mol L− 1 KCl0.15 mol L− 1 CaCl2	0.0375 mol L− 1 Na2S4O6	1.60
	AC3	0.05 mol L− 1 SiO2*0.05 mol L− 1 NaHSO4	0.075 mol L− 1 Na2SO4+0.01875 mol L− 1Na2S4O6	1.60

*

SiO₂ is not soluble at room temperature, and its effect can be ignored at room temperature.

Mott–Schottky measurements

The Mott–Schottky measurements were performed using a Gamry (PC-750) electrochemical workstation. A three-electrode cell was used with the Alloy 800 specimen as the working electrode, a saturated calomel electrode (SCE) and a Ag/AgCl electrode as the reference electrode at 21 and 300°C respectively and a platinum electrode as the counter electrode.

For the Mott–Schottky measurements, the scan was started from E _corr, and the direct current scanning potential range was within the passive region with a scanning rate of 50 mV/step to make sure that no Faraday current is generated. An alternating current signal with a frequency of 1000 Hz and peak to peak magnitude of 10 mV was superimposed on the scanning potential. The equivalent circuit for Mott–Schottky measurements comprised a resistance and a capacitance in series. The former represents the charge transfer resistance, and the latter simulates the effect of the space charge layer in the passive films. The Mott–Schottky theory correlates the reciprocal of the square of the space charge capacitance C _SC of a semiconductor with the electrode potential E under depletion conditions. A semiconducting oxide/electrolyte interface may be considered that the C _SC in the oxide film and the Helmholtz double layer capacitance C _H exist in series. The measured capacitance C of semiconducting oxide films is given by the following relationship: ²²

For n type semiconductors 1 For p type semiconductors 2 where q is the elementary charge of electron (q = 1.602 × 10^{− 19} C), ε is the dielectric constant of the oxide, ε₀ is the vacuum permittivity (8.854 × 10^{− 14} F cm^{− 1}), A is the geometrical surface area, N _D is the donor density, E _FB is the flat band potential, k is the Boltzmann constant (k = 1.38 × 10^{− 23} J K^{− 1}) and T is the absolute temperature. The dielectric constants of the oxides of main alloy elements, namely Fe, Ni and Cr, are in the range 11.9-14.2. ²³ In concentrated solutions, the C _H is much larger than the C _SC (∼50 times), ²⁴ and the measured capacitance C is equal to the space charge capacitance C _SC. Accepting that C _H>C _SC, the N _D can be calculated from the slope of the linear part of the C ^{− 2}–E plots. E _FB can beobtained from the extrapolation to C ^{− 2} = 0.

Semiconductivity in Pb containing chemistries

Mott–Schottky plots were performed for Alloy 690 after passivation at corrosion potential for 24 h in solutions at 300°C. When film formation potentials were 0.2-0.6 V(SCE) (Fig. 1a), ¹¹ the passive film exhibited linear Mott–Schottky behaviour, indicating that the capacitance response of the oxide is controlled by the band bending and can be described by the variation of the space charge capacitance under depletion conditions. The Mott–Schottky plots of Alloy 690 passivated in the alkaline and near neutral chemistries display negative slopes (Fig. 1a and b), suggesting that the anodic films are basically p type semiconductors. In contrast, the slopes of Mott–Schottky plots that are measured from the passive layers formed in the acidic environments, depending on the potential range, can be either positive or negative (Fig. 1c), implying that the anodic films are mixtures of n and p type semiconductors.

OPEN IN VIEWER

1

Effect of pH on Mott–Schottky plots of UNS N06690 prepassivated at 300°C in SG crevice chemistries with and without lead contamination ¹¹ . a in alkaline solutions, b in neutral solutions, c acidic solutions

According to equation (2) and Fig. 1a, the lead contamination in the alkaline solution may give rise to an increase in acceptor density of the anodic film. However, lead contamination in the neutral solution may lead to a decrease in acceptor density of the anodic film (Fig. 1b). In an acidic solution (Fig. 1c), the presence of lead increased densities of both acceptors and donors, but to a less extent compared to the changes observed in Fig. 1a. However, it is noted that the donor or acceptor densities cannot be compared by only comparing the linear slopes in Mott–Schottky plots because the variance in linear slope k may also stem from the dielectric constant of the oxide ε), as given in equation (3) 3 For example, an increase in hydroxide and/or hydrogen concentrations in the anodic films containing lead may also affect the dielectric constant of passive films. Unfortunately, quantitative assessment of the effect of hydroxides on the dielectric constant is not attainable because the dielectric constants of hydroxides are lacking.

Semiconductivity in S containing chemistries

Mott–Schottky plots in sulphate only or thiosulphate only solutions are shown in Fig. 2a, and the negative slopes illustrated that the passive layer was p type semiconductors. In NC0, NC1 and NC2 chemistries, the positive slopes showed that the passive film exhibited n type semiconductor behaviour (Fig. 2b), which can be attributed to the oxygen vacancy in the barrier layer. ²⁵ Since chloride ions adsorb mainly on oxygen vacancies and inhibit oxygen atoms from entering the passive film, the defects in the passive film in NC solutions are oxygen vacancies.^25,26 Even in NC2 containing thiosulphate, the defect type was not changed, indicating that the chloride determined the semiconductor characteristics of the passive film even when 0.075 mol L^{− 1} thiosulphate was added to the chloride solution. It is revealed that the semiconductivity was modified from the n type observed in chloride only solution and in a chloride–thiosulphate solution when the concentration ratio of chloride/thiosulphate was comparatively high, to the p type when the concentration ratio of chloride/thiosulphate was low. At a low chloride/thiosulphate ratio, the adsorption and electrochemical reduction of thiosulphate on the electrode surface determined the semiconductivity, and the adsorption of chloride was inhibited. At high ratios, the adsorption of chloride rendered the passive film easy to break down, and a small amount of thiosulphate could adsorb and reduce at these sites, preventing the healing and repassivation of the passive film. This made the initially metastable pits gradually develop to form stable pits.

OPEN IN VIEWER

2

Effect of pH on Mott–Schottky plots of Alloy 800 prepassivated at corrosion potential at 21°C in SG crevice chemistries with and without sulphur contamination. a,b in neutral solutions, c in alkaline solutions

The slopes in AKC chemistries showed that the passive film exhibited p type semiconductor behaviour (Fig. 2c), which can be attributed to a preponderance of cation vacancies on the cation sublattice of the barrier layer. For p type passive film, holes are present in much higher concentration than electrons. The observed non-linearity in Fig. 2 may reflect a voltage dependent donor concentration, although others have suggested non-uniform doping distributions ²⁷ and/or the presence of deep donor states (levels) that increase with applied potential. ²⁸

The Mott–Schottky plots for Alloy 690 in simulated crevice chemistries with or without sulphur species are shown in Fig. 3. In neutral solutions, the passive film showed p type semiconductors, as shown in Fig. 3a. However, the absolute value of the linear slope in sulphate only solution was lower than that in thiosulphate only solution. In neutral solutions with or without sulphur, the semiconductivity was p type, as shown in Fig. 3b. The presence of thiosulphate decreased the absolute value of the linear slope, possibly due to a change in the film composition or a change in the acceptor densities. In alkaline solutions, the semiconductivity was p type, as shown in Fig. 3c, and the presence of thiosulphate in AKC1 and AKC2 also led to a decrease in the absolute values of slopes.

OPEN IN VIEWER

3

Effect of pH on Mott–Schottky plots of Alloy 690 prepassivated at corrosion potential at 21°C in SG crevice chemistries with and without sulphur contamination. a,b in neutral solutions, c in alkaline solutions

Factors that affect semiconductivity

Film composition

(1) Effect of alloy composition

It is quite clear that the alloy composition would affect the composition of passive films significantly; different alloy types may have different passive layers, and hence, the semiconductivity is also affected. Table 4 lists the semiconductivity of pure metals and alloys in various solutions and temperatures. In brief, the composition and semiconductivity of passive films on pure metals, such as Fe, Ni and Cr, were well documented and summarised. It was generally believed that oxides on Fe are n type semiconductors due to the formation of Fe oxides.^29–31 The composition could be considered as a single layer with γ-Fe₂O₃ (Refs. 29 and 30) or considered as two layers with an inner layer of Fe₃O₄, or γ -Fe₂O₃, or a mixture of the two, and an outer layer of an as yet unidentified Fe(III)oxide/hydroxide. ³¹ Passive films formed on pure Cr were also n type semiconductors with a single Cr(OH)₃ layer or inner CrOOH and outer Cr(OH)₃ layers. ³² However, the passive film formed on pure Ni was always revealed as a p type semiconductor due to the formation of inner NiO and outer Ni(OH)₂ layer.^33–35 For the pure metals, the composition of passive films seemed not dependent on the film formation solutions.

OPEN IN VIEWER

Table 4

Semiconductivity of passive film on metals in various conditions

Materials	Condition	Defect density/cm− 3	Temperature	Semiconductivity	Composition	Reference
Fe	deaeratedbuffer solution (pH 8.5)*	3.8 × 1020 at 30°C,11.7 × 1020 at 50°C,64.8 × 1020 at 80°C	Room temperature,29 30, 50 and 80°C 30	n type	γ-Fe2O3	29, 30
Fe (99.7%)	Deaerated buffer solution (pH 8.4)†	1.6-1.8 × 1020 (inner layer)8.3-12 × 1020 (outer layer)	Ambient temperature (22-23°C)	n type	An inner layer of Fe3O4, or γ -Fe2O3, or a mixture of the two and an outer layer of an as yet unidentified Fe(III)oxide/hydroxide	31
Fe (99.99)	Deaerated buffer solution (pH 8.5)*	2.01 × 1020–4.73 × 1020	Ambient temperature	n type		39
Ni	Deaerated pH 8.5 borate buffer solution	8.8 × 1020	Room temperature	p type	Inner NiO and outer Ni(OH)2	33
Ni	Deaerated buffer solution (pH 8.5)*	1.33 × 1021 (30°C) 1.97 × 1021 (50°C)2.08 × 1021 (80°C)	30, 50 and 80°C	p type	…	34
Ni	Degassed 6 mol L− 1 KOH solution	1.08 × 1021	Room temperature	p type	…	35
Cr	Deaerated buffer solution (pH 8.5)*	1.39 × 1020	Room temperature	n type	A single Cr(OH)3 layer or inner CrOOH and outer Cr(OH)3 layers	32
Fe–20Cr	Deaerated buffer solution (pH 8.5)*	…	Room temperature	n type	Cr substituted γ-Fe2O3	40
Fe–18Cr	Deaerated 0.1 mol L− 1 H2SO4 solution	…	Room temperature	p type	…	36
Fe–18Cr	Deaerated borate buffer solution (pH 8.4)‡	…	Room temperature	n type	…	36
304SS	0.5 mol L− 1 NaCl solution (pH 6)	1.4 × 1021, 1.7 × 1021	Room temperature	n type	…	41
316 SS	0.5 mol L− 1 NaCl at pH values of 8.3, 9.5 and 10.5	1-10 × 1020	Room temperature	n type	…	41
316L	Aerated 1 mol L− 1 H2SO4	0.22-0.32 × 1020 (25°C)6.6-8.4 × 1020 (50°C)4-14 × 1020 (75°C)	25, 50 and 75°C	n type	…	42
316L	Na2SO4 with different NaOH concentrations (residual oxygen < 1 ppb)		320°C (pH 5 and 8)350°C (pH 10)	p type(pH 5)p–n type(pH 8)n type(pH 10)		43
AISI 321	Aerated in 0.5 mol L− 1 H2SO4	6.4-8.4 × 1022	25°C	n type	…	42
Alloy 600	0.5 mol L− 1 H3BO3	…	30, 90, 150, 250 and 300°C	n type at 30, 90 and 150°C, and p type at 300°C	Ni rich oxide layers (30, 90 and 150°C)Cr rich oxide layer (300°C)	44
Alloy 600	0.5 mol L− 1 Na2SO4	…	300°C	p type	Ni rich oxide layers	44
Alloy 600	0.1 mol L− 1 NaOH	…	300°C	p type	Ni rich oxide layers	44
Alloy 600	Deaerated simulated pressurised water reactor primary water in a titanium static autoclave§	…	325°C for 500 h	n type to insulating type, depending on the hydrogen partial pressure		45
Alloy 600	Na2SO4 with different NaOH concentrations (residual oxygen < 1 ppb)	…	320°C (pH 5 and 8)350°C (pH 10)	p type(pH 5)n type(pH 8)p type(pH 10)		43
Alloy 690	Na2SO4 with different NaOH concentrations (residual oxygen < 1 ppb)	…	320°C (pH 5 and 8)350°C (pH 10)	p–n type(pH 5)p–n type(pH 8)n type(pH 10)		43
Alloy 800	Deaerated 0.075 mol L− 1 Na2S2O3	…	21°C	p type		3
Alloy 800	Deaerated 0.15 mol L− 1 Na2SO4	…	21°C	p type		3
Alloy 800	Deaerated neutral simulated crevice chemistries containing Pb (pH 6.06-6.17)	2 × 1027–1.3 × 1027	300°C	n type		9
Alloy 800	Deaerated alkaline simulated crevice chemistry¶	…	21°C	p type		46
Alloy 690	10− 4 mol L− 1 NaOH+500 ppm PbO	…	90°C	p type		47
Alloy 690	Deaerated neutral simulated crevice chemistry with or without 0.0022 mol L− 1 PbO #	…	300°C	p type	Ni- and Fe depletedCr-enriched	11
Alloy 690	Deaerated alkaline simulated crevice chemistry with or without 0.0022 mol L− 1 PbO ‖	300°C…	p type	Cr- and Ni depletedFe-enriched		11
Alloy 690	Deaerated acidic simulated crevice chemistry with or without 0.0022 mol L− 1 PbO Δ	…	300°C	p–n type	Ni- and Fe depletedCr-enriched	11

*

0.2 mol L^{− 1} boric+0.05 mol L^{− 1} citric acid+0.1 mol L^{− 1} tertiary sodium phosphate solution.

†

0.15 mol L^{− 1} B(OH)₃+0.075 mol L^{− 1} Na₂B₄O₇.10H₂O.

‡

0.15 mol L^{− 1} H₂BO₃+0.0375 mol L^{− 1} Na₂B₄O₇.

§

1200 mg kg^{− 1} boron and 2 mg kg^{− 1} lithium, but hydrogen partial pressure was fixed at < 1, 30 and 658 kPa.

¶

0.30 mol L^{− 1} NaCl+0.05 mol L^{− 1} KCl+0.15 mol L^{− 1} CaCl₂+0.40 mol L^{− 1} NaOH.

#

0.15 mol L^{− 1} Na₂SO₄+0.30 mol L^{− 1} NaCl+0.05 mol L^{− 1} KCl+0.15 mol L^{− 1} CaCl₂.

‖‖

0.15 mol L^{− 1} Na₂SO₄+0.30 mol L^{− 1} NaCl+0.05 mol L^{− 1} KCl+0.15 mol L^{− 1} CaCl₂+0.40 mol L^{− 1} NaOH.

Δ

0.15 mol L^{− 1} Na₂SO₄+0.30 mol L^{− 1} NaCl+0.05 mol L^{− 1} KCl+0.15 mol L^{− 1}CaCl₂+0.05 mol L^{− 1} NNaHSO₄.

For the alloys, the situation was a little bit complicated because the semiconductivity could be either n or p or p–n type, not only depending on the alloy composition but also on the film formation conditions. For example, Fe–18Cr alloy showed p type semiconductor in deaerated 0.1 mol L^{− 1} H₂SO₄ solution, while it showed n type semiconductor in deaerated borate buffer solution (pH 8.4) at room temperature. ³⁶ In most cases, the Fe/Ni/Cr ratios were believed as the main factors that affect the semiconductivity: it was inclined to p type when the alloys had high Ni content (e.g. Alloy 690), and it was inclined to n type when the alloys had high Fe and/or Cr contents (such as 316 SS and 304SS). However, Fe/Cr/Ni ratios in passive film were also influenced by the solution chemistries because the impurities in solutions could result in a selective dissolution of certain element from the passive film, which was affected by the temperatures. Therefore, these factors that determined the semiconductivity of a passive film were competitive. Anyway, the passive films of a Ni–Cr–Fe alloy should be a mixture of the Fe, Ni and Cr oxides though the ratio of Fe, Ni and Cr was different. Alloy 690 is an Ni based alloy with composition of 58.98Ni–10Fe–30Cr (wt-%), as shown in Table 1. The Mott–Schottky indicated that the passive films on Alloy 690 formed in the alkaline chemistry at 300°C were basically p type semiconductors. When the this alloy was passivated in the alkaline chemistry free of lead, the anodic film was Cr and Ni depleted and Fe enriched. ¹¹ However, the atomic percentage of Ni was high, though the Ni was depleted, about 30-50%, depending on the depth from the surface; in contrast, the atomic percentage of Cr was about 10-20% and that of Fe was 5-30%. The Ni oxide would raise the concentration of the major charge carriers (electron holes) and promote the annihilation of minor charge carrier (electron) via the recombination reaction, resulting in a p type semiconductor. The situation was similar in neutral solutions. In acidic solution, the atomic percentage of Cr and Ni are high, about 30-40% for Cr and 20-50% for Ni; however, that of Fe is low, about 2-8%. The p–n type semiconductor in acidic solution may be related to the comparative Cr and Ni contents.

For the Alloy 800, it was an Fe based alloy with the composition of 43.2Fe–32.78Ni–21.87Cr (wt-%). Therefore, the differences in composition should affect the composition of passive film as well, but not remarkable. As shown in Fig. 2a, the semiconductivity in thiosulphate only or sulphate only solution were p type, which was possibly related to the selective dissolution of Fe and Cr in these two solutions.^37,38 In neutral solutions containing thiosulphate and chloride ions, the passive layer exhibited n type semiconductor. In this case, the adsorption of chloride on film surface was considered as the dominant one; therefore, the adsorption of thiosulphate was limited. In solutions with high chloride ratio, the selective dissolution of Ni took place, resulting in a Cr and Fe enriched film; consequently, the passive film was n type. In alkaline solutions, the dissolution reaction of passive film was different from that in neutral solutions. The preferred adsorption of hydroxyl on the surface led to a selectively dissolution of Cr. As a result, a film with high Ni content possibly rendered the passive film behaved as p type semiconductor.

(2) Effect of hydrogen and hydroxyl

The hydroxyland hydrogen content in the passive films can also affect the semiconductivity. Yang et al. ³⁷ investigated the hydrogen effects on semiconductivity of passive films on 310 SS by galvanostatic polarisation at various cathodic current densities. They found that the presence of hydrogen in 310 SS caused an inversion of conductivity type of a surface film from p to n type in a borate buffer solution (pH 8.45); n type had a high susceptibility to pitting, which could be mainly correlated to the electronic properties of the passive film on 310 SS. For p type semiconductor passive film on uncharged specimens, holes were present in much higher concentration than electrons. It was reported that hydrogen entering the film would be ionised. ⁴⁸ Therefore, for the charged specimens, the generated electrons during ionisation decrease the concentration of holes in the valence band of passivefilm.

Barral et al. ⁴⁹ reported that an increase in the thickness of Ni(OH)₂ over the passive film on pure nickel resulted in an increased total acceptor density of passive films. According to Maximovitch ⁵⁰ and Sikora and Macdonald, ⁵¹ the acceptor densities in passive films on pure nickel decreased with increasing film formation potential. It has been recognised that the amount of hydroxides in passive films decreased with increasing the film formation potential. ³⁵ A similar phenomenon was also observed in passive films on pure chromium.^52,53 The above observations support the statement that the hydroxides contain a higher charge carrier density than the corresponding oxides. Therefore, the higher acceptor density of the anodic film formed at 300°C in Pb contaminated alkaline crevice chemistries was partly due to its higher hydroxides content; the higher donor density of the anodic film formed at 21°C in NC2 was also partly due to its higher hydroxides content. ³

Effect of solution chemistries

The solution chemistries can affect the semiconductivity dramatically, as shown in Table 4. For example, the passive film formed on Fe–18Cr and Fe–20Cr alloy in deaerated buffer solution (pH 8.4-8.5) were proved to be n type semiconductors;^36,40 passive film formed on 304 SS in 0.5 mol L^{− 1} NaCl solution (pH 6) was an n type semiconductor. ⁴¹ However, in deaerated 0.1 mol L^{− 1} H₂SO₄ solution, the passive film on Fe–18Cr alloy exhibited p type semiconductor. ³⁶ Although the reason of how solution chemistries affect the semiconductivity was not discussed in these papers, it was rational to assume that the solution chemistries changed the composition of passive film, and this impact was more obvious at high temperature. Jang et al. ⁵⁴ measured Mott–Schottky plots and photocurrent spectra of pure nickel samples thermally oxidised and electrochemically passivated respectively. They found that the former was an oxide film, while the latter contained a large amount of hydroxide, and the anodic films formed in pH 8.5 borate buffer solution had a higher acceptor density and lower photocurrent response than the samples thermally oxidised. Montemor et al. ⁵⁵ investigated the chemical composition and electronic structure of the oxide films formed on 316L stainless steel and nickel based alloys in high temperature aqueous environments. They found that the semiconductivity of passive film formed on alloy 600 is p type at pH 10 due to the presence of nickel oxide, but it is n type when formed at pH 8 due to the presence of mixed iron–nickel oxide in the outer layer. Kim ⁴⁴ also found that the passive film formed on alloy 600 in 0.5 mol L^{− 1} Na₂SO₄ and 0.1 mol L^{− 1} NaOH were p type semiconductors, which was attributed to the Ni rich oxide layers. However, a Cr rich oxide layer formed on alloy 600 in 0.5 mol L^{− 1} H₃BO₃ at 300°C resulted in a p type passive film, ⁴⁴ which is hard to explain. Overall, p type passive film may result from an Ni rich oxide, whereas n type passive film may result from an Fe or Cr rich layer at low temperature.

In this work, Pb and S are the major considered impurities; hence, their impact on the semiconductivity are discussed below. Pb impurities can incorporate into the anodic film, and it may result in the following defect reactions ¹¹ 4 5 where and are metal vacancies. These defect reactions will give rise to the annihilation of Schottky defect pairs in passive films. 6 When Pb²⁺ ions occupy trivalent ion vacancies in the spinel oxides via the defect reaction (5), oxygen vacancies increase in parallel to maintain electrical neutrality. In a p type semiconductor, oxygen vacancies are the major electron donors and electrons are the minor charge carriers. An increase in oxygen vacancy will accelerate the minor charge transfer rate S _r that brings about the annihilation of pairs of electrons and electronic holes (S _r = k _r [h][e], where [h] and [e] are the densities of hole and electron in semiconductor). Thus, the hole concentration [h], and thereby the acceptor density (N∝ or ∝[h]) in the anodic film, will be reduced. This statement was confirmed by experimental observations in Fig. 1b. The increases in Mott–Schottky slope indicated a decrease in acceptor density in the anodic film that was formed in the Pb contaminated neutral solution, while the surface analyses indicated that lead contamination in solution did not alter the film composition substantially. ¹¹ The situation in the alkaline solution would be complex because of the uncertainty due to substantial change in the film composition.

In the case of S related species, their interactions with the passive layer at room temperature is dependent on solution pH, as pointed out in a recent paper. ⁴ The effect of sulphur at reduced or intermediate oxidation levels changed the composition of the passive film in NC solutions but does not have remarkable impact on the film composition in AKC solutions. ⁴ In neutral crevice solutions, the incorporation of sulphur into the anodic film may result in the following defect reaction 7 This defect reaction will change the donor density in the n type films where oxygen is the major defect, but this reaction may not occur in AKC solutions because the S is not incorporated into the passive film. The adsorption and incorporation of S into the passive layer resulted in the inhibition of hydrolysis reaction (8). The sulphur prevents the adsorption of hydroxyl and hence prevents oxygen from entering the passive film 8 Accordingly, noting that the rate of oxygen vacancy generation at the metal/barrier layer interface, as shown in reactions (9) and (10), is the same as in NC1 solution since it is a function of potential, it is evident that the oxygen vacancies will accumulate in the barrier layer; therefore, the passive film will tend to break down. 9

10

Summary in solutions containing Pb and S

Figure 4 schematically shows the factors that influenced the semiconductivity of a passive film. Temperature, alloy compositions, solution chemistries, film formation time and formation potential are the main factors that determine the semiconductivity. In essence, these factors result in a change in film composition, film structure, film thickness, hydroxyl and hydrogen contents. The effect of Pb and S impurities on the semiconductivity is a result of competition of these factors, as shown in Fig. 5. This is why the presence of lead contamination displays different impacts on the charge carrier densities in the anodic films formed in the crevice environments having different pHs. In the alkaline environment, the dehydration process is more effectively blocked by the lead contamination, as compared with the phenomenon observed in the near neutral chemistry. Besides, the lead contamination in the alkaline chemistry gives rise to a significant increase in Cr content and a decrease in Fe content. These two factors are likely to be dominant, even though a portion of the holes is consumed by reaction (5), resulting in an increase in acceptor density in the anodic film. The level of lead impurities in the anodic films formed in the acidic chemistry is quite low, and its effect on the dehydration process and film composition is very limited. Therefore, only a slight change was observed in the charge carrier densities. Alloy 690, as an Ni based alloy, its semiconductivity in Pb containing solutions mainly depends on alloy composition. High Ni content in this alloy results in p type semiconductors in most cases (route 1) and n type semiconductors in few cases (route 2).

OPEN IN VIEWER

4

Factors influenced semiconductivity of passive film

OPEN IN VIEWER

5

Main factors that determine semiconductivity of Alloys 690 and 800

The effect of S on the semiconductivity is affected by other ions (e.g. chloride and sulphate ions) and solution pH. Therefore, the semiconductivity of alloy 800 mainly depends on the solution chemistries (route 3). In NC solution, the competitive adsorption of S with chloride or sulphate ions determines the semiconductivity as either p or n type (route 4 or 5). In contrast, the adsorption of thiosulphate and chloride ions is inhibited in alkaline solution. This effect is obvious at room temperature. However, at high temperature, the film formation mechanism was different, and S can incorporate into the passive film even in alkaline solutions.

Correlation between semiconductivity and breakdown of passive film

The electron energy level diagrams in various solutions shown in Fig. 6 correspond to a theory proposed by Sato. ²⁴ Figure 6 shows schematically the electron energy level diagram and the electrostatic potential distribution of the n type passive film on Alloy 800 when anodic potentials are applied below and above the breakdown potential in NC solutions. If a small anodic polarisation is applied to the Alloy 800 in solution NC1, the Fermi level in the metal is lowered, producing an anodic potential drop (band bending upwards) in the film without any change of Helmholtz layer potential difference ΔØ_H at the film/electrolyte interface, and the potential drop thus produced in the film is equal to an overpotential applied against the flat band potential. This situation is maintained so long as the Fermi level of the metal lies within the forbidden energy band in the film, as shown in Fig. 6a. The n type passive film in which E _F lies in the upper part of the forbidden energy gap, therefore, is more stable than the p type passive film in which the Fermi level is situated close to the valence band. ²¹ Chloride, as an adsorbed anion in solutions NC1 and NC0, is more electronegative or has a lower valence than the matrix oxide ion, and its adsorption or incorporation is accompanied by the appearance of an electron energy gap of the matrix oxide. Therefore, the adsorbed chloride ions are incorporated in the passive film and may occupy oxygen lattice sites or interstitial sites and give rise to the appearance of localised electron acceptor levels above the valence band in the forbidden energy gap. Figure 6a shows the electron energy level in a thin n type semiconducting oxide film formed in solutions NC1 and NC0 with adsorbed chloride ions under anodic band bending. Below a critical anodic overpotential at which the band bending lowers the Fermi level of the metal to reach the acceptor levels induced by chloride ions in the outermost layer of the film, there will be no change in ΔØ_H at the film/electrolyte interface. Figure 6b shows the electron energy level in an n type semiconducting oxide film when it breaks down. Above the critical overpotential of 0.6 V, tunnelling transfer of electrons in the film with adsorbed chloride ions will occur from the valence band at the film surface through the chloride induced acceptor levels to the Fermi level in the metal, leaving positive holes at the film surface. In the absence of chloride ions, however, no accumulation of positive holes will occur at the film surface unless further anodic polarisation is applied to lower the Fermi level of the metal to the valence band in the outermost layer of the film. Therefore, the critical potential above which the potential dependent dissolution of the film occurs will be less noble at the sites of chloride adsorption or incorporation than that at the sites containing no chloride ions, as shown in Fig. 6a. In the presence of both thiosulphate and chloride, the electron energy level increase is due to the combined effects of these two ions. Figure 6c shows the electron energy level of the passive film in solutions NC2 and NC3 when a small anodic polarisation is applied. The adsorption or incorporation of chloride and S^x in the outermost passive film may occupy more oxygen lattice sites or interstitial sites and give rise to the appearance of more localised electron acceptor levels above the valence band in the forbidden energy gap compared to solution NC1. Figure 6d shows the electron energy level of an n type passive film when it breaks down in solutions NC2 and NC3. Above the critical overpotential of ∼0 V, tunnelling transfer of electrons in the film with adsorbed chloride and thiosulphate ions will occur from the valence band at the film surface through the chloride induced acceptor levels to the Fermi level in the metal, leaving more positive holes at the film surface than that in solution NC1. As a result of the positive surface charge, the dissolution rate of the passive film will increase due to the adsorbed chloride ions. This will result in a local thinning of the film until a steady state is reached, where the local film dissolution current is equal to the ionic current through the film. When NC2 and NC3 are compared to NC1, thiosulphate seems to increase the localised electron levels in the passive film, resulting in an easier breakdown of passive films in solutions containing thiosulphate. The passive film behaves as an n type semiconductor in NC solutions, and this type of passive film is more resistant to anodic dissolution than p type passive films; n type films are easy to breakdown. ²¹ Furthermore, dislocations or linear defects are present in Alloy 800. Maurice et al. ⁶³ found that dislocation areas can form between the crystalline grains of the passive film formed on single crystal surfaces using an atomic force microscope at the nanometre scale. It is reasonable to assume that some defects should also be present in the passive film. In addition, hydroxyl and hydrogen profiles are present in the outer layer of passive films.^3,4 This creates additional electron levels located somewhere within the forbidden gap. The dislocation induced electron levels in solid oxides constitute a thin cylindrical band along the dislocation network, which may go through the film. If this type of band lies below the Fermi level, it will be fully occupied by electrons. However, under anodic band bending conditions where the Fermi level becomes lower than the dislocation band at the film surface, dislocations in the outermost layer of the film will be positively charged due to electron transfer from the dislocation band to the metal Fermi level. Since hydrogen in passive films is in the form of a proton (H⁺), the lost electron in the film acts as donor for the film. ³⁷ The adsorption or incorporation of chloride ions will introduce an increased number of dislocations in the passive film, and the combination of the dislocation band with the chloride induced donor levels will further facilitate the positive charge accumulation on the film surface by anodic polarisation.

OPEN IN VIEWER

6

Illustration of correlation between semiconductivity and breakdown of passive film on Alloys 800 and 690 based on electron energy levels a, c before film breakdown for n type semiconductors, b, d when film breakdown for n type semiconductors, e before film breakdown and f when film breakdown for p type semiconductors ⁴

Figure 6e and f shows schematically the electron energy level diagram and the electrostatic potential distribution in the p type film within the passive region and above the transpassive potential. As shown in Figs. 1a and b, 2a and c and 3, p type passive films are formed in alkaline solutions for Alloy 800 and in neutral and alkaline solutions for Alloy 690, due to a preponderance of cation vacancies on the cation sublattice of the barrier layer. For p type passive films, holes are present in much higher concentration than electrons. In AKC solutions, due to the adsorption of –OH on the Alloy 800 film interface, the adsorption of chloride and thiosulphate is weakened. Therefore, these two ions do not have an obvious impact on the electron energy level. Figure 6e shows the electron energy level of the passive film when a small anodic polarisation is applied. The Fermi level in the film is elevated (band bending downwards). Below a critical anodic overpotential, there will be no change in ΔØ_H at the film/electrolyte interface. The film begins to transpassively dissolve when E _c < E _F, as shown in Fig. 6f. Sato found that the n type oxide passive film on iron is electrochemically stable against anodic polarisation, whereas the passive film of p type oxide on nickel is relatively unstable and is subject to transpassive, potential dependent dissolution at comparatively less noble potentials. ²¹ This was experimentally verified: for n type passive films formed on Alloy 800 in chloride solutions, there were always some peaks in the polarisation curve relating to the metastable pits,^3,4 and the current was suddenly increased when the film broke down; however, for the p type passive films on Alloys 800 and 690, the current was gradually increased when the film suffered transpassive dissolution. ⁶⁴

Overall, it is generally believed that the stability of an n type semiconductive oxide film is related to the density of donors that are mainly oxygen vacancies. A higher donor density indicates a more defective film structure and lower resistance to breakdown.^62,65 However, the acceptors in the p type semiconductive oxide films are various metallic vacancies. Acceptors resulting from different metallic cation vacancies will be involved in the passive films formed on different metals, even in those formed on the same metal under different conditions. Therefore, the relationship between the resistance to breakdown of a p type semiconductive oxide film and the charge carrier density will be more complex than that observed in an n type semiconductive oxide film. Thus, the stability of a p type semiconductive oxide film is unlikely to be determined solely by the acceptor density. The nature of acceptors plays an important role in the passivity of alloys. For instance, if lead is incorporated into the oxides via reactions (4) and (5), it will not only affect the charge carrier densities but also alter the electronic structures of anodic films. This change can hinder the formation of spinels or cause the decomposition of these oxides, resulting in a profound impact on the passivity. The replacement of Cr³⁺ and Fe³⁺ in spinel oxides with Pb²⁺ via reaction (5) will reduce acceptor densities in anodic films, as demonstrated by Mott–Schottky and photoelectrochemical measurements of the anodic films formed in near neutral crevice environments where the impact of lead incorporation on the composition of anodic film is insignificant.