The profiles of in-depth composition of passive films at 2.0 V on the two nuances of inconels 600 and 690 are represented in Figure 1. The atomic density of internal region in contact with the alloy is constituted essentially with oxide of chrome, while the external region of the film is mainly composed with oxide of nickel. The oxide of iron in low quantity is either in the internal region in the case of the passive films formed in the inconel 600 or in the external region if it is the film formed in the inconel 690. We remark that the interface between the two regions (external and internal) is not abrupt. The comparison of the films formed on the two inconels shows a quantitative difference concerning the thickness and the atomic density of the oxides. This indicates that the natures of the films are intimately linked to the chemical composition of the alloy. On the inconel 690 richer in chrome, there is a film mostly composed of oxide of chrome. Even though nickel is a basic element of this alloy, its oxide present inside the external region of the film is represented by a very thin thickness (2 to 3 mono-stra- tum).

On the other hand, concerning the inconel 600 (15%Cr) both regions have practically the same thickness. Concerning the atomic density, the difference registered between the nickel and the chrome oxides is practically the same for both alloys. Figure 1 shows that inside the internal part of the film the number of chrome atoms in the oxide is practically equal to 8.7 at∙nm⁻². This value corresponds to that of type Cr₂O₃. The number of nickel atoms in the oxide external position is about 6 at∙nm⁻².

In our study, the line of the current/potential curves is made at the speed of 1 mV/s from the cathode potentials to the anode potentials. In order to lessen the oxides gathered in the air, a pre-polarization cathode (5 mA∙cm⁻²) preceded the potential dynamic study for about 5 minutes (values are given in relation to the calomel).

To determinate the role of the iron in the nickel-based materials; we proceeded to the preparation by fusion to plasma of a series of pure alloys. The potential-dynamic study led on the binaries Ni-15%Cr, Ni-30%Cr on one hand and the binaries Ni-15%Cr-8%Fe, Ni-30%Cr-8%Fe on the other hand allowed us to put in evidence the influence of the addition of the iron. It is suitable to make the remark that the last two alloys represent the versions of high purity of alloy 600 and 690. In the analysis Figure 2 where we represent the curves of polarization obtained in the same experimental conditions on the four alloys, many points are to be underlined.

The addition of chromenickel provokes an increase of the current of passivation in the high potentials area; that is to say in the region of the trans-passivation. This phenomenon linked to the dissolution of the chrome, at around 0.6 V, is more remarkable in the case of the binary alloy Ni-30%Cr where an increase of the current of passivation is very important, 10 times higher than that of the Ni-15%Cr.

The addition of iron (8%) in the binary alloys Ni-15%Cr and Ni-30%Cr provokes an important diminution of the current of passivation in the high potentials area. This is due to the passivating character of the iron in this potential area. This result shows clearly that iron exerts a great role on the passivity of alloy 600 and 690.

On one side, Figure 2 indicatesthat the presence of iron provokes a slight increase of the current of passivation in the negative potential area (around −0.7V). This reflects the active behaviour of the iron in this potential area.

Finally, it appears that the increase of the current of passivation in the negative potential area due to the activation of the iron is weaker in the case of the alloys from an industry. This can be attributed to the actions of the impurities present in the industrial inconels (Si, P, B, Mn, S, ...). Nevertheless, the joint or specific action of this impurity remains hard to establish.

The films firstly formed in autoclave are studied in balance potential reach after 2 hours in a pH 9.2 boric tetra borate-acid buffer solution. Values corresponding to the different oxides show that the balance potentials of the oxides formed on the inconels 600 (−268 mV) and 690 (−269 mV) are on the one hand identical and on the other hand electronegative.

In order to obtain the appropriate information on the effect of the frequency and the capacity in a big domain of frequency, an investigation by impedance measurement has been led on the passive films formed at different potentials. The domain of study in frequency should be largeenough (from 10⁵ Hz to 10⁻³ Hz) in order to get to the resistances of the solution and the polarization respectively at high and low frequency. Let us note that in this case, the measurements are made at fixed potential; potential of the formation of the film.

The Bode representation presents a linear variation in a large domain of frequency (10³ Hz; 10⁻³ Hz). At high frequencies (10⁵ Hz; 10³ Hz), the impedance remains constant and equal to the resistance of the solution Rs which is about 40 Ω∙cm² as shown in Figure 3. In the case of alloy 690, richer in chrome, the phenomenon happens from 0.7 V. This shows that the linear behaviour is also controlled by the chemical composition. In fact the increase of the current at high potential due to the dissolution of the chrome around 0.6 V; contributes enormously to the value of the impedance. In the case of the films formed to the potentials where the linearity in Bode representation is verified in a big domain of frequency, the resistance, the real part of the impedance and the capacity are linked to the frequency by expressions (1) and (2).

The order of size of these parameters a (10³) and b (10) shows that the third term of the expression (1) only becomes important at very low frequencies.

where C₀ is the capacity to F = 1 Hz and P (slope) a parameter characterizing more or less the big dispersion in frequency.

Dependence in frequency of capacity and resistance has been observed in different works [6] - [15] and many causes have been proposed such as the non-uniform distribution of the current, the state and rigidity of the surface, the residual faradic processes, and so on and so forth… The linear variation between the resistance and the reverse of the frequency; R = f(1/F) and the reverse capacity and log F; C = f(1/logF) have been observed in the different works concerning the dispersion in frequency of the passive films and oxides [6] - [15] . In order to provide chemical and physical explanation to these mathematical variations, many interpretations have been put forward. Van Geel [14] suggested that this is due to the film variable conductivity because of its non stoechiometry. In harmony with this hypothesis Young [11] proposed a model where the film can be represented by an electrical circuit composed of infinite RC elementary circuits parallel in series with an exponential variation of the resistivity in relation to the distance through the film. This hypothesis has also been proposed by Wood and

Pearson [8] for the interpretation of the dependence infrequency of the conductivity of the anodic oxides films. From another point of view, Gomes and all [16] - [25] interpreted the linear variations on the basis of the dielectric relaxation phenomena in the film due to the contribution of the dipoles in the double stratum. The relations R(1/F) and 1/C(Log F) have been obtained by an extension of the Gevers model [26] [27] where the dielectric constant is not real but rather complex. They have shown that the linear variation of the resistance as related to the reverse of the function R = f(1/F) can be deduced from the one that links 1/C and Log F from Kramers-Kronig relations [17] [18] . The dispersion in frequency observed in this study can be interpreted on the basis of the Young model considering a variation of the resistivity of the film in relation to its thickness. This interpretation is easily justified by the fact that the films formed on inconels are composed of different oxides that have different resistances. In this model, Young considers a circuit composed of infinite RC elementary circuits parallel in series and neglects the variation of the dielectric constant according to the frequency. The exponential variation of the resistivity in relation to the thickness of the film is expressed as follows:

The total C capacity is given by the relation:

The comparison of Equations (3) and (4) with those already established (1) and (2) allow an identification of the parameters (a), (P) and.

(5)

We remark that the ratio of slopes (a) and (P) is constant and equals 4 and 9.2 in decimal logarithm representation.

Thus, linear changes in the resistance obtained in the high frequency domain in the case of formed passive films are in conformity with the relationships established by Young. However curvilinear variations obtained in the field of very low frequencies have never been observed before. Because most of the work on the frequency dispersion has been limited to the field of high and medium frequencies where the variation of the resistance as a function of the reverse of the frequency is checked, curvilinear variations at very low frequencies may be due either to a change in dielectric constant as a function of the assumed zero frequency at high frequencies: Model Young, or to the slow response of deep loads under a very low frequency alternating excitation potential.

Capacity measurements carried out at 1580 Hz frequency on the films formed on the alloys 600 and 690 at 0.3 V, are represented in Figure 4.

This capacities are representative either of an electronically semi-conductor behaviour relating to a type P conductor (negative slopes) [28] . Figure 4, where we represented C⁻² in relation to U potential, shows that the action of chromium is revealed by the appearance of a capacitive peak located in the area of the more negative potentials and characteristic of a semi-conductivity of P type.

The action of iron on the capacitive behaviour of nickel is exerted, contrary to that of chromium, in the more positive potential area (Figure 5) where iron manifests a semi-conductivity of n type. We observe that alloy Ni-8%Fe enables the transition between the capacitive behaviour of nickel and that of iron. However the simultaneous addition of iron and chromium seems to be indispensable to the development of a passivation film capable of manifesting both types of semi-conductivity (n and p). So the results obtained by the study carried out on the ternary alloys Ni-15%Cr-8%Fe and Ni-30%Cr-8%Fe, represent respectively the high purity versions of 600 and 690 alloys and reveal a behavior similar to that of the films formed from those of industrial origin (Figure 4). The purity factor that differentiates the two materials appeared to have little effect on the behaviour of the films.

Knowledge of the resistance values and the reverse of the fixed frequency capacity can establish a comparative investigation of the resistive and capacitive properties of passive films in relation to their potential for the

formation and chemical composition. The choice of frequency is important for this investigation because it is necessary to pick out the values of resistance of very low frequency. This provides access to a more representative resistance of the film; it is sometimes called polarization resistance when related to corrosion. As regard capacity, the comparative study can be made on the basis of 1/C₀ values.

The representation R to 10⁻² Hz and 1/C₀ as related to the potential of formation in the case of the films formed on the 600 and 690 alloys and the pure nickel (Figure 6) shows that the variations R and C are very similar to those of nickel indicating that the external stratum formed on the three materials are very close.

The decreasing of R and C as related to the increasing of the potential of formation becomes much accentuated from 0, 6 V whatever the material is. The resistance as well as the reverse of the registered capacity in the case of the alloy 690 is more important than those of alloy 600 and nickel.

The photo electrochemical behavior of passive films is largely influenced by their semiconducting properties, therefore, studying photocurrent helps to better understand their electronic structure.

Photocurrent measurement in relation to the length of the wave determines the nature of fundamental transitions, the energy of “gap” E_g and to study the optical behaviour of the passive film close to the edge of the absorption.

The photo-electrochemical responses obtained from the potential of balance on the thick oxides films formed in autoclave are represented in Figure 7.

We remark that the photo response of alloy 600 is relatively higher than that of alloy 690. These results are of course in perfect accordance with the fact that the chemical composition of the films that develop on alloys 600 and 690 differs deeply. The films formed on alloy 690 contain more chromium oxide. For the nickel-based alloys, the gap values of the different oxides obtained by applying Equation (6) are clearly different. The determination of the energy of the gap E_g (forbidden band) supposes the knowledge of the relation interpreting the variation of the photocurrent in relation to incidental energy. The quasi-linear variations obtained on the different passive films (Figure 7) show that the photo-electrochemical behaviour of the latter can be described by the model of Gartner [29] . In this photocurrent conditions, I_ph is linked to the incidental energy hv by the relation (6):

where e represents the elementary charge, Φ₀ the flux of photons incidents and W the thickness of the zone of space charge. The Gartner model admits that the photocurrent is generated only in the zone of space charges, beyond this region the pairs of electron-pit formed recombine. Consequently only a very thick region is necessary for the photo-electrochemical effect to be possible. The analysis of the experimental results through Equation (6) allows us to reach the value of the energy of gap E_g and the nature of (n) transitions .The most precise method is the one that introduces the notion of quantic output η defined as the ratio between the numbers of photo charges created and the number of received photons. In this condition Equation (6) becomes:

The obtained results show that the value of n verified experimentally (n = 2) is representative of indirect transitions. This is in harmony with the different works made on the stainless steel [30] - [35] and iron [6] .

The values of the energy gap determined for the two alloys are substantially identical 600 (2.19 V) and 690 (2.25 V).

The analysis of the region at the edge of the gap shows the presence of an Urbach [35] queue. It is an exponential variation (Figure 7).

The photocurrents measured on alloys 600 and 690 represented in Figure 8 show also two distinct regions with a quantitative difference on the global photocurrent. The existence of these two regions shows up the photo-electrochemical properties of the oxides of iron and chrome.

The capacitive study developed on the pure alloys either binary or ternary permitted to point out the role of iron and chromium on the semiconducting properties of passive films on these materials.

It appears in Figure 8 that in the case of alloy 690, the weak potential photo-electrochemical response is much more important than the one obtained at high potential whereas it is the reverse situation that appears concerning alloy 600.

These results show in fact that the photo-electrochemical responses are well controlled by the biggest or weakest quantity of oxides of iron and chrome which constitute passive films. Moreover, the moving of the minimum toward the more negative potentials, where the potential of flat band of iron oxide is located, in the case of the alloy 600 can be attributed to the fact that the oxide formed on the latter is richer in iron oxide than the one formed on alloy 690.

The results obtained on the oxide films formed on both types of alloy are represented in Figure 9 where we remark that in relation to the applied potential, the photo current representation in the case of inconel 690 reveals also the existence of two regions separated by a potential of about 0 V and characterized by the peak of amplitude of about 0.4 V. These variations are similar to those obtained by measurement of capacity. The absence of the second region at low potential in the case of the oxide formed on inconel 600 already obtained by measurement of capacity confirms the fact that this oxide is on a chemical point of view very rich in oxide of iron.