Development of a naphthenic acid corrosion model

A model has been developed to predict and help prevent corrosion when processing high TAN crudes.

ERIC VETTERS
ProCorr Consulting Services

Viewed : 4597

Article Summary

High total acid number (TAN) crudes often sell at a significant discount due to concerns about naphthenic acid corrosion (NAC) – both real and perceived. In some systems, high corrosion rates are experienced with relatively low TAN feedstocks, while other systems without extensive alloy upgrades seem to handle high TAN feeds with few problems. This unpredictability, as well as the localised nature of NAC, has created an air of mystery around NAC and has caused many refiners to take a very conservative approach towards processing high TAN crudes. The challenges of predicting naphthenic acid corrosion along with new developments in understanding the mechanism led ProCorr Consulting Services to analyse the available literature corrosion data and to turn that data into the Tancorr naphthenic acid corrosion model.

NAC is generally understood to occur in parts of the crude and vacuum unit operating above 450°F (232°C) when the TAN value exceeds some threshold value without the presence of a high molybdenum alloy such as 316 or 317L stainless steel. TAN, which is a measure of the acid content, is used as an indicator of NAC potential; however, it not considered to be a reliable predictor of NAC by itself. This temperature range also coincides with that of high temperature sulphur corrosion, meaning that at least two simultaneous corrosion mechanisms can be occurring simultaneously with the potential for significant interaction between mechanisms.

The basic NAC corrosion reaction is shown in Equation 1. Because the iron naphthenate corrosion product is oil soluble, it has not been thought to form a protective scale like the sulphidation corrosion reaction depicted in Equation 2, which is illustrated using H2S as the reactive sulphur compound. Sulphidation occurs at similar temperatures to NAC and is more generalised in nature with less susceptibility to high shear stress. The iron sulphides formed from sulphidation corrosion are not oil soluble and tend to form a scale on the metal surface, which has been shown to provide some level of inhibition against NAC. Equation 3 shows how naphthenic acids can also react directly with iron sulphide, potentially reducing the inhibitive effects of iron sulphide scales against direct naphthenic acid attack on the metal surface:

2RCOOH + Fe → Fe(OOCR)2 + H2         (1)

H2S + Fe→ FeS + H2                   (2)

2RCOOH + FeS ⇒ Fe(OOCR)2 + H2S     (3)

Many theories exist to explain why TAN alone is not a good predictor of corrosion. The most commonly promoted theories are related to the composition of the naphthenic acids and the presence of sulphur in the system. The acid composition claims are usually either that the TAN test is non- specific and detects things other than naphthenic acids or that some acids are non-corrosive or may even inhibit corrosion. While there is a limited amount of testing on specific molecules that seems to indicate that different individual acids cause corrosion at different rates, there is nothing that definitively links differences in crude oil corrosivity to differences in the nature of the actual acid species present, and there is certainly no way to measure the parameters likely to make a difference in the corrosivity of individual acid molecules found in crude oil.

There is actually good evidence to support the inhibitive effect of sulphur on NAC. The idea has been that the iron sulphide corrosion product layer formed from sulphidation corrosion, which usually occurs in the same locations as NAC, inhibits diffusion of naphthenic acids to the metal surface, which thus reduces the corrosion rate due to NAC. In work by Craig,1 the presence of H2S reduced the overall corrosion rate of mineral oil and naphthenic acid mixtures by approximately 80-90% compared to similar systems without H2S. This interaction between NAC and sulphidation is complicated by the fact that naphthenic acids can also react with iron sulphide scales to dissolve the scale.

More recently, a joint industry programme (JIP) at Ohio University discovered the presence of a thin iron oxide layer under an outer iron sulphide layer. This iron oxide layer provided additional corrosion protection and was determined to result from the thermal degradation of iron naphthenate corrosion products. Iron naphthenates are not thermally stable and have been shown to break down to form a separate iron oxide layer under the outer iron sulphide layer.2 It is likely that the size of the iron naphthenate molecule (Fe(OOCR)2) hinders diffusion out of the iron sulphide scale enough to allow some thermal decomposition.

This formation of an iron oxide scale adds yet another layer of complexity to the mechanism by adding at least two new reactions to the mechanism:

Fe(RCOO)2 → FeO + CO2 + RCOR        (4)

4FeO → Fe3O4 + α-Fe                 (5)

The net result is that a complex set of interacting scale forming and scale destroying reactions are simultaneously occurring. This complex corrosion mechanism is depicted in Figure 1, which shows the development of a dual layer protective scale. Alloying elements, such as chromium, add still more complexity as the resulting Cr containing corrosion products are incorporated into the scale. The overall corrosion rate depends on the relative rates of these different reactions, which in turn depend on the fluid composition, the process conditions and the metallurgy.

With this mechanism, a number of distinct corrosion regimes can be expected. At low TAN values, the formation of any oxide scales will be minimal and sulphidation corrosion will dominate. At some level of TAN, the interactions between NAC and sulphidation become significant and a maximum inhibition level is reached. As TAN increases relative to the sulphur level, at some point NAC and the attack of iron sulphide scales by naphthenic acids begins to dominate. Figure 2 depicts how this scale inhibition might look as a function of TAN at a constant sulphur level.
The percentage inhibition in Figure 2 is based on the following overall corrosion rate equation:

CRoverall = [CRS + CRNAC](100-%I)/100     (6)

CRS and CRNAC are the sulphidation and naphthenic acid corrosion rates independent of any interactions between the mechanisms, and %I represents the interaction between the two mechanisms expressed as a % inhibition. The iron sulphide scale formed from sulphidation corrosion naturally limits the steady state corrosion rate, and that effect is already incorporated into the Modified McConomy curve, which is the most widely used method of predicting sulphidation corrosion.

At low TAN, there is limited oxide layer formation, so inhibition levels are low. As the TAN increases, the formation of oxide scale increases, which increases inhibition until a maximum is reached. Beyond some TAN level, the negative impact of TAN on the scale starts to outweigh the benefit of the oxide layer and inhibition starts to decrease. If the TAN relative to the sulphur level in the stream is high enough to disrupt the iron sulphide scale formation, the actual corrosion rate can be higher than the sum of the predicted NAC and sulphidation rates (that is, percentage inhibition becomes negative).