4

Types of CO 2 Corrosion Damage

CO 2 corrosion may manifest itself as general thinning or localised attack. Localised

corrosion is characterised by loss of metal at discrete areas of the surface with
surrounding areas remaining essentially unaffected or subject to general corrosion.
These discrete areas may take various geometrical shapes. Thus, circular depressions
usually with tapered and smooth sides are described as pits. Stepped depressions
with a flat bottom and vertical sides are referred to as mesa attack. Other geometrical
forms of localised corrosion include slits (sometimes referred to as knife line), grooves
etc. In flowing conditions localised attack may take the form of parallel grooves
extending in the flow direction; this phenomenon is known as flow induced localised
corrosion.

4.1. Localised Corrosion of Carbon Steel

CO 2 corrosion can appear in three principal forms, pitting, mesa attack or flow
induced localised corrosion.
Pitting can occur over the full range of operating temperatures under stagnant to
moderate flow conditions. The susceptibility to pitting increases and time for pitting
to occur decreases with increasing temperature and increasing CO 2 partial pressure.
D e p e n d i n g on the alloy composition there exists a t e m p e r a t u r e range with a
m a x i m u m susceptibility for pitting [5].
Inspections of sweet gas wells have indicated that localised corrosion, including
pitting, often occurs preferentially at certain depths (i.e. in certain temperature ranges).
Generally 80-90°C is a temperature range where pitting is likely to occur in sweet
gas wells. Pitting may arise close to the dew point temperature and can relate to
condensing conditions. There are no simple rules for predicting the susceptibility of
steels to pitting corrosion.
Mesa type attack is a form of localised CO 2 corrosion u n d e r m e d i u m flow
conditions [6]. In such attack, corrosion results in large flat bottomed localised damage
with sharp steps at the edges. Corrosion damage at these locations is well in excess
of the surrounding areas.
The conditions most likely to lead to mesa attack are those under which carbonate
films can form but are not strongly stable. Film formation begins around 60°C and
thus mesa attack is much less of a concern at temperatures below this. If the general
filming conditions are borderline then local variations in flow or metallurgy or both
may be enough to de-stabilise films. This type of localised attack results from local
spalling of carbonate scales after reaching a critical scale thickness [7-9]. This local
spalling occurs due to intrinsic growth stresses in the scale [10]. Spalling of the scale
exposes underlying metal which then corrodes and may reform surface scale. On
regaining a critical thickness the newly formed scale can crack and spall again
producing another step.
 Types of CO2 Corrosion Damage
Spalling of scale particles or flakes relieves the stress in the scale adjacent to and
around the spalled area. Therefore, this scale remains attached to the surface and
can protect it from localised attack. As a result, the flat bottomed pits obtain sharp
edges. Mesa attack may also simply result from self sustaining galvanic coupling
between protective and non-protective corrosion films.
Flow induced localised corrosion (FILC) in CO 2 corrosion starts from pits a n d /
or sites of mesa attack above critical flow intensities. The localised attack propagates
by local turbulence created by the pits and steps at the mesa attack which act as flow
disturbances. The local turbulence combined with the stresses inherent in the scale
may destroy existing scales. The flow conditions may then prevent re-formation of
protective scale on the exposed metal.

4.2. Localised Corrosion of Carbon Steel Welds

Localised corrosion of carbon steel welds in CO 2 corrosion systems has been

experienced by many operators. It is a complex problem because it is dependent
partly on the environment (and the nature of any carbonate scale formed), partly on
the metallurgy and composition of the carbon steel and the weld and partly on the
geometry of the weld profile (local turbulence).
Initially, preferential attack may arise from galvanic differences across a weld due
to compositional or microstructural differences between the deposited weld metal,
the parent steel and file heat affected zone (HAZ).
The location and morphology of the preferential corrosion is influenced by a
complex interaction of many parameters including the environment, the operating
conditions, the parent ,;teel composition, the deposited weld composition, the welding
procedure and the initial surface state. Changes in any one of these parameters can
cause a significant difference in the weldment corrosion behaviour.
Changing the composition of the weld metal relative to the parent steel can make
the weld metal more, or less, susceptible to preferential attack. Similarly, changing
the grade of parent steel can affect the behaviour of the weld metal but, in conjunction
with the welding procedure, the parent steel composition will also determine the
microstructure of the HAZ and therefore influence the susceptibility to preferential
attack in that region.
The welding procedure will directly influence the HAZ microstructure, but will
also affect the degree of dilution of the weld metal by the parent steel and the
composition at the fusion line of the weld. The presence of welding slags, oxide
films and inclusions increase the complexity of the weld corrosion phenomenon.
It is extremely important to note that a weld consumable selected to avoid
preferential corrosion in one environment could exacerbate the problem in another.
For example, c o n s u m a b l e s c o n t a i n i n g 1% Ni or 0.6% Ni plus 0.4% Cu as
recommended for seawater injection systems may cause problems if used under
certain conditions in sweet hydrocarbon environments [11]. Rapid corrosion of the
weld metal has occurred in some instances while HAZ attack has also been observed.
The window of conditions under which this problem occurs has yet to be accurately
defined. However, in the majority of cases, failures have occurred at temperatures
approaching conditions under which protective scales are expected to form (70-80°C).
 CO2 Corrosion Control in Oil and Gas Production reDesign Considerations

The risk of preferential weld corrosion can be minimised by conducting laboratory

tests on the relevant weldment under simulated service conditions using appropriate
electrochemical monitoring techniques, including galvanic coupling through zero
resistance ammeters. It should be noted that although laboratory studies have
generally been successful in simulating weld corrosion problems in other situations
than CO 2 corrosion service, in some instances (such as with higher nickel contents)
cathodic weld metal behaviour has been observed in the laboratory, but anodic
behaviour in service, which may be due to the difference in the initial surface state.
Weldment corrosion behaviour must, therefore, be confirmed by monitoring in
service. The same monitoring techniques can be used, ideally in combination with
other techniques such as ultrasonic wall thickness measurements.
The effects of inhibition (and biocide treatments) on weldment corrosion must
also be considered. Although inhibition can be an effective means of controlling
preferential weld corrosion, inhibitor adsorption can be influenced by weld metal
composition and, in some cases, protection is not achieved. Again, inhibitor tests on
weldments under simulated service conditions can be used to select an appropriate
inhibitor formulation.
The theory of why the scale breaks down at the weld is a combination of:

Local turbulence because the weld root protrusion disturbs the flow and eddys
then break up the scale.

The chemistry of the weld is slightly different from the adjacent metal and for
some reason (e.g. carbide structure) the scale is not as protective.

Solving the problem is not easy. Steps which can be taken include:

• Specifying a maximum root penetration of 0.5 mm.

Using filler metals for the root run with alloying additions of copper and
nickel (e.g. ISO:E51 4 B 120 20 (H) AWS:E7018-G) typically used for welding
so-called weathering steels. Low weld silicon contents are also suggested,
probably < 0.35%, since a few practical problems have been experienced in
the past with weld Si contents of around 0.5% or more. A problem with Si is
that recovery across the arc depends upon the arc length and the local shielding
(i.e. on the joint design, welding position etc.). Thus, the same electrode can
give an appreciable range of Si in the weld deposit with different welders or
joint geometry. However, < 0.35%Si should generally be achievable.

Detailed laboratory testing simulating flowing conditions to select the correct

combination of filler and inhibitor for the given conditions. (Testing is
particularly recommended for operations above 70°C).