STRESS INVOLVED CORROSION MECHANISMS.

The interaction of stress, either applied directly or resulting from residual stresses accompanied, by an aggressive environment can result in large decreases in the expected long termmechanical properties of materials. Importantly, the evidence for corrosion processes in these cases in the form of corrosion products is minimal. Examples of these processes include stress corrosion cracking (SCC), corrosion fatigue, fretting and hydrogen assisted environmental cracking or hydrogen embrittlement (HE). These all result in unexpected failure or cracks and other defects appearing in the component..

STRESS CORROSION CRACKING (SCC).

This is an important mechanism which initiates cracks in a material which if left undetected eventually produces failure in a metal. The process has been extensively researched. SCC depends on both an aggressive environment and a TENSILE stress. The tensile stress is required to open up cracks in the material. The stress can be either directly applied or residual in form. An example of residual stress resulting in stress corrosion cracking of brass after deep drawing is shown in the figure above. Large hoop residual tensile stresses are present after the deep drawing process. If an aggressive environment is placed on the brass, cracks will appear due to the conjoint action of the stress and the environment. In its early days this was called "season cracking" as it occurred during the monsoon season in India on brass bullets and cartridge shells due to ammonia from rotting vegetation rendering them useless as they stress corrosion cracked. Similarly "caustic embrittlement" is due to sodium hydroxide in combination with residual stresses on steels.

Two types of cracks occur for SCC, either intergranular or transgranular. Usually the density of cracks increases with depth as the stress increases in the material as the crack grows. One possible method of telling SCC from intergranular corrosion is the density of cracks as a function of depth. For intergranular corrosion the density decreases, but for SCC the density increases.

Specific combinations of environment with materials are required for SCC. Brasses do not SCC in chloride environments while stainless steels do. Stainless steels do not crack in ammonia environments but brasses do.

Stress Corrosion Cracking Mechanisms.

Regions where stress corrosion cracks are viable can be related to the potentiodynamic scans for passive materials. Two regions are possible, one at the start potential for passive films or the pit nucleation potential and a second at the active to passive transition at much lower potentials. It should be noted that at both these potentials the passive film is somewhat unstable. In one case it is just on the verge of forming and in the second case it is just on the verge of breakdown. Therefore SCC could be viewed as amplification of instability in passive films. Importantly, SCC can be viewed as an ANODIC process as anodic current is required for SCC to occur. It is this fact that separates it from hydrogen embrittlement which has many of the same features but is cathodic in control.

Stress corrosion cracking can be separated into two distinct regions, crack initiation and crack propagation.

1. Crack Initiation.

Cracks can be initiated by several mechanisms:-

a. Mechanical Features.

Cracks will often initiate at features such as scratches, nicks or dents on the surface of the metal. In this case the local environment or local stress conditions favor enhanced dissolution, poor formation of a passive film or in-situ damage to a protective film.

b. Local galvanic cells initiating dissolution.

Local corrosive effects dominate the process where the local galvanic cell locally dissolves one phase of the material. This will also localize the stress on the material. The crack in this case may initiate in a transgranular mode or an intergranular mode. One example of the latter would be intergranular cracks during SCC of sensitized stainless steels.

c Pitting type crack initiation.

The pitting potential is related to formation of its and there is some correlation between the pitting potential and the potential for stress corrosion cracking. A 10:1 ratio of pit depth to width was suggested to be needed for a pit to initiate a crack. The local environment in a pit may also be important in the crack growth process. For example, the environment in a pit may favor crack growth by intergranular crack growth. Several studies were made employing a pit as an effective crack in the surface to be used in linear elastic fracture mechanics approach. These have met with some success. However electrochemical effects can nullify this approach.

d. Initiation at stress induced phenomena.

Slip lines intersecting the surface can have a double effect. One is to provide local anodes as the site is very active. The second is to rupture passive films on the surface and locally form dissolving regions.

Once a crack has initiated, then it will grow. As pointed out above the growth mechanism may not be the same as the initiation process. Several mechanisms were proposed to explain the observed features of SCC crack growth.

2. Crack Growth Mechanisms.

One important feature for SCC cracks is that some show clear evidence of stopping and starting. The cracks do not continually grow to failure. The mechanisms proposed try and take this factor into account.

a. Film Rupture Mechanism.

The tensile stress ruptures films at the crack tip and the crack grows rapidly from the bare metal exposed until the crack tip can repassivate in some cases or grow slowly to failure in other cases. Unfortunately, SCC cracks often have significant features on their faces which should be removed by local dissolution effects. In addition the crack plane in transgranular failure is often not the active slip system in the metal. Others have suggested a similar model in which the film is formed by a tarnish process. Intergranular corrosion is proposed to occur by preferential oxidation of the grain boundaries.

b. Film Cleavage Mechanism.

In this mechanism, the surface film grows and may increase in internal stress with thickness. This combined with the applied tensile stress induces brittle failure in the film which propagates across into the metal and provides a period of crack growth. The loss of the film stress and plasticity in the metal than blunts the crack and stops it growing to give periods of crack growth followed by rest while the film grows back to the conditions for cleavage.

c. Adsorption Induced Cleavage.

During the electrochemical process atoms are absorbed on the surface that weakens the bonds. The stress to initiate a crack then decreases and a crack grows until it is blunted by plastic deformation or grows out of the adsorbed region.

d. Adsorption Induced Plasticity.

The adsorption of specific ions in this cases reduces the critical resolved shear stress for dislocation mobility. Dislocations can then move locally under the influence of the tensile stress. This is different from the above model where the cohesive strength was decreased but not the resolved shear stress as in this case.

e. Atomic Surface Mobility.

At the crack tip atoms move away from the tip and vacancies move in by a surface diffusion process coupled with electrochemical activity. There is not a lot of support for this mechanism at present.

f. Corrosion Tunnel Model.

In this case, corrosion tunnels are formed by active corrosion alone emerging slip lines. When sufficient metal is removed in the tunnels, then the undissolved regions between them fracture by ductile overload and some crack growth occurs until it is plastically blunted and the process starts again. A later modification suggested the tunnels were slots on an atomic scale which can only be formed in stainless steels by SCC conditions.

Many of these mechanisms cannot be verified easily by experimental methods. It is suspected that there is no universal model for SCC growth and initiation. Each specific combination of metal environment and stress can produce different fracture evidence. In general SCC fracture surfaces look brittle in nature with evidence of non continual crack growth. Unfortunately, cross sections of samples show some evidence of local ductility or plastic flow as well. this tends to ad to the confusion surrounding mechanistic studies of the process.

Planar slip occurs in stainless steels which fail in SCC by a transgranular mechanism. These alloys are also low stacking fault energy alloys. The intersection of slip bands with passive films is thought to grow cracks.In addition, preferential dissolution in high chloride contents occurs along slip bands. therefore transgranular crack propagation is favored. However, for the sensitized stainless steel, the grain boundaries are favored and intergranular SCC cracks were found.

Measurement of SCC.

Modern measurement of SCC is based on fracture mechanics principles to produce a value of K1SCC. the plain strain fracture toughness under stress corrosion cracking conditions. Fig 1 shows how the value of plane strain fracture toughness under stress corrosion cracking conditions is established. The curve has three distinct regions.

Region 1

A vertical region in which the crack has initiated and grows very rapidly then decreases in crack growth rate. This is where the K1scc determination is made. It is the stress intensity at which a crack will initiate for the metal and environment under test.

Region 2

In region 2 a steady state rate of crack growth is present. The crack is increasing in length and does so over a reasonable range of stress intensity before entering region 3.This is a region of stable crack growth.

Region 3

In this region the fracture toughness of the material is reached and the crack is now unstable as its crack growth rate increases very rapidly with increasing stress concentration to failure.

To obtain the data shown in the figure above, pre-cracked samples are used. A typical four point bend sample is shown in the figure below. In this case the material has a notch in it. Prior to stress corrosion testing a crack is grown from the notch by fatigue. When the crack has reached the desired length the sample is ready for stress corrosion testing.

The stress intensity can be calculated from standard elastic fracture mechanics approach such as:

K=  f(Y) (a)1/2

where:- K- stress intensity; s - stress applied:f(Y) - geometric factor

a- crack length.

If the load conditions are known and the crack length is measured during the test for a fixed sample geometry, K can be calculated for the stress and crack length. The crack length against K data can then be plotted as above to determine K1scc for individual metals.

Several methods of stress application are available. The test described above is a Rising Step Load Test. The load on the samples immersed in solution is increased at regular time periods of between one and four hours. When the load decreases more than 10% then a crack is assumed to have initiated. At this point the load to initiate a crack is known along with the crack length when the crack started to grow. K1scc can then be calculated.

Alternatively, a single cantilever test can be run. In this case a similar sample is loaded at one end. The load is a constant static load. After a given time period, often 1000hrs, the crack is measured to see if it has grown. Several samples at different loads must be run concurrently. K1scc is determined as the minimum load at which crack growth occurs. Again the crack length is the pre-cracked length.

Many other tests are available, such as C ring test and residual strength tests. In these a C shaped sample is loaded and cracks initiation measured. In a residual strength test, a sample is loaded to 75% of the UTS and left in the environment for 30 days. The sample is then fractured and the fracture stress compared to an unexposed sample. The decrease in failure load is a measure of the SCC susceptibility of the metal. Another similar test is the slow strain rate test. A cylindrical sample in an environment is pulled in a tensile test at very slow strain rates. The low strain rate enables the environment to initiate SCC damage. The decrease in mechanical properties between exposed and unexposed samples at the same strain rate is again a measure of the SCC susceptibility. A double cantilever bend test can also be used. The figure below shows the typical sample.

The dimensions are 1 in by 1in by 4 in long. The two bolts threaded into the top are tightened against each other and the material cracked in air. As the crack runs down the sample, the deflection required at the top is met and so the crack then stops running. This then is a measure of K1c for the material. The cracked sample is then placed in the environment of interest and the crack monitored to see if it grows and to what length. After the crack stops growing, the value of K1scc is measured. The value of this is that a single sample provides the measurement of stress corrosion cracking on a quantitative scale.

Effect of Materials Processing on SCC.

It was determined by many researchers that the processing of a material has a significant effect on the stress corrosion resistance of a material. The figure below shows the nomenclature for indicating the orientation of material with respect to processing parameters.

The grain structure after rolling plate or sheet is shown on the figure. The directions indicted as L,S and T are related to the rolling process used to manufacture plate or sheet. The L or longitudinal direction is parallel to the rolling direction. The transverse direction is normal to the longitudinal direction as is the short direction. The T direction is the width of the material while the short direction is the material thickness. For sample orientations the following scheme is used. The first letter indicates the direction which is 90o to the plane of the crack and the second letter indicates the direction of crack growth. The plane shaded would be a crack of the type ST. A sample cut as shown would crack in the ST orientation. This orientation is often the one with the lowest stress corrosion cracking resistance when the intergranular mechanism of SCC is operating. The reason for this is that the ST orientation allows the crack the shortest path through the material completely along grain boundaries with no requirement to propagate transgranularly.

Environments Producing SCC

Aluminum alloys - chlorides, seawater.

Austentitic stainless steels - neutral halides, hot halides, concentrated caustics above 120C.

Ferritic stainless steels - Ammonium nitrates and chlorides, hypochlorites.

Carbon ferritic steels - caustic NaOH solutions greater than 50C, calcium, ammonium and sodium nitrate solutions, seawater.

Nickel chrome alloys - chlorides above 200C, low pH (4) with oxidizers.

Monels - HF acids

Cu-Zn (brasses) - ammonia vapors and salts.

Titanium alloys - organic solvents, methanol.

Prevention

1. Remove or reduce stress - stress relief anneal, thicker sections.

2. Change environment - add inhibitors.

3. Change material - look at data to find better material.

4. Cathodic protection will avoid the problems but remember that hydrogen embrittlement is initiated by cathodic conditions.

CORROSION FATIGUE

Corrosion fatigue is the conjoint action of a cyclic stress and a corrosive environment to decrease the number of cycles to failure in comparison to the life when no corrosion is present. The basic role of the corrosive environment is to decrease the life of the component. Fig 1 indicates the typical data seen for fatigue of ferrous and non ferrous materials. It is a stress against the log of the number of cycles to failure curve, also called an S - N curve. For ferrous materials a distinctive knee is found. For the non ferrous materials and austentitic steels, no knee is found on the curve. When a corrosive environment is present, then the lives to failure often decrease. For the ferrous material, the curve shape is changed as the knee disappears and no distinct threshold stress range exists below which there is no failure.

Features of Fatigue Failure.

Fatigue is a surface initiated process in that the fatigue cracks usually initiate on the exposed surface of a smooth sample. The crack initiation time may be from 25 to 50% of the number of cycles to failure. The initiation stage is called stage 1 in the fatigue process. It is usually accompanied by plastic

deformation at 45o to the applied stress axis as shown in figure 2. These are called permanent slip bands. The crack length in stage 1 fatigue is usually very small.

At the end of stage 1, the crack macroscopically propagates at 90o to the stress axis in stage 11 of the process. At this point multiple slip processes are occurring which lead to blunting of the crack tip and the formation of striations on the fracture surface. The striations are at 90o to the crack growth directions and look like waves on the ocean surface in form. The striations represent one burst of crack growth but may take more than one cycle to form. Striations are characteristic of the fatigue process.The time form initiation until the last cycle is taken up by crack growth in stage 11 of fatigue, so it may take from 75 to 50% of the fatigue life.