Artificial Pit Experiments to Investigate the Growth and Repassivation of Macroscopic Corrosion Pits in Stainless Steels
Tero J. Hakkarainen and Pekka Pohjanne
VTT Manufacturing Technology, Finland
email: Tero.Hakkarainen@vtt.fi
Two different artificial pit configurations were used to investigate the conditions required for continued growth of macroscopic corrosion pits in stainless steels. The growth of the pits was activated to reproduce the conditions in growing 'natural' pits in oxidizing environments. It is demonstrated that the effects of the main variables affecting the conditions for growth and repassivation of open pits can be investigated quantitatively, including e.g. temperature, electrode potentials and composition of the bulk solution. It is concluded that for continued growth of corrosion pits with a large opening to the bulk solution, a strongly oxidizing environment is required and that sulfate ions are fairly inefficient in stopping the growth of propagating pits.
Before corrosion pits lead to failures, they usually must grow to macroscopic dimensions, i.e. their depth and diameter must be of the order of millimeters. The overwhelming majority of investigations on pitting of stainless steels is, however, related to the initiation of pits or to their eventual repassivation when they still are very small. Less attention has been paid to the requirements for growing macroscopic pits to repassivate. One reason to this may be that reproducible controlled macroscopic pits are difficult to produce. Also, experiments on pitting of stainless steels are usually conducted in solutions where chloride ions are the predominant anions, e.g. in plain NaCl solutions or in sea water. Thus, detailed information is lacking on e.g. the effects of sulfate ions in the bulk solution on pit growth and repassivation when their amount is comparable to or exceeds that of chloride ions.
In the following, experiments are described, where two artificial pit configurations are used to investigate the effects of some of the main variables on the possibility of continued growth of macroscopic pits.
Pit with inert walls
One artificial pit configuration ('type A') consisted of a cylindrical cavity with inert walls and a flat bottom formed by the steel sheet under investigation, Fig. 1. Within such pits, the dissolution during stable pit growth is expected to be uniform over the entire surface exposed. Thus, true current densities at the dissolving steel surface can be measured and correlated with e.g. temperature, control potential or estimated composition of the pit solution at the dissolving surface. The diameter of the these pits was 2 mm, and their original depth 2.5 mm.
Figure 1. Artificial pit configuration of 'type A', pit with
inert walls.
Drilled pit
The other pit configuration ('type B') was a small drilled hole on a horizontal specimen surface, Fig. 2. The original diameter of the these pits was 0.5 mm and their original depth about 0.4 mm. The diameter of the specimen discs was 15 mm and their thickness 2...4 mm. This configuration differs from the fist one in some important aspects. Firstly, the prevalent current density over the dissolving metal surface will probably not be uniform, because the walls of the pit will also be attacked and the current density is expected to depend on the distance from the pit mouth. Secondly, in this arrangement an area of open surface of the specimen is exposed to the bulk solution. Thirdly, the original volume of type B pits is only about 1/100 of the volume of type A pits.
Figure 2. Artificial pit configuration of 'type B', drilled
pit.
At the beginning of each experiment, pit growth was activated to simulate the case of real growing pits. One way to do this was to inject a small amount of 'simulated pit solution' into the pit when the specimen was immersed into the bulk solution and was anodically polarized to a high potential (+800 mVAg/AgCl). The simulated pit solution was prepared by dissolving about 50 g of type AISI 316L stainless steel into 200 ml of 10 M HCl, which resulted in a practically saturated (at room temperature) metal chloride solution with metal cations in the same ratios as in the steel [1]. This activation method was used mainly for the pits with inert walls (type A).
Another method was to fill the pit with chromic chloride crystals (CrCl3·6H2O) before immersing the specimen into the bulk solution. The amount of chromic chloride is enough to fill the pit volume with about 3.5 M CrCl3 solution. In this case, the specimen was also anodically polarized to +800 mVAg/AgCl immediately after immersing. This method was applied for the drilled pits (type B).
After activation, the anodic pit current was allowed to 'stabilize' before eventually starting to change the value of a test variable.
Materials and procedures
Two austenitic stainless steels were used in the experiments. One was of type UNS S31254, with 20.0% Cr, 21.5% Ni, 6.2% Mo, 0.73% Cu, 0.22% N and 0.018% C. The other was of type AISI 316L, with 16.9% Cr, 10.7% Ni, 2.55% Mo and 0.03% C. Both steels were in the form of 2 mm sheet.
All experiments were started with an activation of the pit and a stabilizing period for pit growth at 70°C and +800 mVAg/AgCl. The high anodic activation potential was chosen to produce initially a saturation current density during active growth of type A pits. The variables in the tests included the temperature, the electrode potential (controlled with a potentiostat and a reference electrode outside the pit, at a distance of about 5 mm from the pit opening), and the composition of the bulk solution. All isothermal tests were made at 70°C, the tests with continuously decreasing temperature in the interval 70...25°C. When investigating the effects of the electrode potential, the potential was either decreased at a constant rate from an initial value of +800 mVAg/AgCl towards less noble potentials or switched to some predetermined constant potential.
In most experiments, the bulk solution was either a plain 0.2 M NaCl solution (7100 mg/L Cl-), a plain 0.13 M NaSO4 solution (12800 mg/L SO42-), or a mixture of these two solutions. In some of the experiments, the chloride ions in the bulk solution were partially substituted by sulfate ions during the experiment by a slow removal of part of the test solution and an equally slow addition of sulfate solution, repeating this several times if necessary to produce the wanted sulfate to chloride ratio. The relatively high concentrations of chloride and sulfate ions were used in order to keep the potential drop (IR-drop) in the bulk solution at reasonable values. The sulfate content of the sulfate solution was chosen to result in an equal electrical conductivity to that of the plain chloride solution, rather than in equal molarity. In some experiments, also chloride solutions with either higher (0.5 M NaCl, 17750 mg/L Cl-)or lower (0.1 M NaCl) chloride contents were used.
Pit activation
In the case of type A pits (cf. Fig. 1), the pit activation period was essentially similar both for pit activation by simulated pit solution and for pit activation by chromic chloride crystals. It took 20 to 40 minutes for the anodic current to become stabilized to a constant value of about 3 mA. For the two steels, the initial transient period was somewhat different, as shown in Fig. 3, but the final steady state current value was the same for both steels. In the case of steel AISI 316L, the current increased almost immediately to a peak value of several milliamperes, then decreased abruptly to a value of the order of 1 to 2 mA, and increased again slowly to the steady state value. For steel UNS S31254, the initial current increase to a peak value was much slower, and the steady state value was reached almost immediately after the peak. The pit activation behavior of type A pits was not affected by the composition of the bulk solution (i.e. chloride or sulfate).
An inspection of the specimens used in type A pits after the activation period revealed a very even attack of the entire exposed surface of steel AISI 316L, but shallow 'dimples' close to the border of the exposed area (the PTFE 'ring') in steel UNS S31254.
Figure 3. Examples of initial transients and stabilization
of pit growth of type A pits (inert walls, cf. Fig. 1) when activated
by the simulated pit solution in 0.2 M NaCl bulk solution at 70°C
and +800 mVAg/AgCl..
The activation process for type B pits (cf. Fig. 2) was faster and less reproducible than that for type A pits. Also, in this case no constant steady state current value was observed for either steel. The current values were typically of the order 2 to 4 mA almost immediately after the start of the polarization, in the range 2 to 3 mA during the first 5 to 10 minutes, and could approach 10 mA in less than 1 hour. Drilled pits in steel UNS S31254 could be properly activated in the plain chloride solution and in a solution with a sulfate to chloride molar ratio of 0.67, but only occasionally in a solution with a sulfate to chloride ratio of 2, and not at all with higher sulfate to chloride ratios. In contrast, drilled pits in steel AISI 316L could be activated for a period of 20 to 30 minutes even in the plain sulfate bulk solution.
Effect of control potential
When the control potential of type A pits was lowered from the initial activation towards less noble potentials during stable pit growth, the current remained constant to lower potentials in bulk solutions with higher electrical conductivity (ion content), than in bulk solutions with lower the electrical conductivity Fig. 4. Experiments where the potential was switched after the activation to a constant lower value showed that at a current of about 2.5 mA the pit growth in the steel UNS S31254 remained approximately constant until the end of the 2 hour test period, whereas an 'initial' current of 2.0 mA remained stable only 25 minutes before starting to decrease rapidly. For steel AISI 316L, a current of about 1.5 mA started to decrease at an increasing rate after about 1 hour, whereas a current of about 2.3 mA survived the 2 hour test periods without a trend to decrease markedly.
Figure 4. Examples of polarization curves for type A pits (inert walls, cf. Fig. 1) in plain chloride bulk solutions. Steel AISI 316L. Potential scan rate -1 mV/s.
1) 0.1 M NaCl bulk solution (3550 mg/L Cl- )
2) 0.5 M NaCl bulk solution (17750 mg/L Cl- )
On a continuous potential scan towards less noble potentials, the current for steel UNS S31254 decreased to small values (< 100 A) at a control potential of about -100 mVAg/AgCl when the scan rate was -60 mV/min or higher, and at about 0 mVAg/AgCl when the scan rate was -6 mV/min. For steel AISI 316L the corresponding potentials were about -150 mVAg/AgCl regardless of the scan rate.
Immediately after the potential control was switched off during steady state pit growth, the open circuit potential of type A pits (pit bottom potential) was about -100 mVAg/AgCl in the case of steel UNS S31254 and about -150 mVAg/AgCl in the case of steel AISI 316L.
Effect of temperature
When the temperature was gradually lowered during steady state pit growth in the 0.2 M NaCl solution, an abrupt decrease of the current was observed for steel UNS S31254 in the temperature range 52 to 54°C, Fig. 5. For steel AISI 316L, no such drastic current decrease was seen in tests where the final temperature was about 25°C. Instead, the current from the type A pit decreased gradually from the value 3 mA at 70°C to about 1 mA at 25°C.
In the specimen of curve A in Fig. 5, crevice attack on the specimen
surface below the PTFE holder was observed after the test.
Figure 5. Effect of decreasing temperature on the anodic current for the steel UNS S31254. Anodic polarization to +800 mVAg/AgCl , rate of temperature decrease about 15°C/h.
A) type A pit
B) type B pit.
Effect of bulk solution
For pits of type A ('inert walls', Fig. 1) in tests with a total duration up to several hours, the composition of the bulk solution had a minor effect only on either the steady state current value or the polarization behavior. In an experiment where the pit was first activated for 1 hour at +800 mVAg/AgCl in plain sulfate bulk solution and the control potential was then switched to +250 mVAg/AgCl , the current for steel UNS S31254 decreased gradually in about 3 hours from the 'initial' value of 2.4 mA to about 1 mA. In a similar experiment in plain chloride bulk solution, the current remained at the level 2.6 to 2.8 mA for the entire test duration of 2 hours at +250 mVAg/AgCl. The polarization curves from +800 mVAg/AgCl to -300 mVAg/AgCl at a rate of -1 mV/s were very similar in both plain chloride and plain sulfate solutions.
Type B pits (drilled holes, Fig. 2) in steel AISI 316L remained active in the plain sulfate bulk solution for a short period only. In a bulk solution with a sulfate to chloride ratio 40 (116 mg/L Cl- + 12600 mg/L SO42-), however, the pit remained active (current about 2 mA) until the end of the 6 hour test period, Fig. 6. At the end of the latter test, the volume of the pit was about 20 times its original volume, as calculated from the integrated current passed during the test.
Figure 6. Results from isothermal constant potential tests on type B pits in steel AISI 316L in two different bulk solutions. 70°C , +800 mVAg/AgCl .
Sulfate: 0.13 M NaSO4 (12800 mg/L SO42- )
Sulfate/chloride = 40: 116 mg/L Cl- + 12600 mg/L
SO42-
Because of difficulties in pit activation at high sulfate to chloride ratios of the bulk solution, the experiments on drilled pits in steel UNS S31254 had to be started in bulk solutions where the amount of chloride anions was comparable to or exceeded that of sulfate anions. The sulfate to chloride ratio was increased by slowly changing part of the bulk solution during the test. With this method, plain sulfate bulk solution was not attained in the tests made. In one test, started in plain chloride solution, the sulfate to chloride ratio was increased to 10 in about 90 minutes. At the end of the 5 hour test, the pit was still actively growing at a current of about 3.6 mA. By the time the sulfate to chloride ratio of the bulk solution had reached the value 10, the pit volume had increased to a value 12 times its original size and at the end of the test the pit size was 29 times the original one.
Composition of pit solution
The solution inside a growing chloride induced corrosion pit in stainless steels is known to be a concentrated metal chloride solution [e.g. 2-5]. As a saturated chloride solution, the metal cations occupy a volume about 30 times larger than as solid steel, i.e. at least 30 times more metal cations are being produced by the dissolution reaction than can possibly remain inside the pit at a steady state pit growth.
In the simulated pit solution used to activate the type A pits, the relative contents of the metal cations are expected to be 'correct' from the very beginning of the experiments for the AISI 316L steel. In the 2.5 mm deep pit, the dissolution of a steel layer less than 0.1 mm thick will produce enough metal cations to fill the entire pit with a saturated metal chloride solution. At the steady state current of 3 mA the current density in the type A pit is about 100 mA/cm2, which means that 0.1 mm steel will be dissolved in less than 50 minutes. Since it is reasonable to expect that due to a concentration gradient within the pit the average concentration is roughly one half only of the saturation value, and since the cations are being produced at the metal surface in the correct ratios, it can be presumed that when the pit current has been stabilized the composition of the pit solution in the vicinity of the steel surface will also for the steel UNS S31254 correspond the one in 'real' pits. A sufficient, possibly even an excess, amount of chloride anions is added to the pit during the activation procedure.
The activation behavior (cf. Fig. 3) of the pits and the profiles of the pit bottoms after the activation stage suggest that in type A pits the entire exposed surface of steel AISI 316L has been activated practically immediately, whereas the attack of steel UNS S31254 has started as separate small pits, the growth of which has produced a solution able to activate the rest of the pit bottom.
The initial volume of type B pits will have increased by 20% when a charge of about 0.13 mAh (8 mAmin) has passed. In typical experiments, this took less than 3 minutes. The amount of steel dissolved is then enough to fill the pit volume about 10 times with new cations. Thus it is reasonable to expect the relative cation contents in the pit to be correct already from very early stages of the experiments.
Even when the pit volume is multiplied during an experiment in a bulk solution with a high sulfate to chloride ratio, the ratio of sulfate to chloride anions in the pit may not fully correspond to that in the bulk solution. A flux of anions into the pit will be needed to retain electroneutrality in the increasing pit volume. This flux is, however, significantly lower than the flux of cations out of the pit. Mass transport properties of the various anions will determine whether the 'original' chloride ions tend to stay within the pits indefinitely.
Nevertheless, it is evident that at advanced stages of the experiments made in bulk solutions with high sulfate to chloride ratios, the amount of sulfate ions must have greatly exceeded that of chloride ions inside the pits of type B.
Potentials and IR-drops
Results like those in Fig. 4 show that the initial control potential during the activation of type A pits has been high enough to result in a saturation value of the steady state current. When this is the case, the current value is independent of the control potential. The results also indicate that significant IR-drops have been present in all experiments with current values of the order of milliamperes. The solution composition, and thus also the electrical conductivity and the IR-drop, inside the pit at the same saturation current value are not expected to be significantly affected by the chloride content of the bulk solution within the range used in the experiments. The variations and the main part of the potential drops observed at the constant saturation current values must therefore have their origin outside the pit.
An IR-drop correction to the results can be made to give an approximately constant 'true' value for the transition potential from current saturation to 'activation control' for each steel, independent of the electrical conductivity of the bulk solution. This corrected potential will represent the true potential of the pit bottom. For the results in Fig. 4 the R values for an IR-drop correction would be of the order 70 to 80 ohms for the 0.5 M NaCl bulk solution and about 160 ohm for the 0.1 M NaCl bulk solution.
Temperature
The results (cf. Fig. 5) indicate that there exists a temperature, specific for each steel, below which the very rapid dissolution required for the growth of open pits is no more possible. Although experiments at different control potentials and with different chloride contents of the bulk solution to confirm this have not yet been made, it is believed that the transition temperature is independent of these variables, provided their values are large enough not to restrict the possibility of pit growth for other reasons. For the steel UNS S31254 this temperature is about 53°C, for the steel AISI 316L it is below room temperature.
A continued anodic current flow at temperatures below the rapid decrease at the transition temperature shows that crevice corrosion can take place at temperatures substantially lower than proper pitting (cf. Fig. 5)
Practical implications
It can be concluded from the results that very high anodic current densities at the dissolving pit surface are required for continued growth of open macroscopic corrosion pits. For large pits this implies high total currents and large potential drops (IR-drops) in the bulk solution outside but in the vicinity of the pit, even in solutions of fairly high electrical conductivity (high content of ions). A rapid growth of large pits, leading to perforations in a very short times, is therefore expected to be possible in strongly oxidizing environments only.
On the other hand, pitting probably can proceed at a slower rate and in less strongly oxidizing environments in occluded cavities, out of which the ion transport is essentially slower than from pits with large openings.
Once chloride induced pitting has started, it may under otherwise favorable conditions continue even when the sulfate content of the bulk solution is made an order of magnitude larger than that of chloride. A plain sulfate bulk solution will, however, not support a continued growth of pits.
The artificial pit configuration of type A can be used to investigate quantitatively the dependence of the current density inside growing open pits on potential and temperature. This pit configuration is not, however, suitable for investigating the effects of changes in the bulk solution during pit growth.
The effects of the composition of the bulk solution and the temperature on the growth and repassivation of chloride induced open pits can be quantitatively investigated using the artificial pit configuration of type B.
For continued growth of corrosion pits with a large opening to the bulk solution, a strongly oxidizing environment is required.
Sulfate ions are fairly inefficient in stopping the growth of propagating pits.
1. T. Hakkarainen, "Anodic Behaviour of Stainless Steels
in Simulated Pit Solutions". In Corrosion Chemistry within
Pits, Crevices and Cracks, EA. Turnbull, Editor, HSMO, London
1987, 17-26.
2. T. Suzuki, M. Yamabe, and Y. Kitamura, "Composition of Anolyte Within Pit Anode of Austenitic Stainless Steels in Chloride Solution". Corrosion, 1973, 29(1), 18-22.
3. J. Mankowski and Z. Szklarska-Smialowska, "Studies on
Accumulation of Chloride Ions in Pits Growing During Anodic Polarization".
Corros. Sci., 1975, 15, 493-501.
4. T. Tsuru, K. Hashimoto, A. Nishikata and S. Haruyama, "Mass
Transport and Chloride Ion Complexes in Occluded Cell". Materials
Sci. Forum, 1989, 44&45, 289-298.
5. H.S. Isaacs, J.-H. Ho, M.L. Rivers and S.R. Sutton, "In-Situ
X-Ray Microprobe Study of Salt Layers during Anodic Dissolution
of Stainless Steel in Chloride Solution". J. Electrochem.
Soc., 1995, 142(4), 1111-1118.