Corrosion Resistance of High Alloyed Weldments

S. Schultze*, K. Schilling*, J. Göllner**, G. Posch***

*Landesmaterialprüfamt Sachsen-Anhalt, Germany, Magdeburg **Otto-von-Guericke-Universität, Germany, Magdeburg ***Böhler Schweißtechnik, Austria, Kapfenberg

1 Introduction

The corrosion resistance of high alloyed steels are primarily influenced by the service condition but also by the “history” of the material. The whole chain of manufacturing of the material plays thereby an important role. Microstructural changes, increased internal stresses and surface modifications which especially caused by welding are therefore often responsible. For investigation of their influence on the corrosion resistance standardized corrosion tests but also electrochemical methods can be used. Using practical examples this article explains how these methods can be used for describing the corrosion resistance of weldments and for failure mode analysis.

2 Methods for characterization of corrosion behaviour of high alloyed weldments

Most of the standardized corrosion tests are developed for characterization of rolled material and are used for general statements regarding the corrosion resistance, for checking of the delivery status of semifinished products and for comparison of different base materials. They are also used for investigation of weldments but often without taking into account welding inherited features [1] and separate weld test sample preparation. Additional the transfer of the test results achieved in standardized corrosion tests into real service conditions are in most cases not applicable although test solution and temperatures are comparable [1-4].

2.1 Pitting corrosion testing

For characterization of the pitting corrosion resistance the determination of the Critical Pitting Temperature (CPT) standardized in ASTM G-48 is used [5]. Therefore an Iron-(III)-chloride-solution (FeCl3) is used and the CPT is determined by stepwise temperature increase of 2,5 respectively 5 Kelvin after aging for 24 respectively 72 hours. A new method for a quick, accurate and reproducible determination of the CPT is electrochemical noise measurement. Thereby a continuous temperature increase is applied and the change in charge is measured and calculated from the electrochemical noise [6, 7].

The pitting corrosion – and repassivating behaviour can also be described by establishing of current density-potential-curves and determination of characteristic values like pitting corrosion- and repassivating potential. This can be done by using different electrolytes at different temperatures and polarization velocities.

2.2 Intercrystalline corrosion testing

Various test procedures can be applied for intercrystalline corrosion (IC) testing [811]. These methods (Strauss-test, improved Strauss-test, Streicher-test, Huey-test) are based on aging of the test specimens in an definite test solution at defined test

II-1560-05

II-C-306-05

temperature at holding time. As result the depth of attack after bending of the specimen or the mass lost is measured.

Beside these conventional test methods a method called electrochemical potentiodynamic reactivation (EPR) can be applied. This method was established especially to describe the evidence of IC-susceptibility of rolled austenitic strip [1215]. During this test the potential range between the free corrosion potential and the passivation region will be applied. The method is based on the faster reactivation of voids, for example chromium depleted zones, within the passive layer of the former passivated material. For interpretation of measuring data the current density resulting from the passivation and reactivation loop is used. Additional a metallographic interpretation of the corrosion attack, especially in case of weldments, is necessary.

3 Corrosion resistance of special modified weldments

3.1 Pitting corrosion behaviour of different aftertreated weldments

During welding the formation of tempering colors could not completely prevented. Investigations regarding the influence of variation in the welding procedure on the pitting corrosion resistance of high alloyed steels of type X5CrNiMo17-12-2 (DIN- No. 1.4404) and X1NiCrMoCuN25-20-5 (DIN-No. 1.4539) [16] showed, that for example Argon as a root protecting gas has a beneficial influence on the pitting corrosion behavior, but for best results the complete removal of the tempering colors is necessary. In picture 1 the steel surface after different surface treatments is shown.

Picture 1: light microscopic shots of different treated steel surfaces

Similar effects of an aftertreatment are also noticeable in case of weldments. (Tab. 1.). Best results are achieved by pickling. Brushed surfaces appear metallic clean, but the measured pitting corrosion potentials are the lowest. Grinded surfaces show pitting corrosion potentials between pickled and brushed condition.

CrNi-steel brush 100 Grinded with zirconium- emery abrasive disc 220 Blasted with emery 250 Pickled 400

Tab. 1: Pitting corrosion potential of different aftertreated cap layers of high alloyed joints (base material: X5CrNiMo17-12-2 (1.4404); weld metal: SG-X2 CrNiMo18-16-5 (1.4440, 18/17 E);). Test condition: 0,1 N NaCl, 23 °C

- 2 -

II-1560-05

II-C-306-05

3.2 Corrosion resistance of Superduplex weldments

Within a running research project at the University of Magdeburg the joining of superduplex stainless steel (SDSS) with overalloyed filler materials (stainless steel, nickel base) will be optimized [17]. Also different welding processes like TIG, pulsed MAG and SAW are investigated. For a fast characterization of achieved weldments the CPT was determined using electrochemical noise measurements. It could be shown, that the CPT increased from the first layer to the third layer from 66°C to 76°C

3.3 Resistance against intercrystalline corrosion

Various investigations using the EPR method for assessment of the IC resistance of high alloyed steels showed also, that this method is also capable to descripe the sensitization of the material [18-24]. In case of weldments the influence of the thermal weld heat cycle on the microstructure especially in the heat affected zone and in further consequence on the IC resistance has to be considered. EPR measurements do not distinguish between base material, heat affected zone and weld metal. Therefore the measured current densities have to be critical reviewed and an additional metallographic evaluation has to be applied, because an enforced attack within the heat affected zone can not be dedected by using only EPR. Investigations on weldments of base material type X6CrNiTi18-10 (Din-No. 1.4541) using non-stabilized filler material and aged at 350°C for one year in H2-containing agent showed at the first run no different IC-resistance compared to a non-aged weldments. But it was obvious, that an increased selective corrosion process within the ferrite took place and an increased current density was measured. Light microscopic examination after EPR showed also, that especially the weld metal had a different solution behavior compared to the base material. This could also be documented by using atomic force microscopy (AFM) (picture 2).

.

Base material

Weld metal

Picture2: ex situ AFM-shot of dissolution near the fusion line of the weldment in the root pass after EPR testing. Stabilized base material ( X6CrNiTi18-10) non-stabilized filler (E-19 9 B 20)

After EPR-testing there is a difference in elevation between base material and weld. Due to the selective corrosion process within the weld, the surface appears compared to the base material totally fissured and the contribution of the weld to the measured current density is dominant.

II-1560-05

II-C-306-05

4 Application of the methods for failure analysis

4.1 Local corrosion on stainless steel tubes

Tubes made of duplex stainless steel (DSS) X2CrNiMoN22-5-3 (DIN-No. 1.4462) showed severe local corrosion attack within few days in service. The tubes are used for transportation of a high chloride containing agent at 30°C, pressure of 20 bar and a flow rate of 150 m³/h. The chemical composition of the tube material was within the range described in related standards. Also the ferrite content (46%) was within the specified limits. For failure analysis current density-potential curves using the correspondent agent were established for describing the pitting corrosion resistance and the repassivation behaviour. Additional specimens of the corroded tubes were taken out and compared with reference specimens of the same base material. The investigation pointed out a more sensitive tube material compared to the reference material. The potential of the ascent of the current density-potential curve of the tube material was much more than 100 mV lower than the comparable potential of the reference material (Tab. 2).

Tab. 2: Pitting corrosion potential of tube material compared to reference material; same agent as in real condition; temperature: 30°C

Tube material as delivered	300
Tube material grinded	620
Reference material	1250

Applying metallographic and microanalytic investigations dark precipitates on ferrite/austenite grain boundaries within the sensitive base material, mainly aluminium-silicon-oxides were found. In contrast no evidence for such precipitates was detected in the weldment and from optical inspection it was obvious that no corrosion attack took place within the weld.

4.2 Pitting corrosion caused by partially removed oxide layers

Austenitic containers made of X5CrNi18-10 (1.4301) showed severe pitting corrosion near weldments after pressure testing using drinking water (picture 3). According to the specification 2 layer TIG-weldments with steel brush cleaning afterwards have to be established.

Picture 3: Corrosion of a weldment after contact with drinking water

II-1560-05

II-C-306-05

Picture 4a shows a closer look to a corroded weldment and points out the location from which a cross section was taken. Near the borderline between weldment and base material various pits were detected. In picture 4b a metallographic prepared cross section through a pit is shown. Except of insufficient penetration in the root area no further microstructural defects such as carbide formation within the heat affected zone caused by overheating were detected. Also the chemistry of the base material and the drinking water (chloride concentration: 60 mg/l) showed no evidence for possible corrosion susceptibility.

a) b)

Picture 4: macro picture (a) and macrosection (b) of a specimen taken out from the container

One detail of the weld surface is shown in picture 5. Within the heat affected zone a thick, dark oxide layer is visible, which wasn´t removed by brushing. The thickness of the layer is about 100 µm.

Picture 5: brushed weldment with an oxide layer in the heat affected zone

The corrosion was caused by:

• Insufficient removal of oxide layers (tempering colours from yellow to blue (thickness between 10 – 100 nm) and thick, dark gray layers with thichnesses up to 100 µm. Such layers prevent the formation of uniform passive layers, which can prevent corrosion and crevice corrosion can take place.
• Impurities of iron particles on the surface, caused by using wrong steel brushes and other soiling
• Stagnation of the water within the container over weeks. Such conditions can cause pitting corrosion, especially crevice corrosion under thick oxide layers

4.3 Ferrite corrosion within a tank in the food industry

A sparkling water tank made of X5CrNi18-10 (DIN-No. 1.4301) leaked after 4 months in use near horizontal weldments. These weldments were established using TIG with a filler of type X2CrNi19-9 (DIN-No. 1.4316). The chemical composition of base metal and filler material were within the specification, but metallographic investigations

II-1560-05

II-C-306-05

(picture 6) pointed out some selective corrosion processes of the ferritic phase (dark regions in picture 6) within the transition region between the weldment and a pit. Detailed investigations pointed out that this corrosion process was enabled by transforming the ferrite into sigma phase due to the weld heat input.

Picture 6: microstructure, etched using V2Aetchand, M 200 : 1, dissolution of delta ferrite (dark)

Due to the microstructural changes local chromium depletion occurs with a significant drop in the corrosion resistance. The high chloride content of the sparkling water (450 mg/l) caused in further consequence corrosion attack. This corrosion could have been prevented only by an improved material selection.