INTERGRANULAR STRESS CORROSION CRACKING IN CARBON STEELS

In our line of work, we have seen a wide range of failures over the years, from simple overloads, to complicated, multi-cause, interrelated failures. One of our more frequently encountered failure mechanisms is stress corrosion cracking (SCC). SCC is an environmentally-induced mechanism which occurs in certain materials under the combined influence of tensile stress and a specific chemical agent. A wide range of materials are subject to SCC, from exotic, high-strength titanium alloys, to ordinary, carbon steel.

Intergranular stress corrosion cracking (IGSCC) in carbon steels, sometimes referred to as caustic embrittlement, has been well known for many years. Historically encountered in turn-of-the-century riveted boilers, the cracking in those cases was caused by accumulation of caustic chemicals at the rivet crevices due to small leaks. Although IGSCC can often be avoided by system design and careful control of the process environment, it still occurs, often as the result of upset process conditions or minor design/fabrication details, as the following two examples illustrate.

Low-Pressure Steam Line

After 30 years of uneventful service, a low-pressure steam supply line to a paper machine suddenly developed a recurring problem with through-wall cracks and subsequent leaks at the circumferential butt welds. The steam line was fed from a main header, which operated at 50 psig and 500 ^oF. The steam temperature in the line was lowered to 400 ^oF through a spray atemperator fed with boiler feedwater. The cracking was occurring 100 feet downstream from the atemperator. Repair welds were also cracking, sometimes within a few days of the repair.

A section of the line containing several welds, cracks, and repairs was sent to our laboratory for analysis. Our examination revealed that the cracks had initiated on the steam side (which is not always the case!), in the weld metal and adjacent heat-affected zones. A metallographic evaluation revealed that the cracks were intergranular, with a multiple, branching morphology. This clearly identified the failure mechanism as IGSCC (see photos at top of page).

The cracking was attributed to probable upset conditions in feed- water chemistry, leading to accumulation of caustic chemicals in the steam line downstream of the atemperator. Service stresses plus residual stresses from welding, in combination with the aggressive caustic conditions, resulted in the localized IGSCC at the welds.

Our recommendations were to identify the cracked locations with nondestructive testing and replace the affected sections with new material, using welding procedures which minimize residual stresses. Replacement with the same grade carbon steel material (which would be the most cost-effective choice) would be acceptable, provided the upsets in feedwater chemistry could be avoided in the future.

Carbon Steel Heat Exchanger

A newly fabricated carbon steel heat exchanger failed within a few weeks of service. The heat exchanger was a vertical shell and tube design, used to recover heat from a process fluid to convert feedwater into steam. The process fluid was on the tube side at a temperature of 550 to 600 ^oF; the feedwater was on the shell side and was converted into steam at 375 psig and 440 ^oF.

The heat exchanger shell and tubesheets were fabricated of ASME SA-516, Grade 70, carbon steel and the tubes were ASME SA-214 carbon steel. The tube ends were seal welded at the tube- sheets. Leaks had developed in the bottom tubesheet at the tube seal welds (see photo right). No cracking was found at the tube seal welds at the upper tubesheet.

Metallographic examination of a failed tube end revealed intergranular multiple branching cracks in the seal weld and tubesheet material around the holes. The cracks were characteristic of IGSCC (see photo below).

The presence of a crevice between the tube and the tube- sheet hole was the primary contributory factor for causing the cracking. The crevices contained alkaline mineral deposits (pH greater than 9.0 in this case). The reported startup procedure included use of boiler water and plant process water for heat exchanger input. The crevices tended to concentrate the alkaline mineral deposits, creating the caustic condition which led to the IGSCC.

We recommended that the fabrication procedures be modified to incorporate a combination of rolling and seal welding on the tubesheet joints. The rolling would eliminate the crevice between the tube and hole, thereby preventing accumulation of alkaline deposits. As an additional precaution, we also recommended that the welding procedure include appropriate steps to minimize residual stresses.

Recently, a client called us with a problem: Severe localized corrosion, in the form of deep grooves next to the welds, had reduced the effective wall thickness by more than 50 percent in one of his black liquor evaporators.

The evaporator shell was constructed of carbon steel, while the internal vapor separator components were constructed of Type 304 stainless steel. The severe corrosion damage was confined to the stainless steel material. A stainless steel, welded "tee" section was submitted for metallographic examination.

The tee section was fabricated from 1/8-inch thick plates, with a fillet weld on one side only. The corrosion grooves were confined to the weld heat-affected zones and were parallel to the fillet weld (see photo left).

Our examination showed that the corrosion loss was due to intergranular corrosion; the grain boundaries were corroding away in the weld heat-affected zone, allowing grains of the metal to spall off. No significant corrosion was found in the weld metal itself, or in the base metal beyond the heat-affected zone (see photo below, right).

The deterioration was a classic case of intergranular corrosion caused by sensitization. Sensitization is a condition that can occur in Type 304 austenitic stainless steels when they are heated to temperatures in the 1,000 to 1,200 ^oF range (in this case, as a result of welding). At this temperature, the chromium and carbon in the matrix combine to form chromium carbides, which concentrate along the grain boundaries. With the chromium tied up in the carbides, the surrounding metal is left with a lowered chromium content. This chromium depletion results in a loss of corrosion resistance, which leads to intergranular corrosion upon exposure to the service environment.

We recommended replacement of the Type 304 stainless steel with a material resistant to sensitization. One option was to use the low carbon grade of the material, Type 304L. This material has improved resistance to sensitization, because with its lower carbon content, fewer chromium carbides can form, leaving more of the chromium in the matrix. Another option was to replace the Type 304 with Type 321 or Type 347 stainless steel, which are resistant to sensitization, and offer resistance to pitting and general corrosion comparable to Type 304 stainless steel.

ACCELERATED AGING

Our clients sometimes ask us to predict the useful life of a product they are manufacturing or using. We usually think of our motto "find a way to do it," while at the same time warning that such predictions are difficult to make.

The bottom line is that it is normally much too expensive and time consuming to do a wide spectrum of tests for all the conditions one might encounter. Thus, selecting the appropriate properties and designing the critical tests becomes essential.

In order to accurately predict long-term behavior and time-to- failure, good knowledge of the expected environmental and stress conditions is needed. Only then can an effective program of accelerated tests be developed. In addition, good knowledge of the research and trade literature on the materials is useful; this helps us focus on a few critical properties.

The following examples illustrate this process of sharpening the focus to "zero in" on the critical properties.

A polyvinyl chloride coating was being used on copper cables exposed to a variety of outside environments; our client wanted information on how to optimize the protection afforded by the coating. In this case, our literature review indicated that filling with carbon black was critical in protecting against ultraviolet light exposure, and that particular antioxidants, such as zinc salts, were effective against moisture and oxygen attack. Our review indicated that these materials were already added in sufficient amounts to commercially available plastic stock. Consequently, we were able to focus our investigation on the aspect of maintaining a uniform coating thickness and minimal porosity, rather than on extensive testing for the weatherability of the material.

In another project, our client had asked that we test an encapsulant material for metal under a wide variety of harsh conditions. However, an initial quick test in boiling water, followed by a weekend soak in ambient-temperature water, revealed rust migration to the outside surface. This indicated that cracks had already developed in the encapsulant, probably during the molding process. Thus, the more detailed testing under the harsher conditions was not necessary.

In another case, rubber discoloration appeared early in the expected life of a product; the client asked that we determine the cause. We started with detailed chemical analyses, then proceeded with a wide range of accelerated aging tests, including exposure to high-pressure oxygen and various levels of moisture, light, ozone, and heat. The aging tests and chemical analyses showed that a particular additive, in combination with specific environmental conditions, was responsible for the degradation.

Jeff Link, joined MEI-C in February, 1997 as a senior engineering technician. Jeff is a graduate of Oregon State University with a Bachelor of Science degree in Physics. Jeff has over 10 years experience in empirical stress measurements, strain-gage instrumentation, data acquisition, and PC-based instrumentation.

Shara Sundberg, after eight years of service on our office staff, decided to become a full time homemaker and resigned in February 1997. Shara and Ross Sundberg are expecting their first child in April 1997. Our best wishes to Shara and Ross for a happy parenthood!