An investigation was conducted to gain insight on the fracture resistance of carbon steels used in refinery service containing stress oriented hydrogen induced cracking (SOHIC) damage. The investigation included: (1) the development of SOHIC damage in the base metal and heat affected zone of both a conventional and “HIC Resistant” grade of ASTM A516-70 steel, (2) fracture resistance testing of the produced defects after outgassing, and (3) comparison of the data and resulting fracture surface morphologies between the two steels. The depth and character of the SOHIC attack observed on the four systems as a result of the environmental exposure indicated the base metal region of “HIC Resistant” steels to be the most susceptible, followed by the HAZ region of conventional steels. CTOD testing indicated similar toughness of the SOHIC damaged material to a conventional fatigue precracked sample in the absence of hydrogen charging. The fracture surface morphology of the SOHIC was different on the two steels and its appearance was related to differences in microstructure.

Keywords:fracture toughness, KIc, crack tip opening displacement, CTOD, fitness-for-service, carbon steel, hydrogen induced cracking, HIC, stress oriented hydrogen induced cracking, SOHIC, sulfide stress cracking, SSC, sour service, hydrogen sulfide, H2S.

INTRODUCTION  Top

Much of the equipment in petroleum operations is exposed to aqueous process environments containing H2S. A large percentage of this equipment is fabricated from carbon steel and can exhibit susceptibility to wet H2S cracking. In 1984, rupture of an amine absorber pressure vessel at a refinery near Chicago, Illinois led to increased industry concern regarding the wet H2S cracking of carbon steel equipment in refinery service [1]. Failure was attributed to a combination of sulfide stress cracking (SSC) at the toe of a circumferential weld and stress oriented hydrogen induced cracking (SOHIC) which initiated at the tip of the SSC crack and propagated along the fine grained heat affected zone (HAZ) region adjacent to the base metal in the through-thickness direction.

Resistance to SOHIC is of paramount importance. Shallow surface flaws produced by SSC or other defects which possess an inherent stress concentration are common. With adequate SOHIC resistance, these cracks are unlikely to grow to a size which will cause leakage or vessel rupture. As in the incident described above, SOHIC has been more commonly observed in the fine grained HAZ region adjacent to the base metal. This has been observed in the field and in the laboratory on conventional low strength carbon steels used in refinery service and in the laboratory on the advanced (“HIC Resistant”) steels. However, SOHIC susceptibility of “HIC Resistant” steels has not been restricted to the weldments. These steels have also demonstrated SOHIC susceptibility in the base metal [2, 3]. It has been shown that decreases in sulfur content, while decreasing susceptibility to HIC, may result in an increase in the susceptibility to SOHIC [4].

Little to no data is available in the literature governing the fracture resistance of low strength refinery plate steels containing SOHIC damage. Furthermore, toughness data which has been developed on refinery plate steels has typically been limited to LT or TL orientations. SOHIC by definition, is a mechanism of cracking which orients itself and propagates in the through-thickness direction (TS or LS orientations). Propagation of through-wall cracks in the LT or TL directions via SOHIC appears unlikely. Hence, the goal of this program was to obtain information regarding the fracture toughness of refinery steels containing SOHIC damage with propagation in the through-thickness direction.

EXPERIMENTAL PROCEDURES

Materials

Two materials were chosen for evaluation in this study. A conventional ASTM A516-70 steel, to represent existing refinery equipment, and a “HIC Resistant” ASTM A516-70 steel, to represent material which might be used in new construction or repair of existing equipment. The “HIC Resistant” was processed for improved resistance to HIC (i.e. ultra-low sulfur and calcium treated for sulfide shape control). The chemical compositions and mechanical properties of the materials are provided in Tables 1 and 2, respectively.

The microstructures of the materials are presented in Figure 1. The structure of both materials consist of pearlite in a ferrite matrix. Microstructural banding is present in both materials but is far more pronounced in the conventional steel. The plate thickness of both materials was 12.7 mm (0.5 inch).

TABLE 1. CHEMICAL COMPOSITIONS

TABLE 2. MECHANICAL PROPERTIES

Conventional A516-70 (200 X)                      “HIC Resistant” A516-70 (200 X)

FIGURE 1 - Microstructures of the two A516-70 steels evaluated in this program.

Welding

In order to assess the susceptibility and subsequent SOHIC fracture toughness of the HAZ in both materials, samples from each material were welded according to the parameters detailed in Table 3. A welding joint geometry was particularly chosen to produce an HAZ which was oriented parallel with the through thickness direction. A schematic of the joint geometry is provided in Figure 2.

TABLE 3. WELDING PARAMETERS

FIGURE 2 - Butt weld joint geometry utilized in this study.

Double-Beam Sample Preparation and Creation of SOHIC Damage

Double-beam (DB) specimens, modified by the addition of an EDM notch, were used to produce the pre-existing SOHIC damage which was further characterized in this program. The details concerning the use of the DB method have been published by Kane, Cayard and Prager [5]. NACE International has also recently assigned a task group (T-8-23) charged with the standardization of a SOHIC test method which most likely will include the use of the DB test methodology. A schematic of the notched DB specimen is shown in Figure 3. The overall DB length was 30.5 cm (12.0 inch) by 12.7 mm (0.5 inch) thick by 3.8 cm (1.5 inch) wide.

Duplicate DB specimens were machined from the base metal and across the weld from both materials. The DB specimens were oriented such that the length dimension was perpendicular to the plate rolling direction. With this orientation, the applied stress was perpendicular to both the welding and plate rolling directions. This orientation was found to exhibit the most susceptibility to cracking in prior studies [5, 6]. To promote the initiation of SOHIC damage during test, EDM slots were machined across each sample on the tension face 2.2 mm (0.085 inch) deep using a 0.2 mm (0.008 inch) diameter wire. The resulting root radius was 0.13 mm (0.005 inch). In the case of the welded specimens, the slot was centered in the HAZ region.

The DB specimens were loaded to a deflection which would produce an outer fiber stress of 236 MPa (34.2 ksi) on a smooth beam. This stress level corresponded to 90 percent of the specified minimum yield strength of ASTM A516-70. A high stress level was chosen to increase the likelihood of producing the SOHIC environmental flaws. The stressed specimens were then fully immersed for seven days in NACE Standard TM0177-90, Method A [7] solution consisting of 5 percent NaCl, 0.5 percent glacial acetic acid dissolved in distilled water and saturated with H2S gas at one atmosphere and 23 C (75 F). The pH of this solution was approximately 3.3.

Following the seven day exposure, two single-edge-notched three-point bend (SENB3) fracture toughness specimens were sectioned from each beam as shown in Figure 4, for a total of four specimens per condition. The dimensions chosen corresponded to the square section specimen detailed in ASTM E1290 [8] (12.7 mm x 12.7 mm by 7.62 cm [0.5 inch by 0.5 inch by 3.0 inch]). The sides of each specimen were polished to reveal the extent of SOHIC propagation which occurred during the pre-exposure. Triplicate baseline specimens from the base metal and HAZ of each material were also machined for comparison purposes.

FIGURE 3 - Double-beam specimen design chosen to produce the SOHIC damage.

FIGURE 4 - Sectioning of SENB3 specimens from DB specimen.

CTOD Toughness Testing

The baseline specimens from the base metal and weldment were fatigue precracked and tested per the specifications of ASTM E1290. The crosshead displacement rate was 1x10-3 mm/sec. The CTOD testing of the SOHIC damaged specimens was also conducted in accordance with ASTM E1290, except no fatigue precracking was performed. The cracks evaluated corresponded only to the SOHIC damage produced during the pre-exposure. All testing was conducted in air at ambient temperature. In the case of the specimens exposed to the sour environment, the mobile hydrogen was allowed to outgass at ambient temperature for a minimum of 7 days prior to testing. Hence, the toughness values reported represent non-hydrogen charged values. The toughness of hydrogen charged material could be substantially reduced.

RESULTS AND DISCUSSION

Results of the SOHIC Pre-Exposure

SOHIC was produced to varying degrees in the base metal and HAZ specimens from both materials as a result of the pre-exposure. Photomicrographs of the SOHIC produced in the base metal of the conventional and “HIC Resistant” steels are shown in Figure 5 and 6, respectively. Both materials exhibited short in-plane blister cracks ahead of the EDM notch which linked-up in the through thickness direction. However, the physical link-up of the small cracks had some subtle differences. Link-up in the “HIC Resistant” steel occurred on a nearly straight path in the through-thickness direction, whereas the link-up in the conventional steel occurred on a jagged path with discrete steps equivalent to the manganese-pearlite band spacing. It appeared from these differences that the link-up occurred in the ferrite phase. Since the “HIC Resistant” steel was less banded, it possessed a more continuous path for the link-up of the SOHIC.

Photomicrographs of the SOHIC in the HAZ region of the conventional and “HIC Resistant” steels are shown in Figure 7 and 8, respectively. Similar features discussed above were observed in the HAZ region of the two steels. The jagged nature of the SOHIC extension in the conventional steel is particularly evident in Figure 7. The path of the SOHIC extension in the HAZ of the conventional steel was also interesting. It tended to work itself into the fine grained HAZ structure adjacent to the base metal as opposed to the coarse grained region adjacent to the weld.

The physical depth of SOHIC damage was measured on each of the exposed specimens after the fracture toughness testing. The extent of damage was broken into three distinct regions including (1) the EDM notch depth, (2) the continuous SOHIC ligament (i.e. fully connected) and (3) the intermittent SOHIC ligament. The three distinct regions are shown schematically in Figure 9. The results of these measurements are presented graphically in Figure 10. The two systems which developed the largest amount of continuous SOHIC extension were the “HIC Resistant” steel / base metal and conventional steel / HAZ combinations. The continuous SOHIC extension in the base metal of the conventional steel and HAZ region of the “HIC Resistant” steel was limited to less than 0.5 mm (0.020 inch).

Based on the depths of SOHIC observed, it appeared the most susceptible region for SOHIC in conventional steels is in the HAZ region of weldments, particularly the fine grained HAZ region adjacent to the base metal. In contrast, the “HIC Resistant” steel exhibited a higher susceptibility to SOHIC in the base metal. This same behavior was observed in a previous study by Kane et al [2]. In another study [6], propagation of SOHIC from an SSC toe crack in a fillet weld was found to be restricted to the HAZ in conventional A516-70 steels (i.e. arresting in the underlying base metal) whereas the SOHIC propagated past the extent of the HAZ into the underlying base metal of the “HIC Resistant” steels.

The implication of these data are significant. In some cases, “HIC Resistant” steel may actually pose a greater threat to overall equipment integrity than conventional steel due to its increased susceptibility to SOHIC. Furthermore, it may be appropriate to consider the varying susceptibility of different steels to base metal or weldment cracking when developing equipment inspection plans.

FIGURE 5 - Resulting SOHIC produced in the base metal of the conventional A516-70 steel. (50 X)

FIGURE 6 - Resulting SOHIC produced in the base metal of the “HIC Resistant” A516-70 steel. (50 X)

FIGURE 7 - Resulting SOHIC produced in the HAZ region of the conventional A516-70 steel. (50 X)

FIGURE 8 - Resulting SOHIC produced in the HAZ region of the “HIC Resistant” A516-70 steel. (50 X)

FIGURE 9 - Three distinct ligaments which make up the total crack depth.

FIGURE 10 - Notch, continuous and intermittent ligament depths as a result of the pre-exposure.

CTOD Fracture Toughness

The results of the baseline fracture toughness of the four systems evaluated are shown in Figure 11. Note these values represent crack growth in the through-thickness direction, or short-transverse (ST) orientation. Toughness values in the transverse-longitudinal (TL) or longitudinal-transverse (LT) orientations can be much lower. The toughness of the “HIC Resistant” steel was significantly higher than the conventional steel and related to the reduced carbon, phosphorous and sulfur levels of the “HIC Resistant” steel. The toughness of the HAZ region of the conventional and “HIC Resistant” steels was reduced to 65 and 85 percent of the base metal toughness, respectively.

Based on the depth of cracking which resulted from the environmental exposure of the DB specimens, the only material / condition combination of the four evaluated which consistently produced a deep enough crack to yield a valid fracture toughness measurement was the base metal of the “HIC Resistant” steel. The results are presented in Figure 12. Note the comparison relates to the behavior or fracture toughness of fatigue precracked baseline material to material containing SOHIC environmental flaws. By definition, the SOHIC flaw is not completely connected and still possesses an intermittent ligament as previously discussed (see Figure 9). The measured toughness of the material with SOHIC damage was slightly higher than the baseline values by approximately 8 percent.

The increase in toughness associated with the SOHIC flaw can be explained by observing the behavior of the load versus crack opening deflection curves characteristic of each of the two data sets (see Figure 13). The baseline curve resembles that typical for a maximum load plateau CTOD measurement. However, the material possessing the SOHIC damage exhibits an initial peak in the curve upon load-up which quickly decreases. From this point the curve flattens out and begins to increase again resulting in a second peak in the curve. The initial peak, or knee in the curve, relates to the link-up of the intermittent SOHIC ligament to produce a fully planar defect. The remainder of the curve resembles the behavior of the baseline material with a fatigue precrack. Hence, the additional energy required to initially link-up the intermittent SOHIC ligament possibly results in slight crack tip blunting or a larger plastic zone which acts to increase the measured toughness. This load versus crack opening behavior was also noted on the HAZ specimens from the conventional steel. However, the size of the initial peak was less pronounced as a result of the smaller intermittent SOHIC region (see Figure 10).

FIGURE 11 - Baseline toughness values measured on the four material / condition combinations.

FIGURE 12 - Behavior of the baseline fatigue precracked material to the environmental SOHIC material.

FIGURE 13 - Load versus mouth opening displacement response for the SOHIC flaw in comparison to the fatigue precracked material.

Fracture Surface Morphologies

The fracture surface morphologies of the intermittent SOHIC damage on both the conventional and “HIC Resistant” steels were examined under a scanning electron microscope (SEM). As shown in Figure 14, the intermittent SOHIC was highly textured or “woody” in appearance and consisted of quasi-cleavage with some traces of microvoid coalescence. The texture was produced by a combination of the heavily banded microstructure and delamination along sulfide inclusions. This was textured appearance was expected due to the jagged nature of the SOHIC extension observed in Figures 5 and 7, discussed earlier in this paper. The “woody” appearance of the SOHIC observed on the conventional steel duplicated that reported by McHenry et al [1] in conjunction with a sour amine absorber pressure vessel failure. The intermittent SOHIC observed in the base metal of the “HIC Resistant” steel is shown in Figure 15. In this case, the fracture surface produced was relatively flat. The in-plane laminations are visible and interconnected by cleavage of the surrounding ferrite as previously suspected.

FIGURE 14 - Intermittent SOHIC fracture surface morphology observed in the base metal of the conventional steel. (50 X)

FIGURE 15 - Intermittent SOHIC fracture surface morphology observed in the base metal of the “HIC Resistant” steel. (50 X)

IMPLICATIONS AND SIGNIFICANCE OF FINDINGS

Based on the findings of this and previous work on this subject, several implications for equipment integrity were drawn. The poor fracture toughness of the intermittent SOHIC damaged region ahead of a discrete (i.e. continuous) flaw should be considered when performing a fitness-for-service analysis since the overall behavior was governed by the total length (i.e. discrete plus the intermittent crack length). Secondly, the increased susceptibility of “HIC Resistant” steel to SOHIC in the base metal may actually decrease overall equipment integrity as compared to conventional steel, particularly in the consideration of attachment welds. Furthermore, users may wish to consider the varying susceptibilities of different steels to cracking in the base metal when developing inspection plans.

Based on the results of the research program described herein, the following conclusions were made:

1. The most susceptible material / condition for SOHIC evaluated in this study was the base metal of “HIC Resistant” steel followed by the HAZ region of the conventional steel. The high susceptibility was believed to relate to the distinct ferrite fracture path present in both material / condition combinations.

2. Methods to measure the fracture resistance of steel with pre-existing SOHIC damage were developed and successfully applied to the base metal of a “HIC Resistant” A516-70 material.

3. The intermittent SOHIC damaged zone ahead of a discrete SOHIC crack has been shown to have poor fracture toughness. The behavior of the flaw was governed by the total length including the discrete and intermittent extension.

4. The baseline fracture toughness of the HAZ region on both the conventional and “HIC Resistant” steels were below the respective base metal values.

5. The measured toughness of the material with SOHIC damage was slightly higher (8 percent) than the baseline values using fatigue precracked specimens. The slight increase was attributed to the additional energy required to link-up the intermittent damage to produce a fully planar flaw.

6. The SOHIC damage produced separate, but identifiable fracture surface morphologies. The conventional steel produced a textured, “woody” appearance whereas the “HIC Resistant” steel produced a relatively flat fracture with link-up of the in-plane cracks by cleavage in the ferrite phase.

1. H.I. McHenry, D.T. Read and T.R. Shives, “Failure Analysis of an Amine-Absorber Pressure Vessel,” Materials Performance, August 1987.

2. R.D. Kane, M.S. Cayard and M. Prager, “Evaluation of Advanced Plate Steels for Resistance to HIC and SOHIC in Wet H2S Environments,” Proceedings of the Second International Conference on the Interaction of Steels with Hydrogen in Petroleum Industry Pressure Vessel and Pipeline Service, The Materials Properties Council, Inc., New York, NY, Vol. 1, 1994.

3. R.D. Merrick and M.L. Bullen, “Prevention of Cracking in Wet H2S Environments,” CORROSION/89, Paper No. 269, NACE International, Houston, TX, (1989).

4. M. Iino, “Influence of Sulfur Content on the Hydrogen Induced Fracture in Linepipe Steels,” Metallurgical Transactions, Vol. 10A, November 1979.

5. R.D. Kane, M.S. Cayard and M. Prager, “Test Procedures for Evaluation of Resistance of Steels to Cracking in Wet H2S Environments,” CORROSION/94, Paper No. 519, NACE International, Houston, TX, (1994).

6. M.S. Cayard and R.D. Kane, “Characterization and Monitoring of Cracking of Steel Equipment in Wet H2S Service,” CORROSION/95, Paper No. 329, NACE International, Houston, TX, (1995).

7. NACE Standard TM0177, “Laboratory Testing of Metals for Resistance to Sulfide Stress Cracking in H2S Environments”, NACE International, Houston, TX, (1990).

8. ASTM E1290-93, “Standard Test Method for Crack-Tip Opening Displacement (CTOD) Fracture Toughness Measurement,” Vol. 3.01, ASTM, Philadelphia, PA, 1995.

The information presented in this paper is the direct result of internally funded work from InterCorr International, Inc. The authors of this paper would like to acknowledge the support and technical contributions of Mr. Ronald W. McNeil and Mr. Dwain T. McLean who conducted the research work.