Paper No.
10294
2010
MERCURY LIQUID METAL EMBRITTLEMENT OF ALLOYS FOR OIL AND GAS PRODUCTION AND PROCESSING Raymundo Case and Dale R. McIntyre ConocoPhillips Production Assurance Technology Bartlesville, OK ABSTRACT Mercury is a natural component of certain hydrocarbon reservoirs, so exposure to liquid mercury can occur in oil and gas production and processing plants. This paper reports liquid metal embrittlement (LME) test results for a variety of common oilfield and processing plant alloys exposed to liquid Hg across a range of temperatures. Test methods used include slow strain rate testing and C-ring tests. A fracture mechanics approach to derive critical stress intensities for the onset of Hg LME is suggested. Test results are presented for ASTM A516 Gr 70 carbon steel, ASTM A193 Gr B7 low alloy steel, AISI type 317 stainless steel, AISI type 410 stainless steel, Types 2205 and 2507 duplex stainless steels, Gr 2, Gr 5 Titanium alloys, UNS N10276, UNS N04400 and UNS A95086. Anomalous results were obtained in slow strain rate testing of duplex stainless steels in Hg. The implications of these anomalous results are discussed.
©2010 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole, must be in writing to NACE International, Publications Division, 1440 South Creek Drive, Houston, Texas 77084. The material presented and the views expressed in this paper are solely those of the author(s) and are not necessarily endorsed by the Association.
1
INTRODUCTION
Trace amounts of mercury occur naturally in some hydrocarbon reservoirs, and when mercury is encountered in oil and gas production and processing operations it must be contained, removed and eliminated from the product stream. One of the possible problems in mercury-containing production environments is mercury liquid metal embrittlement of the production and processing equipment. Liquid metal embrittlement (LME) is a form of environmental cracking which occurs when structural alloys come in contact with liquid metals of various sorts. The attack normally involves penetration of the liquid metal along the grain boundaries of the structural alloy and consequent loss of structural strength and ductility. Unlike other forms of environmental cracking, such as stress-corrosion cracking or sulfide stress cracking, liquid metal embrittlement may occur in the absence of net tensile stresses in the structural alloy. All alloys are not prone to LME by all liquid metals; rather, there are specific pairs of liquid metals and structural alloys which have sufficient mutual solubility to result in grain boundary penetration. Mercury is unique in that it is liquid over a wide range of ambient temperatures common to oil and gas production and processing environments. Some alloys, such as brass and aluminum, are well known to be highly susceptible to Hg LME whereas the Hg LME susceptibility of other alloys has not been thoroughly studied. In this work a series of laboratory tests were performed to assess the risk of mercury LME in a number of alloys commonly used in oil and gas production and processing equipment. EXPERIMENTAL PROCEDURE AND RESULTS
Three to five standard SSR specimens were machined from each material listed. The standard specimens had a 1.0 inch (25.4 mm) long gage section with a 0.150 inch (3.8 mm) diameter. The shoulder diameter of the specimen was nominally 0.25 inch (6.4 mm) and an overall length of 2.75 inches (69.9 mm). The dimensions were in accordance to ASTM G129. Uniaxial tension tests were conducted with duplicate tensile specimens procured for the materials 2205 duplex stainless steel and solution-treated-and-aged Ti-6AI-4V. Elongated tensile specimens for uniaxial tension specimens had a 1 inch (25.4 mm) gage with a 0.150 inch (3.8 mm) diameter. The shoulder diameter of the specimen was nominally 0.25 inch (6.4 mm) with an overall length of 5 inches (127 mm). Slow Strain Rate Testing. ASTM G129 - "Slow Strain Rate Testing to Evaluate the Susceptibility of Metallic Materials to Environmentally Assisted Cracking" tests were conducted with some modifications. All specimens were cleaned with toluene and rinsed in acetone prior to testing. Tests in air were followed by tests in liquid mercury or mercury vapor. In all liquid Hg tests the gage section was completely immersed in the liquid metal. However, for the tests in Hg vapor, only the bottom shoulder of the specimen was covered with liquid mercury. In all tests the temperature was 140°F with a strain rate of 4x10-6 S·l. For each SSR test, crosshead movement was initiated once the temperature reached equilibrium. Results were compared to a baseline test from each material conducted in air at the same strain rate and temperature.
At the conclusion of each SSR test, the gage section and fracture surface of each specimen was examined at 40x magnification with a light microscope for indications of environmental cracking. 2
Constant Load Testing. Constant load testing was performed per a modified ASTM G49 method, utilizing a proof ring assembly with the gage section of the tensile specimen covered by plastic tubing that contained the liquid mercury environment. The tubing was clamped over the shoulder of each specimen. The specimens were measured to the nearest 0.001 inch and then cleaned with toluene and acetone prior to testing. Using a syringe, distilled liquid mercury was injected into the space between the specimen and the plastic tubing. The temperature was then increased to 140°F and equilibrated. At the test temperature, the specimens were stressed to the required load. Duplicate tensile specimens of each material were stressed to 110% of the actual yield strength (AYS) of the material. Two additional Ti-6AI-4V specimens were stressed to 50% AYS. The test conditions were maintained for 30 days or until failure. Following the 30-day exposure, the gage surface of the specimens was cleaned and visually examined for cracking at 10x magnification. Alloys in the test matrix included: • • • • • • • • • • •
ASTM A516 Gr 70 carbon steel ASTM A193 Gr B7 Cr-Mo heat treated alloy steel AISI Type 317 austenitic stainless steel AISI Type 410 martensitic stainless steel Type 2205 duplex stainless steel Type 2507 duplex stainless steel Commercially pure Gr 2 Titanium High-strength Gr 5 Ti-6AI-4V alloy C-276 (UNS NI0276) Aluminum alloy Type 5083 Alloy 400 (UNS N04400)
These last two, Al type 5083 and Alloy 400 were included as controls SInce their high susceptibility to Hg LME is well known. ASTM G129 slow strain rate tests and ASTM G49 constant-load tests were conducted on specimens of the alloys listed above, at 140°F, in mercury liquid. Some alloys were also tested in mercury vapor. The test temperature of 140°F was chosen because it is a common temperature encountered in oil and gas production environments and in refinery condensers. The constant load tests are judged by fail-no fail criterion; the presence of any cracking whatsoever is considered a failure. The slow strain rate tests are rated by the ratio of the time to failure and percent elongation in the test environment versus these properties when measured in air. In addition, the slow strain rate test specimens are examined after the test for evidence of environmental cracking on the fracture surfaces and along the gauge length. RESULTS AND DISCUSSION After 720 hours on test at 140°F, stressed to 110% of actual yield stress, no failures were observed on the following materials: • •
ASTM A516 Gr 70 carbon steel AISI Type 31 7 stainless steel o Liquid 3
• • •
o Vapor Or 2205 duplex stainless steel Commercially pure Or 2 Titanium High strength Or 5 Ti-6AI-4V, solution treated and aged
The time-to-failure (TTF) and percent reduction in area (RA) ratios of materials which are resistant to environmental cracking typically exceed 0.80, assuming that environmental cracking is not observed on the fracture surface or gauge length. Alloys with good slow strain rate test results in these tests are as follows:
Table 1 - SSR Results Indicating Resistance to Hg Attack Tests Conducted at 140°F in Liquid Hg except as Noted Time-to-Failure Ratio Alloy ASTM A516-70 carbon steel 1.03 AISI 317 SS in liq. Hg 1.07 AISI 317 SS in Hg vapor 1.06 Or 5 Ti-6AI-4V in Hg vapor 1.16 Commercially Pure Or 2 Ti 0.99 C-276 (UNS N10276) 0.91 ASTM A193 Or B7 in liq. Hg 1.03 ASTM A193 Or B7 in Hg vap. 1.04 Or 2507 duplex SS in liq. Hg 1.00 Or 2507 duplex SS in Hg vap. 1.08 AISI 410 SS in liq. Hg 0.91 AISI 410 SS in Hg vapor 0.87 *Small secondary cracks observed along the gage section
Reduction-in-Area Ratio 0.89 0.98 1.03 1.14 0.98 0.78 1.00 1.04 0.90* 1.05 0.98 0.99
Materials which gave SSR test results indicative of susceptibility to Hg attack were as follows:
4
Table 2 - SSR Results Indicating Susceptibility to Hg Attack Tests Conducted at 1400 F in Liquid Hg except as Noted AlloX Alloy 400 (UNS N04400) Gr 2205 Duplex SS Gr 5 Ti-6AI-4V in Liq. Hg Al 5083 in Liq. Hg Al 5083 in Hg vapor
Time-to-Failure Ratio 0.83 0.83 0.59 0.11 0.13
Reduction-in-Area Ratio 0.47 0.46 0.19 0.11 0.20
These results show that there is a useful suite of alloys which show practical immunity to Hg embrittlement under the conditions tested. These materials are: • • • • • •
ASTM A516 Gr 70 carbon steel ASTMA193 Gr B7 alloy steel AISI 317 austenitic stainless steel AISI 410 martensitic stainless steel C-276 (UNS N1 0276) Commercially Pure Gr 2 Titanium
Materials tested in Hg vapor showed either no effect or a reduced effect compared to liquid mercury. Grade 5 Ti-6AI-4V and 2507 duplex stainless steel showed no reduction in properties when tested in mercury vapor but noticeable effects when tested in liquid mercury. (On 2507 duplex stainless steel, the reduction in properties was small but on Gr 5 Ti-6AI-4V the effect of liquid mercury exposure was profound.) Aluminum alloy 5083 showed better properties in Hg vapor than in Hg liquid, but it was still very prone to cracking. This is an unexpected result; even highly susceptible aluminum alloys are not expected to crack in Hg vapor. The SSR test apparatus is heated from the outside, so it was not clear whether the cracking observed in Hg vapor was influenced by some condensation on the gauge length of the specimens. Alternatively, earlier work was done at ambient temperatures or colder, to simulate cold-box conditions in gas processing plants, so the appearance of cracking at 140°F in Hg vapor may signal a temperature threshold. Anomalous results were observed for the duplex stainless steels and high-strength titanium. Grade 2205 duplex stainless steel is widely used in North Sea production applications where mercury exposure is known to occur. To date no failures by LME have been observed. The constant load tests showed no failures, and both austenitic and martensitic stainless steels showed no sensitivity to mercury so it is surprising that an austenitic-ferritic stainless steel would develop a susceptibility to Hg. However, considerable cracking was observed on the fracture surfaces of the slow strain rate specimens (Figures 1 through 3). The reduction in area ratio for 2205 duplex stainless steel (Table 2) was quite similar to that of Alloy 400 (UNS N04400), a material with a significant history of failures in Hg service. Therefore the appearance of the Hg attack on 2205 in these tests points out an unexpected vulnerability not evident from previous tests work or service experience. Alloy 400 (UNS N04400) suffers most problems in the heavily cold-worked condition; since 2205 shows similar TTF and RA ratios, it may be that the vulnerability of 2205 to Hg is specific to applications with heavy cold work.
5
Grade 2507 duplex stainless steel gave very high ratios of TTF and RA in the liquid mercury exposures. However, there was some evidence of secondary cracking on the gage length. This suggests that 2507 duplex stainless steel is not prone to cracking in Hg except at very high plastic strains. The high-strength titanium alloy Ti-6AI-4V in the solution-treated and aged condition passed a constant-load test at 110% of yield stress. However, the slow strain rate (SSR) specimens showed severe loss of ductility when exposed to liquid mercury (Figures 4 and 5). Since the static test results are good and the dynamic (SSR) test results are poor, it would appear that the integrity of the surface oxide film is of crucial importance for the alloyed titanium. Static or quasi-static applications will probably show no effect. However, there would appear to be concern for dynamic applications where the protective surface oxide film could be cracked or breached in some way. Ti-6AI-4V showed very good results in Hg vapor, so limiting Ti-6AI-4V applications to those where Hg vapor only is expected should give good service. Surface treatments to thicken and toughen the oxide film, such as air or steam oxidation, should be beneficial provided the fatigue resistance is not reduced. Alternatively, other heat treatments or other high-strength titanium alloys, without aluminum, might be researched to find combinations less sensitive to Hg. CONCLUSIONS 1) The following common alloys appear to be immune to embrittlement by both mercury liquid and vapor under the conditions tested: ASTM A516 Gr 70 carbon steel ASTM A193 Gr B7 alloy steel AISI 317 austenitic stainless steel AISI 410 martensitic stainless steel C-276 (UNS NI0276) Commercially Pure Gr 2 Titanium 2) Aluminum alloy 5083 and Alloy 400 (UNS N04400) showed significant embrittlement in the tests, in agreement with reports of service failures of these alloys In mercury. 3) Duplex stainless steels and high strength Ti-6AI-4V alloy gave ambiguous test results, whose implications are discussed in more detail in the body of the report. RECOMMENDATIONS 1) Avoid the use of aluminum alloys and non-chromium bearing nickel alloys (Alloy 400) in mercury service. 2) Review applications of high-strength aluminum-bearing titanium alloys in liquid mercury service. Avoid situations involving high plastic strains, high dynamic loads, or damage to the alloy's protective oxide surface film. Use of such AI-bearing Ti alloys in Hg vapor service should be successful.
6
3) Use of duplex stainless steels in Hg service has been successful in the past. However, it would be prudent to avoid highly cold-worked duplex stainless steels in liquid Hg service until these test results are better understood. ACKNOWLEDGEMENTS
We are grateful to Dr. Russell D. Kane, formerly of Honeywell International, now of iCorrosion Inc., and to Dr. Elizabeth Trillo, also formerly of Honeywell International and now of Southwest Research Institute, for conducting the slow strain rate and uniaxial tension tests described herein.
7
Figure 2 - Close-up of 2205 duplex stainless steel SSR specimen tested in liquid Hg at 140oF. Note the beads ofHg on the slow growth zone at right. Compare with Figure 1.
8
Figure 3 - Light photomicrograph of a cross-section through the Hg-exposed 2205 duplex stainless steel specimen's slow growth zone (Figure 2). The slow growth zone's profile extends from top right to top center. Note the secondary crack at lower right. Unetched.
Figure 4 - Fracture surface of a Ti-6AI-4V specimen tested in liquid Hg at 140oF. Note the slow growth zone with Hg droplets on it at center right. Note the longitudinal crack in the shear plane at center.
9
Figure 5 - Light photomicrograph showing a cross-section through the slow growth zone in the Ti-6AI-4V specimen tested in liquid Hg (Figure 4). The fracture profile is at top. Note the longitudinal secondary cracking. Unetched.
10
