Electrochemical Methods to Detect Susceptibility of Ni-Cr-Mo-W Alloy 22 to Intergranular Corrosion D.D. GORHE, K.S. RAJA, S.A. NAMJOSHI, VELIMIR RADMILOVIC, ALFREDO TOLLY, and D.A. JONES Alloy 22 (UNS N06022), a Ni-Cr-Mo-W based alloy, is a candidate material for the outer wall of nuclear waste package (NWP) containers. Even though the alloy is highly stable at low temperatures, it could undergo microstructural changes during processing such as welding and stress relieving. Formation of topologically close-packed (TCP) phases such as , P, , etc. and Cr-rich carbides could make the material susceptible to localized corrosion. Hence, it is important to correlate the microstructural changes with the corrosion resistance of the alloy by nondestructive and rapid electrochemical tests. In this investigation, different electrochemical test solutions were used to quantify the microstructural changes associated with aging and welding of the wrought alloy 22. The results of double-loop (DL) electrochemical potentiodynamic reactivation (EPR) tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution indicated Cr depletion during initial stages of aging of wrought alloy 22. Results of EPR tests in 2 M HCl 0.01 M KSCN solution at 60 °C correlated well with the Mo depletion that occurred near TCP phases formed during aging of both weld and wrought alloy 22 materials. The EPR test results were compared with standard chemical weight loss measurements specified by ASTM standard G-28 methods A and B.
ALLOY 22 (UNS N06022), a Ni-Cr-Mo-W based alloy, is a candidate material for the outer wall of nuclear waste package (NWP) containers.[1–4] Even though alloy 22 in wrought form is considered to have phase stability at the operating temperatures (200 °C) of the repository, exposure to elevated temperatures during fabrication processes such as welding and stress relieving could cause alteration of microstructure. Welding causes formation of dendrites, segregation of Mo, W in the interdendritic region, and formation of topologically close-packed (TCP) phases such as , P, , etc. in both the weld and the heat-affected zone (HAZ).[3,5–7] Increased corrosion rates of aged Ni-Cr-Mo-W alloys were observed because of the sensitized microstructure. The sensitization of Mo-rich nickel-base alloys is found to be different from the sensitization of common austenitic stainless steels and other Ni-Cr alloys such as INCONEL* 600.[9–20] *INCONEL is a trademark of INCO Alloys International, Huntington Woods, WV.
Sensitization of austenitic stainless steel resulted in depletion of chromium adjacent to the chromium-rich M23C6 carbides[21,22] formed along the grain boundaries.
D.D. GORHE, Graduate Research Assistant, and K.S. RAJA and S.A. NAMJOSHI, Research Assistant Professors, are with the Department of Metallurgical and Materials Engineering, University of Nevada, Reno, NV 89557. Contact e-mail: [email protected]
VELIMIR RADMILOVIC, Staff Scientist, and ALFREDO TOLLY, Postdoctoral Fellow, are with the National Center for Electron Microscopy, Lawrence Berkeley National Laboratory, Berkeley, CA 94720. D.A. JONES, formerly with the Department of Metallurgical and Materials Engg., Univ. of Nevada, Reno, NV 89557, is deceased. This article is based on a presentation made in the symposium “Effect of Processing on Materials Properties for Nuclear Waste Disposition,” November 10–11, 2003, at the TMS Fall meeting in Chicago, Illinois, under the joint auspices of the TMS Corrosion and Environmental Effects and Nuclear Materials Committees. METALLURGICAL AND MATERIALS TRANSACTIONS A
In the Ni-Cr-Mo alloys, sensitization was observed to result in the depletion of molybdenum near the TCP or M6C precipitates.[8,23] When the sensitized Ni-Cr-Mo alloys were exposed to a reducing environment, the Mo-depleted regions were preferentially attacked, and in the oxidizing environment, the TCP phases were dissolved, giving rise to the corrosion rate. Raghavan et al. and Cieslak and coworkers[5,6,7] observed the TCP phases containing only the nominal chromium as that of the bulk chemistry, whereas Hodge and Kirchner and Hodge reported that the phase was enriched with Cr and the suggested phase was (Ni, Fe, Co)3(W, MO, Cr)2. There are no reports available on the depletion profiles of the aged alloy 22. These microstructural changes could alter the mechanical as well as corrosion properties of alloy 22. The lids of the NWP containers would not undergo any heat treatment after their closure. The projected time required for the waste to become safe through radioactive decay is 10,000 years. Therefore, it is imperative to develop a testing technique to evaluate the microstructural changes that occur during fabrication or exposure to service conditions, so that the integrity of the waste package container can be predicted. ASTM G28 is the standard test method of detecting susceptibility to intergranular corrosion in nickel- and chromiumbased alloys in wrought conditions. Method A is based on weight loss of the material in ferric sulfate-sulfuric acid solution, while method B is in mixed acid-oxidizing salt solution for 24 hours. These tests are time consuming and destructive in nature. Hence, there is a need to find alternative tests, which should be rapid, nondestructive and able to be carried out in situ. An electrochemical potentiodynamic reactivation (EPR) test would be a better alternate for these weight loss tests and can be used for in-situ evaluation using a suitable electrochemical cell and potentiostat assembly. ASTM G108 specifies a standard procedure for conducting EPR tests for 304 and 304L stainless steels. However, there are no standard test solutions and test procedures for Ni-based alloys. VOLUME 36A, MAY 2005—1153
It is generally believed that formation of TCP phases that occur by diffusion of substitutional atoms may not result in a sharp depletion profile of alloying elements impairing corrosion resistance. This speculation is based on the Cr depletion observed in stainless steels because of the formation of chromium carbides. The deep Cr depletion profile in general has been attributed to the difference in diffusivities of interstitial carbon and Cr. However, it has also been reported that formation of TCP phases such as sigma also results in Cr depletion, especially in low-carbon stainless steel welds. Recently, Turnbull et al. developed a novel electrochemical technique to detect localized corrosion susceptibility of duplex stainless steel welds associated with the alloy-depleted zone due to formation of sigma phase. Substitution of larger atoms such as Mo and W would result in slower diffusion of solute atoms. When formation of the TCP phase is thermodynamically possible, depletion of alloying elements can be anticipated because of the difficulty in diffusion. In alloy 22 also, it is likely that formation of TCP phases could result in a deep depletion profile of alloying elements, causing preferential corrosion attack. The depletion of alloying elements could be of two types: (1) Cr depletion and (2) Mo and W depletion. It should be noted that Cr depletion would result in a decrease in corrosion resistance to oxidizing environments, while Mo depletion would result in lower corrosion resistance in reducing environments. Therefore, a single test procedure may not detect both Cr depletion and Mo depletion. Individual test solutions and procedures are required to be developed to detect the level of individual alloy element depletion. Therefore, the objectives of this investigation were as follows: (1) to develop an EPR test solution and test procedure to detect chromium depletion at the grain boundaries and interdendritic regions (if any) in alloy 22 wrought and weld metals, respectively; and (2) to develop an electrochemical test solution and test procedure to detect molybdenum depletion in the matrix due to the formation of TCP phases.
II. EXPERIMENTAL A. Wrought Material Sixteen-millimeter-thick plates of alloy 22 in mill-annealed condition were received from Haynes International (Kokomo, IN). The chemical composition in wt pct is 0.003C, 0.03Si, 0.002S, 0.005P, 0.24Mn, 0.15V, 0.98Co, 3.53Fe, 2.94W, 13.37Mo, 21.68Cr, and 56.5Ni. The 11-mm diameter and 11-mm thick cylindrical specimens were machined out of the plate with the longitudinal axis parallel to the rolling direction. After thorough washing in soap, running water, and acetone, these specimens were aged at 610 °C, 650 °C, 700 °C, 750 °C, 800 °C, and 850 °C for 0.1, 1, 20, and 100 hours in an electric resistance furnace. The furnace atmosphere was not purged with any gas. The temperature was controlled within 1 °C. For comparison, specimens were tested in the mill-annealed (as-received) condition also. Insulated copper wire covered with glass tube was soldered to the specimen for the electrical connection and mounted in epoxy resin, exposing 1 cm2 of surface area. 1154—VOLUME 36A, MAY 2005
The working surface of the specimens was polished with a series of emery papers, down to 600 grit, thoroughly washed with soap water, and rinsed in distilled water immediately before testing. B. Weld Material The 38-mm-thick alloy 22 (UNS N06022) plates (chemical composition in wt pct 0.002C, 0.04Si, 0.001S, 0.009P, 0.14Mn, 0.171V, 0.01Co, 2.8Fe, 2.68W, 13.91Mo, 20.62Cr, 0.01Cu and Rem Ni with double V edge preparation were butt welded with similar composition filler material, class ERNiCrMo-10, using the multipass gas tungsten arc welding process. The 12.5-mm-size cubic specimens were cut from the welded material for electrochemical evaluation. The weld specimens were given the following three different thermal aging treatments: (1) 760 °C for 1 hour, (2) 760 °C for 140 hours, and (3) 700 °C for 24 hours. For comparison, as-welded specimens and as-received (mill-annealed) wrought specimens were also investigated. An insulated copper wire with a glass shield was soldered to the specimen surface for the electrical connection. The specimens were mounted in epoxy resin, exposing only the top surface, which contained a single weld bead in the longitudinal direction. The weld metal surface was metallographically polished with a series of emery papers, down to 600 grit. The specimen surface did not include the heat-affected wrought material, and the exposed area was 1.5 cm2. C. Electrochemical Testing 1. Test solution for Cr depletion Electrochemical reactivation tests were carried out in a fivenecked 1-L flask with a specimen test electrode and two Pt auxiliary (counter) electrodes (Ingold Electrodes, Boston, MA) on opposite sides of the working electrode, for more uniform current distribution. A saturated silver/silver chloride (SSC) electrode was used as the reference electrode with a solution bridge and Luggin probe. For each test, the solution electrolyte was de-aerated for 1 hour prior to specimen immersion with a pure-nitrogen purge, which continued throughout the test. Single-loop (SL) electrochemical reactivation tests were carried out in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution at 30 °C. Initially, the specimen was passivated at 400 mV (Ag/AgCl) for 120 seconds and the potential was scanned in the cathodic (reverse) direction from 400 mV to the initial corrosion potential at a rate of 0.5 mV/s. The charge transferred during the reactivation was calculated and considered for quantifying the degree of sensitization occurring on the aged specimens. The charge was not normalized with grain boundary (GB) area, as there was no significant change in grain sizes between the specimens aged at 650 °C to 750 °C. The grain size was measured by the procedure described in ASTM standard E 112. In double-loop (DL) reactivation tests, the potential was scanned from the open circuit corrosion potential to 400 mV (SSC) and reversing the potential back to the initial corrosion potential at a scan rate of 0.5 mV/s. The peak current recorded during forward scan was denoted as Ia, and the peak current during reverse scan (if any, depending on the sensitized condition) was denoted as I r . The ratio I r /I a indicated degree of sensitization. In general, the ratio varied from 0 to 1. Increased METALLURGICAL AND MATERIALS TRANSACTIONS A
ratio indicated higher degree of sensitization. Some tests were repeated at scan rates of 0.167 and 1.67 mV/s. Electrochemical reactivation tests were carried out on both wrought and weld specimens in the following test solutions: 0.5 M H2SO4 0.01 M KSCN 0.05 M H2SO4 0.003 M thioacetamide (CH3CSNH2) 0.5 M H2SO4 0.003 M thioacetamide (CH3CSNH2) 0.5 M H2SO4 0.01 M thiourea (NH2CSNH2) ASTM G-28B solution (23 pct H2SO4 1.2 pct HCl 1 pct FeCl3 1 pct CuCl2) (f) ASTM G-28B solution with 0.1 pct KSCN (g) 1 M H2SO4 0.5 M NaCl X KSCN (X 0.01, 0.02, and 0.05 M).
(a) (b) (c) (d) (e)
In general, the test solution was a mixture of an acid and an activator. The selection of the activator was based on the literature results. KSCN is a standard activator used in solution specified by ASTM G108 for detecting Cr depletion in stainless steels. Fang et al. recommended thioacetamide for detecting sensitization in 308 stainless steel. The activators generally contain sulfur-bearing free electron pairs. The species with free electron pairs could be easily adsorbed on the metal surface, which results in lowering of the metalmetal bond strength. If the passive film on the metal surface is intact, such adsorption may not be easy. Weakening of metal-metal bond strength and formation of a dipole between the metal and sulfur ions facilitate easy dissolution of the metal. As passivation is delayed at Cr-depleted regions, the activators preferentially attack these regions during reactivation tests. Double loop electrochemical reactivation tests were carried out using the preceding solutions at 30 °C. The open circuit corrosion potential of alloy 22 specimens in ASTM G-28 solution was greater than 450 mV (SSC). In this solution, the cyclic polarization was carried out by scanning the potential from corrosion potential to an apex potential where the maximum current density was 5 mA/cm2 (about 1200 mV (SSC)). 2. Test solution for Mo and W depletion Double loop EPR tests were carried out on the aged wrought and weld materials in the following test solutions at 30 °C and 60 °C: (h) (i) (j) (k) (l) (m)
0.5 M H2SO4 1 M HCl 0.01 M KSCN 1 M HCl 2 M HCl 0.01 M KSCN 1 N HNO3 2 M HCl 0.01 M KSCN 1 M H2SO4 2 M HCl 0.01 M KSCN 2 M HCl 0.01 M Fe2(SO4)3 0.01 M KSCN
An SSC electrode was used as the reference electrode. The potential was scanned from the corrosion potential to 400 mV and back to an initial potential at a rate of 0.5 mV/s. Based on the DL-EPR test results, the 2 M HCl 0.01 M KSCN solution was considered for further evaluation using the SL-EPR test procedure at 60 °C. D. Chemical Weight Loss Test Chemical weight loss measurements were carried out on the aged wrought materials following the procedure given in ASTM G-28 test method B and test method A. The aged cylindrical specimens were suspended in the boiling solution METALLURGICAL AND MATERIALS TRANSACTIONS A
of 23 pct H2SO4 1.2 pct HCl 1 pct FeCl3 1 pct CuCl2 for 24 hours in method B, and for method A, the specimens were suspended in a boiling solution of ferric sulfate 50 pct sulfuric acid and the weight losses were recorded. All the tests were duplicated or triplicated and average values are reported. E. Transmission Electron Microscopy The 0.15-mm-thick wafers were cut from 11-mm-diameter, aged specimens and mechanically polished using finer grit emery papers down to a thickness of about 60 to 80 m. Threemillimeter discs were punched out of the 11-mm-diameter blank and were further thinned for electron transparency either by electrochemical twin jet polishing or by dimpling followed by ion milling. Electropolishing of 3-mm-diameter discs of alloy 22 was carried out in 5 pct perchloric acetic acid solution at room temperature by applying 50 V. Specimens were observed under a PHILIPS† CM 200 microscope using 200 kV, †PHILIPS is a trademark of Philips Electronic Instruments Corp. Mahwah, NJ.
and energy-dispersive X-ray analysis was carried out using 1.2-nm beam size. III. RESULTS Figures 1(a) through (d) show the microstructures of the alloy 22 specimens after electrolytically etching them in 1 N H2SO4 0.5 N NaCl solution at 1.2 V for about 60 seconds. The mill-annealed specimen showed many etch pits and annealing twins, but it did not show any GB precipitates (Figure 1(a)). The specimen aged at 750 °C for 1 hour did not show much precipitation along the grain boundaries (Figure 1(b)). The specimen aged at 700 °C for 15 hours showed increased etching of the grain boundaries, indicating possible GB precipitation; however, the precipitation was discontinuous on some grain boundaries (Figure 1(c)). Aging at 700 °C for 100 hours caused wider and almost continuous etching of the grain boundaries, as shown in Figure 1(d). The wider etching could be attributed to the precipitation of -phase particles. Specimens aged at 800 °C for 100 hours (figure not included) showed precipitation inside the grains in addition to intergranular precipitation. In short, as aging time and temperature increased, precipitation increased. A. EPR Tests for Cr Depletion The EPR test results in solutions listed from (a) through (g), which were mainly different combinations of sulfuric acid and an activator, are briefly discussed here. The EPR test results in 0.5 M H2SO4 0.01 M KSCN solution were not reproducible. Other researchers also observed similar results for alloy 825. The addition of activators such as thioacetamide and thiourea did not show any reactivation peaks for the specimens aged at 650 °C and 700 °C. Faint reactivation peaks could be observed for specimens aged at 750 °C. The results were not reproducible in these solutions also, and no correlation between the aging time and the reactivation current ratio (Ir /Ia) could be observed. Figure 2 shows one such result for wrought-aged alloy 22 in 0.05 M sulfuric acid 0.003 M thioacetamide solution. VOLUME 36A, MAY 2005—1155
Fig. 1—(a) Microstructure of alloy 22—mill annealed. (b) Microstructure of alloy 22 aged at 750 °C for 1 h. (c) Microstructure of alloy 22 aged at 700 °C for 15 h. (d) Microstructure of alloy 22 aged at 700 °C for 100 h.
Fig. 2—DL-EPR test results of alloy 22 specimens in 0.05 M H2SO4 0.003 M CH3CSNH2. Arrows indicate the direction of scan.
No activation or reactivation peaks could be observed in ASTM G-28B solution. Electrochemical tests carried out in ASTM G-28B solution cannot be considered as reactivation type tests. Nevertheless, the cyclic polarization tests in the G-28 solution could indicate the pitting corrosion resistance of the material after different heat treatments. More positive corrosion potentials of the specimens were observed because of 1156—VOLUME 36A, MAY 2005
the highly oxidizing nature of the solution. Specimens aged at 700 °C for 1 and 20 hours showed more active corrosion potentials compared to those of the specimen aged for 100 hours at 700 °C and the mill-annealed specimen. Passive current density was not affected by the aging treatments. Pitting was not observed on any of the specimens. The electrochemical tests in ASTM G-28 solution delineated the aged specimens only by the difference in their corrosion potentials. The difference in corrosion potential narrowed with longer aging time, as observed in the case of the mill-annealed specimen and the specimen aged at 700 °C for 100 hours. Addition of 0.1 pct KSCN as an activator to the G-28 solution also did not give any good result to quantify the microstructural changes. The solution of 1 M H2SO4 0.5 M NaCl 0.01 M KSCN showed good correlation between different aging temperatures and times for wrought specimens, which were reproducible. The results are discussed in Section 1. 1. Wrought specimens aged at 610 °C The results in this section are only for wrought and aged alloy 22 specimens in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution. Figure 3 shows the SL test results for alloy 22 specimens aged at 610 °C for 1 and 20 hours. The reactivation peak currents are nearly the same for both aged conditions. METALLURGICAL AND MATERIALS TRANSACTIONS A
Table I. Charge Associated with Reactivation Currents during SL-EPR tests of Alloy 22 Aged Specimens in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN Solution Aging Condition (°C/h)
Fig. 3—Results of SL-EPR tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN for alloy 22 aged at 610 °C.
610/1 610/20 650/1 650/20 650/100 700/1 700/20 700/100 750/1 750/20 750/100
Charge Accumulated (Coulomb/cm2) 3.16 5.66 9.66 5.98 1.61 3.67 3.98 1.92 3.48 9.34 6.52
103 103 103 103 103 103 103 103 103 104 104
Reactivation Current Ir (A)
Potential Corresponding to Ir (mV)
7.97 106 9.74 106 3.17 105 2.05 106 6.11 107 1.48 104 1.51 105 3.97 106 3.01 105 2.21 106 no Ir
87.9 78.9 90.8 75.9 78.9 52.9 60.9 60.9 66.9 66.9 —
Fig. 4—Reactivation current ratio of DL-EPR test in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN for alloy 22 aged at different conditions.
Correspondingly, the charge associated with reactivation is also similar for both specimens. Results of DL-EPR tests also showed that both reactivation and activation currents increased with increasing aging time, though the ratio of re-activation current to activation current (I r /I a) did not show a uniform trend. Figure 4 summarizes the reactivation current ratio for the various times and temperatures of aging from DL-EPR tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution. Clearly, the reactivation current increased as the aging time increased from 1 to 20 hours, indicating possible increase in Cr depletion at this aging temperature, but it again decreased as aging time increased to 100 hours. The charge accumulated during reactivation-assisted dissolution of the alloying-element depleted zone during SL-EPR tests is summarized in Table I for the various times and temperatures of aging. 2. Wrought specimen aged at 650 °C Figure 5 shows the SL-EPR test results of alloy 22 aged at 650 °C for 1, 20, and 100 hours. It can be seen that the reactivation peak current decreased with increasing aging time. The DL-EPR results for this aging temperature, as summarized in Figure 4, also follow the same trend as that observed in SL tests (decreasing peak current ratio with increasing aging time). Multiple tests were carried out and good reproducibility could be observed. Both single- and DL-EPR results reveal that the GB Cr-depleted zones could be healed by recovery of Cr, due to diffusion from bulk of the grain, during prolonged aging at or above 650 °C. METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 5—Results of SL-EPR tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN for alloy 22 aged at 650 °C.
Fig. 6—Results of SL-EPR tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN for alloy 22 aged at 700 °C.
3. Wrought specimens aged at 700 °C Figure 6 shows the SL-EPR test results of alloy 22 aged at 700 °C for 1, 20, and 100 hours. These specimens also showed results similar to those shown by specimens aged at 650 °C. The reactivation peak current decreased with increasing aging time. It was observed that during DL-EPR tests (Figure 4), not only did the I r /Ia ratio decrease with time, but the magnitudes of the currents during forward and reverse VOLUME 36A, MAY 2005—1157
potential sweep also decreased with aging time. The peak current during the forward sweep indicates the critical current required for passivation. A decrease in this critical current indicates the recovery of the corrosion resistance of the material. Increased critical current for passivation during initial stages of aging at relatively higher temperatures indicates that Cr depletion occurred due to the formation of Cr-rich precipitates along the grain boundaries, deteriorating the corrosion resistance. However, prolonged exposure to high temperature resulted in recovery of Cr in those depleted zones as both forward and reverse current peaks decreased. 4. Wrought specimens aged at 750 °C Figure 7 shows the SL-EPR test results of alloy 22 aged at 750 °C for 1, 20, and 100 hours. The reactivation peak current decreased with increasing aging time, as in the case of specimens aged at 650 °C and 700 °C. The specimen aged at 750 °C for 100 hours did not show any reactivation charge, indicating almost complete Cr recovery. The DL-EPR results support the results of SL tests (Figure 4). Both the I r /Ia ratio and the individual magnitudes of the currents during forward and reverse potential sweep decreased with increasing aging time for this aging temperature. There was no reactivation peak current observed for the specimen aged for 100 hours at 750 °C (not shown in figure). Figure 8 shows the DL-EPR
plots of the specimens aged 750 °C for 1 hour and the millannealed (as-received) specimen. The mill-annealed specimen also did not show any activation or reactivation current, indicating that there is no Cr depletion in the mill-annealed condition. It can be seen from Table I, for specimens aged above 610 °C, that the charge released during SL-EPR tests decreased with aging time. These results indicate that there could be a Cr-healing effect with increasing aging time if the aging temperature was higher than 610 °C. Figure 9 shows the comparison of SL-EPR test results in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution for alloy 22 specimens aged for 1 hour at different temperatures. The reactivation current increased as the aging temperature increased, indicating increased Cr depletion, which can be attributed to faster diffusion rates at higher temperature. 5. Effect of KSCN concentration and potential sweep rate Figures 10(a) and (b) showed the effect of concentration of KSCN. Increasing the concentration of KSCN did not show any additional benefit. The higher concentration of KSCN suppressed the reactivation peak in the specimen aged at 650 °C for 100 hours, as can be seen from Figure 10(a). A concentration of KSCN of 0.01 M was observed to give reproducible and discernible results. Figure 11 showed the effect of the scanning rate on the alloy 22 specimen aged at 700 °C for 1 hour. Decreasing the scanning rate to 0.1667 mV/s decreased the magnitude of peak currents. A slower scanning rate of 0.1667 mV/s allowed more time for passivation, which resulted in a decreased magnitude of the peak current.
Fig. 7—Results of SL-EPR tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN for alloy 22 aged at 750 °C; 100 h specimen did not show any reactivation peak.
6. Wrought specimens aged at different temperatures for 0.1 hour During welding, the material would undergo thermal cycling for different lengths of time. However, exposure to peak temperature will be of very short duration. In order to simulate the weld HAZ condition, Alloy 22 specimens were exposed to different temperatures for a short duration of 0.1 hour. The DL-EPR tests were conducted on these specimens, as described previously. Table II summarizes the reactivation peak current ratios observed for these specimens. In general, the reactivation peak current increased with an increase in aging temperature when the
Fig. 8—Results of DL-EPR tests in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN for alloy 22 aged at 750 °C for 1 h and for mill-annealed (asreceived) specimen; mill-annealed specimen did not show any activation or reactivation peak.
Fig. 9—Comparison of SL-EPR test results in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution for alloy 22 specimens aged for 1 h at different temperatures.
1158—VOLUME 36A, MAY 2005
METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 10—(a) DL-EPR test results of wrought alloy 22 in 1 M H2SO4 0.5 M NaCl 0.05 M KSCN. (b) DL-EPR test results of wrought alloy 22 in 1 M H2SO4 0.5 M NaCl 0.02 M KSCN.
lead to precipitation of secondary phases and associated alloy element depletions. B. EPR Tests for Mo and W Depletion
Fig. 11—Effect of scanning rate on passivation behavior of alloy 22; wrought specimen aged at 700 °C for 1 h.
Table II. Summary of Ratio of Reactivation to Activation Current Ratio (I r/Ia) of the Alloy 22 Specimens Aged at Different Temperatures for 0.1 h; Test Solution: 1 M H2SO4 0.5 M NaCl 0.01 M KSCN (NRP No Reactivation Peak) Reactivation Current Ratio (Ir /Ia)
Aging Temperature, °C
650 700 750 800 850
NRP 0.369 0.305 0.417 0.323
NRP 0.497 0.564 0.623 0.615
NRP 0.9 0.318 — 0.432
— — 0.9 — 0.366
— 0.589 0.522 0.52 0.46
specimens were aged for 0.1 hour, even though no direct correlation between the peak current ratios and the aging temperatures could be observed. The important information obtained in this series of tests was that exposure to temperatures in the range 650 °C to 850 °C for short durations could result in secondary phase precipitation. A similar temperature-time history could be anticipated during quenching of the material from the solution annealing temperature. Delayed cooling and a slightly extended temperature excursion in the range of 650 °C to 850 °C could METALLURGICAL AND MATERIALS TRANSACTIONS A
1. Wrought specimens aged at 650 °C Figure 12 shows the results of SL-EPR tests in 2 M HCl 0.01 M KSCN solution at 60 °C. The reactivation current peak and the associated charge transfer during SL-EPR increased with the increase in aging time for the specimens aged at 650 °C. The DL-EPR test results, as summarized in Table III, also showed similar trends as the SL-EPR test results. In DL-EPR tests, current peaks during both forward and reverse potential sweeps increased with increasing aging time. The increase in critical current density for passivation with increasing aging time showed decreased corrosion resistance of the material because of aging. This behavior is opposite that observed in the 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution. 2. Wrought specimens aged at 700 °C Figure 13 shows some typical results of SL-EPR tests in 2 M HCl 0.01 M KSCN solution at 60 °C. The SL-EPR results clearly showed an increase in charge transfer with an the increase in aging time. The ratio between reactivation peak currents (from DL-EPR tests) increased with aging time; however, the difference was not significant. Comparison of either forward peak currents (critical current for passivation) or reactivation peak currents (from Table III and Figure 14) for different aging times at 700 °C gave good resolution between the aging conditions. The activation peak current increased significantly with the increase in aging time. 3. Wrought specimens aged at 750 °C The results of SL-EPR and DL-EPR tests in 2 M HCl 0.01 M KSCN solution at 60 °C followed similar trends as those of specimens aged at 650 °C and 700 °C. Figure 14 summarizes the activation peak currents and reactivation peak currents observed on all the aging conditions of alloy 22 in 2 M HCl 0.01 M KSCN solution at 60 °C. The activation and reactivation current increased with the increase in time. The I r /I a ratio increased marginally with increasing aging time. A comparison of either forward peak currents or reactivation current peaks between the different aging times at VOLUME 36A, MAY 2005—1159
Fig. 12—Results of SL-EPR test in 2 M HCl 0.01 M KSCN at 60 °C for alloy 22 aged at 650 °C.
Table III. Activation and Reactivation Peak Currents in DL-EPR Tests in 2 M HCl 0.01 M KSCN Solution at 60 °C for Alloy 22 Wrought Specimens Aged at Different Conditions Heat-Treatment Conditions Temperature (°C)/Time (h) 650/1 650/20 650/100 700/1 700/20 700/100 750/1 750/20 750/100 800/100 850/100
Activation Current (A)
Reactivation Current (A)
0.000251 0.00269 0.00497 0.00316 0.00842 0.0134 0.0022 0.014 0.0239 0.052 0.11
0.000192 0.00218 0.003712 0.00207 0.00365 0.008063 0.002372 0.007435 0.0121 0.022 0.034
Fig. 14—Comparison of activation and reactivation current from DL-EPR tests of aged wrought alloy 22 in 2 M HCl 0.01 M KSCN at 60 °C. As aging time is increased, the activation and reactivation current increased at particular aging temperatures.
or elemental depletion conditions can be expected for two different aged conditions. For example, the specimen aged at 700 °C for 100 hours could show a similar GB structure as that of the specimen aged at 750 °C for 20 hours. 4. Wrought specimens aged at 800 °C and 850 °C The DL-EPR results for alloy 22 aged at 800 °C and 850 °C for 100 hours (Table III) also showed similar results. In these specimens, precipitation of TCP phases was found to occur along the grain boundaries and within the grain as well (figures not shown). Therefore, the magnitudes of both activation and reactivation current peaks were very large compared to other aging conditions. C. Chemical Weight Loss Measurements by ASTM G 28A and B Methods
Fig. 13—Results of SL-EPR test in 2 M HCl 0.01 M KSCN at 60 °C for alloy 22 aged at 700 °C.
750 °C gave good resolution between the aging conditions of 1, 20, and 100 hours. Table III shows the average activation and reactivation current from DL-EPR tests in 2 M HCl 0.01 M KSCN at 60 °C for alloy 22 aged at different temperatures and times. In general, both the activation and reactivation peak currents increased with the increase in aging temperature and time, indicating the deteriorated corrosion resistance of the material. The magnitude of the dissolution current cannot be considered as a signature of a specific heat treatment condition, as similar microstructural 1160—VOLUME 36A, MAY 2005
Table IV summarizes the detailed comparison of the observations after ASTM G-28 method A and B tests. Table V gives the average corrosion rates of the wrought and aged alloy 22 in ASTM G-28B and ASTM G-28A methods. In general, ASTM G-28 method B resulted in higher corrosion rates than method A in almost all the aging conditions (except for 650 °C, 1 and 20 hours, and 700 °C for 1 hour). The increase in corrosion rate with aging time was exponential in the case of method B, whereas method A showed a marginal increase in corrosion rate with increase in aging time at all the aging temperatures. However, at the short aging period (1 hour in this case), both methods A and B showed corrosion rates of the same order of magnitude. The pits formed on the specimens (aged for longer aging times at higher temperatures) in method B solution were large, deep, few in number, and mainly at the edges, and as aging time increased, the pits also appeared at the top and bottom surfaces. The specimen aged at 750 °C for 1 hour did not show any visible attack, but as the aging time and temperature increased, the pit depth and density of pits increased. Pits formed in method A solution were pinhole type, large in number, very shallow in depth, and mainly at curved surfaces, but as aging time increased, the pits also appeared at the top and bottom surfaces. The mill-annealed specimen after the ASTM G28-A test showed no attack. METALLURGICAL AND MATERIALS TRANSACTIONS A
Table IV. Summary of the Results of Visual and Microstructural Observations of Specimens after Chemical Weight Loss Tests with ASTM G 28 Methods A and B Comparison in ASTM G-28 A and B Weight Loss Test Heat Treatment 650 °C/1 h 650 °C/20 h 650 °C/100 h
700 °C/1 h 700 °C/20 h
ASTM G-28 A
ASTM G-28 B
corrosion rate (CR) 0.55 mm/year,dark and large number of triple junctions observed, no visible attack on outer surfaces CR: 0.86 mm/year, dark GB can be easily seen, small, shallow pits on curved surfaces, small attack on top and bottom surfaces CR: 1.23 mm/year, pinhole type small and shallow pits, large in number, uniformly distributed over the curved surfaces and also on the top and bottom surfaces, GB was clearly revealed with mild attack on some grains CR: 0.47 mm/year, GB revealed as very dark continuous network, visible attack only at top and bottom surfaces CR: 1.65 mm/year, pinhole type shallow pits, large in number, on curved surfaces, dark continuous GB, few fallen grains
CR: 0.1326 mm/year, hard to find any attack on GB, no visible attack on outer surfaces CR: 0.599 mm/year, GB not clearly visible, minor attack on curved surfaces, no attack on top and bottom surfaces CR: 86.3 mm/year, large and deep pits, few in number, microstructure showed severe attack on the bulk of the grains plus falling of some grains CR: 0.4034 mm/year, discontinuous light GB network, minor visible attack CR: 87.85 mm/year, very large, deep pits concentrated mainly at edges, light continuous GB network, four to five attacked grains but not fallen CR: 95.83 mm/year, deep, large pits on curved surfaces, edges and on top and bottom surfaces, attack inside grains also, few fallen grains CR: 1.21 mm/year, light and discontinuous GB network, hardly visible attack on outer surfaces CR: 80.27 mm/year, large number of fallen grains, large and deep pits at the edges and top and bottom surfaces CR: 101.67 mm/year, few fallen grains but large number of attacked grains, small, deep pits on the curved surfaces and top and bottom surfaces but large and deep pits at the edges
700 °C/100 h
CR: 3.86 mm/year, pinhole type shallow pits, large in number on curved surfaces and few on top and bottom surfaces, dark continuous GB network, few fallen grains
750 °C/1 h
CR: 0.78 mm/year, dark, deep GB attack, hardly visible attack on outer surfaces
750 °C/20 h
CR: 1.95 mm/year, very few grains attacked inside but not fallen, pinhole type shallow pits on curved surfaces
750 °C/100 h
CR: 4.05 mm/year, attack on few grains, pinhole type shallow pits on curved surfaces and at the center of top and bottom surfaces
Table V. Average Corrosion Rates of Aged Alloy 22 after 24-Hour Immersion in Boiling Acid Test According to ASTM G-28 Method A and ASTM G-28 Method B Aging Condition (°C/h) Mill annealed 650/1 650/20 650/100 700/1 700/20 700/100 750/1 750/20 750/100
Average Corrosion Rate (mm/year) in ASTM G-28A Test
Average Corrosion Rate (mm/year) in ASTM G-28B Test
0.1437 0.5582 0.8567 1.2214 0.4781 1.6466 3.8603 0.7765 1.9454 4.5428
0.0995 0.132 0.599 86.33 0.4034 87.84 95.82 1.21 80.26 101.67
Figure 15 shows the scanning electron microscopy (SEM) micrograph of the unpolished surface of the alloy 22 specimen aged at 750 °C for 1 hour after the method A test, and Figure 16 shows the SEM micrograph of the alloy 22 specimen aged at 750 °C for 1 hour after the method B test. A clear difference in degree of attack on the GB of these specimens can be observed. It is interesting to note that even METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 15—SEM micrograph of the unpolished surface of wrought alloy 22 aged at 750 °C for a 1 h specimen after the ASTM G 28-A test. Grain boundaries are visible, which indicates the attack at the GBs.
though the corrosion rate was higher after the method B test than after that of method A, the attack on the GB was more pronounced by test method A. SEM micrographs of the alloy 22 specimen aged at 650 °C for 1 hour after the G28-A test VOLUME 36A, MAY 2005—1161
Fig. 16—SEM micrograph of polished surface of wrought alloy 22 aged at 750 °C for a 1 h specimen after the ASTM G 28-B test. The GB is not visible, which indicates no attack on the GB.
showed triple junctions as well as some discontinuous grain boundaries (not shown here). However, the same specimen after the G28-B test did not show any GB attack. The solution of test method A preferentially attacked grain boundaries, whereas the general corrosion attack was observed in the solution of method B. The solution of method A has more concentration of sulfuric acid than that of method B. As sulfuric acid solution is used for detecting Cr depletion, it can be considered that the Cr-depleted regions were attacked in the method A tests. Figure 17 shows the SEM micrograph of the specimen aged at 750 °C for 20 hours after the ASTM G28-B test. Severe intergranular attack and a few fallen grains could be observed, which led to a high corrosion rate. A good correlation between ASTM G-28-B test results and EPR results in 2 M HCl 0.01 M KSCN at 60 °C could be obtained when only the magnitudes of the activation or reactivation current peaks were compared with the respective ASTM G28-B corrosion rates. Figure 18 illustrates an example of this comparison, where the DL-EPR reactivation peak currents of alloy 22 specimens aged at 750 °C and 700 °C were compared with the results of ASTM G28 method B tests. The good correlation between these two tests indicates that the EPR test can be a better evaluation method for in-situ applications. Moreover, it is reasonable to assume that the EPR test could give better reproducible results than that of the weight loss measurements. The reason for this assumption is that weight loss measured after boiling acid immersion testing pertains not only to the attack on the alloying element-depleted zone, but also to the dropout of the precipitates when the adjacent area was completely attacked, which could vary from specimen to specimen. Such experimental variations are not present in the case of electrochemical testing. D. Transmission Electron Microscopy Analysis of the Precipitates Composition profile analyses were carried out on thin specimens (thinned by ion milling or electrochemical polishing) using transmission electron microscopy–energy-dispersive 1162—VOLUME 36A, MAY 2005
Fig. 17—SEM micrograph of an alloy 22 specimen aged at 750 °C for a 20 h specimen after the ASTM G28-B test. The micrograph shows severe intergranular attack.
Fig. 18—Comparison of the corrosion rate in the ASTM G28-B method with the reactivation current in the 2 M HCl 0.01 M KSCN solution at 60 °C.
Fig. 19—Chemical composition profile across a GB precipitate of an alloy 22 specimen, thermally aged at 750 °C for 20 h. The GB precipitate shows a decrease in Cr content. Both Cr and Mo depletion could be observed adjacent to the precipitate. The inside figure shows the location of the precipitate from where the profile is taken.
X-ray (TEM-EDX). Figure 19 shows a typical composition profile observed across a GB precipitate of the alloy 22 specimen aged at 750 °C for 20 hours. The precipitate was enriched with Mo and W and was about 100-nm wide. CorMETALLURGICAL AND MATERIALS TRANSACTIONS A
respondingly, the vicinity of the precipitate was depleted with Mo. The width of the depleted region was about 125 to 180 nm from the interface of the precipitate/matrix. The decrease in Cr content was observed inside the GB precipitate and in the adjacent area as well. Another type of GB precipitate containing enriched Mo, Cr, and W content was observed in another specimen of alloy 22 aged at 750 °C for 20 hours. Figure 20 shows the elemental profile across such a GB precipitate. Depletion of the Mo content near the GB precipitate is clearly seen, and it was measured to be 24 pct less compared with the matrix composition. The Cr profile is relatively flat near the GB. However, the compositional heterogeneity between the GB precipitate and the adjacent matrix is predominant, which could lead to localized corrosion attack. Figure 21 shows a composition profile observed across a GB precipitate of the alloy 22 specimen aged at 750 °C for 100 hours. For all the line profiles, the X-ray intensity of each element has been normalized with the total sum of intensities of all major alloying elements at each point of the line scan. The precipitate was enriched with Mo and W, and was about 185-nm wide. Correspondingly, the vicinity of the precipitate was depleted with Mo and W. The width of the depleted region was about 180 to 230 nm from the interface
Fig. 20—Chemical composition profile across a GB precipitate of alloy 22 aged at 750 °C for 20 h specimen. This profile is from the same specimen as in Fig. 19. However, the GB precipitate here shows an increase in Cr content. The inside figure shows the location of the precipitate from where the profile is taken.
Fig. 21—Chemical composition profile across a GB precipitate of the alloy 22 specimen, thermally aged at 750 °C for 100 h. The GB precipitate shows a decrease in Cr content. Both Cr and Mo depletion could be observed adjacent to the precipitate. METALLURGICAL AND MATERIALS TRANSACTIONS A
of the precipitate/matrix. The minimum content of the Mo was about 8.9 wt pct, indicating that this region contained 30 pct less Mo than that of bulk content. It is interesting to note that the Cr profile is flat at the interfaces. Figure 22 shows the precipitates found in alloy 22 specimen aged at 750 °C for 100 hours. The particular precipitate was analyzed as 27.9Ni, 14.8Cr, 49.6Mo, 4.2W, 2.2Fe, and 0.9Co (wt %). The precipitate is highly enriched with Mo. In addition, small precipitates were observed to be present in different orientations. The depleted Mo profile supports the EPR results in the 2 M HCl 0.01 M KSCN solution at 60 °C, which showed large reactivation peaks. E. Weld Metal The as-weld microstructure reveals the dendritic structure after electrochemical etching of the specimen at 1.2 V for 60 seconds in 0.5 M sulfuric acid solution. The TCP phase as the terminal solidification constituent in the as-welded condition was not revealed even though other researchers have observed it.[5,6,7] The microstructure of the weld metal aged at 760 °C for 140 hours revealed GB precipitates along with interdendritic precipitates. The DL-EPR tests carried out at 30 °C in the 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution (used for detecting Cr depletion) did not show either an activation or reactivation peak current for any of these specimens. The specimens showed passivity during forward potential sweep without revealing a critical current peak prior to passivation. The reverse potential sweep after reaching the potential of 400 mV (SSC) also did not show the reactivation peak at 30 °C. The observed results indicated that the Cr profile across the dendrites and across the grain boundaries was flat and no depletion occurred due to formation of secondary phases. Cieslak and co-workers[5,6,7] also reported similar results based on the TEM-EDX analyses of the weld dendrite structure of alloy 22. The results of the DL-EPR test for weld aged alloy 22 in 2 M HCl 0.01 M KSCN at 30 °C were interesting. The weld 760/140 hours specimen showed a very large activation peak and the weld 760/1 hour specimen showed
Fig. 22—TEM micrograph of an alloy 22 specimen aged at 750 °C for 100 h showing precipitates in different orientations. VOLUME 36A, MAY 2005—1163
a very small activation peak. This result encouraged us to do the tests in the same solution but at a higher temperature of 60 °C. Figure 23 shows the typical results of DL-EPR tests in 2 M HCl 0.01 M KSCN solution at 60 °C. A minimum of three experimental runs were carried out for each aged condition of the specimens. Table VI summarizes the peak current ratios (I r /I a) of EPR tests. The table gives minimum, maximum, and average values of the ratios of the peak currents among the test runs. The alloy 22 weld specimen aged at 760 °C for 140 hours showed very high activation and reactivation current, indicating high Mo depletion. Figure 24 shows the microstructure of weld metal aged
at 760 °C for 140 after potentiostatic etching at the potential corresponding to the reactivation current peak. The grain boundaries and interdendrites were preferentially attacked. The absence of reactivation peaks at 30 °C and the preferential attack at 60 °C in the same solution indicate the presence of the activation energy barrier for corrosion. At present, the mechanism is not exactly known, but it is possible that adsorption of sulfur species could increase with temperature at the alloy element depleted regions above a threshold limit, causing metal-metal bond weakening and leading to corrosion. IV. DISCUSSION A. Wrought Metal
Fig. 23—DL-EPR test results for an alloy 22 weld specimen in 2 M HCl 0.01 M KSCN at 60 °C.
Table VI. Reactivation Current to Activation Current Ratios of Weld Specimens during DL-EPR Tests in 2 M HCl 0.01 M KSCN Solution at 60 °C Aged Condition Mill-annealed wrought As-welded 700 °C for 24 h 760 °C for 1 h 760 °C for 140 h
Minimum Ir/Ia NP 3.4 104 2 104 1.97 104 0.227
Maximum Ir /Ia
Average Ir /Ia
NP 5.43 104 3.98 104 2.47 104 0.322
NP 4.27 104 2.94 104 2.28 104 0.29
Fig. 24—Interdendritic and intergranular attack on an aged (760 °C for 140 h) weld after potentiostatic etching in a 2 M HCl 0.01 M KSCN at reactivation peak potential. 1164—VOLUME 36A, MAY 2005
Generally, it is expected that the degree of sensitization increases with increasing aging time. Huebner et al. observed a corrosion rate increase with increasing aging time for alloy 22, which was attributed to the depletion of Mo. It should be noted that the increased corrosion rates of sensitized austenitic stainless steels, INCONEL 600, and alloy 825 were attributed to Cr depletion. The increase in corrosion rate with increasing aging time depends on the aging temperature. Pan et al. observed a deeper and narrower Cr depletion profile in alloy 825 during the initial stages of aging at 700 °C, and with increasing aging time, the Cr depletion profile became shallower and broader. Lopez et al. correlated EPR test results in 2 M H2SO4 0.5 M NaCl 0.01 M KSCN solution to the Cr depletion in duplex stainless steels. According to Kruger et al., EPR test results in the sulfuric acid KSCN solution can be better correlated with the width of the Cr-depleted zone rather than with the depletion zone depth for Ni-Cr-Fe alloys. Similarly, in this investigation also, the reactivation current ratio in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution can be correlated to the Cr-depleted regions formed during initial stages of aging. Without carrying out analyses of the Cr profile across the grain/precipitate boundaries, it is difficult to understand if the reactivation current ratio was more sensitive to the Cr depletion zone width or zone depth. Kruger et al. observed that 14 pct Cr content was the threshold for intergranular attack for a nickel alloy containing 16 pct Cr. Pan et al. observed that for alloy 825, the critical Cr content below which intergranular corrosion attack could be observed was 19 pct against a bulk Cr content of 22 pct. Therefore, it is reasonable to assume in the present study that the EPR tests were more sensitive to regions with less than 19 pct Cr. The DL-EPR results show that during aging at 610 °C, the precipitation process increased the Crdepleted zone size with increasing aging time from 1 hour to 20 hours but again decreased from 20 to 100 hours. Aging at 650 °C, 700 °C, or 750 °C initially caused very deep and narrow Cr-depleted zone, which was replenished with diffusion of Cr from the bulk of the grain with increasing aging time. This process resulted in a shallower and wider Cr depletion profile. If the Cr depletion profile shows higher Cr content than the detection limit of the EPR test, no reactivation would be observed. Generally, SL-EPR test results support the DL-EPR test results. Published literature attributes the increased corrosion rate of aged Ni-Cr-Mo-W alloys to the dissolution of Mo-rich METALLURGICAL AND MATERIALS TRANSACTIONS A
phase and associated Mo depletion.[8,12–14] However, this study on alloy 22 specimens using modified electrochemical reactivation tests shows that Cr depletion occurs during the initial stages of thermal aging. In the absence of extensive microstructural studies with elemental profile analyses, the Cr depletion could be attributed to three possible reasons: (1) formation of Cr-rich carbides, (2) formation of Cr-rich ordered phase of type Ni2Cr, and (3) formation of Cr-rich (Ni, Fe, Co)3(Cr, Mo, W)2 type -phase precipitates. Even though the carbon content of the alloy is very low, precipitation of carbides is possible as the kinetics of carbide precipitation were observed to be faster than -phase precipitation. Pan et al. observed extensive carbide precipitation for alloy 825 with 3 times more carbon content than alloy 22 at similar temperatures. When the Mo content is more than 6 pct, it is expected that the predominant carbide phase is M6C. Interestingly, both the M6C and phase showed almost similar chemistry,[5,9,23] and Hodge showed that the phase has about 40 pct more Cr than the bulk. Therefore, it is possible that these Cr-rich carbides could act as nuclei for other secondary precipitates. Precipitation of short-range-ordered phases such as Ni2Cr during low-temperature exposure could act as nuclei for precipitation of other phases or could result in Cr partitioning, even though this ordering occurs homogeneously. Segregation of S, P, and Si along the grain boundaries during aging also could contribute to the anodic dissolution current. However, in DL-EPR tests, such current predominantly occurs during the forward sweep, and therefore, the reactivation current during the reverse sweep can be attributed to only the Cr-depleted zone. The present investigation showed precipitation of secondary phases in wrought alloy 22 occurring in two stages: stage 1—nucleation of Cr-rich precipitates during the initial stages of aging (aging at 610 °C and initial stages of aging at 650 °C, 700 °C, and 750 °C for 1 hour); and stage 2—precipitation of Mo-rich -phase particles (long aging times ( 1 hour) at 650 °C to 750 °C) during which time the Cr-depleted zone was replenished. The TEM-EDX results on the specimen aged at 750 °C for 100 hours also showed evidence of flat Cr and depleted Mo profiles adjacent to the GB precipitate. The EPR test results in 2 M HCl 0.01 M KSCN solution showed increased activation and reactivation current peaks with an increase in aging temperature and time for most aging conditions. The change in the reactivation current ratio was not significant enough to correlate it to available ASTM standard test results. However, if the current peaks during forward and reverse sweeps are considered independently for different aged conditions, it could be observed that the current peaks increased with an increase in aging conditions, as seen in Figures 12 through 14. The current peak during forward sweep (Ia), which is also denoted as the critical current for passivation, is an indication of the ability of the material to passivate. For example, the critical current density for passivation of 304 SS in 0.5 M H2SO4 was observed to be two orders higher than that of alloy 22. The critical current was observed to increase with the aggressiveness of the environment and to decrease with the ability of the material to passivate. Therefore, the increase in critical current for passivation clearly indicates a loss in corrosion resistance of the material with aging. In summary, the EPR results in 1 M H2SO4 0.5 M NaCl 0.01 M METALLURGICAL AND MATERIALS TRANSACTIONS A
KSCN solution indicate that during the initial stages of aging, Cr-rich precipitate forms, which results in Cr depletion, and as the aging time is increased, TCP phases such as the phase, which is enriched in Mo with a possible composition of Ni7Mo6 or Ni3Mo2, form, resulting in Mo depletion and increased activation and reactivation currents in 2 M HCl 0.01 M KSCN solution at 60 °C. The TEM-EDX data also supported these results. ASTM weight loss tests The ASTM G-28 method A test employs sulfuric acid ferric sulfate solution to determine the susceptibility to intergranular corrosion. In general, the attack on the specimens heat treated at different temperatures of 650 °C and 700°C for 1 hour was greater in method A than that in method B. The I r /I a ratio was also high for these specimens in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution and numerically the same at 0.81 and 0.83, respectively. The specimen aged at 750 °C for 1 hour showed less of a corrosion rate in method A than that in method B. The I r /I a ratio (0.52) for this specimen in the preceding solution was also less as compared with specimens aged at 650 °C for 1 hour and 700 °C for 1 hour. Thus, the weight loss results also support the electrochemical results. The microstructure observation also revealed greater GB attack after the method A test (Figure 15 and 16). At lower aging temperatures and shorter aging times, the occurrence of Cr depletion could be anticipated and the metal becomes susceptible to intergranular corrosive attack. A solution containing sulfuric acid and a suitable activator is generally used for detecting Cr depletion as in the case of type 304 stainless steel. Therefore, method A, which contains sulfuric acid (oxidizing environment), could be a more suitable test in these aging conditions. Method B solution contains 23 pct sulfuric, but it also contains hydrochloric acid (predominantly reducing type). In method B, Cl ions play a very important role in the destabilization of the passive film, and hence, the corrosion rate in method B can be mainly attributed to the role of acidified Cl ions. At higher aging time and aging temperature, TCP phases form and Mo, W depletion takes place. Hence, all 100-hour heat-treated specimens showed a high corrosion rate in method B as compared with method A. The specimen aged at 700 °C for 1 hour showed a low corrosion rate in method B as compared to method A. However, with an increase in aging time, the corrosion rate increased at a faster rate after the method B test as compared with method A. This suggests that at shorter aging times, the Cr depletion plays a major role, and after longer aging times, Mo depletion predominates. Though the corrosion rates of the specimens aged at 700 °C for 100 hours and 750 °C for 100 hours after method B test were higher than those of the specimens aged at 700 °C for 20 hours and 750 °C for 20 hours, respectively, after the method B test, the difference was marginal. This suggests the 20 hours of aging time was sufficient to cause predominantly Mo depletion at higher temperatures. However, at a lower temperature (650 °C), longer than 20 hours of aging time was found to be required to cause significant Mo depletion. This is evident from the weight loss in method A and the EPR results of alloy 22 specimens aged at 650 °C for different aging times. For the specimens aged at 650 °C for 1 hour and 20 hours, VOLUME 36A, MAY 2005—1165
the weight loss and reactivation current ratio matched each other. Overall, the ASTM G-28 test methods also support the occurrence of Cr depletion at shorter aging times and lower temperatures and the occurrence of Mo depletion at longer aging times and higher aging temperatures. B. Weld Metal In weld metal specimens, both in as-weld and aged conditions, no reactivation peaks could be observed in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution. The Mo segregation at the interdendrites and formation of Mo-rich TCP phases has been observed in the weld metal.[5,6,7] The high-temperature TCP phases were retained due to terminal solidification products of the weld metal.[56,67] During aging of the weld metal, no chromium depletion could be observed as the precipitation reaction proceeded with the already available retained intermetallic phases or Mosegregated regions. Stage 1 of the precipitation reaction, which was observed in wrought metal, was absent in the case of the weld metals. The weld specimens aged at 760 °C for 140 hours showed very high ratios of reactivation to activation current as compared to other aged specimens. When comparing the current ratios, it could be observed that as-welded specimens showed relatively higher values than the specimens aged at 700 °C for 24 hours and 760 °C for 1 hour. However, if the current peaks during forward and reverse sweeps are considered independently for different aged weld specimens, it can be observed that the current peaks increased with aging conditions, as seen in Figure 23. In this present study, as-welded specimens showed relatively lower peak currents during both forward and reverse sweeps. The observation of the reactivation current peak for the as-welded specimen is an indication of the presence of Mo- and W-depleted regions. The depletion of these elements could occur at the dendrite cores and in the vicinity of the TCP phases. Summers et al. reported a 2.7 vol pct of TCP phases in the as-welded condition. The volume of TCP phases increased to about 5 pct when aged at 760 °C for 1 hour or 700 °C for 20 hours and to about 30 pct after aging for 100 hours at 760 °C. The reactivation current behavior could be directly correlated to the formation of TCP phases and associated Mo depletion, as the current during reverse sweep increased only marginally during the initial stage of aging at 760 °C and increased substantially after prolonged aging, which followed the volume fraction of TCP phases present in the material. The depletion of Mo and W at the dendrite core may not have caused preferential attack, as the content of these elements would have been above a threshold level. At present, the minimum level of Mo or W required to resist the corrosion attack by 2 M HCl 0.01 M KSCN solution is not known. V. CONCLUSIONS The EPR testing of aged wrought materials in 1 M H2SO4 0.5 M NaCl 0.01 M KSCN solution at 30 °C detected Cr depletion, while EPR testing in 2 M HCl 0.01 M KSCN solution at 60 °C detected Mo depletion. Aging at or above 650 °C for long durations ( 1 hour) replenished the Cr depletion as the reactivation current peaks decreased with increasing aging time. An increase in the aging 1166—VOLUME 36A, MAY 2005
temperature and time increased the corrosion rate in ASTM G-28 method B tests, which could be attributed to the TCP phase formation and associated Mo and W depletion. The weight loss measurements could be correlated to the DL-EPR tests in 2 M HCl 0.01 M KSCN solution at 60 °C. The EPR tests of aged wrought alloy 22 suggested two possible steps of precipitation reactions first: an unidentified Cr-rich phase precipitates along grain boundaries followed by a second reaction of formation of Mo-rich phase. ACKNOWLEDGMENT This work was sponsored by the Office of Civil Radioactive Waste Management, United States Department of Energy, under the cooperative agreement with the University and Community College system of Nevada. The authors are grateful to Drs. Gerald Gordon, Frank Wong, and Raul Rebak for their interest and support. REFERENCES 1. T.S. Edgecumbe Summers, M.A. Wall, M. Kumar, S.J. Mathews, and R.B. Rebak: Scientific Basis for Nuclear Waste Management XXII, Materials Research Society Symposium Proceedings, D.J. Wronkiewicz and J.H. Lee, eds., Materials Research Society, Warrendale, PA, 1999, vol. 556, pp. 919-26. 2. R.B. Rebak, T.S. Edgecumbe Summers, and R.M. Carranza: Scientific Basis for Nuclear Waste Management XXIII, Materials Research Society Symposium Proceedings, R. W. Smith and D.W. Shoesmith, eds., Materials Research Society, Warrendale, PA, 2000, vol. 608, pp. 109-14. 3. T.S. Edgecumbe Summers, R.B. Rebak, T.A. Palmer, and P. Crook: Scientific Basis for Nuclear Waste Management XXV, Materials Research Society Symposium Proceedings, B.P. McGrail and G.A. Cragnolino, eds., Materials Research Society, Warrendale, PA, 2002, vol. 713, pp. 45-52. 4. Yucca Mountain Science and Engineering Report, Rev. 1, Feb. 2002 (DOE/RW-0539, Rev. 1 [www.ymp.gov/documents/ser_b/index.htm]). 5. M.J. Cieslak, T.J. Headley, and A.D. Romig Jr.: Metall. Trans. A, 1986, vol. 17A, pp. 2035-46. 6. J.S. Ogborn, D.L. Olson, and M.J. Cieslak: Mater. Sci. Eng., 1995, vol. A203, pp. 134-39. 7. M.J. Cieslak, G.A. Knorovsky, T.J. Headley, and A.D. Romig, Jr.: Metall. Trans. A, 1986, vol. 17A, pp. 2107-16. 8. U.L. Heubner, E. Altpeter, M.B. Rockel, and E. Wallis: Corrosion, 1989, vol. 45, pp. 249-58. 9. D. Peckner and I.M. Bernstein: Handbook of Stainless Steels, McGrawHill Book Co., New York, NY, 1977, pp. 4-53. 10. A. Turnbull, P.E. Francis, M.P. Ryan, L.P. Orkney, A.J. Griffiths, and B. Awkins: Corrosion, 2002, vol. 58, pp. 1039-48. 11. H. Samans, A.R Meyer, and G. F. Tisinai: Corrosion, 1966, vol. 22, pp. 336-41. 12. R.B. Leonard: Corrosion, 1969, vol. 25, pp. 222-28. 13. M.A. Streicher: Corrosion, 1976, vol. 32, pp.79-85. 14. F.G. Hodge and R.W. Kirchner: Corrosion, 1976, vol. 32, pp. 332-39. 15. M. Raghavan, B.J. Berkowitz, and J.C. Scanlon: Metall. Trans. A, 1982, vol. 13A, pp. 979-84. 16. U. Heubner and M. Kohler: Werkstoffe Korr., 1992, vol. 43, p. 181. 17. D.S. Dunn, Y.-M. Pan, and G.A. Cragnolino: Corrosion, NACE, Houston, TX, 2000, paper no. 00206. 18. G.S. Was and R.M. Kruger: Acta Metall., 1985, vol. 33, pp. 841-54. 19. R.M. Kruger, S.F. Claeys, and G.S. Was: Corrosion, 1985, vol. 41, pp. 504-12. 20. H.M. Tawancy: J. Mater. Sci., 1996, vol. 31, pp. 3929-36. 21. Z. Fang, Y.S. Wu, L. Zhang, and J.Q. Li: Corrosion, 1998, vol. 54, pp. 339-45. 22. N. Lopez, M. Cid, M. Puiggali, I. Azkarate, and A. Pelayo: Mater. Sci. Eng. A, 1997, vol. 229, pp. 123-28. 23. F.G. Hodge: Corrosion, 1973, vol. 29, pp. 375-83. 24. R.M. Kruger and G.S. Was: Metall. Trans. A, 1988, vol. 19A, pp. 2555-62. METALLURGICAL AND MATERIALS TRANSACTIONS A
25. C. Gabrielli: Technical Report No. 004/83, Solartron Instruments, Farnborough, United Kingdom, 1980. 26. Y.M. Pan, D.S. Dunn, G.A. Cragnolino, and N. Sridhar: Metall. Trans. A, 2000, vol. 31A, pp. 1163-73. 27. K. Miyata and M. Igarashi: Metall. Trans. A, 1992, vol. 23A, pp. 953-59.
METALLURGICAL AND MATERIALS TRANSACTIONS A
28. K.S. Raja, D.D. Gorhe, S.A. Namjoshi, and D.A. Jones: 11th Int. Conf. on Environmental Degradation of Materials in Nuclear Power Systems— Water Reactors, Stevenson, Washington, Aug. 10–14, 2003, American Nuclear Society, IL, pp. 248-58. 29. K.S. Raja and D.A. Jones: Corrosion, 2003, NACE, Houston, TX, paper no. 03396.
VOLUME 36A, MAY 2005—1167