Microstructure and properties of post weld heat treated2.25Cr1Mo weld metal
E-L. Bergquist, L. Karlsson, M.Thuvander, E. Keehan
ESAB AB, P.O Box 8004, SE-402 77 Gothenburg, Sweden
Abstract
In the present investigation, 2.25Cr1Mo weld metals deposited with manual metal arc (MMA) welding and submerged arc welding (SAW) have been compared with respect to their tolerance to long post weld heat treatment (PWHT). Mechanical properties including the impact transition temperature have been determined and the microstructures have been characterised using light optical microscopy and field emission gun scanning electron microscopy.
As expected, the tensile strength and the hardness of the weld metals decrease with increased tempering time. Of greater interest is the impact toughness results. Analysing the 47 J impact transition temperature, the MMA weld metal withstands long duration PWHT better than the SAW weld metals. The microstructures suggest that this could be an effect of differences in bainite morphology and grain size, characteristics governed by the cooling rate and the chemical composition.
Keywords
2.25Cr1Mo, post weld heat treatment, microstructure, tensile strength, impact toughness
Introduction
Large welded components in 2.25Cr1Mo steel are likely to be repaired several times during their service life. Each time a post weld heat treatment (PWHT) will be made in order to relieve stresses and improve mechanical properties. In certain locations, the weld metal might be subjected to multiple PWHT. As it is well known that long duration PWHT eventually leads to degradation of especially the impact toughness [1], it is of great interest to better understand the effects on microstructure and properties.
The study is focussed on explaining differences in tolerance to long duration heat treatment observed when comparing weld metal produced with manual metal arc welding (MMA) and weld metal produced with submerged arc welding (SAW).
Materials
A multilayer MMA weld deposited with the stick electrode ESAB OK 76.26 (AWS A/SFA 5.5 E9018-B3) was included in the investigation together with multilayer welds deposited with SAW and the consumable combination ESAB OK Autrod 13.20SC / ESAB OK Flux 10.63 (AWS A/SFA 5.23 F8P8-EB3-B3).
The assembly of the joints and the welding parameters were, with exception for the welding speed, in accordance with AWS SFA 5.5 [2] and AWS SFA 5.23 [3], respectively. The welding parameters are shown in table 1.
Table 1: Welding parameters
| Sample designation | Welding method | Electrode ∅ (mm) | Polarity | Current (A) |
|---|---|---|---|---|
| MMA 0.9 | MMA | 4 | DC+ | 170 |
| SAW 1.3 | SAW | 3 | DC+ | 450 |
| SAW 2.3 | SAW | 4 | DC+ | 550 |
| Sample designation | Voltage (V) | Welding speed (mm/s) | Heat input (kJ/mm) | Calculated t 8/5 (s) |
|---|---|---|---|---|
| MMA 0.9 | 27 | 3.9 | 0.9*) | 10.6 |
| SAW 1.3 | 28 | 10 | 1.3**) | 13.3 |
| SAW 2.3 | 28 | 6.7 | 2.3**) | 25.6 |
*) Thermal efficiency factor k=0.8 **) Thermal efficiency factor k=1.0
One weld from each group was kept in the as welded condition while the other welds were tempered as shown below. The two step heat treatment routines were chosen on the basis of actual customer requests.
- Heat treatment routine A: 690+5°C, 1h.
- Heat treatment routine B: 630+5°C, 33h directly (without cooling) followed by 690+5°C, 10h.
- Heat treatment routine C: 630+5°C, 33h directly (without cooling) followed by 690+5°C, 15h.
The heating rate above 400°C was restricted to maximum 220°C/h and the cooling rate down to 400°C was restricted to
Doc. II-A-168-05
Table 2. Chemical composition of all-weld metal. All results in wt.% unless otherwise stated.
maximum 250°C/h.
Experimental procedures
The chemical composition of the weld metals was determined using optical emission spectroscopy for all elements except S, C, O and N for which LECO combustion equipment was used.
Cross sections were cut and prepared using standard metallographic techniques. Two different etchants were used to reveal the microstructure. For light optical microscopy (LOM) the colour etchant Beraha I was considered most suitable, while for FEG-SEM (Field Emission Gun Scanning Electron Microscopy) a mixture of CuCl2, HCl, picric acid, ethanol and water was used.
Tensile test specimens with circular cross section and diameter 10 mm were machined from the centre of the joints in the longitudinal direction. Thereby, the tested material was all-weld metal. The tensile strength was determined at room temperature in accordance with EN 10002-1.
The hardness of the weld metals was determined using the Vickers method with 10 kg load. Series of indentations were made on polished cross sections along lines starting in primary weld metal close to the top surface and finishing close to the root surface. The spacing of the indentations was 0.5 mm.
The Charpy-V impact toughness was determined at five temperatures using five test pieces at each test temperature. Standard sized 10x10 mm specimens were used. Testing was performed in accordance with ASTM E23. The temperature at which the average impact toughness fell below 47 J was determined (47J-ITT). The 47 J level was chosen based on customer requests.
Results
Chemical composition
The chemical composition of the weld metals are presented in Table 2. Minor differences in the Mn, V, Ti and Al contents of the weld metals are noted. The difference in Mn-content will be discussed later. The differences in V, Ti and Al contents are considered insignificant in this case.
Microstructure
The microstructural features considered important, were very similar when comparing the last welded bead and reheated regions of the welds. This was in particular the case for the longer two-step PWHTs of greatest interest in this study. Only microstructures from the last welded bead will therefore be presented in the following sections.
| Element | MMA 0.9 | SAW 1.3 | SAW 2.3 |
|---|---|---|---|
| C | 0.071 | 0.068 | 0.071 |
| Si | 0.35 | 0.29 | 0.33 |
| Mn | 0.81 | 0.66 | 0.71 |
| P | 0.007 | 0.007 | 0.007 |
| S | 0.004 | 0.003 | 0.002 |
| Cr | 2.29 | 2.22 | 2.21 |
| Ni | 0.05 | 0.06 | 0.06 |
| Mo | 1.08 | 0.98 | 0.96 |
| V | 0.023 | 0.011 | 0.010 |
| Nb | 0.008 | 0.007 | 0.006 |
| Al | <0.001 | 0.020 | 0.019 |
| Ti (ppm) | 100 | 20 | 20 |
| B (ppm) | <1 | 1 | 2 |
| O (ppm) | 350 | 390 | 250 |
| N (ppm) | 110 | 130 | 80 |
As welded condition
All three weld metals are largely bainitic but with some important differences when comparing the finer details. The microstructure of the MMA weld as well as the SAW 1.3 sample consists of packages of elongated and oulined platelets (Figure 1). These are interpreted as indiviual bainitic ferrite platelets outlined by retained carbon-enriched austenite. Carbides may possibly also be present between the platelets.
The bainitic ferrite platelets appear slightly thinner in the MMA sample compared to the SAW 1.3-weld (Figures 1a-b and 1c-d). In the SAW 2.3 sample a generally coarser and less elongated morphology is seen (Figure 1e-f). The residual austenite/carbides are nearly continuously surrounding the platelets in the MMA sample whereas the platelets are somewhat less well defined in the SAW 1.3 sample. In the SAW 2.3 sample the residual austenite and/or carbides appear mostly as separated formations.
Carbide precipitation is observed inside many of the platelets in the MMA weld, but not in the two SAW samples.
Post weld heat treated condition A: 690°C, 1h. After 1 h heat treatment at 690°C, the retained austenite has started to decompose into aligned but discontinous carbides along the platelet boundaries (Figure 2). A fairly large amount of intragranular precipitates has formed in all samples.
Post weld heat treated condition B: 630°C, 33h + 690°C, 10h.
After this two-step heat treatment the individual bainitic ferrite platelet boundaries are no longer easily resolvable and mainly the sheaf bondaries appear in the microscope. The retained austenite outlining the bainitic ferrite platelets has largely disappeared.
c) SAW, 1.3 kJ/mm, LOM d
e) SAW, 2.3 kJ/mm, LOM f) SAW, 2.3 kJ/mm, FEG-SEM
Figure 1. The as welded microstructure is mainly bainitic in all three welds. An example of intra-platelet precipitation can be seen in Figure b (arrow)
High magnification FEG-SEM studies revealed that the intragranular precipitation is very sparse in the MMA sample after PWHT routine B. It is richer in the SAW 1.3 sample but
Doc. II-A-168-05
b) SAW, 1.3 kJ/mm
c) SAW, 2.3 kJ/mm
Figure 2. FEG-SEM micrographs of the weld metals after 1 h tempering at 690°C. Aligned carbides are decorating the platelet boundaries.
richest in the SAW 2.3 sample. Large carbides are observd at the bainite sheaf boundaries in all three samples.
Post weld heat treatment condition C:630°C, 33h+690°C, 15h
The rearrangement of carbides from the bainite platelet boundaries to the bainite sheaf boundaries has proceeded further. Light optical micrographs of the as welded microstructure and the microstructure after the longest heat treatment illustrate the rearrangement (Figure 3).
FEG-SEM studies showed that the intragranular precipitation is relatively sparse in the MMA sample. In both SAW samples, the density of precipitates has decreased compared to after heat treatment routine B. Large carbides are observed in all samples.
Tensile strength
The MMA weld metal has higher proof and ultimate tensile strength than the SAW weld metals (Figures 4 and 5). This difference remains, although at a lower strength level, also
Doc. II-A-168-05
after PWHT. The proof strength and the ultimate tensile strength have decreased for all samples after PWHT routine A, and even more after PWHT outine B. Comparing PWHT routine B and routine C, the tensile strength is more or less constant.
1000
900
800
700
b) MMA, 0.9 kJ/mm, tempered
Rp0,2 (MPa)
600 500 400 300
200 100 0
Figure 4. Proof strength
1000 900
800 700 600 500 400 300
Rm (MPa)
d200 100 0
Figure 5 Ultimate tensile strength
Hardness
As expected, the hardness result mirrors the tensile strength, the MMA weld metal being harder than the two SAW weld metals (Figure 6). All three weld metals soften significantly with heat treatment.
400
350
| MMA, HI=0.9kJ/mm | |
| SAW, HI=1.3kJ/mm | |
| SAW, HI=2.3kJ/mm |
d) SAW 2.3 kJ/mm, AW e) SAW 2.3 kJ/mm, tempered
Figure 3. LOM-micrographs of the microstructures in the as-
welded condition (left column) and after heat treatment
routine C (right column). The microstructures in the tempered
condition appear coarser than the microstructures in the as
Hardness (HV10)
300
250
200
150
100
welded condition as a consequence of carbide rearrangement. 50 0
| AW | 1h, 690C | 33h, 630C+ | 33h, 630C+ |
| 10h, 690C | 15h, 690C | ||
| Figure 6. Average hardness. | |||
Doc. II-A-168-05
Charpy-V impact toughness Table 3. Average Charpy-V impact toughness test results.
The average results of the Charpy-V toughness at each combination of test temperature and PWHT are presented in table 3 and as 47J-ITT in Figure 7.
In the as welded condition, the upper shelf values for the MMA weld metal and the SAW 1.3 weld metal are similar. The SAW 2.3 weld metal has the highest as welded upper shelf values. All weld metals have significantly higher upper shelf values after heat treatment, but the increase is smaller for the MMA weld metal.
At first, the 47J-ITT improves with heat treatment for all three samples. After the longest PWHT however, the 47J-ITT of the SAW samples have increased while the 47J-ITT of the MMA sample is unchanged compared to after PWHT routine B.
20
10
0
-10
-20
-30
-40
-50
-60
-70
Post weld heat treatment
Figure 7. 47J impact transition temperatures of as welded and tempered weld metals.
Discussion All three weld metals are largely bainitic. In the as welded condition more or less continuos films of retained austenite and/or carbides separate the bainitic ferrite platelets in the MMA weld and in the lower heat input SAW sample. The retained austenite/carbides appear as more blocky formations in the higher heat input SAW weld and the individual ferrite platelets are not as clearly outlined.
Carbide precipitation is frequently observed inside the platelets of the MMA weld but not in any of the two SAW samples. The presence of intra-platelet carbides suggests that lower bainite to some extent has formed in the MMA-sample. This may be a combined effect of the chemical composition and the cooling rate. The higher Mn-content in the MMA weld metal lowers the Bs-temperature approximately 20°C compared to the SAW weld metals according to the commonly used equation [4].
Bs =830-270wt%C-90wt%Mn-37wt%Ni-70wt%Cr-83wt%Mo (°C)
47 J transition temperature (C)
| PWHT routine | Test temperature | Average test results (J) | ||
|---|---|---|---|---|
| (°C) | MMA | SAW | SAW | |
| 0.9 kJ/mm | 1.3 kJ/mm | 2.3 kJ/mm | ||
| As | RT | 56 | 67 | 81 |
| welded | 0 | 38 | 49 | n.t |
| -20 | 36 | 23 | 51 | |
| -30 | n.t | n.t | 45 | |
| -40 | 21 | 16 | 27 | |
| -60 | 14 | 7 | 21 | |
| A: 1h 690°C | RT | 132 | 217 | 222 |
| -20 | 74 | 194 | 160 | |
| -30 | 81 | 90 | n.t | |
| -40 | 44 | 22 | 101 | |
| -50 | n.t | n.t | 44 | |
| -60 | 36 | 9 | 26 | |
| B: 33h 630°C + 10h 690°C | RT | 160 | 238 | 241 |
| -20 | 109 | 165 | 188 | |
| -40 | 54 | 95 | 135 | |
| -50 | 52 | 43 | 86 | |
| -60 | 23 | 15 | 32 | |
| C: | RT | 150 | 204 | 237 |
| 33h | -20 | 105 | 113 | 186 |
| 630°C + | -30 | n.t | 99 | 156 |
| 15h | -40 | 57 | 49 | 23 |
| 690°C | -60 | 37 | 14 | 17 |
At the same time, the higher cooling rate of the MMA weld gives a higher degree of undercooling before the transformation actually takes place. The actual difference in bainite transformation temperature may therefore be even larger. A lower bainite transformation temperature will promote thinner and longer bainitic ferrite platelets, an increased number of platelets within a sheaf and higher dislocation density in the bainitic ferrite [5]. Further, a lower bainite transformation temperature increases the likelihood for lower bainite formation. These features probably explain the slightly higher tensile strength of the MMA weld metal.
The occurrence of high temperature embrittlement in 2.25Cr1Mo steels is well known to be an effect of carbide growth [1]. When the carbides have reached a critical size, they will act as nucleation sites for cleavage fracture. It is therefore the largest carbides that control toughness [5]. The present investigation also shows that the time period allowed at a defined temperature before toughness drops may vary. As illustrated in Figure 7 for the 47 J impact transition temperature, the MMA weld metal withstands long duration PWHT better than the SAW weld metals.
During the short post weld heat treatment routine A (1h at 690°C), the carbon-enriched residual austenite decomposes into ferrite and cementite (M3C). Aligned carbides replace the continuous austenite/carbide films at the platelet boundaries and fine carbides, probably M2C, precipitate inside the matrix.
The situation becomes more complex for the longer two-step heat treatment routines B and C. The cementite, not being an equilibrium phase, will start to dissolve [1, 5, 6]. Carbon and alloy elements from decomposing austenite and dissolving cementite diffuse to preferential nucleation sites, such as the sheaf boundaries, forming thermodynamically more stable carbides.
As already discussed there are indications that the MMA weld metal has transformed to bainite at a lower temperature than the SAW weld metals. This means that smaller grains and a larger grain boundary area will be present in the MMA weld metal. A larger number and smaller cementite particles will therefore form in the MMA weld metal when the retained carbon-enriched austenite decomposes. Due to formation of lower bainite, cementite will also to some extent precipitate inside the bainitic ferrite platelets in the MMA weld metal and not only at the platelet boundaries as in the SAW weld metals. During tempering, M3C is enriched in chromium [5, 6, 7]. The enrichment will proceed until M3C is saturated with chromium or until alloy carbides are precipitated and starts to attract chromium. With further tempering, nucleation of the alloy carbide M7C3 occurs at the sheaf boundaries and in the vicinity of M3C [5, 6]. During growth, M7C3 depletes M3C in chromium [7]. At the carbon content of the weld metals in this study, only M7C3 is stable and M3C will eventually disappear.
The growth rate of M7C3 is depending on diffusion of alloy elements and carbon. A larger grain boundary area will, when the material is tempered result in formation of smaller and more numerous grain boundary carbides. The duration of growth to a critical size will thereby be extended compared to a material with larger and less numerous carbides. Further, in the case of lower bainite, chromium must be transported from intra-platelet cementite (M3C) to the sheaf or grain boundaries before contributing to the growth. This requires time and will further extend the time needed for carbides to grow to a critical size. The growth rate will therefore be higher in the SAW weld metals where the M3C to a higher extent are located at the plate/sheaf boundaries.
Following the discussion above three possible interacting explanations to the observed variation in PWHT response may be listed:
- Differences in the as welded grain size and the bainite morphology (shown by LOM and FEG-SEM).
- Differences in the as welded carbide size distribution.
- Differences in alloy element distribution at the time M7C3 nucleates.
Doc. II-A-168-05
The latter two contributing factors are likely effects of differences in grain size and bainite morphology. Detailed TEM-studies of carbide type, size distribution and composition have been initiated and results will be presented elsewhere.
Conclusion
The influence of long duration post weld heat treatment on 2.25Cr1Mo type weld metals has been studied. Weld metals produced by MMA and SAW at various heat-input levels have been tested with respect to tensile strength, impact toughness, hardness and microstructure in both as welded and heat treated conditions.
As expected, the tensile strength and the hardness of the weld metals decrease with increased tempering time. However, analysing the 47 J impact transition temperature, the MMA weld metal is shown to withstand long duration PWHT better than the SAW weld metals. This is considered to be a consequence of differences in the as welded grain size, bainite morphology and carbide size distribution governed by cooling rate and weld metal chemical composition. This differences all affect carbide growth during PWHT and thereby the time allowed before toughness deterioration.
References
- K. Tamaki et.al.,Temper embrittlement in HAZ of Cr-Mo steels, Welding in the world, vol 43, no 2, 36-48 (1999)
- AWS SFA 5.5, Specifications for welding rods, electrodes and filler metals.
- AWS SFA 5.23, Specifications for low-alloy steel electrodes and fluxes for submerged arc welding.
- R.W.K. Honeycombe, H.K.D.H. Bhadeshia, Steels. Microstructure and properties, 2:nd edition. p 133. Edvard Arnold, London (1995)
- H.K.D.H. Bhadeshia. Bainite in steels. IOM Communications Ltd. 2001
- C.D Lundin, K.K Khan. Effect of heat treatment on the elevated temperature properties of a 2 ¼ Cr-1Mo submerged arc weldment. WRC Bulletin 405, 1995.
- R.C. Thompson, H.K.D.H. Bhadhesia. Changes in chemical composition of carbides in 2.25Cr-1Mo power plant steel. Part 1: Bainitic microstructure. Material science and Technology. Vol. 10, March (193-203) 1994
Doc. II-A-168-05