Phase balance in 9%Cr1%Mo steel welds

TWI Bulletin, January/February 1991

Roger joined TWI in January 1988 as a Research Metallurgist. Before this he gained a degree in metallurgy at the University of Leeds, going on to study superplasticity in titanium alloys for his PhD.

His work in the Stainless Steels and Corrosion Section of the Materials Department includes the study of creep resisting steels and general stainless steel failure investigations.

Delta ferrite has a number of detrimental effects on the mechanical properties of 9%Cr1%Mo steel welds. Roger Panton-Kent summarises the work carried out at TWI to identify favourable compositions and welding conditions to avoid high weld metal and HAZ ferrite levels.

There is considerable interest in the development of 9%Cr1%Mo ferritic steels for high temperature service in both conventional and nuclear steam plant applications. In recent years, modified 9%Cr1%Mo steel, ASTM A213 T91, has been increasingly used for heat exchanger tubing. The compositional specification of modified 9%Cr1%Mo steel is compared with that of conventional 9%Cr1%Mo alloy in Table 1, the main differences in the new alloy being the addition of about 0.20%V, 0.08%Nb and 500ppm N.

Table 1: Composition requirements of standard and modified grades of 9%Cr-1 %Mo steel, wt %

*    ASTM A387/A387M G9
**  ASTM A387/A387M G91

High strengths result from the formation of a fine dispersion of Nb (C, N) precipitates during tempering, and enhanced creep properties arise from the establishment of stable M ₂₃C ₆ type precipitates containing V. ^[1] As a result of higher allowable stresses at elevated temperature, material cost savings can be realised. In addition, modified 9%Cr1%Mo steel exhibits good resistance to oxidation. Consequently, this steel has emerged as a candidate material for pressure vessels and steam generators of fast breeder reactors.

Although 9%Cr1%Mo steels normally have fully martensitic microstructures, their chemical compositions are close to the borderline of delta ferrite formation. ^[2] And so, weldments which are subjected to elevated temperatures and rapid cooling frequently contain small amounts of delta ferrite in both the weld metal and heat affected zone (HAZ). It has been established that delta ferrite has a number of effects on the properties of steels. It can cause reduction of creep ductility and toughness of weldments ^[3,4] but may act to restrict HAZ grain growth, and reduce creep crack propagation rate. ^[5] This being so, it is important that the effects of varying welding conditions and material compositions should be understood to allow control of delta ferrite formation.

Effects of composition

The Cr _eq factor was in fact developed to describe the ferrite/martensite balance in 9%Cr1%Mo steels covering a fairly wide compositional range, whereas the Kaltenhauser ferrite factor was originally developed for low carbon 12%Cr steels. Calculated Cr _eq and FF, and observed proportions of delta ferrite in manual metal arc weld pad samples produced at TWI are shown in Fig.1. ^[8,9] It is evident that in both cases ferrite level increases with increasing compositional factor. However, while Cr _eq may be of value over a broad range of material analysis, it is less satisfactory for the more restricted range of weld metal composition that may be produced using commercially available MMA consumables. On the other hand, the FF value gives a relatively good prediction of the delta ferrite content, and it can be seen that weld metal with a FF of approximately 7.5 or less would be expected to have a fully martensitic microstructure.

Effects of weld procedure

Over and above the effects of composition, work at TWI ^[8] has shown that the percentage of delta ferrite in 9%Cr1%Mo weld metal and HAZs is also dependent upon welding procedure. The situation is complex. In the general case, transformation of delta ferrite to austenite during welding is promoted by lower cooling rates. ^[3] Therefore, higher heat input welds, which have lower cooling rates, would be expected to result in decreased delta ferrite contents. Moreover, higher preheat temperatures, which also reduce cooling rates, give rise to further reductions in delta ferrite levels. Conversely, low heat inputs and preheat temperatures should produce high delta ferrite contents. This behaviour was found with single pass weld metal deposited using an MMA consumable with FF of 11.6, as in Table 2, while an extreme example of a high ferrite weld metal microstructure, resulting from rapid cooling of a thick section electron beam weld, is shown in Fig.2.

Table 2: Ferrite contents of bead-on-plate welds

Parent material FF = 9.3
Consumable FF = 11.6

Second, the influence of heat input must be related to the steel composition. With highly ferritic material, i.e. material with a high FF, ferrite is stable over a wide temperature range. Consequently, while ferrite forms during heating, the amount possibly increasing with increasing arc energy, the reverse transformation to austenite takes place at relatively low temperature and hence is retarded. As a result, retained ferrite levels may tend to increase with increasing heat input. On the other hand, low FF material produces less ferrite at peak temperatures, but the reverse transformation to austenite occurs at higher temperatures, and the net result could be a reduction in the final ferrite content.

Clearly, the effects of composition and arc energy depend on the relative rates of the austenite-to-ferrite and ferrite-to-austenite reactions. At present, neither can be predicted. For weld metal, it can generally be expected that final ferrite level will decrease at higher arc energy, but this may not be the case for the surrounding HAZ. Hence, for a given parent steel, appropriate welding trials are necessary to determine the overall effect of change in welding conditions. Further study is necessary to clarify the situation so that material behaviour can be predicted.

Conclusions

• Weld metal and HAZ delta ferrite contents in modified 9%Cr1%Mo steels are dependent on composition and welding conditions.
• The Kaltenhauser ferrite factor provides a good indication of final weld metal ferrite content for a given welding condition.
• In general, increasing arc energy and preheat temperature cause a reduction in weld metal delta ferrite level. However, higher heat input and preheat temperature may give rise to increased delta ferrite in the surrounding high temperature HAZ, and trials are necessary to establish the effect of changing welding procedure for individual parent steels.

In contrast, the delta ferrite content of the high temperature HAZ of bead-on-plate deposits on a steel with FF = 9.3 was found to increase with increasing heat input and increasing preheat temperature ( Table 2). [8] Figure 3 shows a typical duplex martensitic ferritic microstructure in a high temperature HAZ region close to a weld fusion boundary. Similar increases in HAZ delta ferrite levels with arc energy have been reported for 12%Crl%Mo welds. [10].

To account for such a disparity in ferrite forming tendency between weld metal and HAZs, two factors must be considered. First, the ferrite in weld metal is residual from solidification, and within broad limits the ferrite level at the end of solidification is independent of arc energy and solidification rate. Hence increasing arc energy affects only the high temperature decomposition into austenite, and a general trend of lower final ferrite at room temperature with slower cooling would be anticipated. The ferrite in the HAZ, however, forms in the solid state from austenite on heating, and, if this reaction is sluggish, it is in principle possible for the extent of transformation to be reduced by welding at low heat input. Hence, reduction in arc energy decreases the final ferrite content because there is less time to form ferrite during heating to high temperature.

References