Welding martensitic stainless steels

TWI Bulletin, December 1977

by T G Gooch, BSc, MSc(Eng), PhD, MIM, MWeldI, DIC

For many engineering applications, steels containing about 12%Cr show an attractive combination of properties. The alloys are hardenable, and can be heat treated to a range of strength levels, while offering good corrosion and oxidation resistance. A number of these 'martensitic stainless' steels have been developed, and the present article is intended to summarise the welding characteristics of the grades most commonly encountered in practice.

Definition of martensitic stainless steel

Before considering the welding behaviour of martensitic stainless steels, it is as well to define the essential features of this type of alloy. Dealing with the term 'stainless' first, a stainless steel is essentially an alloy based on iron, but containing sufficient chromium to give it the property of 'passivity'. The phenomenon of passivity is associated most commonly with aluminium and its alloys. Aluminium is a reactive material but on exposure to environments containing some oxygen, a thin, tenacious, rapidy self-healing film of aluminium oxide is formed on the surface and this 'passive' film effectively protects the underlaying metal from further attack. Stainless steels are rendered corrosion resistant by a similar mechanism (1). At about the 10-12%Cr level, the steel contains sufficient chromium to form a passive film, probably of chromium oxide, on the metal surface. The passive film can be formed in a wide range of environmental conditions and, in the passive state, the corrosion rate of the material is greatly reduced. Hence a steel containing around 10-12% chromium or above is 'stainless', and for most duties, can be expected to corrode less rapidly than steels with lower chromium contents. At the same time, it should be noted that chromium is a reactive element and if for any reason the passive film is absent, stainless steels can corrode faster than plain mild steel.

Because of the corrosion resistance offered, alloys with about 12%Cr will be economically attractive for a variety of engineering and other applications. It therefore becomes necessary to consider the effect that a chromium content of 10-12% has on the metallurgical characteristics of a steel. A plain mild steel containing perhaps 0.2% carbon will norally show a 'ferritic' microstructure, with the carbon present as carbides in one form or another (e.g. pearlite or bainite). On heating such a material, it transforms to another phase, austenite, at above the Ac ₁ temperature and, if cooled slowly from the austenite range, will transform back to a ferrite/carbide structure. Although the final transformation product may not be identical with that initially existing, it will be soft in general mechanical terms and have comparable mechanical properties. If the same steel is cooled sufficiently rapidly from the austenite range, it will transform to martensite, i.e. the steel is hardened. Martensite is essentially a hard, brittle phase, so that in the fully hardened condition, the material has high strength, but quite low ductility and toughness. Then heat treatment at temperatures below the Ac ₁ can be used to temper and soften the steel, the carbon once more forming carbides and, by altering the time and temperature of heat treatment, a range of mechanical properties can be achieved.

The addition of 12%Cr to the same 0.2%C steel to improve the corrosion resistance has one overriding effect on the transformation behaviour of the steel. With 12%Cr, extremely slow cooling from the austenite range is required to get transformation to a ferrite/carbide structure. For practical purposes, the steel is fully hardenable and even cooling which is regarded as slow in most production conditions, will still give a fully hardened, martensitic material ^[1-3] ( Fig.1). Hence, steels containing about 12%Cr, a little carbon and only minor amounts of other elements are not only stainless, they are also martensitic, and this is the type of material considered below.

General welding behaviour

Overall, the physical characteristics of martensitic stainless steels are not too dissimilar to those of plain mild and low alloy steels, so that conventional steel welding techniques are applicable, apart, perhaps, from preferred arc lengths tending to be shorter and currents lower with the former type of material. ^[2] Expansion behaviour is also similar and jigs etc. to control distortion with plain steel will be directly usable for martensitic stainless steels.

From the point of view of obtaining sound, crack free welds in 12%Cr alloys, attention must be paid to two problems in particular, namely hydrogen induced cold cracking and a form of solidification cracking. The necessity of avoiding hydrogen cracking dominates the welding of martensitic stainless steels and the following is primarily concerned with this problem as it affects the most common grades of material. Solidification cracking is less widely recognised and, at the present time, factors determining its occurrence have not been well defined. The problem is described below, the observations made being based on the author's experience rather than on any comprehensive study.

Hydrogen cracking

It is widely recognised that hydrogen embrittles steels, and that hydrogen picked up during a welding operation from sources such as moisture in the consumable or on the plate, hydrogenous matter in an electrode coating, etc. can act to cause cracking as a weld cools to room temperature and, indeed, for some time after cooling is completed. The general probelm of hydrogen cracking is well considered in a Welding Institute publication; ^[4] for the present purpose it is sufficient to note that the rtisk of cracking depends on the joint restraint, on the amount of hydrogen present, and on the susceptibility of the weld regions which are heated to a sufficiently high temperature to undergo metallurgical transformation.

Joint restraint can be fully quantified only rarely, but in general, fit-up should be as good as possible, while heavy sections are more likely to crack than thin components. The hydrogen 'potential' of a welding operation can be more readily assessed and there is ample evidence ^[4] that the likelihood of cracking can be reduced greatly by measures such as drying manual metal arc electrodes at high temperature to remove as much hydrogenous material as possible, or by ensuring that wire electrodes are clean and dry.

Overall, hard transformed microstructures are more sensitive to hydrogen embrittlement than soft ones and with carbon-manganese and lean alloy structural steels, cracking often can be avoided by controlling the thermal cycle experienced during welding so that soft, non-susceptible microstructures are formed. ^[4] If a preheat is applied, or if a high arc energy is used, the cooling rate from the austenite range after welding can be reduced so that transformation to a non-artensitic product occurs, with associated reduction in cracking susceptibility. This control of transformation product is extremely effective in avoiding hydrogen cracking in such steels. Unfortunately, the same measure is not applicable to martensitic stainless steels, since for practical purposes no matter how slow the cooling rate is after welding, these alloys will give a fully hardened structure that is sensitive to hydrogen induced cracking. Hence, the high hardenability of martensitic stainless steels assumes major significance in the welding of these materials: they simply cannot be welded without forming a microstructure which is highly susceptible to hydrogen cracking ( Fig.2).

With plain carbon and carbon-manganese structural steels, the critical relationship between material composition, joint geometry, process hydrogen potential, preheat and heat input is well understood ^[4] and sufficient information exists for it to be possible in most circumstamces to predict welding procedures giving freedom from cracking. For martensitic stainless steels, however, the situation has been less well explored and virtually all cases some judgement is necessary unless, of course, directly relevant experience exists. In other words, if doubt exists in a particular situation, a realistic procedural test should be carried out to establish a safe welding procedure.

Plain 12% Cr alloys

Welding procedure

Consideration is given first to alloys containing about 10-14%Cr and 0.1-0.15%C. Such materials are exemplified by BS 1449: Part 4: 1970 Grade 403 S12, and the American AISI 410 specifications, and represent the major tonnage of martensitic stainless steels made. A number of points must be appreciated regarding welding procedures for such alloys and are equally relevant to other grades of steel considered below.

Even though the transformation behaviour of martensitic stainless steels is not affected significantly by cooling rate, preheat and the control of interpass temperature are the most commonly employed weapons against hydrogen induced cracking. ^[2,5-7] It must be borne in mind, however, that the aim of preheat is not to change the transformation occurring in the weld area but to hold the joint at a sufficiently high temperature for the embrittling effect of hydrogen to be negligible, and for hydrogen diffusion out of the material to take place. A schematic illustration of the thermal cycle thus experienced during welding is shown in Fig.3. With straight 12%Cr steels, a preheat of 250°C is generally advisable (T ₁ ); higher preheat levels of up to 350-400°C have been employed, ^[5-8] but these are not normally necessary as hydrogen does not cause appreciable embrittlement above about 250°C. ^[4] This is shown in Part I of Fig.3. In Part II, welding is carried out with the preheat temperature being maintained as an interpass temperature.

Following welding in Part II, ( Fig.3), one of two courses of action is open. First, as in IIIa, the joint can be held at temperature (T ₂ ) for a period long enough for hydrogen diffusion out of the weld to take place. ^[2] Hydrogen diffusion rates in different materials are reasonably well quantified, and charts exist so that estimates can be made of the holding time necessary for hydrogen to diffuse out of welds of varying geometry over a wide temperature range. ^[4] Following removal of hydrogen by diffusion, the weldment can be cooled to room temperature, and in most cases a post-weld heat treatment will be necessary to temper and toughen the hardened weld area (T ₃ ): this will be typically at 720°C, as shown in Part IV. ^[2,3] ) Alternatively, if it is inconvenient to maintain the weld at temperature for a time after welding, the assembly can be cooled to T ₂ and then given an immediate post-weld heat treatment without prior cooling to room temperature. ^[6] The heat treatment will both temper the material, and allow rapid hydrogen diffusion. It will be noted that in neither route IIIa or IIIb is the weld area permitted to cool to room temperature while appreciable amounts of hydrogen are present.

Whether route IIIa or IIIb is selected, it is imperative that control of joint temperature during welding be maintained. The point is that, on cooling, the austenite phase will be retained down to a temperature M _s when it will start transofrming to martensite; however, transformation to martensite will not be completed until further cooling to a temperature M _f has occurred, and the weld must be cooled to close to or below the M _f temperature in stage III. ^[2,7] This is for two reasons. First, hydrogen diffusion from austenite is extremely slow, so that if temperature T ₂ is well above M _f , much of the weld area will remain austenitic: hydrogen will not diffuse out, but will remain in the weld area and can cause cracking when complete transformation to martensite occurs on final cooling in route IIIa. Second, if route IIIb is adopted, but the weld does not cool to the M _f , the final heat treatment will be on material containing an appreciable amount of austenite. Given sufficient time at elevated temperature, the austenite will transform to a ferrite/carbide structure, ^[3] and this means that one of two situations will arise.

With thin materials, where the heat treatment is less than about 1hr, the austenite may not transform completely during heat treatment, and hydrogen may not be evolved. In this case, virgin martensite will form on final cooling with a significant hydrogen content when cracking may well result. Alternatively, and more usually, the austenite will transform during the heat treatment, but unfortunately the resultant isothermal transformation product will have low toughness and, especially in heavy sections, the joint mechanical properties may be seriously impaired. ^[2]

It will be evident from the foregoing that although straight 12%Cr steels are sensitive to hydrogen cracking, the problem can readily be overcome by suitable preheat and post-heat procedures. The major problem is that the welding procedure hinges on some knowledge of the M _s and M _f temperatures and these temperatures will depend on the alloy composition. The latter temperature is perhaps the more important, but is difficult to measure experimentally so that little information exists for various alloys. A number of workers have derived empirical formulae for defining the M _s temperature in terms of the material compositions. ^[3,9-11] but these have been based on lower alloy steels than the 12%Cr grades and are therefore rather inaccurate as far as martensitic stainless steels are concerned. From published data on 12%Cr steels, together with other work, the following relationship has been obtained for martensitic stainless steels, ^[12] and is used by the author:

M _s = 540 - [497%C + 6.3%Mn + 36.3%Ni + 10.8%Cr + 46.6%Mo]°C

For practical purposes, the M _f temperature can be taken as 100°C below the M _s . In the author's experience even if this is above the real M _f , the amount of austenite retained is too low to have any practical effect. Referring now to Fig.3, temperature T ₂ can be taken as M _s - 100°C, the M _s being calculated as above for the particular analysis material being welded.

One further point on welding procedure should be made. The time allowed for hydrogen diffusion out of a completed weld will depend on the material thickness and, if heavy section mateial is involved, the time required may be prohibitively long. ^[4] If examination of hydrogen diffusion data indicates this to be the case, route IIIb in Fig.3 may be attractive. Alternatively, it may be possible to hold the joint at intermediate stages during welding to permit hydrogen to diffuse out: the distance hydrogen has to diffuse in a partially completed weld is less than in a completed joint, and the holding time necessary for hydrogen removal can be reduced greatly. Finally, in some situations, it may be of benefit to complete welding, cool to T ₂ to obtain full transformation to martensite, and then heat to a slightly higher temperature to speed up the hydrogen diffusion rate. The temperature range 450-500°C should, however, be avoided, as this can cause loss of toughness and corrosion resistance. ^[13-15]

Effect of composition

With the compositional range embraced by straight 12%Cr steels, it should be recognised that differences in material susceptibility can exist between different specifications or even different casts. At present, it is not possible to be definitive regarding the effect of composition but, for practical purposes, it appears that the major factor is the carbon content. ^[2,5] The higher the carbon content, the more susceptible is the martensite to hydrogen embrittlement. If the carbon level is below 0.1%, it is normally feasible to weld 12%Cr steels without preheat or post-weld heat treatment, at least in material thicknesses up to about 6mm, ^[16] provided that care is taken to use well dried consumables and avoid hydrogen pick-up as far as possible. At the other extreme, if the carbon level is around 0.2% or above, preheat and post-weld heat treatment are essential. Needless to say, most commercial plain 12%Cr alloys fall between 0.1% and 0.2% carbon so that, unless suitable prior experience is available, the welding procedure should be planned as above.

Choice of weld metal

If full joint strength is required from a weld in a straight 12%Cr steel, it will be necessary to employ matching composition consumables and the above welding procedure is appropriate. When using matching composition consumables, a significant difference must be noted between martensitic stainless steels and more commonly encountered structural steels. In the latter case, weld metals employed normally have a lower carbon content than the parent material and they therefore transform during welding at a higher temperature than the parent material. As commented above, hydrogen is much more soluble in austenite than in ferrite, so as a structural steel weldment cools and the weld metal transforms, it is possible for hydrogen to move from the weld metal into the heat affected zone (HAZ). Thus, when a weld in a structural steel cools to room temperature, the risk of cracking is highest in the HAZ, because the higher carbon content tends to give amore susceptible microstructure than exists in the weld metal, and possibly also because the hydrogen level is higher. With martensitic stainless steels and matching composition fillers, however, this will not be the case; the weld metal and HAZ will transform over a similar range of temperature, and hydrogen rejection into the HAZ will not occur, resulting in a higher proportion of the total hydrogen remaining in the weld metal than with structural steels. ^[2]

Hence, if matching fillers are employed for a 12%Cr steel fabrication, a particular risk of weld metal cracking must be recognised ( Fig.4) and this being so, it is imperative that consumables used for 12%Cr steels should give the lowest possible hydrogen contents. ^[17] With manual metal arc electrodes, for example, high baking temperatures (350°C and above) are definitely to be preferred.

If full mechanical strength is not required from the weld area, austenitic stainless steel consumables containing typically 23%Cr and 12%Ni can be used and, in fact, such weld metals probably represent more common practical use thanthe matching fillers. Austenitic fillers offer two advantages when welding martensitic stainless steels. ^[2,4,17] First, the proof stress of the weld metal will be low in relation to that of the parent material, so that residual welding stresses (and hence the risk of cracking) are reduced. Second, and of greater importance, the higher solubility of hydrogen in austenite means that hydrogen remains in the weld metal (and austenite has little susceptibility to hydrogen embrittlement) rather than diffuse into the HAZ and cause cracking.

Given austenitic fillers, the risk of cracking is reduced so that preheat and interpass temperature requirements can be relaxed over the procedures illustrated in Fig.3. How far they can be relaxed cannot be predicted with precision but, typically, a preheat of 150°C should be adequate to avoid cracking. In fact, common industrial practice is to maintain a welding procedure as in Fig.3, with 250°C preheat and use austenitic consumables as a 'belt and braces' approach.

One aspect of austenitic fillers which should be borne in mind is the fact that in the as-deposited condition they will contain perhaps 10% of the high temperature δ-ferrite phase to prevent solidification cracking. During the final post-weld heat treatment, this ferrite can transform into sigma phase, causing embrittlement in the weld metal. ^[1,2] With most 12%Cr fabrications, final heat treatment times are not sufficiently long for sigma formation to be of practical significance, but it should be considered if for any reason heat treatment times of, say, ten hours are envisaged. In this case, the author prefers not to employ the common 23Cr/12Ni type, but rather the lower alloy 18Cr/10Ni or perhaps 18Cr/10Ni/3Mo consumables. These do not have such high tolerance for dilution as do 23Cr/12Ni fillers,and may give some martensite in the weld metal, but the risk of cracking is low because the bulk of the weld metal remains austenitic. In the author's opinion, the risk of embrittling sigma formation is lower with such a weld metal than with the more common 23Cr/12Ni types.

Another approach is for the martensitic steel to be buttered with 23Cr/12Ni consumables and then heat treated. Dilution of the buttering runs will reduce the weld metal alloy content, so that sigma formation should not be a hazard,and the weld can be completed using austenitic filler with no further heat treatment necessary. The drawback to this approach is that if a defect is found on final NDT, it may be necessary to cut out so much weld metal that the buttering is removed: the repair will give an as-welded HAZ in the martensitic material, and assessment will be necessary of the relative advisability of leaving such an HAZ or applying heat treatment, depending on the service requirements.

Low carbon martensitic alloys

Because of the high risk of hydrogen induced cracking in welding conventional 12%Cr steels, alloys have been developed having lower carbon contents. If the carbon level of a 12%Cr steel is reduced with no other composition almodifications, the material becomes increasingly ferritic, so that the range of properties obtainable through choice of tempering temperature with a martensitic alloy are lost. To overcome this, materials have been developed with about13%Cr, less than 0.1%C, but containing about 4%Ni to obtain a fully hardened (in the sense of a fully martensitic) structure. The most common one also contains about 0.5%Mo for improved corrosion resistance, particularly in marine applications. ^[18-21] .

There is no doubt that the risk of hydrogen cracking is lower with 13/4 alloys and similar materials than with the straight 12%Cr grades. Most welding ^[20] is carried out with a preheat and interpass of 150°C and it is normally found sufficient on completion of welding merely to cover the item with asbestos blankets and allow it to cool, rather than apply a definite holding stage for hydrogen removal by diffusion. Higher preheat levels have been used, but it must be appreciated that the presence of nickel greatly depresses the M _s and M _f temperatures and under cooling rates appropriate to welding, the M _s is typically around 220°C. Obviously, it will not be satisfactory to hold a completed weld at, say, 250°C prior to post-weld heat treatment as considered in Fig.3 above. Nickel also lowers the Ac ₁ temperature, so that final heat treatment for tempering will be carried out at a temperature rather lower than for the plain 12%Cr alloys, typically in the range 600-640°C.

The 13/4 alloys are welded most commonly with matching composition fillers and, as noted above, a risk of weld metal cracking rather than HAZ cracking predominates. In fact, weld metal cracking appears the major problem with these steels. For most purposes, electrode drying at about 220°C minimum appears adequate, given preheat and reasonably controlled cooling as above. However, there is little doubt that higher drying temperatures are of benefit and,provided that the electrode coating can withstand it, baking at 450°C is advisable. The reduction in cracking risk thus obtainable is illustrated by the fact that if high temperature drying is employed it is often feasible to weld even restrained assemblies in 13/4 material with no preheat or control of cooling after welding whatsoever. Needless to say, before such a practice is adopted, realistic procedural testing is vital.

As with plain 12%Cr alloys, 23/12 fillers are applicable and the risk of sigma formation is reduced by the fact that heat treatment of 13/4 alloys is normally at around 620°C, rather than 720°C. At the same time, some consideration of filler composition may be necessary if heat treatment times are protracted.

With martensitic stainless steels having low M _s and M _f temperatures, some debate exists as to whether or not the weld area should be cooled to around the M _f after each run. ^[19-20] There is no doubt that if this is done, welding time and associated costs can be increased greatly. It is probable, nonetheless, that this measure is preferable because if interpass temperatures are alloed to rise too far, a coarse microstructure can be formed, ^[22] especially in matching composition weld metal. On final transformation, this tends to cause some loss in toughness and also higher susceptibility to hydrogen cracking. ^[23] However, the question of optimum interpass temperature for any given steel composition has not been resolved unambiguously and interpass temperature is definitely a variable which should be given full heed in procedural testing.

High carbon creep resisting alloys

In contrast with 13/4 materials, creep resisting martensitic stainless steels commonly contain significant amounts of carbon, perhaps 0.2%, as well as other elements added to form carbides and give good creep resistance. In consequence, the materials are highly susceptible to hydrogen embrittlement and, as a class, represent some of the most difficult materials to weld from the hydrogen cracking viewpoint. ^[24,25] .

Because of their composition, the M _s and M _f temperatures are commonly quite low and this places the welder in the invidious position of having to cool a joint after welding to ensure complete transformation to martensite, even though this involves coolingto temperatures where hydrogen embrittlement becomes a significant hazard. Thus, referring to Fig.3, the alloys commonly are welded with a preheat of 250°C, but, taking the 12%Cr/0.2%C/Mo/W/V alloy as an example, they are permitted to cool to about 180°C after welding to ensure complete martensite formation.If route IIIa is being selected, careful holding for hydrogen removal is essential and it is certainly unwise to cool the weld area to below 150°C if appreciable hydrogen remains after welding. With heavy section material, say25mm, this may involve extremely lengthy holding times, unless measures considered above to reduce hydrogen diffusion times are applicable.

As with other martensitic stainless steel grades, austenitic fillers can be employed and will reduce significantly the risk of cracking, although these do not, of course, preclude the necessity for careful post-weld cooling to about180°C. However, such consumables may not be appropriate for high temperature service on three counts. First, the coefficient of thermal expansion of the weld metal will differ from that of the martensitic parent material, so that if service under fluctuating temperatures is encountered, a risk of thermal fatigue exists, with failure taking place down the fusion boundary where thermally generated strains are highest. Second, carbon migration may take place during high temperature operation from the HAZ into the weld metal because of the higher chromium content in the weld metal. This leaves a carbon depleted zone adjacent to the fusion boundary with concomitant loss of mechanical strength in this area. Finally, at low temperatures in the creep range, the weld metal simply may not be strong enough to match the properties of the parent material. Needless to say, if matching fillers are specified, stringent attention to all sources of hydrogen must be paid.

Precipitation hardening grades

Because of the wide range of material compositions that have been developed to give precipitation hardening, only general comments can be made. For an appraisal of the welding metallurgy of representative alloys, the reader is referred to a WRC Bulletin. ^[26] However, the alloys share the starting point of reducing the carbon content followed by the use of a precipitation hardening treatment to obtain high strength in the material. ^[2,26,27] Because the carbon level is so reduced, normally below 0.1%, the risk of hydrogen cracking is diminished and, with alloys such as 17/4 PH ^[26-29] and FV 520B, ^[27-28] welding can be carried out without preheat, provided attention is paid to consumable hydrogen levels. At the same time, a risk of hydrogen cracking does exist, and, again, procedural tests may be necessary to establish thatwelding procedures without preheat are, in fact, safe.

One rider is that precipitation hardened grades are normally welded eiter in the solution treated ^[2] or over aged ^[29] conditions. This is because in these heat treatment conditions, the material strength is minimal, and the danger of cracking is reduced compared with material which has a stiff matrix, with a high yield strength.

It will be appreciated that the metallurgical strengthening reactions in precipitation hardening martensitic alloys take place at temperatures below the Ac ₁ . This being so, precipitation will take place, during welding, in the HAZ some distance away from the weld, in material that is heated to below the Ac ₁ temperature. ^[2,27] The extent of precipitation at any point will depend on the precise themal cycle experienced and the original material condition, and either strengthening or over ageing and softening may ensue. Hence, if uniform properties are required across a weld in precipitation hardening material, a full heat treatment after welding is obligatory. In some situations, only a simple precipitation treatment after welding will be practicable. Because of their high alloy content, precipitation hardening steels generally have low M _f temperatures. Welds must therefore be cooled to around room temperature prior to precipitation heat treatment, and preferably between each run, to ensure complete transformation. If this is not done, the precipitation treatment will be on a largely austenitic material, and the full strengthening effect will be lost.

Weld metal solidification cracking

For the vast majority of fabrications, hydrogen cracking is the dominant probelm with martensitic stainless steels but, from members' queries to The Welding Institute, it has become apparent that with a number of alloys solidification cracking can occur ( Fig.5). While this can be expected to depend on material composition, the presence of residual elements etc., it appears ^[30] that cracking is most pronounced when the weld metal composition is such that substantial amounts of the high temperature δ-ferrite phase above, say, 25%, are retained down to room temperature. This, it will be noted, is in sharp contrast with the case of common austenitic stainless steels, where ferrite is regarded as almost universally beneficial in suppressing solidification cracking. ^[2] .

The cause of the problem is not clear, but it has been suggested that cracking occurs after solidification, perhaps as a result of low ductility of the two phase structure over some temperature range between the solidus and normal ambient. In this respect, similar cracking has been found ^[31] in the HAZs of welds in martensitic stainless steels of composition such that appreciable amounts of ferrite are formed during the welding cycle. Comparable cracking has been observed ^[32] also in duplex austenite/ferrite stainless steel weld metals of high ferrite content.

For practical purposes, it appears desirable that the ferrite level when welding martensitic stainless steels should be restricted to about 15% maximum in situation where solidification cracking can occur, as in highly restrained assemblies or under high heat input/high travel speed conditions. It may be noted at the same time that such a restriction on ferrite level is generally desirable from the point of view of optimising mechanical properties, since ferrite does tend to reduce toughness of the weld metal. ^[2,26,29] .

Summary

The origin of martensitic stainless steels and the salient metallurgical and welding characteristics have been described. The major cracking problem which is likely to arise in practice is hydrogen induced cracking in both the heat affected zone and weld metal, although weld metal solidification cracking can also occur. Factors determining the occurrence of each have been considered. Hydrogen cracking can be avoided by control of preheat, interpass, and post-heat temperatures, with consumable hydrogen potential being minimised. Solidification cracking appears primarily associated with high weld metal ferrite levels. It has been emphasised that if directly relevant prior practical experience is not available for a given application, realistic procedure testing is essential. Typical welding conditions for commonly encountered grades of steel that can be used as a starting point for procedural testing are given in the Table.

Suggested welding conditions for common grades of steel

References

1. Colombier, L. and Hochmann, J. 'Stainless and heat resisting steels'. Edwards Arnold (Publishers) Ltd, London, 1967.
2. Castro, R. and de Cadenet, J.J. 'Welding metallurgy of stainless and heat resisting steels'. Cambridge University Press, Cambridge, 1974.
3. Irvine, K.J., Crowe, D.J. and Pickering, F.B. JISI, 1960, 195, August, 386-405.
4. Coe, F.R. 'Welding steels without hydrogen cracking'. The Welding Institute, Abington, 1974.
5. Linnert, G.E. Metals Eng, Quarterly, 1967, November, 1.
6. 'Code de bonne pratique pour le soudage des aciers inoxydables'. Inst. Belge de la Soudure, Brussels, 1st ed. 1970.
7. Kreischer, C.H. et al. Weld. Res. Suppl., 1961, XXVI, II, November, 489s-496s.
8. Brouwer, G. Philips Welding News, 1963, 14, 2, 14-26.
9. Payson, P. and Savage, C.H. Trans. Amer. Soc. Met., 1944, 33, 261-275.
10. Steven, W. and Haynes, A.G. JISI, 1956, 183, August, 349-359.
11. Grange, R.A. and Stewart, H.M. Trans. Amer. Inst. Min. Met. Eng., 1946, 167, 467-490.
12. Gooch, T.G. and Clarkson, P. unpublished work.
13. Woolman, J. and Mottram, R.A. 'Properties of the En steels'. e, BISRA, Pergamon Press, 1969.
14. Gill, G.M. and Sweeney, B. J. Brit. Steel Cast. Res. Ass., 1964, 82, December.
15. Truman, J.E., Perry, R. and Chapman, G.N. JISI, 1964, 202, 2, 745-756.
16. Muller, R. Schweisstechnik (Berlin), 1975, 25, 8, 359-361.
17. Fedorov, V.G. et al. Weld. Prod., 1974, 21, 7, 35-36.
18. Küpfer, R. Sulzer Tech. Rev., 1966, 4, 173-183.
19. Anon. G.F. Information, Georg Fischer, 1968, 17-26.
20. Gysel, W. ibid. I/1970, 37-53.
21. Clarke, G.R. and Haynes, B. Proc. 20th Annual SCRATA Conf. 1974, paper 8.
22. Koroleva A.G. et al. Auto. Weld., 1970, 23, 7, 6-9.
23. Fedorov, V.G. et al. Weld. Prod., 1976, 23, 12, 2-4.
24. Kauhausen, E. et al. Briot. Weld. J., 1960, 7, 12, 693-707.
25. Eaton, N.F. and Glossop, B.A. Met. Con. and Brit. Weld. J., 1969, 1, 12, 6-10.
26. Harkins, F.G. WRC Bulletin, 1965, No. 103, February.
27. Morley, J.I. and McWilliam, J.A. Brit. Weld. J., 1960, 7, 10, 651-658.
28. Kaae, J. Brit. Weld. J., 1968, 15, 7, 333-344.
29. Linnert, G. Weld. J., 1957, 36, 1, 2-27.
30. Castner, H.R. ibid., 1977, 56, 7, 193s-220s.
31. Vagi, J.J. and Martin D.C. ibid., 1956, 35, 3, 137s-144s.
32. Honeycombe, J. and Gooch, T.G. Welding Institute Members Report, 13/1976/M, May, 1976.