Skindeep improvements - laser beam surface modification

TWI Bulletin, May/June 1994

Paul Woollin joined TWI in 1992 after leaving the University of Cambridge, where he obtained BA and PhD degrees in metallurgy and materials science, and subsequently worked as a post-doctoral research associate. He began at TWI as a project leader in the stainless steels and corrosion section and in 1994 was appointed manager of the newly formed stainless and non-ferrous metals section. His university work involved the study of fatigue in nickel based superalloys and the fracture behaviour of local brittle zones. He is currently investigating the corrosion fatigue performance of structural steels, laser treatment of high alloy austenitic materials and the cracking behaviour of duplex stainless steels in sour environments.

The service properties of a material can be greatly improved by surface modification. Paul Woollin outlines the effects of power beam treatment of different alloys, especially by laser.

Developments in laser and electron beam technology have led to the possibility of localised modifications to the microstructures of a range of materials. Such modifications can lead to improved service properties in the surface layers of a component, whilst leaving the bulk properties essentially unchanged. There are a number of mechanisms by which these changes can be brought about, but all depend on the ability to manipulate the laser/electron beam accurately, and on the high power density of the beam.

When a laser/electron beam is absorbed by a material, rapid heating takes place within a small volume and, if the exposure time is short, rapid cooling follows. Cooling rates of around 10 ⁶K/sec are achievable. Therefore, by scanning a beam over the surface of a component, a rapid thermal cycle can be applied to a thin layer of material at the surface of the substrate, without any significant heat treatment of the bulk. By controlling the peak temperature and cooling rate, and by making additions to the liquid pool, a range of mechanical and corrosion properties can be developed in the surface layer.

Modification without melting

When the peak temperature in the surface remains below the metal melting point, power beam processing simply acts as a means of localised heat treatment. The rapid thermal cycle associated with use of a power beam is particularly suited to hardening of iron based systems by austenitising and then cooling quickly, to suppress equilibrium transformation, and promote diffusionless transformation to martensite. Similar transformation can be induced in other alloy systems but is rarely of technological significance.

The response of a steel to surface hardening in this way is dependent upon the carbon content. ^[1,2] At low carbon levels, high temperature transformation to delta ferrite may occur. If ferrite is retained on cooling, the hardness of the steel surface will be lower than if it were fully martensitic. The resultant martensite will also be relatively soft at low carbon levels. With higher carbon levels, the hardenability increases, but, above approximately 0.4%C, the time required for austenitisation and homogenisation of the carbon content increases, because of the need to dissolve an increasing quantity of iron carbide. To increase the time in the austenite field, more heat input is required, which consequently reduces the cooling rate. The result is an increased tendency to retain austenite after cooling or to allow diffusion controlled transformation.

Both of these lead to a reduction of the hardness of the treated layer. Figure 1 gives an example of the hardness profile achievable in medium carbon steel by this process. The very high surface hardness (almost 1200 HV0.1) relative to that of the base steel (around 250 HV0.1) and the limited depth of the hardened layer (about 100µm) are particularly noteworthy. Typical applications of this technique are hardening of cylinder liners and saw teeth. ^[1,2]

Modification involving melting

Surface alloying
Perhaps the most obvious means of surface modification through power beam melting is to make a chemical addition to change the composition of the surface layer. This can be done by alloying with a gas or a solid, which dissolves in the melt, or by addition of particles which do not dissolve fully and remain discrete.

Firstly, the composition of the liquid pool may be altered by use of an appropriate shielding gas during laser surface treatment. In particular, carbon or nitrogen can be added to the liquid pool (or removed) in this way. For example, steels may be carburised, for increased hardness and wear resistance, by use of methane gas, although attempts to achieve a similar effect without melting have proved unsuccessful, as a consequence of the low solid state diffusion rate of carbon.

Nitrogen alloying also has a number of potential uses. Ongoing research at TWI has demonstrated an increase in the pitting corrosion resistance of superaustenitic stainless steel through use of laser surface treatment with N ₂ trailing gas (see Table). Nitrogen addition to titanium alloys produces hardening through formation of TiN phase.

Alloying with a solid material can be achieved either by prior deposition of a film, which is then incorporated into the surface by power beam melting, or by addition of a powder to the liquid pool. Hardening of some steels may be achieved by addition of alloying elements ( e.g. Ni, Cr) to modify the kinetics of the equilibrium phase transformation and hence promote martensite formation. Examples include Ni-P layers electrolytically deposited and Cr layers flame sprayed on to cast irons. When melted into the surface, these layers give improved wear resistance, through the increase in hardness, and improved corrosion resistance, through the change in the surface chemistry. ^[1] Similar work has shown that laser surface alloying of Cr into the surface of carbon steel and Mo into 304 grade austenitic stainless steel can promote large increases in corrosion resistance. ^[3]

The range of alloy composition attained in this work included 5-80%Cr in the carbon steel and 0-10%Mo in the stainless steel. Complete mixing and uniform dispersal of the alloying elements were observed. The depth of the layer produced was typically of the order of a few hundred micrometres. The practical success of corrosion resistant layers of this type would, however, be critically dependent upon attaining complete surface coverage. Any gaps in the layer would tend to behave anodically with respect to the bulk of the surface and would be attacked rapidly in corrosive media.

Another means of increasing surface hardness, which is applicable to a range of materials operating at high temperatures, is incorporation of hard, non-metallic particles, such as WC, TiC and SiC. This is most conveniently done through addition of a powder to the liquid pool. By control of the surface melting parameters it is possible to avoid dissolving the particles. Examples of the practical use of this technique include surfacing of extrusion dies, hot working tools and gas turbine components. ^[1]

Surface alloy redistribution
Cast alloys and weld metals with significant alloy contents can suffer from considerable segregation during solidification. This can lead to inhomogeneous properties and is particularly significant in materials designed for high corrosion resistance, such as stainless steels and nickel based alloys. Segregation leads to formation of alloy lean regions which have considerably lower resistance to attack than the bulk composition would indicate. Figure 2 shows a corrosion tested sample of as-solidified superaustenitic stainless steel meeting UNS S31254. This illustrates preferential corrosion of the dendrite cores, which were alloy lean as a result of the rejection of key alloying elements ( i.e. Cr and Mo) into the interdendritic regions during solidification. Figure 3 quantifies the effect of this segregation on the pitting corrosion resistance of superaustenitic stainless steel remelted using the tungsten inert gas (TIG) process. ^[4] A reduction in critical pitting temperature, from 65°C for the parent material to 30-45°C for the remelted material, was observed. Such segregation in as-cast or as-welded material can be reduced by a homogenisation heat treatment but this is time-consuming and often impractical.

The extent of segregation depends on a number of variables, including the alloy system ( e.g. as described in Ref.19, it generally occurs to a lesser extent in nickel based alloys than stainless steels), the solidification mode and the cooling rate. If solidification occurs under equilibrium conditions, i.e. slowly, the extent of segregation is controlled by the shape of the phase diagram and can be calculated relatively easily using the Scheil equation. ^[5] The composition of solid forming at a given point in the solidification process is given by:

C _s= kC _o(1 - f _s) ^k-1

where
C _s = solid composition
C _o = mean alloy composition
f _s = volume fraction of solid
k = equilibrium partition coefficient.

Figure 4 illustrates the calculated composition profile along a volume element which solidified with solute partitioning as described by the Scheil equation. This model assumes complete diffusion of alloying elements within the liquid phase and no diffusion in the solid. Allowance for only limited liquid diffusion alters the shape of the calculated concentration profile somewhat, but does not affect the difference between the composition of the first and last material to solidify. However, the final extent of segregation after cooling does depend on the extent of solid state diffusion.

For stainless steels solidifying to primary ferrite, diffusion coefficients of solute elements are relatively high and much homogenisation can occur during cooling under typical arc welding conditions. By comparison, nickel base alloys and stainless steels solidifying to primary austenite show negligible homogenisation during cooling. ^[6] However, under non-equilibrium conditions ( i.e. rapid cooling), undercooling at the tip of solidifying dendrites leads to a modified segregation pattern. The extent of segregation generally decreases as the local cooling rate increases. ^[7-11]

A consequence of this is the ability to reduce near-surface segregation in as-cast/welded nickel base alloys and stainless steels, which solidified as primary austenite, through power beam surface melting. ^[10,11] By increasing the travel speed of the beam over the surface of the segregated material, the cooling rate can become sufficiently high to reduce segregation and improve the corrosion resistance of the as-cast/welded material to a level comparable with that of homogeneous parent material ( Fig.5). It has been demonstrated that, at high beam travel speeds, the partitioning coefficients of Cr and Ni in an austenitic stainless steel tend towards unity. ^[12]

Surface microstructural refinement

As the cooling rate of a liquid metal is increased, the scale of the resulting solidification structure tends to decrease. This effect is closely linked to the variation of the extent of segregation, as described previously. Both phenomena are a result of dendrite tip undercooling. ^[8,9] Figure 6 shows an example of fine scale structures formed during rapid solidification of an electron beam treated superaustenitic stainless steel. Refinement of the microstructure in a similar way, using a laser beam, has been shown to enhance the high temperature oxidation resistance of some stainless steels. ^[13] The mechanism by which this improvement was achieved was related to the role of grain boundary diffusion in the formation of the protective oxide film. Improvements in wear resistance of cast irons, as a result of microstructural refinement by power beam surface melting, have also been observed. ^[1]

Modification of the surface phase balance

Remelting and rapid solidification can also be used to bring about changes in the distribution of phases present in a surface layer, by suppressing diffusion controlled transformations. Although a similar effect can sometimes be achieved without the need for melting, the short time spent at high temperature limits the extent of solid phase re-solution of second phases and subsequent compositional homogenisation. However, when surface melting is achieved, rapid mixing and therefore homogenisation, can occur.

Perhaps the most straightforward application of this principle is the martensitic hardening of the surface of iron based components, in particular, cast irons. In general, these materials cannot be successfully surface hardened unless melting has taken place to homogenise the carbon content. Such surface hardening allows excellent wear resistance to be imparted to otherwise cheap and easily manufactured components. A similar effect is achieved through use of the TIG remelting process, ^[2] but power beams offer the potential advantages of greater speed and smaller volume of modified material.

Surface melting and rapid cooling can also be used to dissolve undesirable second phase particles and either hold them in solid solution or create a finer, more beneficial distribution. For example, austenitic stainless steels, which have been sensitised to intergranular corrosion by grain boundary carbide precipitation, can be desensitised by laser surface remelting, leading to a restoration of the corrosion resistance of the material. ^[14,15] A similar effect has been reported by laser surface annealing without melting. Improved pitting corrosion resistance has also been achieved in solution annealed austenitic stainless steel through laser surface melting. This effect was attributed to a modification of the distribution of sulphide particles, which act as corrosion pit initiation sites. The resulting sulphides were smaller and more finely divided than in the original material, as a result of reprecipitation under rapid cooling conditions, and, possibly, vaporisation of some of the original sulphides. ^[3]

Rapid cooling can have a pronounced effect on the shape of the high temperature phase diagram. Non-equilibrium conditions may alter both the phase balance developed during solidification and the extent of subsequent solid state phase transformations. This has important consequences, for example in austenitic stainless steels. The microstructures developed under the cooling conditions associated with arc welding of stainless steels can be predicted using the Schaeffler diagram or a modification of it. ^[16] However, under rapid solidification conditions, microstructures may be developed which do not conform with the predictions of the Schaeffler diagram. ^[17,18]

Stainless steels which solidify as primary austenite are essentially unaffected by rapid solidification, the room temperature structure being predominantly austenite regardless of cooling rate. However, steels which solidify as primary ferrite show a strong link between cooling rate and the final microstructure. ^[17,18] Steels which solidify as ferrite, but transform to approximately 85-100% austenite during cooling, show one of two effects under rapid solidification conditions. Solid state transformation to austenite may be suppressed, leading to higher than normal final ferrite levels. On the other hand, if the cooling rate is high enough, solidification directly to primary austenite may occur, as a consequence of undercooling in the melt. This leads to retention of a higher austenite content than the Schaeffler diagram would predict. Stainless steels which solidify as primary ferrite and transform to 85% austenite when arc welded, tend to demonstrate only the first of the aforementioned effects under rapid cooling, i.e. solidification remains to primary ferrite and solid state transformation to austenite is suppressed, leading to higher than usual ferrite levels in the final structure.

These non-equilibrium phase transformations have been widely noted as a side-effect of power beam welding but they also have potential applications in the surface treatment of stainless steels whose corrosion resistance depends upon the phase balance.

Surface resolidification without crystallisation

Perhaps the greatest degree of structural change which may be obtained through rapid solidification is development of an amorphous structure, i.e. the prevention of crystallisation. Such non-crystalline structures have potentially better wear and corrosion resistance than equivalent crystalline materials. Amorphous structures may be developed in appropriate alloy systems at cooling rates of around 10 ⁶ K/sec or faster, but only if the conditions required for nucleation of crystals can be avoided. However, power beam treatment cannot lead to development of an amorphous state if the material crystallises with the same structure as the substrate. In this case, epitaxial growth will always occur. ^[1] Amorphous, metallic glass structures have been produced, e.g. FeB on steel and NiNb on Nb, ^[2] but the difficulties associated with their formation on a large scale, suggest that 'crystal-free' surfaces would be extremely difficult to ensure. Consequently, the enhanced corrosion properties of amorphous structures are likely to be particularly difficult to exploit, although wear resistance would be more tolerant to the presence of crystalline regions.

Summary

The structure and properties of a metallic surface may be modified in a range of ways, by means of rapid heating and cooling, using a laser or electron beam. In particular, improvements in wear and corrosion resistance have been demonstrated in a range of materials, by modifying only the outer skin of the component, usually a few hundred micrometres in depth, whilst leaving the bulk essentially unchanged.

Although similar effects can be attained by other means in some cases, power beams offer potentially greater speed and smaller volume of modified material than the more conventional methods. At present the practical use of power beam treatment is limited to the simplest processes, but there is an expanding range of techniques with potential industrial applications.

References