T. Sourmail

Introduction

Stainless steels represent the most diverse and complex family of all steels. The list of their applications is endless: from the harsh environments of the chemical, oil production and power generation industries to street furniture or automotive trims without forgetting most cutlery, they are used either for decorative purposes and/or for their excellent resistance to corrosion.

Stainless steels are stainless because a protective layer spontaneously forms on their surfaces and reduces the rate of corrosion to almost negligible levels. Under normal conditions, this layer heals very rapidly if scratched, so that if stainless steels only suffered from uniform corrosion, they could survive for literally millions of years (Nature, 2002:415, Newman, p743).
It is generally agreed that stainlessness is obtained for additions of about 12 wt% of chromium or more, although corrosion rates continuously reduce with increasing chromium contents from 0 to this limit.

12% Cr-steels are resitant to atmospheric corrosion but are useless in acids such as HCl or H2SO4 where they exhibit a corrosion rate even greater than plain carbon steels. As will be discussed later, corrosion resistance can be greatly enhanced above that of a basic 12% Cr steel by further addition of Cr and/or use of other alloying elements such as Ni, Mo, N etc.

Corrosion resistance is of course not the only design criterion: materials cost considerations will typically favour basic, cheaper grades (12%Cr-0.1%C) rather than heavily alloyed steels. Mechanical properties must also be taken into account, as must fabrication difficulties (machinability, deformability, weldability, etc.). The number of grades is therefore seemingly infinite, with a large number of standard compositions to which manufacturers add their proprietary variants.

These are usually divided in four or five classes on the basis of their microstructures:

martensitic or transformable stainless steels
ferritic stainless steels
austenitic stainless steels
duplex stainless steels

to which some add precipitation hardened stainless steels, although they are themselves divided into martensitic/ferritic/etc. so that they could be included in the above categorisation.

Corrosion resistance

Iron is not stable in most of the environments in which it is used. Like most non-noble metals, it does not naturally occur in its elemental form and has to be extracted from oxides. In service, the tendency is for re-formation of these oxides. Hence the rust that can be observed on most unprotected steel components.

Materials scientists and chemists often take recourse to thermodynamics to quantify and compare the stability of different phases, oxides, etc. In this case however, thermodynamics alone gives an erroneous representation of the problem: most stainless steels are used in conditions in which the dissolution of the metal is energetically favorable.
Instead, they owe their environmental resistance to the formation of a chromium oxide film which acts as a kinetic barrier: atomic transport through this layer is so slow, that the steel can be considered inert.

The passive film formed at the surface of stainless steels is extremelly thin (1-2 nm). Its chromium content depends on the bulk content, and in general, increases with the latter. The overall corrosion resistance is also enhanced as the chromium content is raised.

Corrosion can nevertheless occur if the passive film breaks down, locally or uniformly. This can happen by different mechanisms depending on the conditions of use. The most common types of corrosion are:

Uniform corrosion of stainless steels can occur in acidic or hot alkaline solutions. It results in uniform loss which can easily be predicted and allowed for. As mentionned in the introduction, uniform corrosion is very slow when the metal is in the passive state.
General corrosion resistance is increased with increasing chromium content, but other elements can be detrimental. In particular, sulfur in solid solution is believed to make passivation more difficult (M. Schütze ed., Corrosion and Environmental Degradation, Wiley-Vch, Chap 3) and therefore is generally undesirable for good corrosion properties.
Unfortunately, sulfur makes welding considerably easier (D. T. Llewellyn, Steels: Metallurgy and Applications, 1992, Butterworth-Heinemann) and also improves machinability. In the case of welding, sulfur appears to modify the surface tension of the weld pool and therefore alters its shape significantly. Austenitic grade 316 with sulfur content lower than 0.007 wt% tend to have a high width-to-depth ratio while higher sulfur contents lead to a narrower, deeper weld pool (specifying the sulfur content of 316L for welding).
Some of the standard grades contain a quantity of sulfur delibarately greater than the typical 0.003 that can otherwise routinely be achieved with modern steel-making processes (the free machining grades).
Nickel significantly improves the general corrosion resistance of stainless steels, by promoting passivation. The austenitic stainless steels series therefore possesses a corrosion resistance superior to that of martensitic or ferritic stainless steels (no nickel), particularly with mineral acids.
Pitting corrosion is the result of the local destruction of the passive film and subsequent corrosion of the steel below. It generally occurs in chloride, halide or bromide solutions. If a fault in the passive layer or a surface defect results in the local destruction of the former, dissolution of the steel underneath leads to a build up of positively charged metallic ions, which in turn causes negatively charges (e.g. chloride ions) to migrate near the defect. Even in a neutral solution, this can cause the pH to drop locally to 2 or 3, and can prevent regeneration of the passive layer.
In the passive condition, the current density is in the scale of nanoamperes/cm2; in the pit, however, it may be above 1A/cm2. Similarly, the concentration in chloride ions can be thousands of times greater than that in the solution.

Schematic illustration of pitting corrosion.

The above figure illustrates the process: the anodic dissolution of the steel leads to introduction of M+ in solution, which causes migration of Cl- ions. In turn, metal chloride reacts with water following:
```
M+Cl- + H2O -> MOH + H+Cl-
```
This causes the drop of pH mentioned earlier. The cathodic reaction, on the surface near the pit follows:
```
O2 + 2H2O -> MOH + 4OH-
```
While the propagation phenomenon is well understood, the mechanism of pit initiation is still debated. The initiation of pitting has long been associated with the presence of MnS inclusions which are difficult to avoid in the steel making process. It has recently been shown that these inclusions are surrounded by a Cr depleted region which is believed to cause the initiation (Nature, 2002:415, Ryan et al, p770).

The pitting resistance of a stainless steel is affected by its composition. Increasing the Cr content, or adding Mo or N both enhance the pitting resistance, though they are not equally potent in this respect. For comparison purposes, an index is often used to represent the combined effect of these elements:
```
pitting index=Cr+3.3Mo+16N
```
where Cr, Mo and N are given in weight percent.
One obvious environment where pitting corrosion is of concern is marine applications. AISI type 316 (an 18Cr-12Ni austenitic stainless steel with 2-3% Mo) is often the material of choice in this case, although the severe conditions met in offshore platforms, for example, call for heavily alloyed steels with up to 6% Mo (for example 254SMO, Avesta Sheffield).
Use of stainless steels in offshore platforms.
Street furniture is another case where pitting resistance might be relevant, particularly in colder areas where salt de-icing is common.
Sensitisation is one of the corrosion mechanisms which causes widespread problems in austenitic stainless steels, particularly in welding. This problem can be so severe as to cause grain decohesion, as shown in the picture below.

Grain decohesion due to intergranular corrosion, photo courtesy M. Shimada (Shimada et al, Acta Mater., 2002:50, p2331).

In normal conditions, austenitic stainless steels are given a high-temperature heat-treatment, often called solution-treatment, which, as its name indicates, aims at obtaining a single f.c.c. solid solution (austenite).
At lower temperatures however (roughly, 800 C and below), this solution is not stable: in the basic AISI 304, the carbon content exceeds the solubility limit in austenite. This is due to the presence of chromium, which forms stable carbides.

In austenitic steels, the chromium carbides are M23C6 (see Sourmail, Mater. Sci. Techn. 2001:17, p1 for more details about this phase), in which M stands for Cr, Fe and possibly some Mn and Mo if present in the alloy. Because the formation of these carbides involves long-range diffusion, it can be avoided by quenching the steel after the solution-treatment.

If the steel is subsequently reheated, precipitation of M23C6 may occur. Because nucleation is considerably easier on grain boundaries, there is a very strong tendency for localised precipitation. Precipitation on other defects (twin boundaries, dislocations, inclusions..) does occur, although after longer exposure at high temperatures. Depending on the composition, M23C6 may be found on grain boundaries after only a few minutes at temperatures around 750 C. The range of temperature in which sensitisation occurs is bounded in the lower temperatures by the very slow kinetics of diffusion controlled transformations, and in the higher temperatures, by the fact that chromium depletion is less pronounced at higher temperatures. As illustrated below, the conditions in which a steel is sensitised vary with temperature and time.

The time and temperature dependency of sensitisation, after Mayo, Mater. Sci. Eng. A, 1997:232, p129.

When austenitic stainless steels are welded, any metal which has been melted usually cools fast enough to prevent carbide formation. However, in the metal adjacent to the fusion zone, the so-called heat-affected zone (HAZ), the temperature changes might be such that sensitisation occurs.

Remedies

Various solutions can be implemented to avoid sensitisation:
- The first one is obviously to reduce the carbon content of the material so as to limit the precipitation of M23C6. This approach defines the AISI L grades, such as 304L and 316L, which have lower carbon content than their standard counterparts. For both these steels, the maximum acceptable carbon content is reduced to 0.03 wt% (from 0.08 for the corresponding standard grades).
  
  The effect of carbon content on the kinetics of sensitisation. After (Gooch, Weld decay in austenitic stainless steel, The Welding Institute, 1975)
- Another `similar' solution consists in introducing carbide formers which have an even greater affinity for carbon than chromium. These include Nb, Ti, V or Ta. Steels containing these elements (or a combination) are said to be stabilised (with regard to grain boundary precipitation of M23C6).
  Grades 321 (Ti stabilised) and 347 (Nb stabilised) represent the most common stabilised austenitic stainless steels. In welding applications, grade 321 is not used as a filler metal because titanium does not transfer well accross a high temperature arc. 347 is therefore used as a filler metal when joining components made out of 321 or 347 (the latter being seldom used as parent material).
  To obtain stabilisation, it is not sufficient to add Nb or Ti. A stabilisation heat-treatment must be performed to ensure formation of TiC or NbC. This is usually performed by maintaining the steel for 1 or more hours at temperatures around 900 C. At lower temperatures, M23C6 may form faster than TiC or NbC.
- In some cases, a solution-treatment can be given to dissolve carbides which may have formed on grain boundaries, after welding for example.
- A variety of other factors impact on the problem, such as grain size and the nature of the grain boundary. It has been shown that sensitisation can be avoided by grain boundary engineering (Shimada et al, Acta Mater., 2002:50, p2331), in which a thermomechanical treatment produces a microstructure with mostly low angle grain boundaries. The latters have a lower energy and are therefore less potent nucleation sites for M23C6. On the other hand, it appears that increasing the orientation randomness of the grain boundaries beyond a treshold can also lead to a reduction in sensitisation (Wasnik et al, Acta Mater. 2002:50, p4587). In this study however, it is possible that the cause is a grain size reduction rather than a change in the nature of the grain boundaries.
Mechanism

As explained earlier, sensitisation is caused by the formation of chromium carbides on grain boundaries. The Cr-rich precipitate draws chromium from the adjacent matrix, which results in the formation of a chromium depleted zone. If the chromium content is below 11-12% in this area, the steel is said to be sensitised.

Schematic illustration of grain boundary chromium carbides precipitation and corresponding Cr profile.

This short description of the problem hides most of its interesting complexity. The first difficulty occurs if one considers the phase diagram austenite/M23C6. This predicts that the chromium content of the austenite in equilibrium with M23C6 is only slightly lower than the bulk composition, which makes sense if one remembers that M23C6 seldom form more than 1% volume fraction.
The reason why the chromium content locally drops to much lower values is to be found in the dynamics of the fluxes of elements diffusing towards the precipitates. This problem is reviewed in details in C. H. Too, MPhil thesis, 2002 and Sourmail et al, ISIJ Int. 2003:43, p1814.
Stress corrosion cracking is an important phenomenon often related to pitting corrosion. It is the result of a combination of corrosion and applied stress. Under these conditions, a crack can progress alternatively through dissolution of its tip or mechanical propagation. Stress corrosion cracking is particularly dangerous because it may take thousands of hours for a crack to nucleate, but considerably less for it to propagate. Dramatic examples of catastrophic corrosion by SCC include the collapse of swimming pools ceilings in 1985 (Switzerland, 12 fatalities), and 2001, (Netherland, no fatalities), both attributed to SCC of type 304 or 316 austenitic stainless steels. It has been suggested that 6%Mo austenitic stainless steels should be used in these environments.
NPL's guide to Stress Corrosion Cracking

Phase equilibria

Main phases

As mentioned in the introduction, the designation stainless steel conventionally implies little more than a 12% Cr content. Most of the stainless steels are based on the Fe-Cr-C and Fe-Cr-Ni-C systems, but other alloying elements are also important.

Iron and its alloys can exist in two crystallographic forms (body centred or face-centred cubic). In pure iron, the f.c.c. structure exists between 910 and 1400 C, the b.c.c. structure below and above this interval (up to the melting temperature of 1539 C).
The importance of this phase-transformation in the metallurgy of steels cannot be overestimated. This transformation allow for a wide range of microstructures to be achieved by controlled heat-treatment. Mechanical properties are essentially related to microstructure, and can therefore be obtained in an extraordinarily large range of strength, toughness, etc.. Stainless steels are routinely produced with strengths from 100 MPa to largely more than 1GPa.

Knowledge of the relative stability of the b.c.c. and f.c.c. structures of iron alloys is therefore of prime concern. The history of stainless steels started with a martensitic grade (12%Cr-0.1%C) in Sheffield, UK and the austenitic 18%Cr-8%Cr in Germany ( more about the history of stainless steels). For this reason, and also because they are most often the major alloying elements, Cr and Ni have long been used as reference to quantify the influence of alloying elements on the b.c.c.<->f.c.c. phase transition: chromium additions tend to stabilise the b.c.c. phase, while nickel additions stabilise the f.c.c. one.

Pseudo-binary (isopleth) Fe-Cr diagram for 0.1% C. Chromium additions stabilise the b.c.c. form of iron, beyond about 18%, Fe-Cr steels no longer undergo the α/γ transition (ferritic stainless steels).

Without carbon, the limit beyond which austenite no longer forms is about 13.5 wt%. However, additions of carbon stabilise the austenite and therefore increase this limit.
Other alloying elements also affect the stability of austenite/ferrite in one direction or another. This has led to their classification as α-stabilisers or γ-stabilisers. The concept of Cr-equivalent (for α-stabilisers) and Ni-equivalent (for γ-stabilisers) is widely used in conjunction with the Schaeffler diagram to quantify their roles:
Cr equivalent = (Cr) + 2(Si) + 1.5(Mo) + 5(V) + 5.5(Al) + 1.75(Nb) + 1.5(Ti) + 0.75(W)
Ni equivalent = (Ni) + (Co) + 0.5(Mn) + 0.3(Cu) + 25(N) + 30(C)
All in weight percent.

Schematic Schaeffler diagram. Although primarily intended for use in welding, this diagram is sometimes used in alloy design as it provides an easy way to estimate the microstructure of a stainless steel.

Modern thermodynamics calculation tools such as Thermocalc or MTDATA based on the CALPHAD method allow more rigourous determination of equilibrium phase diagrams in multicomponent systems.

Second phases

In most grades of stainless steels, alloying elements are present in quantities sufficient to cause precipitation of second phases. Most often the stable carbides, nitrides or intermetallics are of little relevance as they tend to follow a long and complex precipitation sequence.
This is because the kinetics of precipitation are largely controlled by nucleation, and nucleation itself is not necessarily easier for the most stable precipitates.
From the Fe-Cr diagram presented earlier, it appears that typical martensitic steels should exhibit ferrite and M23C6 in equilibrium at 600 C (for example). In practice, this carbide is only found after relatively long ageing. Intermediate phases are, in order, cementite, M2X and M7C3 to finally obtain M23C6.
These sequences are far more complex in heavily alloyed ferritic or austenitic stainless steels such as those used in the power generation industry. Considerable research work is being devoted to predicting quantitatively the precipitation sequences in such alloys. This is mainly because their life expectancy (about 30 years) vastly exceeds the length of laboratory tests.

Martensite formation

Most stainless steels have a high hardenability, meaning that reconstructive austenite to (ferrite + carbides) transformation is unlikely to happen unless the steel is cooled particularly slowly.
The most important features of these alloys are therefore the martensite start (Ms) and finish temperatures (Mf). For martensitic steels, the range [Mf-Ms] should be above the room temperature to ensure fully martensitic structure. On the contrary, the [Ms-Mf] range of austenitic stainless steels is often well below 0 C, which is why austenitic steels are used in cryogenic applications. Cold work can cause martensitic transformation to an extent with depends on the deformation and on the alloying composition. Heavily alloyed austenitic steels with up to 20% Cr and 25% Ni are fully stable.

The mechanisms of martensite nucleation are reasonably well understood and there are a number of models which predicts the Ms temperature with an acceptable accuracy.

Categories

On the basis of their main microstructural features, grades of stainless steels are typically divided into four categories:

Martensitic stainless steels, typified by AISI types 410/420/440 contain about 12% Cr and 0.1% C in its basic composition, leading to a fully martensitic structure at room temperature.
Ferritic stainless steels contain larger amounts of Cr which stabilise the ferritic structure.
Austenitic stainless steels, such as AISI type 304 typically contain 18% Cr and 8% Ni which stabilises the austenitic structure. The large susbtitutional content depresses the martensitic transformation temperatures well below 0 C.
Duplex stainless steels, whose microstructure is approximately made up of 50/50 ferrite/austenite.

Martensitic Stainless Steels

These steels still undergo the b.c.c./f.c.c. transformation of iron, although the range of austenite stability is reduced.
As for conventional steels, mechanical properties can be considerably altered by heat-treatments. Typical heat-treatments consist of austenitisation at a temperature suitable for dissolution of carbides. Stainless steels have a high hardenability, that is to say, reconstructive transformations are considerably slowed by the presence of Cr, so that a fully martensitic structure can be achieved without a severe quench. Oil or water quenching are nevertheless used with large sections so as to ensure martensite formation throughout.

Typical compositions cover 12 to 18 Cr and 0.1 to 1.2 C (wt%). As with other martensitic steels, a balance must be sought between hardness and toughness. An untempered martensitic structure typically has high hardness/yield strength but a low toughness and ductility (although the exact values depend on the carbon content). In many conditions, these are used after a tempering treatment between 600 and 750 C, which result in a lower hardness but improved toughness.

In some applications such as cutlery, surgical intruments etc., high strength is desirable and toughness/ductility of little concern. A lower temperature tempering is then used to retain most of the strength. AISI type 420 (0.15-0.4C, 1.0Mn, 1.0Si, 0.04P, 0.03S, 12-14Cr all max wt%) is a typical composition for such applications. Its proof strength in quenched and tempered condition can be in excess of 1.2 GPa. For type 440C tempered at 300 C, the proof strength can reach about 2 GPa.

The table below shows the composition and typical use of AISI standard martensitic grades:

Standard grades of martensitic stainless steels. Comparison of grade specifications .
AISI grade	C	Mn	Si	Cr	Ni	Mo	P	S	Comments/Applications
410	0.15	1.0	0.5	11.5-13.0	-	-	0.04	0.03	The basic composition. Used for cutlery, steam and gas turbine blades and buckets, bushings..More.
416	0.15	1.25	1.0	12.0-14.0	-	0.60	0.04	0.15	Addition of sulphur for machinability, used for screws, gears etc. 416 Se replaces suplhur by selenium. More
420	0.15-0.40	1.0	1.0	12.0-14.0	-	-	0.04	0.03	Dental and surgical instruments, cutlery.. More
431	0.20	1.0	1.0	15.0-17.0	-	1.25-2.0	0.04	0.03	Enhanced corrosion resistance, high strength. More
440A	0.60-0.75	1.0	1.0	16.0-18.0	-	0.75	0.04	0.03	Ball bearings and races, gage blocks, molds and dies, cutlery, More
440B	0.75-0.95	1.0	1.0	16.0-18.0	-	0.75	0.04	0.03	As 440A, higher hardness
440C	0.95-1.20	1.0	1.0	16.0-18.0	-	0.75	0.04	0.03	As 440B, higher hardness

In additions to the standard grades, a large number of alloyed martensitic stainless steels have been developed for moderately high temperature applications. Most common additions include Mo, V and Nb. These lead to a complex precipitation sequence. A small amount (up to 2 wt%) of Ni is added which improves the toughness.
The 12Cr-Mo-V-Nb steels are used in the power generation industry, for steam turbine blades operating at temperatures around 600 C. Current research focusses on achieving service temperatures of 630-650 C under a stress of 30 MPa.

Ferritic stainless steels

Ferritic stainless steels: contain typically more chromium and/or less carbon than the martensitic grades. Both changes act towards stabilisation of ferrite against austenite so that ferrite is stable at all temperatures. Therefore, ferritic stainless steels cannot be hardened by heat-treatments as is the case of martensitic ones. They exhibit lower strength but higher ductility/toughness. Typical application may include appliances, automotive and architectural trim (i.e. decorative purposes), as the cheapest stainless steels are found in this family (aisi 409).

Some standard grades of ferritic stainless steels. Comparison of grade specifications .
AISI grade	C	Mn	Si	Cr	Mo	P	S	Comments/Applications
405	0.08	1.0	1.0	11.5-14.5	-	0.04	0.03	0.1-0.3 Al
409	0.08	1.0	1.0	10.5-11.75	-	0.045	0.045	(6xC) Ti min
429	0.12	1.0	1.0	14.0-16.0	-	0.04	0.03
430	0.12	1.0	1.0	16.0-18.0	-	0.04	0.03
446	0.20	1.5	1.0	23.0-27.0	-	0.04	0.03	0.25 N

High chromium ferritic stainless steels such as 446 are sensible to the so-called '475 C embrittlement', which is caused by the decomposition of the Fe-Cr solid solution in two phases, Fe and Cr-rich respectively. Around 475 C and below, and for Cr contents greater than about 25 wt%, this decomposition is spinodal and typically exhibits wavelength below 10 nm. As the decomposition occurs, a continuous increase of hardness is observed: for example, the hardness of an Fe-28Cr steel can increase by more than 300 Hv over an exposure 10,000 at 450 C (Ishikawa et al., Mater. Trans. JIM, 36:1995, p16-22). This results in a severe drop of impact toughness and ductility.
Addition of Ni appear to accelerate the spinodal decomposition and raise the maximum temperature at which it is observed. When post-weld heat-treatment is not possible, welding of ferritic stainless steels is usually done with a metal filler containing Ni, and there is therefore the possibility of weld embrittlement.

Austenitic stainless steels

Austenitic stainless steels: these stainless steels owe their name to their f.c.c. crystallographic structure. Typical compositions in the early 20th century were 18Cr-8Ni. The austenite in these alloys was only stable because of the relatively large carbon content, and modern equivalent usually contain up to 10.5 Ni.
These steels are often in metastable conditions at room temperature or below, and while the reconstructive formation of ferrite is not of concern, the formation of martensite can be. Most grades have a martensite start temperature (Ms) well below 0 C. However, cold work can result in formation of martensite at temperatures higher than Ms (this result in the sample becoming magnetic, while a fully austenitic structure is not). The impact of deformation on the stability of the material is conveniently quantified by the M_d,30 temperature, the temperature at which the structure is 50% martensitic for 30% deformation.

The presence of nickel improves considerably the corrosion resistance when compared to the martensitic and ferritic grades.
AISI type 304 is the basic 18/8 austenitic stainless steel, so widely used that it accounts for about 50% of all stainless steel production. Other standard grades have different preferred applications; for example, type 316 which contains up to 3 % Mo, offers an improved corrosion resistance, in particular, improved pitting corrosion resistance, which makes it a material of choice for many marine applications (off-shore platforms etc..), but also for coastal environments (more on stainless steels in architecture).
In severe conditions however, 316 is not sufficient and special steels such as 254 SMO are used (example: steels used in offshore oil platforms), which contain up to 6% Mo.

The AISI 300 series and other examples of heat resistant austenitic stainless steels; E1250 is Esshete 1250 (Corus), 254SMO is trademark Avesta Sheffield.
AISI grade	C max.	Si max.	Mn max.	Cr	Ni	Mo	Ti	Nb	Al	V
301	0.15	1.00	2.00	16-18	6-8
302	0.15	1.00	2.00	17-19	8-10
304	0.08	1.00	2.00	17.5-20	8-10.5

310	0.25	1.50	2.00	24-26	19-22
316	0.08	1.00	2.00	16-18	10-14	2.0-3.0
321	0.08	1.00	2.00	17-19	9-12		5 x %C min.
347	0.08	1.00	2.00	17-19	9-13			10 x %C min.
E 1250	0.1	0.5	6.0	15.0	10.0					0.25
20/25-Nb	0.05	1.0	1.0	20.0	25.0			0.7
A 286	0.05	1.0	1.0	15.0	26.0	1.2	~1.9	~0.18	~0.25
254SMO	0.02	0.8	1.0	18.5-20.5	17.5-18.5	6-6.5	~1.9	~0.18	~0.25