Selection of high-temperature abrasion resistant steels for the mining and processing industry

(1)

Selection of high-temperature abrasion resistant steels

for the mining and processing industry

Lucie Gutman

Materials Engineering, master's level 2020

Luleå University of Technology

Department of Engineering Sciences and Mathematics

(2)

Summary

High-temperature abrasion is an expensive issue in industrial fields such as glass and cement production or mining and processing industry. Yet its effects on steel are not well documented. This study investigates and analyses the behaviour of six different steel grades placed in hot abrasive conditions similar conditions encountered in the industry to enables better material selection.

Abrasion tests in a slurry pot were done at room temperature and at 500 ^◦C. Impact and tensile tests were also performed at different temperatures. To complete the mechanical properties evaluation, hardness measurements were executed before and after tempering at 500 ^◦C.

Wear rates assessed at room temperature or at 500 ^◦C, are independent of the mechanical properties of the material. At high temperature, it was shown that wear rates and performance of the steels were influenced by tempering and leading to a unique microstructures for all steel grades investigated and equalize their performances.

To conclude, high temperature wear of the investigated grades does not depend on their mechanical properties, however, it can be influenced by their tempering resistance. As the temperature increase, steel tempers, its mechanical properties decrease and homogenise with other steel grades’ performances, but some grades keep their properties longer at high temperature.

(3)

Acknowledgements

This project was carried out with the help and support of many people and I am thankful to each of them.

First, I would like to acknowledge Dr David Quidort and Dr C´eline Knafou, from ArcelorMittal Industeel for the opportunity to realise this incredible project, for the guidance, feedback and wise pieces of advice. Thank you for letting me carry on my master thesis in the best conditions possible during these troubled times.

I am deeply grateful to Prof. Esa Vuorinen, who always answered my question with wisdom, kindness and patience and also gave me the possibility to realise my master thesis in such an interesting field.

Alexandre Giorgi and Ianis Charleux helped me with equipment, technical questions and daily small troubles, always with smiles, kindness, creating a so good working atmosphere. I express my honest gratitude to you both and I wish you all the best for your future.

I also would like to thanks Dr Bianca Frincu for her willingness to spend time on SEM analysis with me with graciousness and for helpful comments.

Thanks to everyone working at CRMC that help me during this project, you create a nice and welcoming space to work, learn and grow.

Finally, I am grateful for all the help, support and smiles from my friends and family. Through these troubled times of pandemic, you were far away but still there for me, thank you.

(4)

Chapter 1 Introduction

Abrasion and more generally wear are actual concerns in industrial fields. By progressive mass loss of con- cerned machine parts, abrasion decreases their life-length and causes their failure, especially in applications as mining, mineral handling or crushing. In some other industrial applications, temperature and wear coexist.

Temperature increases the rate and severity of the wear. Hence failure occurs even more quickly. Failure has an important cost, not only on new components but also in maintenance and machine downtime.

The environmental aspect is also of importance nowadays. Failure has a considerable impact on the envi- ronment since raw materials might be used for production, as well as non-renewable energy for production, transportation, etc.. To avoid or at least space out costly replacement of components, wear resistance of the parts and careful material selection are a matter of interest.

1.1 Project’s background and motivations

Abrasion resistance at room temperature of a large range of material is a well documented and studied subject. Theory of room temperature abrasion is exposed books such as Wear analysis for engineers [1], whereas Chintha [2], for instance, produced a detailed literature review on the subject in 2019. Abrasion resistance of specific steel was also studied, as in Lindstrom [3], master thesis in which carbide-free steels’

wear resistance is investigated until field tests.

On the other hand, fewer studies were performed on the coupled effect of high-temperatures and abrasion.

The issue exists and is an expensive expense in some industrial fields as cement plants, fibreglass wool production or steel production. For instance, studies were done on sinter plants parts [4, 5, 6, 7], where temperatures can reach 1200^◦C. High-temperature abrasion can be observed on sinter crusher metallic parts that crush hot sinter cake at a temperature between 600 ^◦C and 800^◦C. Abrasion and high-temperatures also coexist in rotary cooler that cool down sinter, in which the abrasive is a mixture of coke and iron ore at 700 ^◦C.

Quench and tempered steels are commonplace in abrasion resistance field. A master thesis, see [8], study their response to coupled effect of high-temperature and abrasion. However, some field tests [4] reported better resistance of multiphases steels in hot abrasives conditions.

Knowing the behaviour of different steel grades when submitted to high-temperature abrasion conditions is crucial to perform a careful material selection. For each situation, each material can respond differently and choosing wisely the grade employed can save time, money and resources. For these reasons, it is important to investigate multiphase steels behaviour in hot abrasive conditions.

1.2 Objectives of the project

The main objective of the project entitled Selection of high-temperature abrasion-resistant steels for the mining and processing industry is to study the behaviour of several ArcelorMittal Industeel steels placed in hot abrasive conditions, in order to be able in the future to select the best steel for each specific application.

A wide collection and analysis of data regarding the behaviour and performances of several steel grades in

(6)

hot abrasive conditions is the first step before the selection of material. Several aspects of the grades will be studied in order to have an overview of their properties and to be able to advise the best steel for the best application.

1.3 Scope of the project

The presented work was performed with Industeel, a subsidiary of ArcelorMittal group, the world’s leading steel and mining company. Six different Industeel’s production facilities produce a large range of steel grades, with carbon, low alloy and stainless steels but also nickel-based alloys, that can be shaped into several types of products: heavy plates, formed and frame cut pieces, forged blocks, clad plates, rings and ingots for forged pieces. The diversity in steel grades and products enables Industeel to fulfil requirements for diverse applications, as high strength steels, steels for moulds and tools, protection steels and wear-resistant steels.

Several Industeel’s steel grades were selected to be studied in abrasive conditions at a temperature that could reach 500 ^◦C. Wear resistance at room temperature and at 500 ^◦C was investigated. The microstructure changes through tempering has also been studied and finally their mechanical properties, at room temperature, at elevated temperatures and through tempering were considered.

(7)

Chapter 2 Wear and abrasion theory

Wear can be defined as the progressive loss of mass and material due to relative motion between a surface and one or several contacting substances [1, 9]. In industrial fields, wear of industrial components can be a severe and costly issue. Hence, there are economical interests to understand wear behaviour and develop wear resistance. Indeed, in industrialized countries, the cost of abrasive wear can reach 4% for the gross national product, as reported in [10]. However, wear does not directly depend on the bulk material mechanical properties, but on the tribosystem properties that influence the significance of the wear and the wear mechanisms involved, thence improving wear resistance is a complex aim [1,11].

Wear rely upon the tribosystem properties. A tribosystem is described with a wear situation and one or several wear mechanisms. There is eight different wear mechanisms, cited in table 2.1a. The dominant wear mechanisms express the way the material is removed or damaged, or the surface changes due to wear. It can be influence by different elements such as the relative motion, the loading, the surface topography, etc.

There are four main wear situations, based on the nature of the motion that causes the wear, summed up in table2.1b. For each situation, there is dominant wear mechanisms and several parameters that influence the severity of the wear [1].

Table 2.1: Wear mechanisms and situations according to Bayer [1]

(a) The eight different wear mechanisms Wear mechanisms

Adhesive wear

Single-cycle deformation Repeated-cycle deformation Electrical discharge

Atomic Chemical Tribofilm Thermal

(b) The four different wear situations Wear situations

Sliding Rolling Impact Abrasion

2.1 Abrasion or abrasive wear

One of the wear situation is abrasion, also called abrasive wear. It is due to the sliding contact between hard asperities and a softer surface. When a load is applied on the asperities that move along a solid surface, hard particles penetrate the softer body and remove material [12, 13, 2]. The appearance of the wear scar, the severity of wear and the nature of the wear mechanisms involved depend on several factors such as the nature of particles, their size and shape, their amount, and the relative hardness of particles and surface. Abrasion is mainly caused by deformation mechanisms (single-cycle and repeated-cycle).

Gates [12], gives a proposition of abrasion classification, which is then used by several authors such as Chavon-Nava [13]. Abrasion can be caused by several abrasion processes that can be divided, according

(8)

to Gate’s classification, into two main groups: Two-body and Three-body abrasion, as presented in figure 2.1. In Two-Body abrasion, the first body is being worn by the second body in motion relative to the first body. In that case, the worn body can be called ”the body” whereas the moving body is ”the counterbody”.

In Three-Body abrasion processes, at the interface between the body and the counterbody a third body such as solid particles, lubricant, wear debris is found. To illustrate the difference between Two-Body and Three-Body abrasion, Chintha [2] takes the example of samples polishing. Two-Body abrasion occurs when a metallic sample is polished on sandpaper, whereas polishing a metallic sample on a polishing cloth with a hard particle suspension is Three-Body abrasion.

The Three-Body abrasion category is divided into two situations; Open abrasion and Close abrasion. In close conditions, the particles (third body) are trapped between the moving surfaces. Whereas in open conditions, the surfaces can be considered as far apart and only the body and the third body are involved in the abrasion process [12, 13].

Finally, the Open Three-Body abrasion gathers different situations regarding the particles. If the particles are fragmented and fractured during the abrasion, then it is high-stress abrasion. If the particles remain unbroken, it is low-stress abrasion. The last option is coarse particles, that cut the surface and remove a large amount of material, in this case, the gouging abrasion is occurring [12,13].

Figure 2.1: Diagram of the different wear processes according to Gates [12]

2.2 Influence of certain factors on abrasion

The wear resistance of a material depends on the tribosystem the material is part of. However, some mechanical properties such as hardness or toughness can be linked to wear resistance of the material in specific wear conditions. As the microstructure directly influences the mechanical properties, it can be interesting to link the microstructure to wear resistance in a specific tribosystem. When submitted to high temperature, materials properties and behaviour change, the temperature is known to increase the severity of abrasive wear and the wear rate [14]. The intensification of abrasion with temperature can be gradual because linked to the hardness of the material, that decreases gradually as temperatures increases. The abrasion can also change considerably above a critical temperature. In that case, a change in wear mechanisms is responsible for the wear evolution, for instance, the matrix can become too soft or oxidation can cause failure [7]. As the temperature can change hardness, toughness and the microstructure of materials, it is important to study changes that can occur to understand high-temperature abrasion phenomenon.

The next parts describe different aspects of the material that can be linked to abrasion resistance. The effects of temperature on these aspects are also discussed in order to understand the temperature role in abrasion.

(9)

2.2.1 Hardness

One of the material properties that can influence the wear rate and intensity is hardness [1,15]. For some materials, in specific configurations, linear correlations between abrasion and hardness can even be found [2].

As Hernandez [16,14] explained, a high hardness means high resistance to penetration under a concentrated load, so to plastic deformation. If the material is resistant to plastic deformation, less material is removed by wear and the wear rate is lower. Bayer [1], also exposed that if the surface of the material is harder than the abrasive, the wear processes change and the wear should be reduced. Even if in some cases, a linear correlation can be found between hardness and wear resistance, Mutton [17] exposed that it is not a rule.

Indeed, using different heat treatments, steels with the same hardness but different composition were obtained and submitted to the same tests. The wear resistance was different even if the hardness of the material was the same. Hence the correlation between hardness and wear rate should be handled with caution.

However, beyond a certain initial hardness, the wear rate is less and less sensitive to hardness. Indeed, medium-hard steel can easily harden during deformation which decreases wear rate by increasing the surface hardness. On the other hand, hard steels hardens less under abrasion, hence the wear resistance is not increased. Chintha [2] highlighted this link between the ability to work harden and the wear resistance.

As temperature increases, hardness tends to decrease gradually down to a plateau value. If the steel is free from carbide forming elements, no secondary precipitation occurs and the hardness decreases considerably above 400 ^◦C. But if the steel contains high amounts of carbide forming elements such as Cr, Mo, or V, secondary precipitation can take place between 400 ^◦C and 600 ^◦C. New carbides, hard precipitates, are formed during the secondary precipitation. The hard precipitates strengthen the matrix and obstruct the dislocation motion. The hardness of the material is retained. [16,14].

2.2.2 Toughness

Another material property that influences the tribosystem is toughness. Indeed toughness can be described as the ability of the material to absorb energy before fracture, the tougher the material, the more difficult it is to remove the matter from it. Hernandez [16,14] reports that if two materials has an equivalent hardness, the toughest has the highest abrasion resistance in case of Three-Body abrasion wear because more energy is required to remove the matter from the steel during the process. Chintha [18] carried on experiment on equivalents hardness, composition and microstructural phases steels in order to determine the role of fracture toughness in abrasion wear resistance. The conclusion was identical: the toughest material resists the best to abrasion.

2.2.3 Oxidation and tribolayer

Oxidation and tribolayer formation also influence wear at high temperature. Tribolayers can be formed with wear debris, oxides and materials on the surface of the material. Tribolayers also called mechanically mixed layers or composite layers. These surface films can in some situations reduce wear as explained in [14, 1].

Winkelmann [5] exposes that material with lower hard phase content tends to form composite layers easier than the ones with more hard phases. If oxides are used to form composite layers that can in some occasions reduce wear. At high temperatures, oxides are formed continuously. If they are weaker than the initial material, the wear is increased. Whereas, if they form a protective layer, stronger than the initial material, wear decreases.

2.2.4 Microstructure

The microstructure is a key parameter for mechanical properties and therefore for wear resistance. Several aspects were investigated through different studies. In 2011, Leiro [19] investigated austempered steels and showed that as the austempering temperature decreased the microstructure was thinner and the wear resistance better. Xu [20] also related a grain size diminution with better abrasion resistance. Decreasing the grain size enables to increase both strength and ductility, which are important properties for wear resistance.

In the case of multi-phase steels, the volume fraction of each phase, as well as the bonding between phases are also important. Hernandez [16,14], highlighted that a higher hard phase content (such as martensite) increases the global hardness of the material and its wear resistance. On another hand, Xu [20] established

(10)

that an increase of martensite volume fraction is beneficial to the wear resistance up to a certain point. Hence it exists an optimal volume fraction of the hard phase that leads to the best wear resistance of the steel.

The presence of micro carbides is described as an obstacle to dislocation motion and so can be considered as improving the wear resistance of the steel according to Hernandez [16,14].

Several authors also mentioned the Transformation-Induced-Plasticity (TRIP) effect as improving wear resistance. The TRIP effect is the transformation of retained austenite into martensite induced by local stress or strain. This transformation, described by Bhadeshia and Honeycombe in [21], is known to improve mechanical properties of the steel, such as ductility or strength. It also allows a higher work-hardening rate and relieves stress concentrations in the microstructure. Xu [20] observed the evolution of retained austenite volume fraction during abrasion. Due to the TRIP effect, the retained austenite content of certain steel decreased during abrasive wear, it also leads to a surface hardening. Leiro [19] recorded that the thickness of the hardened layer increases as retained austenite volume fraction increases. However Gola [15] explained that no correlations were found between retained austenite content and wear resistance. Indeed retained austenite can exist in different morphology: film-like of blocky. The film-like austenite is very stable and does not transform into martensite, whereas blocky austenite is unstable even at room temperature and can easily turn to martensite. For Liu [22], an increase in retained austenite volume fraction is beneficial to the wear resistance, if a TRIP effect can occur. If no transformation can take place, wear resistance of the steel is decreased by retained austenite. In abrasion, a low volume of retained austenite leading to a high hardness is more interesting than a high volume fraction of austenite, that leads to better toughness which is improving impact stresses resistance [15].

2.3 Influence of temperature on microstructure

The influence of temperature on certain material properties was discussed in the previous section, see 2.2.

However, even submitted to high temperature for a long time, tempering of the steel occurs. Tempering is a change of microstructure leading to a change of mechanical properties of the steel. This phenomena can occur in industrial fields: mechanical parts can be submitted to hot abrasive conditions for several minutes, hours or days, then cooled down for a time before being reused. Hence the effect of tempering on steel needs to be understood better in order to be able to cross-reference data with abrasion behaviour.

When submitted to temperatures between 250^◦C and 700^◦C for a certain amount of time, steels are subjected to tempering [23]. Tempering of steel is due to a gap between the initial microstructure (martensite or bainite for instance) and the equilibrium phase. As temperature increases, atomic diffusion also increases, enabling transformations to approach equilibrium phase. Tempering is a diffusional process, its rate and extend are determined by the gap between the initial microstructure and the equilibrium as Bhadeshia explained in [21].

2.3.1 Tempering of martensite

Before tempering, martensite is a body-centered tetragonal single-phase, supersaturated with carbon. The transformation that occurs during tempering turns martensite, a single-phase, into tempered martensite which is part of the iron-iron carbide phase diagram, as it can be read in [23]. The transformation occurs by rejection and precipitation of excess carbon. Tempered martensite is made of dispersion of cementite particles, a hard phase, in a continuous ferritic matrix, a more ductile phase, see [21]. In tempered martensite, the hard phase is made of carbides, however even if cementite is the most stable precipitate, it is not the first carbide to form; transition carbides can form and then dissolve to form cementite, more details can be found in [24].

Tempering sequence

Tempering of martensite is a transformation that occurs in several stages, depending on the temperature. The different steps of the transformation of fresh martensite into tempered martensite, described in the following paragraph, are exposed in details in several books [21,23,24,25].

(11)

Step 1 occurs between 0 and 250^◦C. If the steel is low carbon steel, its martensite start-temperature (Ms) is high, hence the first stage of tempering can occur during the quench on first formed martensite. This phenomenon is called auto-tempering.

During the first tempering stage, the excess carbon present in martensite starts to segregate to defects, such as dislocations or boundaries. If retained austenite is present in the microstructure, excess carbon also starts to partition into it. Tetragonality of the untempered martensite is partly lost. Finally, carbides precipitate in martensite. The temperature increases carbon distribution instability in the matrix, leading to precipitation of transition carbides or directly cementite. In the case of low carbon steels (less than 0,2 wt% of carbon), transition carbides (ε-carbides) can not precipitate. Indeed, as the carbon concentration is low, the (M_s) temperature is rather high, carbon trapped in martensite as enough time during cooling to diffuse to the laths boundaries. So that there is no carbon left in solution to precipitate in carbides during reheating.

Step 2 takes place at a higher temperature: between 200 ^◦C and 300^◦C. At that temperature, the excess carbon present in martensite formed carbide precipitates. In steel containing more than 0,4 wt% of carbon, retained austenite is likely to be present after the first quench. In that case, during step 2 of tempering, the retained austenite decomposes into bainite, made of bainitic ferrite and cementite.

Step 3 leads to a complete loss of martensite tretagonality, martensite turns to a not supersaturated in carbon ferrite. At this point, cementite appears. The nucleation of cementite takes place on to different types of nucleation site. On the one hand, cementite nucleates on boundaries: prior austenite grain boundaries, martensite laths boundaries, or twin boundaries. On the other hand, cementite can appears at the interface between transition carbides and matrix. At the interface, lath-like cementite initially forms and grows as the transition carbide disappears.

Step 4 is the final step of tempering, that occurs as temperature overcome 350 ^◦C. At these temperatures, ferrite plates recrystallises into equiaxed grains. Recovery of dislocation structure is also observed with the elimination of dislocation tangles and a drop in dislocation density. As temperature increases so do the diffusion rate of carbon allowing cementite to coarsen and spheroidise. Spheroidisation decreases the energy of the structure and occurs best near boundaries, where diffusion is easier.

Final cementite particle size is determining for mechanical properties of the material and depending on temperature. As the temperature increases, the diffusion rate increases as well. Cementite grows better and the particle size increases. The phase boundary area between ferrite and cementite decreases, leading to a softer and weaker material, with higher ductility and higher toughness.

Carbide formation

During the tempering of martensite, depending on the temperature, carbides form as described in [24]. The firsts to form, between 100 ^◦C and 200 ^◦C are transition carbides. As temperature increases to 250 ^◦C, cementite starts to precipitate, up to 700^◦C. Finally if the concentration in alloy elements is sufficient in the material, alloy carbides can form at temperatures greater than 550^◦C.

In steels containing more than 0,2 wt% of carbon, transition carbides precipitate from the beginning of tempering. As the temperature increases, transition carbides are replaced by cementite, even if cementite is not the most stable carbide [21, 24]. Indeed alloying elements can be strong carbide forming elements, if their concentration is sufficient, alloying elements can form carbides more stable than cementite. However, in order to precipitate, alloy carbide needs temperatures greater than 500^◦C. The nucleation of alloying element carbides requires substitutional diffusion, which is too slow at temperature below 500^◦. On the other hand, carbon can diffuse through the lattice, with interstitial diffusion. As transition carbides and cementite only require carbon diffusion to nucleate and grow, these carbides will precipitate at lower temperatures compared to alloying element carbides. Coarsening rate also depends on diffusion. Hence, once formed cementite coarsens faster than any alloy carbides since carbon diffusion is faster. Another conclusion is that cementite dispersion will always be coarser than alloy dispersion if both exist at the moment in the steel, as explained in [21].

Alloying element carbide dispersion that replaces cementite at higher temperatures, remains fine and stable even at high temperatures, giving the steel a high strength level. The transformation of coarse cementite dispersion into an Alloying element carbide dispersion is referred as secondary hardening and can happen in three different ways, described in [21,24].

(12)

· First of all, In-situ transformation, refers to the nucleation of alloying element carbides at the interface between cementite and ferrite. In that case, the carbon is provided by the cementite which disappears as alloying element carbides appears. Final dispersion is fine and widely spaced leading to a limited strength contribution.

· Separate nucleation within the ferrite matrix is the transformation way leading to nucleation of alloying element carbides on dislocations. A very fine alloy carbide precipitation is obtained.

· Finally, Separate nucleation on boundaries also lead to alloying element carbides precipitation. In that case, nucleation occurs on prior austenite grain boundaries, original martensite laths boundaries and new ferrite boundaries. In these regions energetically favourable, diffusion is even more rapid through diffusion paths. As diffusion is more rapid, the precipitated are more massive since the ageing process is more advanced.

Bhadeshia exposes the main strong carbide forming elements that can cause secondary hardening [21]. If molybdenum or tungsten is the predominant alloying element, then secondary hardening will be observed.

On the other hand, vanadium, titanium and niobium are so strong carbide formers that they can cause secondary hardening even at low concentration. Their carbide dispersion is also very fine and stable, even at temperatures as high as 700^◦C, providing resistance to ageing and temperature. Finally, nitrogen and boron lead to very stable carbides, more stable than cementite.

Temper embrittlement

Temper embrittlement corresponds to the decrease in toughness than can be observed in some cases. The temperature of tempering and composition can cause that phenomenon. The brittle-to-ductile transition is shifted to higher temperatures by manganese, nickel or chromium, favouring temper embrittlement [23].

Tempering temperatures between 350^◦C and 575^◦C, favour segregation of impurity atoms to prior austenite grain boundaries, causing embrittlement of some steels. The formation of carbides with a critical plate-like shape during tempering at lower temperatures (230^◦C to 370^◦C) can also lead to temper embrittlement of the material, as reported in [24].

2.3.2 Tempering of bainite

On the contrary to martensite which is a single-phase, bainite is composed of bainitic ferrite and cementite.

As bainite start temperature (Bs), is much higher than (Ms), bainite forms at higher temperatures compared to martensite and auto-tempering is unavoidable as explained in [25]. During the bainite formation, carbon partition from supersaturated ferrite to retained austenite and carbides precipitate. Recovery also happens during the formation of bainite [21]. Thanks to auto-tempering, the final bainitic microstructure is less sensitive to heat treatments compared to martensite. In martensite, strength is linked to the supersaturation with carbon, so as the tempering occurs and carbon precipitates, the material softens. However, during bainite formation, carbon precipitates in coarse cementite and so contributes little to the strength. As carbon in bainite does not precipitate during tempering and recovery already occurs during bainite formation, only the transformation of ferrite plate-like into equiaxed ferrite and the spheroidisation of cementite followed by its coarsening can induce softening in bainite during tempering.

Tempering sequence

Changes in the bainitic microstructure with a tempering follow the same logic as in martensitic microstructure case [25]. In tempering of bainite, first, carbon diffuse from supersaturated ferrite into austenite. Carbides precipitate in bainitic ferrite, even during bainite transformation as in upper bainite formation, see [21].

Then, carbides precipitate from enriched austenite and retained austenite decomposes in ferrite and carbides, at temperatures greater than 400 ^◦C.

As holding time increases cementite coarsens. Precipitation of alloying element carbides leads to secondary hardening. However, as cementite is much coarser in bainite than in martensite, it takes much longer to dissolve and the reaction is slower. Finally, recrystallization of the plate-like ferrite into equiaxed structure leads to softening of the material.

(13)

Carbides formation

Formation of carbides in bainite during tempering is exposed by Bhadeshia in [25]. Both cementite and alloying elements carbides are discussed.

Cementite exists in different forms in bainite, one is particles of cementite located at lath boundaries, whereas the other one is a fine dispersion of cementite particles within the laths. Cementite particles form during bainite formation, its coarsening rate during tempering is lower compared to martensite. This particular microstructure gives bainite better stability to tempering, with a fine microstructure kept through longer times.

Tempering allows diffusion of solutes from ferrite into the cementite, which approaches its equilibrium composition. Moreover, as alloy carbide form, the enrichment of cementite decreases until it starts to dissolve.

Even if cementite is not part of the equilibrium system, it is kinetically favoured in bainitic tempering like in martensitic tempering. Indeed, cementite growth does not require long-range diffusion of substitutional solutes, but only interstitial carbon diffusion. Whereas alloying element carbides, more stable than cementite, requires diffusion of substitutional solutes. The transition from cementite to alloying elements carbides can require numerous transition carbides.

No hardness peaks corresponding to secondary hardening are observed on tempered bainitic steels. However, alloying elements carbides precipitates can nucleate in different sites: within the tempered bainite plate, or on heterogeneous nucleation sites such as interfaces between ferrite and cementite.

(14)

Chapter 3 Materials and methods

3.1 Materials

Impact and abrasive resistant steels are a matter of interest in this study. ArcelorMittal Industeel provided six different steel grades to be part of the investigations. Each grade under investigation have particularities, specific chemical composition and a specific fabrication process that leads to their specific mechanical properties, a summary for each steel is presented in the next paragraph. Classic abrasion and impact resistant steels are based on high hardness, however high hardness level decreases shock resistance and also processing ability. Multiphased grades are based are based on their content of retained austenite, which is between 2

% wt and 8% wt, a ductile phase and a hard matrix improve the wear resistance. The different grades are presented in table3.1.

Table 3.1: Six grades provided by ArcelorMittal Industeel

Grade full name Abbreviation Plate thickness [mm] Microstructure Heat treatment

Multiphase 400 A MP 400 A 10

Bainite, Martensite Retained austenite

Micro-carbides

Air quenched

Multiphase 400 B MP 400 B 10

Micro-carbides

Oil quenched

Multiphase 450 A MP 450 A 40

Micro-carbides

Oil quenched Tempered

Multiphase 450 B MP 450 B 10

Micro-carbides

Water quenched

Multiphase-Ti 500 MP Ti 500 10

Bainite, Martensite Retained austenite Titanium carbides

Oil quenched Tempered

Martensitic 500 MA 500 25 Martensite Oil quenched

Tempered

3.2 Mechanical properties testing methods

3.2.1 Abrasive wear testing

To investigate wear resistance of steels, especially to sliding and impact situations, a stirring type machine was used. The stirring machine was designed by Industeel and enables to reproduce field conditions at the

(15)

laboratory scale, the experimental set-up is illustrated in figure 3.1. The sample tested are cuboids fixed on the machine by a screw-nut system, their dimensions are 5 mm x 25 mm x 50 mm. The rotation speed of the specimen is constant and equal to 600 rpm. The sample is put in rotation in an abrasive grit (coke, quartzite, sand, clinker, crushed glass), with or without lubricant, or liquid phase (water, acids). One test is 5 hours of rotations in abrasive grit. Every hour, the sample is taken out from the crucible, washed, dried and weighted so the wear criterion can be calculated according to the equation3.1. Also, the abrasive grit is changed every hour. In this study, the abrasive was dry quartzite.

To reach 500^◦C and perform high-temperature tests, the crucible was heated thank to a strip heater. Hence the sample was placed in the quartzite and the crucible close, then the strip heater was turned on. One hour and 15 minutes were needed for the quartzite to reach 500 ^◦C. After this heating time, the test started for one hour.

WC(i) = ∆M (i)

M₀ × 10000 (3.1)

where,

W_C(i), the wear criterion for a time i M0, the initial mass of the sample, [g]

∆M (i) = M0− M (i), the mass loss for a time i, [g].

Figure 3.1: Schematic view of the experiment set-up [4]

3.2.2 Impact testing

The impact resistance of the different kind of steels was also investigated with Charpy tests. The tests were realised according to standard ISO 148-1 [26] on V-notched sub-sized specimens (7.5 mm x 10 mm x 55 mm) taken along the length direction of the sheet of metal, see figure3.2. Specimens were hit with a 2 mm knife.

Samples were taken in the rolling direction in order to present assess the impact properties in the weaker steel direction. Sub-sized specimens were used since some steel plates under investigation were slightly thinner than 10 mm, in that case, the thickness of the specimen is reduced from standard size 10 mm to 7,5 mm, but the other dimensions such as the notch or the length are kept equal to the standard size’s one as explained Wallin [27]. To be able to compare the results of the different steels, all the specimens were machined in sub-sized ones, even for thicker plates.

(16)

To get the transition curves of the steel grades used in this study, samples were broken at different temperatures from -150 ^◦C to 410^◦C, for each temperature two specimens were tested. To cool down samples, isopentane and liquid nitrogen were used. Once the desired temperature reached, the specimens were im- mersed for 10 min, then they were taken out from the cooling system and hit by the Charpy knife in less than 5 sec, to respect the standard. In case of temperatures higher than room temperature, samples were heated for 10 min in a fluidized bed furnace before being tested in less than 5 sec.

Rupture energy, KV₂is directly recorded by the machine. To get the resilience, the rupture energy must be divided by the section of the samples (in that case: 0.6 cm²), hence the resilience and the rupture energy are linked by the formula 3.2.

KCV2=5KV2

3 (3.2)

where,

KCV2, the resilience, [J.cm²] KV2, the rupture energy, [J].

Figure 3.2: Schematic view of the samples used for impact testing [28]

3.2.3 Tensile tests

Different samples were submitted to tensile tests, realised according to the standard ISO 6892-1 [29] for room temperature tests, and ISO 6892-2 [30] for high-temperature tests. For high temperatures tests, the samples were taken along the width of the steel sheet in order to assess the weakest direction and so the weakest properties of the steel sheet. During the test, two different strain rates were applied to the samples, one before the 0.2 offset yield strength and one after. The change enables to have high precision in the elastoplastic region and to be time-efficient.

In the case of high-temperature tensile tests, the samples were heated up to the desired temperature and held at the temperature for at least 10 min. Heating, holding and testing step took place in the sample furnace, directly adapted on the testing machine. The temperature of the sample was monitored thanks to 3 thermocouples, placed at different locations on the sample. When the temperature of the test is below 600^◦C, the tolerance on the measured temperature compared to the testing temperature is ± 3 ^◦C, and the maximum difference of temperature between the thermocouples of the sample is set at 3^◦C by the standard ISO 6892-2.

In order to perform tensile tests, two different machines, with their configuration were used. In the case of room temperature tests, a Syntech 150 kN machine was used whereas for high-temperature tests, a Zwick 250 kN machine was used. In both cases, the length of the extensometer was 25 mm and the sample dimensions were identical. The strain rates used and the way to evaluate them is different on the machines. At high temperature, the first strain rate is equal to 0.15 mm.min⁻¹and is based on the strain of the extensometer, the second strain rate is equal to 1 mm.min ⁻¹ and is based on the strain of the crosshead. On the Syntec machine, used for room temperature tests, the first strain rate is 0.00025 s⁻¹ and the second is 0.0020 s⁻¹, both are based on the displacement of the crosshead. The testing conditions are summed up in table 3.2.

(17)

Table 3.2: Tensile test conditions, RT is Room Temperature and HT is High Temperature

Machine Syntech 150 kN Zwick 250 kN

Temperature RT HT

First strain rate 0.00025 s⁻¹ 0.15mm.min⁻¹

Base of the first strain rate Extensometer Cross head

Second strain rate 0.0020 s⁻¹ 1mm.min⁻¹

Base of the 2nd strain rate Cross head Cross head

3.2.4 Hardness measurements

For this study, different hardness measurements were performed, on different samples and at different mo- ments of the study. As the purposes were different in each case, several indenters and load were used, hence Vickers, Brinell and also Rockwell were performed.

Before each stirring test, Vickers hardness tests were performed on each side of the sample to detect any machining issue. Vickers hardness was also used to investigate the hardness of some impact samples after tempering treatment, and the hardness of stirring samples after high temperature tests. Tests were realised according to the standard [31]. The indentation was made with a Vickers indenter and a load of 30 kg during 10 sec. The diagonals of the imprint were evaluated with a PRESI software, that also calculated directly the corresponding hardness. Two indents were taken on each side of the sample, hence 4 values were used to get the hardness of each stirring sample.

Brinell hardness measurements were realised, according to the standard [32], on each steel grade under investigation, on as-delivered material and also on each sample that has been tempered. The indenter used was a steel ball of 10 mm diameter, applied on the steel with a load of 3 tons during 15 sec. Three indents were performed on each side of the sample, that means a total of 6 indents were done to access the hardness of the material. The diameter of the imprint was measured with a binocular microscope, in two directions. The average diameter of each imprint was used to calculate the hardness thanks to the formula given in equation 3.3.

HB = 2 × m

π × D(D −√

D²− d²) (3.3)

where,

HB, the Brinell hardness

m, the mass applied [kg], 3000 kg in this study D, the steel ball diameter [mm], 10 mm in this study

d, the imprint diameter [mm].

3.3 Tempering realisation

Tempering resistance is essential in industrial applications of high temperature abrasive resistant steels.

Indeed, some parts can be exposed to high temperatures for several hours, even days, hence the tempering resistance of each steel needs to be investigated. Several tempering, with different conditions in time but also in temperature were realised to investigate different material properties after high-temperature exposure.

In the scope of this study, only tempering effects (exposure to temperatures ranging from 100 to 500 ^◦C) was studied. Whereas annealing, which deals with higher temperatures (from 500 to 900 ^◦C) was left out.

Tempering treatment, also known as ageing, can relief residual stress and decrease hardness of the material and so decrease abrasion resistance see [33].

Tempering effects on hardness, microstructure and impact properties was investigated. In order to investigate the impact properties of the different steels after exposure to different temperatures, temperings were realised directly on sub-sized impact specimens (7.5 mm x 10 mm x 55 mm). Several temperatures were selected:

410 and 500^◦C. One of the tempering was realised in a fluidized bed furnace, the same as the one used to heat up the impact specimens for Charpy test. Two samples of each kind of steel were let in this furnace for

(18)

2 hours at 410^◦C. After, they were air-cooled and tested at room temperature. Two more samples of each kind were treated in the same conditions: 2 hours at 410 ^◦C, in the fluidized bed furnace but tested at 410

◦C. This experiment enables to investigate the effect of cooling. The second tempering (2 hours at 500^◦C) is realised in a Nabertherm furnace. The samples were also air-cooled and tested at room temperature.

The effect of tempering duration on hardness and microstructure were also a mattered of interest. Hence samples were cut in the different steel plates. The sample dimensions were 30 mm x 30 mm and their thickness was the same as the plate’s thickness, summed up in table 3.3. Different durations of treatment were selected: 10 min, 2 h, 8 h and 240 h. As the samples had different thicknesses and were put in a hot furnace, the duration of treatment was different, since the exposure time starts once the sample has a homogeneous temperature. A 10 mm-thick sample takes 10 min to reach and homogeneous temperature, but a 30 mm-thick one takes 20 min more. As all the samples were put in the furnace at the same moment, they were taken out at the different time, except for 240 h duration since in that case 30 min difference can be considered as negligible.

Table 3.3: Dimensions of the samples used for tempering treatments Steel grade Sample dimensions

MP 400 A 30 mm x 30 mm x 10 mm MP 400 B 30 mm x 30 mm x 10 mm MP 450 A 30 mm x 30 mm x 40 mm MP 450 B 30 mm x 30 mm x 10 mm MP Ti 500 30 mm x 30 mm x 10 mm MA 500 30 mm x 30 mm x 30 mm

3.4 Microstructural characterisation techniques

In order to observe the microstructure the samples were cut in the rolling direction so segregated veins can be observed. Then the samples were hot mounted, polished and etched with Nital 3 % solution.

Half of the entire thickness of the plate was investigation. Indeed the microstructure is symmetrical on both side of the mid-thickness. Hence three different areas were studied: Near surface, quarter thickness and mid-thickness. With optical microscope, enlargements from 25 to 200 were used systematically and 1000 enlargement on some interesting details. Optical microscope was used to investigate microstructure changes after tempering. The six grades selected for the project were observed with optical microscopy before any heat treatment and after a tempering carried on at 500^◦C, for 240 hours.

Scanning Electron Microscope (SEM) was also used to study samples microstructure at a smaller scale.

Two grades were selected to be observed with SEM: MP 450 A and MA 500. For each grade, MP 450 A and MA 500, two samples were selected: one tempered at 500 ^◦C for 240 h and a second in as-delivered conditions. The selection of the samples was done to observe the effect of tempering on the two main sorts of microstructures in this study: multiphased martensitic-bainitic and fully martensitic microstructure.

(19)

Chapter 4 Results

4.1 Mechanical properties

4.1.1 Abrasion resistance

Room temperature abrasion tests

Stirring tests (see 3.2.1) were performed to investigate the abrasion resistance of each grade. To have an idea of the repeatability, several tests were performed on the sample grade. As the results were similar for these tests, it was possible to test only one sample for the other grades. One of the outcomes of the stirring test is the wear criterion, that can also be translated into a wear rate in percentage of initial mass lost per hour of test. For each grade, there is a linear correlation between abrasion duration and wear criterion, since the wear rate does not depend on time. The experimental data were linearized with a set intercept equal to zero, since at the beginning of the experiment, the wear is zero. The linearisation of the experimental data ends in a coefficient of determination greater than 0, 99, hence good goodness of fit. That is why the results presented and their interpretation is based on the linearized values.

The results of the room temperature experiments on each grade are recorded in table 4.1. In this table, the final wear criterion is calculated with the formula 3.1 based on the mass lost by the sample after 5 hours of stirring test. The wear rate is the loss of initial mass in percentage per hour of stirring test. It can be seen that out of the six different grades, two have a significant higher wear rate: MP 400 B (with 0, 44%_IM/h) and MP 400 A (with 0, 46%_IM/h). On the other side, MP 450 A has the lowest wear rate at room temperature: 0, 39%_IM/h. MP Ti 500 and MP 450 B have equivalent wear rate of 0, 40%_IM/h, a slightly better performance than MA 500 with 0, 41%_IM/h. However with a wear rate between 0,39 and 0,41%_IM/h, the abrasion resistance, in these conditions, can be considered as equivalent for MP 450 A, MP 450 B, MP Ti 500 and MA 500.

Table 4.1: Results of stirring tests at room temperature

Steel grade Number of

samples tested Final wear criterion Wear rate [%IM/h]

MP 400 A 1 228 0,46

MP 400 B 1 219 0,44

MP 450 A 1 196 0,39

MP 450 B 2 200 0,40

MP Ti 500 1 199 0,40

MA 500 1 204 0,41

(20)

Abrasion tests at 500^◦C

Abrasion resistance of each grade was also studied at 500^◦C. In that case, several samples were studied for MP Ti 500, MP 450 B and MP 400 B, whereas other grades were tested once. A great changeability was observed between the different samples of the same grade, for instance, the standard deviation on final wear criterion between the three samples of MP Ti 500 is equal to 40, with a mean value of 375 after 5 hours testing. As the repeatability of the test at high temperature is so low, it is difficult to rank the grades according to their abrasion resistance at high temperature. A linearisation of the results were done in the same way as it was for room temperature tests, with a set intercept equals to zero. Once more, the goodness of fit was high and the coefficient of determination above 0, 99. Accordingly, even if the repeatability of the test was low at a higher temperature, the linearity of the results was good.

In table 4.2, the results of the 500 ^◦C experiments are summed up. Despite differences in behaviour at room temperature, at 500^◦C the results are close and all the grades seem having similar abrasion behaviour.

MP 450 A, MP 450 B, MP Ti 500 and MA 500 exhibit once more close wear rate, between 0, 72%IM/h and 0, 76%IM/h. Whereas MP 400 B has a higher wear rate (0, 82%IM/h). On the other side, MP 400 A showed the lowest wear rate with 0, 67%IM/h. Nevertheless, as the repeatability of the test is so low, the results should be taken lightly.

Table 4.2: Results of stirring tests at 500^◦C

Steel grade Number of

samples tested Final wear criterion Wear rate [%_IM/h]

MP 400 A 1 337 0,67

MP 400 B 2 411 0,82

MP 450 A 1 374 0,75

MP 450 B 2 361 0,72

MP Ti 500 3 378 0,76

MA 500 1 382 0,76

4.1.2 Impact properties

Transition curves

The brittle-to-ductile transition was studied for each grade under investigation. As the abrasion resistance of the material could be linked to toughness, the rupture energy of the grades at high temperature is a matter of interest to understand high-temperature abrasion phenomenon, Charpy impact tests were so performed until 410^◦C, the maximal temperature of the available material. The main results are summed up in table 4.3and transition curves plotted in figure4.1.

MP Ti 500 does not present a real transition from brittle to ductile, indeed brittle failure occurs from -150

◦C to 410 ^◦C. the rupture energy is low, about 11 J.

MP 400 B, on the other hand, has a strong brittle-to-ductile transition between -80 ^◦C and 150 ^◦C. The upper plateau is about 130 J, and the room temperature rupture energy is 107 J. However, the upper plateau does not last as the temperature increases. Indeed at 410^◦C, the rupture energy drop to 77 J, it is 43 % less than at 200^◦C.

With an upper plateau about 120 J, MP 400 A is just below MP 400 B. Its transition is more gentle and at higher temperatures, from -80^◦C and, 200^◦C. As the transition is less steep, the room temperature rupture energy is lower: only 35 J. At 410 ^◦C, the rupture energy does not belong to the upper plateau any more, but equals 75 J being 34% less than at 200^◦C.

MP 450 B and MP 450 A showed similar transition curves. The transition between fragile and brittle failure occurs from -80^◦C to 100^◦C. Their upper plateau is about 70 J, and their rupture energy at room temperature is similar (47 J for MP 450 B and 54 J for MP 450 A). However, their upper plateaus have some differences. In the case of MP 450 B, the upper plateau last until 300^◦C, and the gap between KV₂(200^◦C) and KV₂(410^◦C) is about 32 %. Whereas the upper plateau of the MP 450 A transition curve is shorter. Indeed at 300^◦C, the rupture energy is already lower than 70 J. The rupture energy at 410^◦C is even lower giving a gap between

(21)

200 ^◦C and 410^◦C of 36 %.

Finally, MA 500 also changes from brittle to ductile behaviour between -80^◦C and 100^◦C, but the transition is much softer since the rupture energy values are lower. At room temperature, the rupture energy is equal to 24 J and the upper plateau is about 50 J. As the upper plateau is low, the drop at 410^◦is less important since the rupture energy is 45 J, being 18 % less than at 200^◦C.

Figure 4.1: Transition curves plotted for each grade from -150 ^◦C to 410 ^◦C

Table 4.3: Main results of impact tests

Steel grade KV₂(20^◦C) [J] KV₂(200^◦C) [J] KV₂(410^◦C) [J]

MP 400 A 35 113 75

MP 400 B 107 134 77

MP 450 A 54 72 46

MP 450 B 47 71 48

MP Ti 500 10 12 -

MA 500 24 55 45

Tempering effect on impact resistance

In addition to the transition curve of each grade, the effect of tempering on the impact properties of the material was also investigated. Three different conditions were tested: samples tempered for 2 h at 410 ^◦C and tested at 410 ^◦C, samples tempered in the same conditions (2 h, 410 ^◦C) but tested after air cooling at room temperature and samples tempered for 2 h at 500^◦C tested at room temperature after air cooling.

The results of the different tests are presented on table4.4and on figure 4.2to enable easy comparison.

Tempering has a different effect on the impact properties of each grade. In the case of MP Ti 500, no matter the tempering or testing conditions (at high temperature or after air cooling), the values of rupture energy were low, about 10 J. Even if after a 2 h long tempering at 410^◦C, the rupture energy was even lower, only 8 J.

In the case of MP 400 A, the rupture energy at 410 ^◦C, with or without holding time is equal, but the aircooling after tempering decrease a lot the toughness of the material. The loss is about 45 J, that being so 60 % of the rupture energy needed at 410 ^◦C, no matter the holding time. The tempering temperature increase from 410 ^◦C to 500 ^◦C causes a small decrease in toughness (3 J).

(22)

MP 400 B showed an increased toughness as with tempering without cooling compared to the same testing temperature without holding time in temperature. However, with air cooling added to the treatment, the rupture energy was lower. Increasing the temperature of tempering from 410^◦C to 500^◦C, linked to a small decrease of toughness at room temperature after treatment.

The toughness of MP 450 A is constant through time, the values are equal with or without tempering.

However, with air cooling, the toughness drops with a loss of 20 J, being 43 % the rupture energy needed at 410 ^◦C after a 2 h tempering at the same temperature. As for tempering temperature increases, the toughness increases significantly.

The tempering treatment at 410 ^◦C of MP 450 B without cooling increase the toughness of the material compared to the sample grade tested at 410 ^◦C, but if the sample is air-cooled before testing, the rupture energy is lower. The third tempering at a higher temperature (500 ^◦C, for 2 h) showed higher rupture energy at room temperature compared to the testing conditions applied to the samples treated at a lower temperature.

MA 500 showed similar responses to tempering. With a moderate increase in toughness as holding time increase, a decrease as air cooling is added after the tempering. However, the difference between the samples tested at room temperature after tempering for 2 h at either 410^◦C or 500^◦C is much bigger. Indeed in the case of tempering for 2 h at 500 ^◦C, the samples tested at room temperature showed the highest values of rupture energy obtained for this grade, about 83 J.

Table 4.4: Results of impact tests perform to investigate tempering effect, H stand for tested at High temperature and R for tested at Room temperature

Steel grade KV2(410^◦C, 2h, H) [J] KV2(410^◦C, 2h, R) [J] KV2(500^◦C, 2h, R) [J]

MP 400 A 75 30 27

MP 400 B 86 60 56

MP 450 A 47 27 44

MP 450 B 58 30 39

MP Ti 500 11 8 -

MA 500 53 28 83

Figure 4.2: Histogram of rupture energy of the different grades after tempering, H stand for tested at High temperature and R for tested at Room temperature

(23)

4.1.3 Tensile properties

For each grade, several samples were submitted to tensile test in order to assess mechanical properties at room temperature and at higher temperatures, being 200 ^◦C, 300^◦C, 400^◦C and 500^◦C. The evolution of ultimate tensile strength (UTS) and yield strength 0.2 (YS) as temperature increase were studied, see figure 4.3. At room temperature, the mechanical properties spread out on a large interval. As the temperature was increased, the initial range vanishes and the mechanical properties decrease. UTS values at room temperature range from 1774 MPa to 1161 MPa, being 583 MPa difference. Whereas at 500^◦C, the UTS values range on a 201 MPa large interval, from 912 MPa to 711 MPa. The same phenomenon occurs with YS values: at room temperature, the properties of the grades are distributed in an interval of 496 MPA, but as temperature is increased, the interval shrinks down to 90 MPa difference.

MP Ti 500 is the grade that shows the highest UTS and YS at room temperature, respectively 1744 MPa and 1350 MPa. UTS of this grade increases slightly with temperature up to 200^◦C. However, at higher temperatures, UTS decreases more abruptly, and will become one of the lowest values of UTS at 500^◦C. On the YS side, the value of YS decreases from room temperature to 200^◦C, stay stable as the temperature reaches 300 ^◦C and then decreases as temperature increases up to 500^◦C. MP Ti 500 shows the lowest value of YS at 500 ^◦C. Hence even if at room temperature, this grade has the highest UTS and YS, as the mechanical properties decrease abruptly with temperatures, the properties at 500 ^◦C are pretty low.

At room temperature, MP 450 B has also good mechanical properties with an UTS of 1606 MPa and an YS equals to 1265 MPa. The evolution of its mechanical properties is similar to MP Ti 500. UTS increases as temperature increases slowly from room temperature until 200^◦C, as the temperature reaches higher values, UTS of the material decreases with a higher rate. However at 500 ^◦C, is the UTS value of MP 450 B not so low in comparison with the other grades, indeed it has the second highest value, 845 MPa. On the other hand, YS decreases slightly between room temperature and 200^◦C and ends up being the highest YS value (1195 MPa) of the tested grades. As temperature increases above 200 ^◦C, YS decreases more radically. At 500 ^◦C,is the YS of all grades very similar to each other , with MP 450 B showing the medium value.

Looking at room temperature UTS ranking, MA 500 is right behind MP Ti 500 and MP 450 B. Its behaviour is very similar, a slight increase between room temperature and 200^◦C, followed by a more abrupt decrease as the temperature reaches 500^◦C. MA 500 UTS value at 500^◦C is the lowest of the tested grades. On the YS side, the values at room temperature and 500 ^◦C are close to MP 450 B ones. However the behaviour in between is slightly different: as temperature increases, from room temperature to 200 ^◦C MA 500 YS decreases faster than MP 450 B YS.

MP 450 A shows another type of evolution with temperature. From room temperature to 200 ^◦C, the UTS increases slightly, from 1364 MPa to 1490 MPa. At 300^◦C, the UTS value is just above 200^◦C value, but still higher than the room temperature UTS. After 300 ^◦C, the UTS decreases with the same rate as MP 450 B.

The sharp decrease starts at a higher temperature compared to previous grades (MP Ti 500, MP 450 B or MA 500), hence even if initial properties are lower, high-temperature properties are averages compared to previous grades. Concerning YS, MP 450 A show a strong increase between room temperature and 200^◦C.

That is why even if YS decreases at higher temperatures, MP 450 A does not present low YS at 500^◦C.

Looking at UTS evolution with temperature, MP 400 B has a behaviour very close to MP 450 A. As temperatures increases from 20 ^◦C to 200 ^◦C, an increase in UTS is observed, followed by a modest decrease.

After 300^◦C, UTS values drop. However, even after 300^◦C, the decrease in UTS is more gentle than other grades’. That leading to the highest UTS value at 500^◦C. MP 400 B YS value is also the highest at 500^◦C.

The decrease in YS as temperature increases is gentle, hence YS at 500^◦C is only 283 MPa less than YS at 20^◦C.

Finally, MP 400 A mechanical properties show evolution with temperature leading to a small gap between room temperature properties and properties at 500 ^◦C, compared to other grades. The results of MP 400 A tensile tests are extracted from Industeel archives. The tests were performed on similar samples, took in similar conditions on a 15 mm thick plate of MP 400 A. In that case, UTS value decreases from room temperature to 100 ^◦C, before increasing slightly from 100^◦C to 300^◦C. Leading to the same situation as for MP 400 B and MP 450 A with a UTS value at 300 ^◦C higher than the UTS value a 20 ^◦C. Once the temperature overcome 300^◦C, UTS values decrease to end with medium values compared to other grades. The MP 400 A YS evolution with temperature is very similar to MP 400 B YS evolution, with a gentle decrease as temperature increases.

Regarding tensile tests results, the different grades can be divided into two groups defined according to their

(24)

behaviour between room temperature and 300 ^◦C. On the one hand, the materials with a shape decrease in UTS once the temperature reaches 200 ^◦C, gathering the grades with the higher mechanical properties at room temperature: MP 450 B, MP Ti 500, and MA 500. One the other hand can be found the grades with a UTS rather stable up to 300 ^◦C, but with a stronger drop in UTS at higher temperatures. In this category, MP 400 A, MP 400 B and MP 450 A present lower mechanical properties at room temperatures, but exhibit smaller gaps between UTS and YS at room temperature and at 500^◦C.

4.1.4 Hardness

Stirring test samples surface hardness

Before any test, the surface hardness of the stirring sample was evaluated in HV 30, with 4 indents, 2 on each face of the sample. After 5 hours of testing, performed according to the procedure described in 3.2.1, the hardness of each sample was evaluated once more, with 4 indents. In table 4.5, is presented the initial hardness of each grade, which correspond to the average hardness of all the samples of that grade. The following column is the average hardness of the stirring samples after 5 hours of testing at room temperature.

Finally, the last column records the mean hardness of the different samples tested for 5 hours at 500 ^◦C.

After room temperature tests, the difference between initial and final hardness is negligible. However, after high-temperature tests, the final hardness is much lower than the initial one. During the abrasion test, MP Ti 500 hardness decreased of 30 % in comparison with its initial hardness. MP 450 B and MA 500 also showed important hardness drop with a loss of 29 % of their initial hardness. MP 400 B lost a bit less, with 24 %. The smallest hardness drops correspond to MP 450 A, with 18 % and MP 400 A with only 10 %.

MP 400 A has the lowest initial hardness, but also the lowest hardness loss due to tempering during the high-temperature abrasion test. On the contrary, MP Ti 500 has the highest initial hardness of the tested grades, but its loss of hardness through tempering is the most important. Another observation is the extent of values. Initial hardness values range from 552 HV30 to 397 HV30, the gap correspond to 155 HV30.

Whereas hardness values measured after high-temperature abrasion tests vary from 385 HV30 to 325 HV30, reducing the difference between extremes to 60 HV30.

Table 4.5: Mean hardness of samples used of stirring tests, before and after 5 h of test

Steel grade Initial hardness [HV 30]

After test at 20^◦C [HV 30]

After test at 500^◦C [HV 30]

MP 400 A 397 402 357

MP 400 B 426 430 325

MP 450 A 454 459 374

MP 450 B 510 509 362

MP Ti 500 552 552 385

MA 500 475 475 336

Tempering effect on hardness

Different tempering treatments were performed on each grade, the temperature selected was 500^◦C and the duration varied from 10 min to 240 h. After air cooling, each sample was submitted to the Brinell hardness test. The main idea is to understand the evolution of hardness through tempering and to compare the different hardness values with the initial hardness of the material. Hardness evolution with tempering time is represented in figure 4.4and the exact values are available in table4.6.

Initially, MP Ti 500 have the highest hardness of all the tested grades, with 492 HB. This grade is also the one with the highest hardness after 240 hours tempering at 500 ^◦C. However, from initial condition to tempered for 240h, the material losses 31 % of its initial hardness, from 492 HB to 342 HB. This loss of hardness occurs as soon as the treatment begins since, after 10 min of tempering at 500 ^◦C, the hardness already dropped to 397 HB, being a loss of 19 % in comparison with initial hardness.

MP 450 B behaviour is similar to MP Ti 500’s. The hardness of the sample treated for 240 h at 500 ^◦C is 33 % lower than the initial hardness. The initial hardness is also pretty high, with 462 HB. However, with

(25)

(a) Ultimate tensile strength evolution with temperature

(b) Yield strength 0.2 evolution with temperature

Figure 4.3: Evolution of mechanical properties with temperature from room temperature to 500^◦C

Selection of high-temperature abrasion resistant steels for the mining and processing industry